US Army Corps

of Engineers
Construction Engineering USACERL Technical Report 99/41
Research Laboratories April 1999

Reliability Centered Maintenance

(RCM) Guide
Operating a More Effective Maintenance Program

Alan Chalifoux and Joyce Baird

This manual outlines a comprehensive method

of organizing an efficient maintenance program
by applying the concepts of Reliability Centered
Maintenance (RCM). RCM combines
professional intuition and a rigorous statistical
approach, and recognizes that different
maintenance strategies apply to different facility
equipment: run-to-failure, preventive, predictive,
and proactive maintenance. The RCM approach
applies these differing maintenance strategies in
an optimal mix, to ensure that facility equipment
is maintained sufficient to accomplish the facility
mission without wasting maintenance labor.
This guide is meant to help maintenance
supervisors, managers, and technicians
organize and operate an efficient and effective
maintenance program in an environment of
maintenance budget cutbacks.

Approved for public release; distribution is unlimited.

USACERL TR 99/41 1

Executive Summary

Maintenance management is a complicated business. Facility maintenance

budgets are continually scrutinized by fiscal managers in a constant effort to
trim dollars. Maintenance managers are under constant pressure to squeeze
every last bit of productivity out of every maintenance dollar.

This manual outlines a comprehensive method of organizing an efficient mainte-

nance program through applying the concepts of Reliability Centered Mainte-
nance (RCM). Combining professional intuition and a rigorous statistical ap-
proach, RCM recognizes that there are different maintenance strategies followed
for different facility equipment: run-to-failure, preventive maintenance, predic-
tive maintenance, and proactive maintenance. The RCM approach applies these
differing maintenance strategies in an optimal mix, to ensure that facility
equipment is being maintained sufficient to accomplish the facility mission with-
out wasting inordinate amounts of maintenance labor baby sitting facility
equipment.

This manual presents the RCM approach for maintenance supervisors, manag-
ers, and technicians to use as a guide in organizing and operating a tight, cost-
effective, lean and mean maintenance program in light of and in spite of the
continual cutbacks in maintenance budgets.
2 USACERL TR 99/41

Foreword

This study was conducted for the Facilities Management Division (EMD) at
Madigan Army Medical Center (MAMC) under Project 4A162720D048, Indus-
trial Operations Pollution Control Technology.; Work Unit Y67, Reliability Cen-
tered Maintenance. The technical monitor was Michael Carico, MAMC-FMD.

The work was performed by the Industrial Operations Division (UL-I) of the
Utilities and Industrial Operations Laboratory (UL), U.S. Army Construction
Engineering Research Laboratories (CERL). The CERL principal investigators
were Alan Chalifaux and Jearldine I. Northrup. Special credit is given to the
National Aeronautics and Space Administration (NASA) for the use of its docu-
ment Reliability Centered Maintenance Guide for Facilities and Collateral
Equipment, December 1996 in the preparation of this report. Walter J. Mikucki
is Chief, CECER-UL-I; Dr. John Bandy is Laboratory Operations Chief, CECER-
UL; and Gary W. Schanche was the responsible Technical Director, CECER-TD.
The CERL technical editor was William J. Wolfe, Technical Information Team.

Dr. Michael J. OConnor is Director of USACERL.

USACERL TR 99/41 3

Contents

Executive Summary ......................................................................................................... 1

Foreword............................................................................................................................ 2

1 Introduction................................................................................................................. 9
Background ......................................................................................................................... 9
Objectives.......................................................................................................................... 12
Approach........................................................................................................................... 13
Scope ................................................................................................................................ 13
Units of Weight and Measure ............................................................................................ 13

2 RCM Definition and Philosophy ............................................................................. 14

Definition ........................................................................................................................... 14
RCM Analysis.................................................................................................................... 14
RCM Principles ................................................................................................................. 15
The RCM Process............................................................................................................. 17
RCM Program Benefits ..................................................................................................... 19
Impact of RCM on a Facilitys Life Cycle ........................................................................... 22

3 RCM Program Components.................................................................................... 24

Reactive Maintenance....................................................................................................... 24
Preventive Maintenance (PM) ........................................................................................... 25
Preventive Maintenance Criteria ..................................................................................................26
Determining PM Task and Monitoring Periodicity .........................................................................26
Condition Monitoring (CM) ................................................................................................ 28
Proactive Maintenance...................................................................................................... 30
Specifications for New/Rebuilt Equipment ........................................................................ 31
Balance........................................................................................................................................32
Alignment.....................................................................................................................................34
Alignment Effects .........................................................................................................................35
Failed-Part Analysis .....................................................................................................................37
Root-Cause Failure Analysis (RCFA) ...........................................................................................37
Reliability Engineering and Reliability Calculations......................................................................38
Rebuild Certification/Verification ..................................................................................................42
Age Exploration............................................................................................................................42
Recurrence Control......................................................................................................................43
4 USACERL TR 99/41

Facility Condition Assessment .....................................................................................................44

4 Use of Condition Monitoring (CM) Technologies ................................................. 45

Introduction ....................................................................................................................... 45
Spot Readings versus Continual Real-Time Data Collection ........................................... 45

5 Vibration Monitoring and Analysis......................................................................... 47

Theory, Applications, and Techniques............................................................................... 47
Basic Vibration Theory .................................................................................................................47
Information Obtained through Vibration Monitoring......................................................................51
Detection Interval/Amount of Data Collected ...............................................................................52
Overall Vibration...........................................................................................................................52
Spectrum Analysis and Waveform Analysis .................................................................................52
Torsional Vibration........................................................................................................................53
Multi-Channel Vibration Analysis..................................................................................................53
Shock Pulse Analysis...................................................................................................................53
Vibration Sensor Mounting (Permanent Installations) ..................................................................53
Laser Shaft Alignment..................................................................................................................54
Limitations ......................................................................................................................... 54
Logistics ............................................................................................................................ 54
Equipment Required ....................................................................................................................54
Operators.....................................................................................................................................55
Available Training .........................................................................................................................55
Cost .............................................................................................................................................55

6 Thermography .......................................................................................................... 56
Theory and Applications ................................................................................................... 57
Limitations ......................................................................................................................... 58
Logistics ............................................................................................................................ 58
Equipment Required ....................................................................................................................58
Operators.....................................................................................................................................58
Training Available .........................................................................................................................58
Cost .............................................................................................................................................59

7 Passive (Airborne) Ultrasonics............................................................................... 60

Theory, Applications, and Techniques............................................................................... 60
Basic Theory of Ultrasonic Detection ...........................................................................................61
Leak Detection .............................................................................................................................62
Electrical Problems ......................................................................................................................65
Mechanical Inspection .................................................................................................................66
Ultrasonic Translators...................................................................................................................69
Limitations ......................................................................................................................... 69
USACERL TR 99/41 5

Logistics ............................................................................................................................ 69
Equipment Required ....................................................................................................................70
Operators.....................................................................................................................................70
Training Available/Required..........................................................................................................70
Cost .............................................................................................................................................70

8 Lubricant and Wear Particle Analysis .................................................................... 71

Purpose............................................................................................................................. 71
Machine Mechanical Wear Condition...........................................................................................71
Lubricant Condition ......................................................................................................................71
Lubricant Contamination ..............................................................................................................72
Standard Analytical Tests .................................................................................................. 72
Visual and Odor ...........................................................................................................................72
Viscosity.......................................................................................................................................73
Water ...........................................................................................................................................73
Percent Solids/Water....................................................................................................................73
Total Acid Number (TAN)..............................................................................................................73
Total Base Number (TBN) ............................................................................................................74
Spectrometric Metals ...................................................................................................................74
Infrared Spectroscopy ..................................................................................................................74
Analytical Ferrography .................................................................................................................74
Special Tests ..................................................................................................................... 75
Glycol Antifreeze..........................................................................................................................75
Karl Fischer Water .......................................................................................................................75
Application......................................................................................................................... 76
Motors, Generators, Pumps, Blowers, Fan...................................................................................77
Gearboxes ...................................................................................................................................77
Chillers.........................................................................................................................................78
Diesel Engines .............................................................................................................................78
Compressors ...............................................................................................................................78
Hydraulic Systems .......................................................................................................................78
Large Reservoirs..........................................................................................................................78
Lubrication Analysis .....................................................................................................................78
Sampling......................................................................................................................................79

9 Electrical Condition Monitoring.............................................................................. 80

Techniques ........................................................................................................................ 80
Megohmmeter Testing..................................................................................................................81
High Potential Testing (HiPot).......................................................................................................81
Surge Testing ...............................................................................................................................81
Conductor Complex Impedance...................................................................................................82
Time Domain Reflectometry ........................................................................................................82
6 USACERL TR 99/41

Motor Current Spectrum Analysis (MCSA) ..................................................................................82

Radio Frequency (RF) Monitoring................................................................................................82
Power Factor and Harmonic Distortion ........................................................................................82
Motor Current Readings...............................................................................................................83
Airborne (Passive) Ultrasonics .....................................................................................................83
Transformer Oil Analysis ..............................................................................................................83
Applications....................................................................................................................... 83
Equipment to be Monitored..........................................................................................................83
Conditions Monitored ...................................................................................................................84
Detection Interval .........................................................................................................................84
Accuracy ......................................................................................................................................84
Limitations....................................................................................................................................84
Logistics ............................................................................................................................ 85
Equipment Required ....................................................................................................................85
Operations ...................................................................................................................................85
Training Available .........................................................................................................................85
Cost .............................................................................................................................................85

10 Non-Destructive Testing .......................................................................................... 86

Techniques ........................................................................................................................ 86
Radiography.................................................................................................................................86
Ultrasonic Testing (Imaging).........................................................................................................87
Magnetic Particle Testing .............................................................................................................88
Dye Penetrant ..............................................................................................................................89
Hydrostatic Testing.......................................................................................................................89
Eddy Current Testing....................................................................................................................89
Location and Intervals ....................................................................................................... 90
Intervals .......................................................................................................................................90
Locations .....................................................................................................................................92
Applications....................................................................................................................... 92
Limitations ......................................................................................................................... 93

11 Conclusions .............................................................................................................. 95

References....................................................................................................................... 96

Distribution
USACERL TR 99/41 7

List of Figures and Tables

Figures

1 Bearing life scatter. ....................................................................................................11

2 RCM logic tree. ..........................................................................................................17
3 Sample RCM system data sheet. ..............................................................................19
4 Sample failure mode sheet. .......................................................................................20
5 Failure mode sheet for bearings. ...............................................................................21
6 Failure mode sheet for stator. ....................................................................................21
7 Maintenance cost trends under an RCM program. ...................................................22
8 Stages of life cycle cost commitment. .......................................................................23
9 Decrease in life of cylindrical roller bearings as a function of misalignment. ............36
10 Sample vibration data................................................................................................50
11 Two plots of vibration data juxtaposed in the same graph.........................................51

Tables

1 Recommended coupled alignment tolerances (General Motors, 1993)....................35

2 Recommended maximum inspection intervals (API 570). ........................................90
USACERL TR 99/41 9

1 Introduction

Background

Maintenance often takes a low priority in the overall operating strategy of a fa-
cility. Maintenance programs are managed and funded by people, and human
nature seems to abide the old tenet, If it aint broke, dont fix it. In facilities
management the definition of broke is extreme. Broke typically means that a
piece of equipment has catastrophically failed (e.g., resulting in a pollution fine),
or (at the very least) that it has failed to the point that it has become an annoy-
ing disturbance in the normal daily operation of a facility.

While few people will argue against the need for performing regular mainte-
nance, few fiscal managers will make the financial commitment to funding main-
tenance programs at a level that will keep a facility well maintained. Fiscal
managers usually assign maintenance programs a very low priority. Compared
to other facility departments, maintenance departments have no real product
and - as such - produce no real income. Many fiscal managers view money spent
on maintenance as money thrown down a black hole. In spite of any life-cycle
proofs to the contrary, fiscal managers look to cut maintenance budgets first
when any other fiscal need arises. Not until they see the bathroom floor flooded
with sewage or swelter in an office working at 85 F for hours do they realize
that something is broken and may need repair.

Fiscal managers continually put maintenance budgets under the closest scrutiny
in an effort to reduce dollars spent on maintenance, while expecting facility per-
formance to remain on a constant par. This forces maintenance supervi-
sor/managers to trim essential (but less obvious) work from their daily agendas.
The most common area trimmed is preventive maintenance, i.e., those mainte-
nance activities performed on facility equipment before equipment failure. The
importance of preventive maintenance is less obvious to those people not inti-
mately familiar with facility equipment and operation. The consequences and
cost of not performing PM only become obvious when it is too late.

Preventive maintenance requires that maintenance personnel pay regular visits

to observe the condition of facility equipment. The most basic tasks on these PM
visits is to take a look at the equipment to see if there are any telltale signs of
10 USACERL TR 99/41

failure or imminent failure. Also, depending on the type of equipment, the main-
tenance mechanic may have a checklist of tasks he has to perform (e.g., draining
a little oil and visually checking for foreign matter or discoloration). In large fa-
cilities such as MAMC, the basic PM task of walking out to a piece of equipment
and giving it a quick look over requires a great deal of time. It also requires that
this time be invested by a trained mechanic; untrained personnel are likely to
miss telltale signs of failure. MAMC-FMD wanted to determine if there was any
means of automating this basic PM inspection activity. By doing so they (or any
other facility) could free up skilled labor for more other tasks, thereby squeezing
more out of every maintenance dollar.

Preventive maintenance is typically performed based on the calendar. Mainte-

nance personnel schedule visits to a particular piece of equipment based on cer-
tain time intervals having elapsed. While certainly better than no PM at all,
calendar-based PM may result in too much time being spent on a piece of equip-
ment. Each visit to a properly functioning piece of equipment takes time away
from other maintenance activities. Numerous visits to a piece of equipment with
no news to report can be regarded as wasted maintenance dollars. Calendar-
based PM, while much preferable to no PM, is not the optimal way to run a PM
program.

Typically, the next step up (from calendar-based PM) is performing PM based on

equipment run time. This method is thought to provide a better means of get-
ting a maintenance mechanic out to a piece of equipment just as it is beginning
to show signs of wear. Intuitively, performing PM based on equipment run time
makes sense. Equipment does not have to be checked repeatedly if it has not
been used. Generally speaking, it is the actual operation of the equipment that
wears it down, so it makes sense to check the equipment after it has run a suffi-
cient amount of time to incur some wear. (However, run time is not a proper cri-
terion for performing PM on all equipment. There are certain pieces of equip-
ment that require visual inspection when they have not been run).

Calendar-based PM assumes that failure probabilities can be determined statis-

tically for individual machines and components and parts can be replaced or ad-
justments can be performed in time to preclude failure. This is not always true.
A common practice has been to replace bearings after a certain number of oper-
ating hours, the assumption being that bearing failure rate increases with time
in service. Figure 1 shows that this assumption is not always true. This figure
shows the failure distribution of a group of 30 identical 6309 deep groove ball
bearings installed in bearing test machines that run the bearings to failure. The
wide variation in bearing life is evident; there is no strong correlation between
bearing operating hours and bearing life.
USACERL TR 99/41 11

Figure 1. Bearing life scatter.

An improvement on calendar-based PM (and the traditional next step up) is con-

dition monitoring (CM). Here the maintenance supervisor/manager defines some
critical parameters that reflect the condition of a piece of machinery (e.g., am-
perage draw, bearing temperature, shaft vibration). Acceptable limits are de-
fined for each of these parameters. Some sort of data acquisition devices are at-
tached to measure the critical parameters and the stream of real-time data are
compared to the limits. Once the limits are exceeded, an alarm is issued so that
a PM visit can be scheduled.

The original focus of the CERL research project was to integrate condition moni-
toring equipment with the computerized maintenance management system
(CMMS) at MAMC. The intent was to collect real-time data and feed it into the
MAMC CMMS. However, once MAMC and USACERL personnel took a hard
look at what data could/should be collected and transferred to the MAMC
CMMC, issues arose that were broader than the technical issues of data acquisi-
tion:

1. Hardware and software technology exists to collect, store, and analyze gigabytes
of data. But does it make sense to replace all other maintenance activities with
data collection? (No, not necessarily.)

2. Is the path of data intensive maintenance management always the best? (No,
data collection, storage, and analysis systems cost money to install, operate, and
maintain. They are not always the most effective maintenance technique.)

3. Is there a downside to taking the man away from the machine, replacing the
regular human visits with electronic CM? (Yes, the human technician may see
something wrong that is not monitored by the CM in place.)
12 USACERL TR 99/41

Given questions (and answers) such as the above, the process of replacing PM
with CM is a complicated one. Any decision to replace calendar-based PM with
CM generates several organizational/policy-related questions regarding a facil-
itys maintenance program. Technical analysis of CM techniques quickly be-
comes overshadowed; workable solutions to the larger issues seem out of reach.

MAMC and CERL found a workable solution in the concept of Reliability Cen-
tered Maintenance (RCM) as developed by the National Aeronautics and Space
Administration (NASA). The RCM approach is a dynamic, ongoing effort, re-
quiring constant review today of the maintenance practices and policies put in
place yesterday. Its basic aim is to increase the reliability of machinery/systems
using a combination of four maintenance techniques: reactive maintenance, pre-
ventive maintenance, predictive maintenance, and proactive maintenance.

This manual is meant to serve as a guide to MEDCOM maintenance personnel

implementing an RCM program. As such, is not a formulaic cookbook that can
be followed mindlessly. A good RCM program requires that maintenance super-
visor/managers and staff be engaged in and constantly thinking about the value
of their present procedures. This manual will hopefully provide some basic di-
rection to the MEDCOM maintenance supervisors/in bringing experience to bear
on the specific facility entrusted to their care.

Objectives

The Facilities Management Division (FMD) at Madigan Army Medical Center

(MAMC) realizes the importance of constant and regular maintenance, and has
pursued and practiced an aggressive maintenance program since its inception.
Its maintenance program was planned and implemented as the facility was be-
ing built, and the maintenance program has been managed and documented
from the beginning using commercial computerized maintenance management
system (CMMS) software. Preventive maintenance is a keystone of the program
and accounts for about 60 percent of the maintenance dollars expended at
MAMC. The objective of this project was to help FMD determine if certain es-
sential, but rote, preventive maintenance efforts could be automated to free up
maintenance workers time.
USACERL TR 99/41 13

Approach

A literature search was done to uncover recent, relevant information in the area
of Reliability Centered Maintenance. A manual was prepared to serve as a guide
to MEDCOM maintenance personnel implementing an RCM program.

Scope

Although this work was done specifically at the request of MAMC, it is impor-
tant to note that RCM is a generally approach to equipment maintenance appli-
cable at many military, industrial, or commercial facilities. However, RCM is not
a formulaic cookbook methodology that can be followed mindlessly. A good
RCM program requires that maintenance supervisor/managers and staff be en-
gaged in and constantly thinking about the value of their procedures. This man-
ual is meant to provide some basic direction to maintenance supervisors in
bringing their experience to bear on the specific facilities entrusted to their care.

Units of Weight and Measure

U.S. standard units of measure are used throughout this report. A table of con-
version factors for Standard International (SI) units is provided below.

SI conversion factors
1 in. = 2.54 cm
1 ft = 0.305 m
1 yd = 0.9144 m
1 sq in. = 6.452 cm2
1 sq ft = 0.093 m2
1 sq yd = 0.836 m2
1 gal = 3.78 L
1 lb = 0.453 kg
1 oz = 28.35 kg
1 psi = 6.89 kPa
F = (C x 1.8) + 32
14 USACERL TR 99/41

2 RCM Definition and Philosophy

Definition

Reliability Centered Maintenance can be defined as an approach to mainte-

nance that combines reactive, preventive, predictive, and proactive maintenance
practices and strategies to maximize the life that a piece of equipment functions
in the required manner. RCM does this at minimal cost. In effect, RCM strives
to create the optimal mix of an intuitive approach and a rigorous statistical ap-
proach to deciding how to maintain facility equipment.

The key to developing an effective RCM program lies in effectively combining the
intuitive and statistical approaches. Intuition and statistics each have strong
and weak points. Intuition is an effective tool when applied judiciously; however,
if applied without serious reflection and review, it results in arbitrary, shoot-
from-the-hip solutions to problems. A rigorous statistical approach has its lim-
its, too. The first limit of the statistical approach is cost. Developing and/or
analyzing an amount of data sufficient to provide a statistical basis is an expen-
sive task. One may also fall into the analysis paralysis pitfall; the more one
delves into a problem the more data it seems is required to solve it. The second
limit of the statistical approach is applicability. Statistics often do not tell the
whole story. Data does not always produce definite trends, since there may be
none.

RCM Analysis

RCM analysis carefully considers the following questions:

What does the system or equipment do?

What functional failures are likely to occur?
What are the likely consequences of these functional failures?
What can be done to prevent these functional failures?

To implement RCM, it is imperative that the maintenance supervisor/manager

and maintenance technician think about their facilities in terms of function.
USACERL TR 99/41 15

That means thinking about facility equipment in terms of systems, subsystems,

components, and subcomponents. This terminology is used throughout this
manual.

RCM Principles

The primary RCM principles are:

1. RCM is Concerned with Maintaining System Functionality. RCM seeks to pre-

serve system or equipment function, not just to maintain a piece of machinerys
operability for operabilitys sake. It should be noted that a common strategy is to
maintain system function through equipment redundancy. Equipment redun-
dancy improves functional reliability but increases system life cycle cost (due to
the increased first cost of installing the redundant equipment). The increased life
cycle cost of installing redundant equipment often eliminates redundancy as the
RCM method of providing system reliability.

2. RCM is System Focused. It is more concerned with maintaining system function

than individual component function. The question asked continually is: Can
this system still provide its primary function if a component fails? (In this exam-
ple, if the answer is yes, then the component is allowed to run to failure.)

3. RCM is Reliability Centered. RCM treats failure statistics in an actuarial man-

ner. The relationship between operating age and failures experienced is impor-
tant. RCM is not overly concerned with simple failure rate; it seeks to know the
conditional probability of failure at specific ages (the probability that failure will
occur in each piece of equipment).

4. RCM Recognizes Design Limitations. The objective of RCM is to maintain the in-
herent reliability of system function. A maintenance program can only maintain
the level of reliability inherent in the system design; no amount of maintenance
can overcome poor design. This makes it imperative that maintenance knowl-
edge be fed back to designers to improve the next design. RCM recognizes that
there is a difference between perceived design life (what the designer thinks the
life of the system is) and actual design life. RCM explores this through the Age
Exploration (AE) process (see Section 3.4.6).

5. RCM is Driven by Safety First, then Economics. Safety must be maintained at

any cost; it always comes first in any maintenance task. Hence, the cost of main-
taining safe working conditions is not calculated as a cost of RCM. Once safety
on the job is ensured, RCM assigns costs to all other activities.
16 USACERL TR 99/41

6. RCM Defines Failure as an Unsatisfactory Condition. (Failure is not an option.)

Here failure is defined as a loss of acceptable product/service quality level, or
failure is defined as a function not being maintained.

7. RCM Tasks Must Produce a Tangible Result. The tasks performed must be shown
to reduce the number of failures, or at least to reduce the damage due to failure.

8. RCM Recognizes Four Maintenance Categories and Uses a Logic Tree to Screen
Maintenance Tasks. This ensures consistency in determining how to perform
maintenance on all types of facility equipment. Each piece of equipment is as-
signed to one of four categories:

a. Run-to-Failure - Under an RCM program, run-to-failure is a conscious deci-

sion reached after analysis of what facility function(s) would be affected by
system failure versus the (life cycle) cost of preventing failure.

b. Calendar-Based Maintenance (PM) - This is the most basic approach. It

schedules tasks based on the time since that task was last performed. It is
the type of maintenance most often performed in Preventive Maintenance
programs.

c. Condition Monitoring (CM) - This maintenance is performed based on predic-

tive testing and inspection. Real-time data is gathered and analyzed as a
way to determine when a piece of equipment requires maintenance.

d. Proactive Maintenance - Efforts in this area of a maintenance program are

aimed at applying the lessons learned from past maintenance experience to
future situations. This includes writing better specifications, precision re-
build, failed part analysis, and root-cause failure analysis. Figure 2 shows
the logic tree used to determine what kind of maintenance should be applied
to each piece of facility equipment.

9. RCM is an Ongoing Process. This is one of the most important characteristics of

RCM. No maintenance procedures escape review. Maintenance personnel
gather data from the successes/failures achieved and feed this data back to im-
prove future maintenance procedures and design of new systems. This feedback
loop is an essential part of the RCM process. This includes: changing old equip-
ment specifications that have been proven inadequate or incorrect, rebuilding
worn/failed equipment to better resist failure, performing failed-part analysis,
and performing root-cause failure analysis.
USACERL TR 99/41 17

Figure 2. RCM logic tree.

Will failure of the facility or equip-

NO YES

Is the item Can redesign solve the problem

expendable? permanently and cost effectively?

YES NO NO YES

Is there a Condition Monitoring (CM) technology Redesign

(e.g. vibration testing, thermography) that will give
sufficient warning (alert/alarm) of impending
NO YES

Is CM cost- and
priority-justified?

NO YES

Is there an effective PM task

that will minimize functional
NO YES

Is establishing redundancy cost-

and priority- justified?
NO YES

Accept Risk of Install Redundant Define PM task Define CM task

Failure Units and schedule and schedule
Run-to-Failure Proactive Maint. Preventive Maint. Predictive Maint.

Source: NASA Facilities RCM Guide, pp 2-3.

The RCM Process

RCM grew out of the aircraft industry in the late 1960s and 1970s. Since many
aircraft equipment failures have disastrous consequences, the basic RCM process
developed was very formal and rigorous. The basic steps in developing a formal
RCM analysis are:
18 USACERL TR 99/41

1. Define the major systems and components. The user defines the systems. Where
systems are extremely complex and this complexity makes analysis difficult, the
user may opt to define subsystems as a means of organizing the problem into
manageable pieces.

2. For each system, define all functions of that system.

3. For each of those functions, define the possible functional failures that could oc-
cur (i.e., what could go wrong that would prevent the system function from occur-
ring).

4. For each functional failure, define all possible failure modes (i.e., each equip-
ment failure could be the cause of the functional failure).

5. For each failure mode, state whether it would be due to improper operation, im-
proper maintenance, or both.

Figure 3 presents a sample RCM analysis sheet (RCM System Data Sheet) that
would be generated in applying a rigorous RCM analysis to a chilled water sys-
tem supplying computer equipment. Figure 3 analyzes the system and the func-
tions it performs. It also lists the functional failures that could occur. Figure 4
presents one of the sample Failure Mode Sheets that would be produced de-
scribing how one of the components of the chilled water system could fail; it lists
the failure modes of the component. Note that Figure 4 is a breakout of one of
the 12 distinct failure modes listed in Figure 3. In a formal and complete RCM
analysis, 11 other Failure Mode Sheets such as Figure 4 would be produced.
Figures 5 and 6 present the root cause failure analyses of two sub-components of
the component analyzed in Figure 4 (motor = component, while stator = sub-
component of motor, and rotor = sub-component of motor in this example).
Figures 5 and 6 represent two of nine sheets detailing how/why sub-components
of the motor would fail.

Review of the sample RCM Information Sheet, the sample component failure
mode sheet, and the two sample Root Cause Failure sheets illustrates how ex-
tensive, time-consuming, and expensive a formal RCM process can become. Due
to the extensive up-front effort involved in producing a formal RCM analysis, it
is recommended that MEDCOM facilities only pursue this level of detail for
those systems where the consequences of failure are catastrophic.
USACERL TR 99/41 19

Building Chilled Water System

Maintenance (M)
Function Functional Failures Failure Modes or Operation (O)
Provide chilled Total loss of flow Motor Failure Both
water at speci- Pump Failure Both
fied flow rate Catastrophic Leak M
and tempera-
Blocked Line M
ture
Valve out of position Both
Insufficient flow Pump cavitation O
Drive problem M
Blocked line M
Valve out of position Both
Instrumentation M
Chilled water tem- Chiller fatigue Both
perature too high Low refrigerant M
Fouled heat exchanger M
Instrumentation problem M
Cooling tower problem M
Valve out of position Both
Figure 3. Sample RCM system data sheet.

RCM Program Benefits

1. Reliability. The primary goal of RCM is to improve equipment reliability. This

improvement comes through constant reappraisal of the existing maintenance
program and improved communication between maintenance supervi-
sors/managers, maintenance mechanics, facility planners, building designers,
and equipment manufacturers. This improved communication creates a feed-
back loop from the maintenance mechanic in the field all the way to the equip-
ment manufacturers.

2. Cost. Due to the initial investment required to obtain the technological tools,
training, equipment condition baselines, a new RCM program typically results in
a short-term increase in maintenance costs (see Figure 7). The increase is rela-
tively short-lived. The cost of reactive maintenance decreases as failures are pre-
vented and preventive maintenance tasks are replaced by condition monitoring.
The net effect is a reduction of reactive maintenance and a reduction in total
maintenance costs. As a by-product, energy savings are often realized from the
use of the CM techniques that are part of any RCM program.
20 USACERL TR 99/41

Electric Motor # 123456

Function: To provide sufficient power to pump 300 gpm chilled water
Component Functional Failure Failure Mode Source of Failure
Insulation contamination
Insulation Failure Excessive current
Stator Motor will not turn Voltage spike

Open winding Phase imbalance

Excessive temperature
Insulation contamination
Motor will not turn Burnt rotor
Excessive current
Rotor
Excessive temperature
Wrong speed Excessive vibration
Imbalance
Fatigue
Improper lubrication
Misalignment
Imbalance
Bearings Motor will not turn Bearing seized
Electrical pitting
Contamination
Excessive Thrust
Excessive temperature
Mainline contact failure
Motor will not turn Bearing seized
Control circuit failure
Motor Controller
Loss of electrical power
Wrong speed VFD malfunction
Cabling failure
Excessive current
Overloads/fuse Motor will not turn Device burned out Excessive torque
Poor connection
Fatigue
Shaft/coupling Pump will not turn Shaft/coupling sheared Misalignment
Excessive torque
Figure 4. Sample failure mode sheet.

3. Scheduling. The ability of a condition monitoring program to forecast certain

maintenance activities provides time for planning, obtaining replacement parts,
making the necessary logistical arrangements (i.e., notifying occupants of equip-
ment downtime) before the maintenance is executed. CM reduces the unneces-
sary maintenance performed by a calendar-based preventive maintenance pro-
gram, which tends to err consistently on the safe side in determining time
intervals between maintenance tasks.

4. Equipment/Parts Replacement. A principal advantage of RCM is that it obtains

the maximum use from the equipment. With RCM, equipment replacement is
based on equipment condition, not on the calendar. This condition based ap-
proach to maintenance extends the life of the facility and its equipment.
USACERL TR 99/41 21

Root Cause of Failure Mode for Electric Motor Bearings

Failure Mode Mechanism Reason Root Cause
Bearing Seized Seal Failure
Contamination
(This includes seals, Cleanliness
shields, lubrication Oil Leak
system, and lock nut Lubrication Insufficient
Procedural
Excessive Procedural
Wrong Type Procedural
Inherent
Metallurgical
Excessive Temp.
Imbalance
Fatigue
Misalignment
Excessive Load
Fit-up
Application
Installation Procedural
Contamination See lubrication
Surface Distress Storage Procedural
Insulation
Electrical
Welding
Figure 5. Failure mode sheet for bearings.

Root Cause Failure Mode for Electric Motors (Electrical)

Age Inherent
Oxidation
Environment Chemical attack
Power quality
Phase imbalance
Overheating Excessive current Short on/off cycle
Low voltage
Overloaded
Stator Insulation resistance Moisture
reading zero ohms
Contamination Environment Improper lube
Process related
Lack of winding
support
Phase imbalance
Fatigue Excessive vibration
Imbalance
Misalignment
Resonance
Figure 6. Failure mode sheet for stator.
22 USACERL TR 99/41

Figure 7. Maintenance cost trends under an RCM program.

5. Efficiency/Productivity. Safety is the primary concern of RCM. The second most

important concern is cost-effectiveness. Cost-effectiveness takes into considera-
tion the priority or mission criticality and then matches a level of cost appropri-
ate to that priority. The flexibility of the RCM approach to maintenance ensures
that the proper type of maintenance is performed when it is needed. Mainte-
nance that is not cost-effective is identified and not performed.

In summary, the multi-faceted RCM approach promotes the most efficient use of
resources. The equipment is maintained as required by its characteristics and
the consequences of its failures.

Impact of RCM on a Facilitys Life Cycle

RCM must be a consideration throughout the life cycle of a facility if it is to

achieve maximum effectiveness. The four recognized major phases of a facilitys
life cycle are:

Planning
Design
Construction

Operations and Maintenance.

USACERL TR 99/41 23

Figure 8 shows that planning (including conceptual design) fixes 2/3 (66.7 per-
cent) of a facilitys life cycle cost. The subsequent design phases fix about an-
other 30 percent of the life-cycle cost, leaving only about 4 percent fixable in the
later phases. Thus, the decision to institute RCM at a facility, including condi-
tion monitoring, will have a major impact on the life-cycle cost of that RCM pro-
gram. This decision is best made during the planning phase. As RCM decisions
are made later in the life cycle, it becomes more difficult to achieve the maxi-
mum possible benefit from the RCM program.

Murphys Law being what it is, it is rare that a complete and well-planned RCM
program is instituted at the planning stage of a project. However, maintenance
personnel need not despair. Even though maintenance is a relatively small por-
tion of the overall life-cycle cost, a balanced RCM program is still capable of
achieving savings of 30 to 50 percent in a facilitys annual maintenance budget.
While these operations and maintenance (O&M) savings may not be the majority
of the facilitys life cycle cost, they are still a significant portion of the yearly op-
erating costs of a facility, and would be well appreciated by any fiscal manager
looking to cut operating costs.

Figure 8. Stages of life cycle cost commitment.

24 USACERL TR 99/41

3 RCM Program Components

An RCM program implements reactive maintenance, preventive maintenance,
condition monitoring, and proactive maintenance in an optimal mix. It also
combines the intuitive approach with the statistical approach in determining
equipment condition.

Reactive Maintenance

Reactive maintenance is referred to by many different names: breakdown main-

tenance, repair, fix-when-fail, and run-to-failure (RTF) maintenance. When ap-
plying this maintenance strategy, a piece of equipment receives maintenance
(e.g., repair or replacement) only when the deterioration of the equipments con-
dition causes a functional failure. The strategy of reactive maintenance assumes
that failure is equally likely to occur in any part, component, or system. Thus,
this assumption precludes identifying a specific group of repair parts as being
more necessary or desirable than others.

The major downside of reactive maintenance is unexpected and unscheduled

equipment downtime. If a piece of equipment fails and repair parts are not
available, delays ensue while the parts are ordered and delivered. If these parts
are urgently required, a premium for expedited delivery must be paid. If the
failed part is no longer manufactured or stocked, more drastic and expensive ac-
tions are required to restore equipment function. Cannibalization of like equip-
ment or rapid prototyping technology may satisfy a temporary need but at sub-
stantial cost. Also, there is no ability to influence when failures occur because no
(or minimal) action is taken to control or prevent them. When this is the sole
type of maintenance practiced, both labor and materials are used inefficiently.
Labor resources are thrown at whatever breakdown is most pressing. In the
event that several breakdowns occur simultaneously, it is necessary to practice a
kind of maintenance triage in an attempt to bring all the breakdowns under con-
trol. Maintenance labor is used to stabilize (but not necessarily fix) the most
urgent repair situation, then it is moved on to the next most urgent situation,
etc. Replacement parts must be constantly stocked at high levels, since their use
cannot be anticipated. This incurs high carrying charges and is not an efficient
way to run a storeroom.
USACERL TR 99/41 25

A purely reactive maintenance program ignores the many opportunities to influ-

ence equipment survivability. However, it can be effective if used selectively and
performed as a conscious decision based on the results of an RCM analysis. This
RCM analysis compares the risk of failure with the cost of the maintenance re-
quired to mitigate that risk and cost of failure (again, refer to Figure 2, RCM
Logic Tree).

Examples of equipment that may be reactively maintained (run-to-failure) are:

non-critical electrical motors less than 7.5 HP, restroom exhaust fans, water
heaters serving bathrooms, lamps in areas where a few burned out lamps will
not pose any safety hazard or affect the use of the area (e.g., halls, cafeterias,
lounges).

Preventive Maintenance (PM)

PM consists of regularly scheduled inspection, adjustments, cleaning, lubrica-

tion, parts replacement, calibration, and repair of components and equipment.
PM is also referred to as time-driven or calendar-based maintenance. It is per-
formed without regard to equipment condition or (possibly) degree of use.

PM schedules periodic inspection and maintenance at pre-defined intervals (in-

tervals based on time, operating hours, or cycles) in an attempt to reduce equip-
ment failures for susceptible equipment. Depending on the intervals set, PM can
result in a significant increase in inspections and routine maintenance; however,
it should also reduce the seriousness and frequency of unplanned machine fail-
ures for components with defined, age-related wear patterns.

Traditional PM is keyed to failure rates and times between failures. It assumes

that these variables can be determined statistically, and that one can therefore
replace a part that is due for failure shortly before it fails. The availability of
statistical failure information tends to lead to fixed schedules for the overhaul of
equipment or the replacement of parts subject to wear. PM is based on the as-
sumption that the overhaul of machinery by disassembly and replacement of
parts restores the machine to a like-new condition with no harmful side effects.
In addition, this renewal task is based on the perception that new components
are less likely to fail than old components of the same design.

Failure rate or its reciprocal, Mean-Time-Between-Failure (MTBF), is often used

as a guide to establishing the interval at which the maintenance tasks should be
performed. The major weakness in using these measurements to establish task
periodicities is that failure rate data determines only the average failure rate.
26 USACERL TR 99/41

The reality is that failures are equally likely to occur at random times and with a
frequency unrelated to the average failure rate. Thus, selecting a specific time to
conduct periodic maintenance for a component with a random failure pattern is
difficult at best.

As stated in Section 2.4, RCM grew out of the aircraft industry in the late 1960s
and early1970s. This early RCM approach is documented in Reliability Centered
Maintenance (Nowlan and Heap 1978), which demonstrated that a strong corre-
lation between age and failure did not exist and that the basic premise of time-
based maintenance was false for the majority of equipment.

In summary, PM can be costly and ineffective when it is the sole type of mainte-
nance practiced.

Preventive Maintenance Criteria

Preventive maintenance criteria should reflect the age-reliability characteristics

of the equipment based on the equipment history. Equipment that has under-
gone PM should show a strong failure rate versus age correlation. There should
be definite data (or at least definite conclusive in-house experience) to bear out
that a piece of equipment will fail at a certain age.

However, whether a piece of equipment should receive PM is not necessarily a

function of that equipments mission criticality. In selecting equipment that
should receive PM, the maintenance supervisor/manager should use the process
shown in Figure 2. The selection process guides maintenance supervi-
sor/manager in selecting the maintenance strategy appropriate to each piece of
equipment.

Determining PM Task and Monitoring Periodicity

This section offers suggestions for selecting equipment monitoring periodicities,

i.e., determining the time intervals between PM visits.

Although many ways have been proposed for determining the correct frequency
of preventive maintenance tasks, none are valid unless the in-service age-
reliability (i.e., failure rate versus age) characteristics of the system or are
known. This information is not normally available and must always be collected
for new systems and equipment. Condition monitoring techniques (e.g., taking
real-time data to determine the health of a piece of machinery) can be used to
help determine equipment condition vs. age.
USACERL TR 99/41 27

Careful analysis of similar kinds of hardware in industry has shown that, over-
all, more than 90 percent of the hardware analyzed showed no adverse age-
reliability relationship. This does not mean that individual parts do not wear;
they do. It means that the ages at failure are distributed in such a way that
there is no value in imposing a preventive maintenance task. In fact, in a some
cases, imposing an arbitrary preventive task increases the average failure rate
because some PM tasks are actually detrimental to machines (e.g., machine dis-
assembly, overgreasing). Of course, one would hope that detrimental tasks are
never intentionally assigned, but overzealous PM technicians have sometimes
done this.

The Mean Time Between Failures (MTBF) is often used as the initial basis for
determining PM interval. This approach is incorrect in that it does not provide
any information about the effect of increasing age on reliability. It provides only
the average age (for a group of components) at which failure occurs, not the most
likely age (for a specific component). In many cases a Weibull distribution, as
used by the bearing industry to specify bearing life, will provide more accurate
information on the distribution of failures.

If good information on the effect of age on reliability is lacking, the best thing
that can be done is to monitor the equipment condition (condition monitoring).
This is explained in the next section.

The goals of a PM visit to a piece of equipment are: (1) to determine equipment

condition, and (2) to develop a trend to forecast future equipment condition. The
following techniques are recommended for setting initial periodicity:

1. Anticipating Failure from Experience. For some equipment, failure history and
personal experience provides an intuitive feel for when to expect equipment fail-
ure. In these cases, failure is time related. Set monitoring so that there are at
least three monitoring PM visits before the anticipated onset of failures. These
three visits will give the maintenance technician enough of a look at the piece of
equipment to become familiar with it. In most cases it is prudent to shorten the
monitoring interval as the wear-out age is approached:

2. Failure Distribution Statistics. In using statistics to determine the basis for se-
lecting periodicities, the distribution and probability of failure should be known.
Weibull distributions can provide information on the probability of an equipment
exceeding some life. For example, bearings are normally specified by their B 10
life; i.e., the number of revolutions that will be exceeded by 90 percent of the
bearings. Depending on the criticality of the equipment, an initial periodicity is
recommended that allows a minimum of three monitoring samples prior to the
28 USACERL TR 99/41

B10 life or, in less severe cases, prior to the MTBF point. In more critical cases, a
B2 life can be calculated and the monitoring interval can be adjusted accordingly.

3. Lack of Information or Conservative Approach. The most common practice in

the industry is to monitor the equipment biweekly or monthly due to lack of in-
formation and poor monitoring techniques. This often results in excessive moni-
toring. In these cases, significant increases in the monitoring interval may be
made without adverse impacting equipment reliability.

When indications of impending failure become apparent through trending or

other predictive analysis methods, the monitoring interval should be reduced
and additional analysis should be done to gain more detailed information on the
condition of the equipment.

Condition Monitoring (CM)

Condition monitoring, also known as predictive maintenance, uses primarily

nonintrusive testing techniques, visual inspection, and performance data to as-
sess machinery condition. It replaces arbitrarily timed maintenance tasks with
maintenance scheduled only when warranted by equipment condition. Continu-
ing analysis of equipment condition monitoring data allows planning and sched-
uling of maintenance or repairs in advance of catastrophic and functional failure.

The CM data collected is used in one of the following ways to determine the con-
dition of the equipment and to identify the precursors of failure:
Trend Analysis. Reviewing data to see if a machine is on an obvious and im-
mediate downward slide toward failure.
Pattern Recognition. Looking at the data and realizing the causal relation-
ship between certain events and machine failure. For example, noticing that
after machine x is used in a certain production run, component ax fails due to
stresses unique to that run.
Tests against Limits and Ranges. Setting alarm limits (based on professional
intuition) and seeing if they are exceeded.
Statistical Process Analysis. If published failure data on a certain ma-
chine/component exists, comparing failure data collected on site with the
published data to verify/disprove that you can use that published data.
USACERL TR 99/41 29

For trending purposes, a minimum of three monitoring points before failure may
reasonably be expected are recommended. Three data points allow one to deter-
mine whether equipment condition depreciates linearly.

CM does not lend itself to all types of equipment or possible failure modes and
therefore should not be the sole type of maintenance practiced. Chapters 5-10
give information on specific CM technologies and instrumentation. For example,
to obtain the total picture of a chilled watered system, a CM effort would have to
collect the following data (one can see how extensive - and costly - this approach
could become):

1. Flow Rates. Chiller water flow would be measured using precision, nonintrusive
flow detectors.

2. Temperature. Differential temperature would be measured to determine heat

transfer coefficients and to indicate possible tube fouling.

3. Pressure. Differential pressures across the pump would be measured to deter-

mine pump performance, and differential pressures across the chiller evaporator
and condenser sections should be measured to determine the condition of the
chiller tubes (i.e., whether they were fouling).

4. Electrical. Motor power consumption would be used to assess the condition of the
motor windings.

5. Ultrasonic Testing. Pipe wall thickness would be measured to determine erosion

and corrosion degradation.

6. Vibration. Vibration monitoring would be used to assess the condition of rotating

equipment (such as pumps and motors). Additionally, structural problems can be
identified through resonance and model testing.

7. Lubricant Analysis. Oil condition and wear particle analysis would be used to
identify problems with the lubricant, and to correlate those problems with vibra-
tion when wear particle concentrations exceed pre-established limits.

8. Fiber Optics. Fiber optic inspections would be used to determine component

wear, tube fouling, etc.

9. Thermography. Thermography scans check motor control centers and electrical

distribution junction boxes for high temperature conditions. High temperature is
indicative of loose connections, shorts, or failing conductor insulation. Piping in-
sulation should be checked for porosities. Here, high temperatures are indica-
tive of failed/failing areas in the pipe insulation.
30 USACERL TR 99/41

10. Eddy Current. Eddy current testing is used to determine and locate leaking
tubes.

11. Airborne Ultrasonics. Airborne ultrasonics indicate air leaking from control sys-
tem piping and compressors.

Proactive Maintenance

A proactive maintenance program is the capstone of RCM philosophy. It provides

a logical culmination to the other types of maintenance described above (reac-
tive, preventive, and predictive). Proactive maintenance improves maintenance
through better design, installation, maintenance procedures, workmanship, and
scheduling.

Proactive maintenance is characterized by the following attitudes:

Maintaining a feedback loop from maintenance technicians to building archi-
tects, engineers, and designers, in an attempt to ensure that design mistakes
made in the past are not repeated in future designs.
Viewing maintenance and supporting functions from a life-cycle perspective.
This perspective will often show that cutting maintenance activities to save
money in the short term often costs more money in the long term.
Constantly re-evaluating established maintenance procedures in an effort to
improve them and ensure that they are being applied in the proper mix.

Proactive maintenance uses the following basic techniques to extend machinery

life:
proper installation and precision rebuild
failed-parts analysis
root-cause failure analysis

reliability engineering
rebuild certification/verification
age exploration
recurrence control.

These proactive maintenance strategies are explained in the following sections.

USACERL TR 99/41 31

Specifications for New/Rebuilt Equipment

Equipment requires proper installation to control life cycle costs and manmade
reliability. Poor installation often results in problems routinely faced by both
maintenance personnel and operators. Rotor balance and alignment, two com-
mon rework items, are often poorly performed or neglected during initial instal-
lation. Adopting and enforcing of precision standards can more than double the
life of a machine. For example, the contract specification for leveling equipment
being installed should include a maximum acceptable slope of the base and the
frame; e.g., a maximum slope of 0.001 in./ft. The specification also should include
the type and accuracy of the instrument used for measuring the slope; e.g., a 12-
in. machinists level graduated to 0.0002 in./ft. After the criteria have been in-
cluded in the contract specifications, the installation should be checked to ensure
that the mechanic has complied with the specification.

Equipment is often procured using inadequate specifications. Existing stan-

dards, often 25 to 30 years old, do not reflect current changes in building tech-
nology, but usually address only general or minimal performance criteria. Addi-
tionally, the life cycle costs and failure histories of families of equipment are
rarely documented for purchasing and contract personnel who (by regulation)
must procure conforming products solely based on initial least cost.

To solve this problem, design engineers must write proper specifications, re-
search (and test if possible) the equipment of different vendors, and document
problems. These specifications should include, as a minimum, vibration, align-
ment, and balancing criteria. If these criteria are included in the plans and
specifications for new construction or major building renovations, they become
part of the contractual documents that the contractor must fulfill. This gives the
building owner some solid recourse to make the contractor prove that the equip-
ment in the building is operating properly and not just that it is operating. At
the turnover of the facility to the owner/tenant, if the contractor does not fulfill
the requirements listed in these vibration, alignment, and balancing criteria
listed in the specifications, then the owner can refuse acceptance of the building
until the contractor does so. For example, rotating equipment has some sort of
shaft(s) and bearings. Vibration analysis can determine if the shafts are aligned
and if the bearings have been damaged in shipment or installation. Finding
alignment and bearing problems before turnover is preferable to the more usual
scenario, in which a certain number of bearings fail after turnover well before
any normal wear or operation would have caused them to fail.

Local companies that specialize in vibration measurement (and work with vibra-
tion daily) and shaft alignment should be consulted to assist in writing the sec-
32 USACERL TR 99/41

tions of the contract specifications that deal with vibration/alignment. That way,
the designer can be sure that all pertinent vibration/alignment criteria are in-
cluded.

Performance testing is another proactive maintenance strategy that must be

conducted. Performance testing can occur in several places: (1) in the factory
prior to shipment, (2) after the equipment is installed and immediately prior to
acceptance, and (3) at the beginning of the daily operation of the equipment, to
establish a performance baseline as the equipment begins operation.

Balance

Bearings are the machine components that support and transfer the forces from
the rotating element to the machine frame. The fact that only 10 to 20 percent of
rolling element bearings achieve their design life results in the perception that
bearings inherently pose a reliability problem. One of the leading causes of pre-
mature rolling element/bearing failure is parasitic load due to excessive forces
imposed by imbalance and misalignment. Parasitic loads result in increased dy-
namic loads on the bearings. The design formulas (SKF, 1973) used to calculate
theoretical rolling element/bearing life are:

a. for ball Bearings:

3
L10 Life Hours =(16,667/RPM) x (C/P)

b. for roller Bearings:

10/9
L10 Life Hours = (16,667/RPM) x (C/P)

Where:

L10 = the number of hours 90 percent of a group of bearings should attain or

exceeded under a constant load (P) prior to fatigue failure

C= the bearing load that will result in a life of 1 million revolutions

P= the actual bearing load, static and dynamic.

As shown, bearing life is inversely proportional to speed and more significantly,

inversely proportional to the third power of load for ball and to the 10/9 power
for roller bearings.
USACERL TR 99/41 33

Balance Calculations

Precision balance of motor rotors, pump impellers, and fans is one of the most
critical and cost effective techniques for achieving increased bearing life and re-
sultant equipment reliability. It is not usually sufficient to perform a single
plane balance of a rotor to a level of 0.10 in/sec, nor is it sufficient to balance a
rotor until it achieves seemingly low vibration levels. When analyzing a piece of
rotating equipment, the vibration technician should take readings in the vertical
and horizontal planes, and at all bearing points. He should also take axial
readings. Precision balance methods should also include the calculation of re-
sidual imbalance.* Residual imbalance calculates the imbalance left in the
piece of equipment once the balancing procedure has been executed. Nothing is
perfect. A piece of rotating equipment will never be balanced so that all of the
mass is distributed evenly around the axis of rotation. As the piece of equipment
rotates the additional force imposed by this extra piece of mass will be slung
around the axis.

The following equation can be used to calculate residual imbalance:

Vr
U M Eq. 1
Ve

Where:

U = amount of residual imbalance

Vr = actual imbalance

Ve = trial mass imbalance

M = trial mass.

This equation can be expressed as:

(Trial Weight) (Trial Weight Radius) (Amplitude After Balance)
Residual Imbalance =
Trial Weight Effect

* Note: The following equations and discussion of permissible imbalance is based on ISO 1940/1, Mechanical Vi-
bration-Balance Quality Required of Rigid Motors (1986).
34 USACERL TR 99/41

Permissible imbalance is related to equipment type and rotor mass. In general,

the greater the rotor mass, the greater the permissible imbalance.

Effect of Imbalance

As discussed earlier, imbalance forces make a major contribution to decreased

bearing life. For example, consider a rotor turning at 3600 RPM with 1 oz. of un-
balance on a 12-in. radius.

Calculating the amount of centrifugal force due to imbalance:

mr ( 2 f )2
F = mA = mr 2 = = 0.102 mrf 2
g

Where:
F = Force
m = imbalance (lb)
A = acceleration (in.2/sec)
r = radius of imbalance (in.)
= rotational velocity (radians/sec)
f = rotational frequency (Hz)
g = 386.4 in./sec.

Substituting 1 oz. (1/16 lb.), 12 in., 3600 RPM (60 Hz) yields:
F = 0.102 (1 16 ) (12 ) (60 ) = 275 lb
2

Thus, 1 oz. of imbalance on a 12-in. radius at 3600 RPM creates an effective cen-
trifugal force of 275 lb. Now calculate the effect of this weight on bearing life.
Suppose that the bearings were designed to support a 1000-lb. rotor. The calcu-
lated bearing life is less than 50 percent of the design life:
Actual L10 Life = (Design Life ) [(1000 1000 + 275)]3
= 0.48 Design L10 Life

Alignment

The forces of vibration from misalignment also cause gradual deterioration of

seals, couplings, drive windings, and other rotating elements with close toler-
ances. Maintenance technicians should use precision equipment and alignment
USACERL TR 99/41 35

methods, (e.g., reverse dial or laser alignment system) to bring alignment toler-
ances within precision standards. Contrary to popular belief, both laser align-
ment and reverse dial indicator equipment offer equal levels of precision; how-
ever, laser alignment is considerably easier to learn and much easier to execute
in the field.

In addition to the alignment specifications, Table 1 contains the additional toler-

ance recommendations.

Alignment Effects

Based on data from a petrochemical industry survey, precision alignment prac-

tices achieve:
average bearing life increases by a factor of 8.0
maintenance costs decrease by 7 percent
machinery availability increases by 12 percent.

Table 1. Recommended coupled alignment tolerances (General Motors, 1993).

Tolerance
Maximum Speed Horizontal & Vertical Angularity (Inch/10
Coupling Type (RPM) Parallel Offset (IN). inch of Coupling Dia.)
Short Coupling 600 0.005 0.010
900 0.0053 0.007
1200 0.0025 0.005
1800 0.002 0.003
3600 0.001 0.002
7200 0.0005 0.001
Coupling with 600 0.005 N/A
Spacer (Meas- 900 0.0018 N/A
urement is per
1200 0.0012 N/A
inch of spacer
length) 1800 0.0009 N/A
3600 0.0006 N/A
7200 0.00015 N/A
Source: NASA RCM Guide, page 3-14.
36 USACERL TR 99/41

Figure 9 presents the effect of misalignment on bearing life of a cylindrical roller

bearing and shows the drastic decrease in component life caused by a few min-
utes (i.e., micro-degrees) misalignment.

Misalignment can also cause a significant increase in the cost of energy con-
sumption. Consider the case below. After re-alignment, the current draw of 460V
motor decreased from 25 to 23 Amp. Assume the motor has a power factor of
0.90, given in the relation:

kW =
(3)1 2 ( AV PF )
1000
Where:
kW = Change in power consumption, kilowatts
A = Change in Amperage draw
V = Rated Voltage
PF = Power Factor.

Figure 9. Decrease in life of cylindrical roller bearings as a function of

misalignment.
USACERL TR 99/41 37

Substituting the values given above:

1.732 (2 Amps )(460 V )(0.90 )
kW =
1000
= 1.43kW

Assuming the motor runs an average of 7,000 hours per year, and the average
cost of power during the year is $0.07/kW-hour:
$ saved per machine train = 1.43 (0.07 )(7000 ) = $700.00/ye ar

Failed-Part Analysis

This proactive process involves visually inspecting failed parts after removal to
identify the cause(s) of their failure. More detailed technical analysis may be
conducted when necessary to determine the root cause of a failure.

Bearings are generally the weakest equipment components since they are sub-
jected to constant wear. Only 10 to 20 percent of bearings achieve their design
life. The root causes of bearing failures may relate to improper installation, poor
lubrication practices, excessive balance and alignment tolerances, or poor stor-
age and handling techniques. Failed-bearing analysis provides methods to cate-
gorize defects such as scoring, color, fretting, and pitting and to relate these
findings to the most probable cause of failure.

Over half of all bearing problems result from contamination or improper instal-
lation. While indicators of contamination normally appear on the internal race
surfaces of bearings, indicators of installation problems generally are evident on
both internal and external race surfaces.

Root-Cause Failure Analysis (RCFA)

In some cases, plant equipment fails repeatedly, and the failures are accepted as
a normal idiosyncrasy of that equipment. Recurring problems such as short
bearing life, frequent seal fracture, and structural cracking are symptoms of
more severe problems. Without performing RCFA, maintenance personnel only
see the symptomatic problems; they can do nothing other than continue with the
frequent repairs. Repeated failures result in high costs for parts and labor and
in decreased customer goodwill and mission support reliability. Further, unreli-
able equipment may pose a personnel safety hazard.

CM methods can identify most equipment faults at an early enough stage to pre-
vent equipment failure. However, CM methods are not directed at discovering
38 USACERL TR 99/41

the underlying reason for the failures. For example, a bearing may fail repeat-
edly because of excessive bearing loads caused by an underlying misalignment
problem. CM would most likely predict a bearing failure and thus allow the
bearing to be replaced before it fails; however, if no one recognizes the misalign-
ment and corrects it, then conditions causing the failure will remain. The fail-
ures will recur and continue to require unnecessary corrective work and cause
undue downtime.

RCFA proactively seeks the fundamental causes that lead to facility and equip-
ment failure. Its goals are to:
find the cause of a problem quickly, efficiently, and economically
correct the cause of the problem, not just its effect
provide information that can help prevent the problem from recurring
instill a mentality of fix forever.

Reliability Engineering and Reliability Calculations

In combination with other proactive techniques, reliability engineering involves

the redesign, modification, or improvement of components or their replace-
ment by superior components. Sometimes a complete redesign of the component
is required. In other cases, upgrading the component metal or adding a sealant
is all that is required.

Mean Time Between Failure

Reliability can be expressed by the following reliability function:

R (t ) = 1 F (t )

where:

F(t) = the probability the system will fail by time t, and

0 < F(t) < 1

F(t) is basically the failure distribution function, or the unreliability function.

If the random variable t has a density function of f(t), then the expression for re-
liability is:
R (t ) = 1 F (t ) = f (t )dt
USACERL TR 99/41 39

Assuming that the time to failure is described by an exponential density func-

tion, then where:

L = the mean life

(
f (x ) = 1/L e t L )
t = the time period of interest
e = the natural logarithm base (2.7183),

the reliability at time t is:

( )
R (t ) = 1/L e t L = e t L

Mean life (L) is the arithmetic average of the lifetimes of all items considered.
The mean life (L) for the exponential function is equivalent to mean time be-
tween failures (MTBF). Thus,

where:
R (t ) = e t M = e F t

F = the instantaneous failure rate

M = the MTBF.

Failure Rate

The rate at which failures occur in a specified time interval is called the failure
rate during that interval. The failure rate (Fr) is expressed as:

In the following example, as a system component fails, it is replaced and data is

taken to record when the replacements occurred:

Fr =
(No. of
Failures )
Total Operating Hours

Component # Failed at Time (Operating Hours)

1 75
2 125
3 130
4 325
5 525
Total Operating Hours 1180
40 USACERL TR 99/41

Fr = 5/1180 = 0.0042

Assuming an exponential distribution, the system mean life or MTBF is:

MTBF = 1/Fr = 1/0.0042 = 236 hours

Reliability Component Relationships

1. Series networks
Reliability R = (R A )(RB )(RC )

If a series configuration is expected to operate for a specified time period, its

overall reliability can be derived by the following expression:
RS = e nt

Where n = the number of components in the series.

2. Parallel networks

Reliability R = RA + RB - (RA x RB) 2-component network

Reliability R = 1 - (1-RA)(1-RB)(1-RC) 3-component network

n
Reliability R = 1 - (1-R) n identical components

3. Series-parallel networks

Network reliability for systems that contain both series and parallel components
is calculated similarly to calculating overall electrical resistance for a network
that has resistors in parallel and series arrangements. As with electrical cir-
cuits, analyze the parallel portions to create an equivalent resistance (reliabil-
ity), and then complete the analysis by combining all components and equiva-
lents serially.

Related Reliability Factors

1. Inherent Availability (Ai)

USACERL TR 99/41 41

a. The probability that a system (or piece of equipment) when used un-
der stated conditions in an ideal support environment, will operate
satisfactorily at any point in time as required.

b. Excludes preventive or scheduled maintenance time and administra-

tive delay time.

2. Achieved Availability (Aa)

a. The probability that a system or equipment, when used under stated

conditions in an ideal support environment, will operate satisfac-
torily at any point in time.

b. Preventive (scheduled) maintenance is included.

c. Excludes logistics delay time and administrative delay time.

3. Operational Availability (Ao)

a. The probability that a system or equipment, when used under stated

conditions in an actual operational environment, will operate sat-
isfactorily when called on.

b. Includes active maintenance time, logistics delay time, and adminis-

trative delay time.

Weibull Distributions

The Weibull distribution is generally used to determine probability of a failure

due to fatigue. The original work was conducted in 1933 and was entitled A Sta-
tistical Theory of the Strength of Materials. The original work was not directly
related to bearings; however, it was modified in 1947 by Lundberg and Palingren
to account for the effect of the lubricant and the fact that not all cracks propa-
gate to the bearing surface (Lundberg 1947).

The importance of the Weibull distribution is that the fatigue behavior of a group
of bearings can be assessed, and that changes in the failure distribution could be
used to identify the introduction of new sources of failure, i.e., changes in oper-
ating condition, lubrication/installation practices, etc.
42 USACERL TR 99/41

The Weibull distribution is:

F(t ) = 1 e (t T )**k

where:
F(t) = the failure probability
T = the point in time at which 63.2 percent of the bearings have failed
k = corresponds to the gradient.

The value of k for bearings is 10/9 for ball bearings and 27/20 for roller bearings.
Given these values, F(t) should fall in the range from 0.07 to 0.60.

To use the Weibull distribution to determine failure probability for bearings, it is

necessary to have a minimum of 10 identical bearings operating under condi-
tions as nearly identical as possible.

Rebuild Certification/Verification

When new or rebuilt equipment is installed, it is essential to verify that it is op-

erating properly. To avoid unsatisfactory operation and early failure, the equip-
ment should be tested against formal certification and verification standards.

Age Exploration

Age Exploration (AE) is a key element in establishing an RCM program. AE

provides a methodology to vary key aspects of the maintenance program in an
attempt to optimize the maintenance process. For example, suppose a vendor
recommends that maintenance technicians open and inspect of a chiller at cer-
tain intervals. During the open and inspect, the technician notes the condition
of various components of the chiller. Ideally, the condition evaluation sheet com-
pleted by the technician is then correlated with performance data from the facil-
ity energy management control system (EMCS), vibration data, and oil analysis
data. As a result of this analysis, the decision may be made to change the inter-
val of the open and inspect until monitored conditions indicate a degradation has
occurred.

However, if there is no data available from the EMCS, the maintenance crews
may perform a simpler form of AE. Using their best professional judgment, they
increase the time period until the next open and inspect. If the chiller shows
no adverse signs of wear at the next opening, the time period until the next
open and inspect is increased some more. Plainly, this type of AE relies solely
USACERL TR 99/41 43

on the experience of the attending maintenance technician(s) to increase the

time intervals between regular heavy maintenance. This simpler form of AE
uses gut feeling to push the envelope of the machine and must be practiced
judiciously. (It probably also warrants assigning the same technician to the de-
vice consistently.)

No matter which AE method is used correlation of machinery condition with

performance data or gut feeling it is essential that data be available to
gauge the success of the AE effort. These data may be as simple as a listing of
the dates the PM inspections were performed and a verbal record of what each
inspection found. To proceed with AE and to take no data is a contradiction; by
definition, the process of AE lengthens the time intervals between regular PM
inspections, and subsequently records the effects of lengthening those time in-
tervals.

Recurrence Control

This section provides a systematic means for dealing with repetitive failures.
Repetitive failures are defined as the recurring inability of a system, subsystem,
structure, or component to perform the required function, i.e.:

repeated failure of an individual piece of equipment,

repeated failures of various equipment within a system or subsystem, or
failures of the same or similar components in several different systems.

Repetitive failures may become evident when reviewing data from these sources:

number of maintenance man-hours by system or component

number of maintenance work orders by system or component
maintenance backlog (by system)
control instruments out of service
number of inadvertent safety system actuations
number of unplanned facility shutdowns
lost mission or production time
forced outage reports
number of accident reports.

The following process for conducting an analysis of repetitive failures is pro-

vided:
44 USACERL TR 99/41

Monitor plant or equipment performance

Identify repetitive system failures

Establish priorities for solution and allocation of resources

Assign problems for analysis
Analyze problems/determine root cause

Recommend corrective action

Select corrective action
Implement selected corrective actions

Evaluate results of implemented corrective actions.

Facility Condition Assessment

A well-rounded and proactive maintenance program incorporates a constant pro-

cess of Facilities Condition Assessment (FCA). FCA is characterized as follows:
FCA is a continuous process wherein the in-house staff analyzes the effec-
tiveness of the maintenance program; it is not a series of discrete independ-
ent events, studies done by outside consultants, or flavor of the month
buzzwords.
FCA is data intensive. In addition to making qualitative judgments based on
experience, the maintenance supervisor/manager must have at his ready
maintenance data and the computer tools (hardware and software) to analyze
that data.

The first characteristic (goal, really) can only be achieved if the facility com-
mander realizes that a good maintenance program is a crucial element to the
successful operation of the facility, and if the maintenance supervisor/manager
instills in his personnel the attitude that their job is important and worth doing
right.

The second characteristic requires that the maintenance supervisor/manager has

sufficient data available to track maintenance labor and material costs, and has
the database tools to slice and dice that data as he finds necessary. He must
have a software tool that is easy to use yet powerful enough to generate the
graphs and charts required, a tool simple enough to use that the maintenance
technician can enter data on a daily basis.
USACERL TR 99/41 45

4 Use of Condition Monitoring (CM)

Technologies

Introduction

A variety of methods are available to assess the condition of systems/equipment,

to determine the most effective time to schedule maintenance:
vibration monitoring and analysis
thermography
passive ultrasonics
lubricant and particle wear analysis (oil analysis)
electrical condition monitoring
nondestructive testing.

Chapters 5 to 10 describe each of the CM technologies, covering the basic theory

of how the technology operates, the purpose of applying the technology, accept-
able applications, the equipment and operators required, available training, and
cost. Note that no attempt is made to provide a current, up-to-date listing of
commercial CM products available; product models and prices change too
quickly. The intent is to provide the reader with a basic understanding of the
available CM technologies so that he can make informed decisions about which
to employ at his facilities. In reviewing the various technologies, the mainte-
nance manager/supervisor should keep in mind that a well-rounded preventive
maintenance program is not built on any one technology.

Spot Readings versus Continual Real-Time Data Collection

Condition monitoring can be performed in two basic ways:

1. One can take spot readings with hand-held instruments.

2. One can install permanent data acquisition equipment and take repeated read-
ings that are stored for future analysis.
46 USACERL TR 99/41

Generally, taking spot readings provides sufficient information for making in-
formed decisions regarding maintenance of facilities. The degradation of facility
equipment is usually not so rapid as to require the up to the minute reporting
that a permanent data acquisition system produces. Usually, the maintenance
technician can keep a log of these spot readings and develop trends from these
logs.

Permanent condition monitoring equipment is expensive to install, and the data

bases created cost money to analyze and maintain. Typically, permanent data-
logging equipment is installed only on super critical equipment used in produc-
tion processes, equipment that - if it goes down - costs the facility money by the
minute when it is not operating. MEDCOM facilities do not usually have
equipment that warrants this expense.

A simple and inexpensive means of acquiring real-time data is to use the existing
direct digital control (DDC) energy management and control system (EMCS) if
the building has one. The major expense of installing a permanent data-logging
system is running the information infrastructure (e.g., cabling throughout the
building and data collection devices). If a building has a DDC EMCS, this major
expense has already been absorbed. However, a buildings DDC system is de-
signed for comfort control and does not contain the diagnostic data points that
would be required by the maintenance department for equipment monitoring. If
a DDC system does exist, the maintenance department will most likely have to
add a few points to the control system to acquire the diagnostic data required for
CM. This will entail some additional software programming.
USACERL TR 99/41 47

5 Vibration Monitoring and Analysis

Theory, Applications, and Techniques

Analysis of system and equipment vibration levels is one of the most commonly
used CM techniques. Vibration monitoring helps determine the condition of ro-
tating equipment and structural stability in a system. It also helps identify noise
sources. (Severely vibrating equipment is noisy.)

Basic Vibration Theory

Vibration is simply the movement of a machine or machine part back and forth
from its position of rest. A weight hanging on a spring is the simplest example of
how vibration works. Until a force is applied to the weight to cause it to move,
we have no vibration. By applying an upward force, the weight moves upward,
compressing the spring. If we released the weight, it would drop below its neu-
tral position to some bottom limit of travel, where the spring would stop the
weight. The weight would then travel upward through the neutral position to
the top limit of motion, and back again through the neutral position. The motion
will continue in exactly the same manner as the force is reapplied. Thus, vibra-
tion is the response of a system to some internal or external force applied to the
system.

With a few exceptions, mechanical troubles in a machine cause vibration. The

most common problems that produce vibration are:
unbalance of rotating parts
misalignment of couplings and bearings

bent shafts
worn, eccentric, or damaged parts
bad drive belts and drive chains
bad bearings
torque variations
electromagnetic forces
48 USACERL TR 99/41

aerodynamic forces
hydraulic forces

looseness
rubbing
resonance.

The amount of time required to complete one full cycle of a vibration pattern is
called the period of vibration. If a machine completes one full cycle in 1/60th of a
second, the period of vibration is said to be 1/60th of a second. The period of vi-
bration is a simple and meaningful characteristic often used in vibration detec-
tion and analysis. Another simple characteristic is the frequency. Frequency is
related to period by the following formula:
frequency = 1/period

Frequency is the inverse of period. In reality, frequency is a measure of the

number of complete vibration cycles that occur in a specified amount of time.

The frequency of vibration is usually expressed in cycles per minute (CPM).

Specifying vibration frequency in CPM makes it easy to relate this characteristic
to another important specification of rotating machinery: revolutions per minute
(RPM). So, if you have piece of machinery that operates at 3600 RPM, you can
expect certain problems to create vibration at a frequency of 3600 RPM. Fre-
quency is sometimes expressed in cycles per second, or Hertz (Hz). The relation-
ship between Hz and CPM is expressed by the following equation:
CPM = Hz 60

Vibration displacement is defined as the total distance traveled from one ex-
treme limit to the other (the peak-to-peak displacement). Peak-to-peak vibra-
tion displacement is usually expressed in mils, where 1 mil equals 1/1000th of an
inch (0.001 in.).

Since a vibrating piece of machinery is moving, it has a velocity. The vibration

velocity constantly changes. At the top limit of the motion the speed is zero since
the weight must come to a stop before it can go in the opposite direction. The
speed of velocity is greatest as the weight passes through the neutral position.
Since the velocity of the part is constantly changing throughout the cycle, the
highest peak is selected for measurement. Vibration velocity is expressed in
inches per second.
USACERL TR 99/41 49

Since vibration velocity is directly related to vibration severity, for the most gen-
eral purpose vibration measurements, it is the preferred parameter for meas-
urement. As a rule of thumb, vibrations occurring in the 600 to 60,000 CPM fre-
quency range are generally best measured using vibration velocity.

Under conditions of dynamic stress, displacement alone may be a better indica-

tion of severity, especially when the machine part exhibits the property of brit-
tleness, the tendency to break or snap when stressed beyond a given limit. For
example, consider a slowly rotating machine that operates at 60 RPM, and that
exhibits vibration of 20 mils peak-to-peak displacement caused by rotor unbal-
ance. In terms of vibration velocity, 20 mils at 60 CPM is only 0.0585 in./sec,
which would be considered good for general machinery and little cause for im-
mediate concern. However, keep in mind that the bearing of this machine is be-
ing deflected 20 mils. Under these conditions, fatigue may occur due to stress
(resulting from the displacement) rather than due to fatigue (caused by the ve-
locity of displacement).

Generally, the most useful presentation of vibration data is a graph showing vi-
bration velocity (expressed in inches/second) on the vertical axis and frequency
on the horizontal axis. By analyzing this data, a trained vibration technician
can ascertain what kinds of problems exist. The trained technician has, in effect,
learned to read vibration signatures; he has learned to interpret what the dif-
ferent peaks in the different frequency ranges indicate. For example, when
analyzing a 3600 RPM pump motor, a peak at 3600 RPM indicates some kind of
mass imbalance. A peak at 7200 RPM (two times the rotational frequency) gen-
erally indicates a bent shaft.

Figure 10 illustrates what vibration data are typically taken. The top part of the
figure shows a schematic diagram of a fan/motor assembly, the points where vi-
bration data are taken (points A,B,C,D), and manual recording of the readings
obtained at those points. The bottom part of the figure shows a graph of the
data. Notice the spikes in the graph. These are what the vibration analyst
learns to interpret.

Figure 11 shows two plots of vibration data juxtaposed in the same graph. The
bottom plot is the vibration data for an engine running properly. The top plot
shows the vibration data for the same engine with a misfiring cylinder. Observe
the notes written by the analyst who has interpreted the peaks. Specific vibra-
tion frequencies indicate how different parts of the motor are operating. When
reviewed by a trained vibration analyst, vibration data are extremely helpful in
determining the health of the machine.
50 USACERL TR 99/41

Figure 10. Sample vibration data.

All rotating machinery will exhibit a certain degree of vibration. The question
then becomes How much is too much? There are no realistic figures for select-
ing a vibration limit, which, if exceeded, will result in immediate machinery fail-
ure. The events surrounding the development of a mechanical failure are too
complex to set reliable limits. However, there are some general guidelines that
have been developed over the years that can serve as general indication of the
condition of a piece of machinery.
USACERL TR 99/41 51

Figure 11. Two plots of vibration data juxtaposed in the same graph.

When setting up a vibration monitoring program that uses hand-held vibration

instrumentation, it is necessary to ensure that the measurements are taken con-
sistently. A slight variation in the location where a measurement is taken on a
piece of machinery can significantly alter its accuracy. This issue becomes espe-
cially difficult to police when several technicians take measurements at different
times on the same piece of machinery.

Information Obtained through Vibration Monitoring

If applied by a trained professional, vibration monitoring can yield information

regarding: wear, imbalance, misalignment, mechanical looseness, bearing dam-
age, belt flaws, sheave and pulley flaws, gear damage, flow turbulence, cavita-
tion, structural resonance, and material fatigue.

The maintenance supervisor/manager must make the decision whether it makes

economic sense to perform this function with in-house labor forces or whether it
should be outsourced to a contractor specializing in vibration monitoring and
analysis. In making this decision, maintenance supervisors/managers should
consider whether they have sufficient in-house labor to dedicate to vibration
monitoring. Vibration monitoring theory is complicated, the equipment is ex-
pensive, and the analysis of the data collected is a skill that must be practiced
regularly. Once he has completed the basic vibration training (costing several
thousands of dollars), a maintenance technician must be committed and allowed
to work at least 3 days per month in vibration analysis if he is to stay competent
with the technology and analysis techniques.
52 USACERL TR 99/41

Detection Interval/Amount of Data Collected

Narrow band vibration analysis can provide several weeks or months of warning
of impending failure. In establishing a vibration monitoring program, one must
first determine how often to take sampling data. Different vibration frequencies
forebode different upcoming failures. The frequency of data collection depends
on machine type and failure category. Typically, it is not cost effective to take
real-time vibration data; spot checking facility equipment once per month (or
once per quarter) with hand-held vibration monitoring equipment usually pro-
vides sufficient warning of impending problems. Facility rotating equipment
(e.g., fans, pumps) does not deteriorate fast enough to warrant continual real
time data collection.

Maintenance technicians should realize that accumulating more data is not nec-
essarily indicative of a better vibration monitoring program. Even after the first
costs of the vibration monitoring and data acquisition system have been ab-
sorbed, there is an overhead associated with data collection. The data must be
analyzed and interpreted. Even with the sophisticated software available to as-
sist the maintenance technician with these tasks, it takes an ongoing time in-
vestment.

Overall Vibration

Overall vibration is defined as the sum of an vibration energy produced across a

filtered band-width. Overall measurements are rarely used in todays modern
maintenance programs. It is typically better to concentrate on specific frequen-
cies.

Spectrum Analysis and Waveform Analysis

Spectrum analysis is the most commonly employed analysis method for machin-
ery diagnostics. In this type of analysis, the vibration technician focuses on
analyzing specific slices of the vibration data taken over a certain range of
CPM. (Section 5.1.1 above explains how vibration data are presented in graphi-
cal form.) Spectrum analysis can used to identify the majority of all rotating
equipment failures (due to mechanical degradation) before failure.

Waveform analysis, or time domain analysis, is another extremely valuable

analytical tool. While not used as regularly as spectrum analysis, the waveform
often helps the analyst more correctly diagnose the problem.
USACERL TR 99/41 53

Torsional Vibration

Torsional vibration is often used to detect the vibration associated with the
measurement of gear vibration and torque. It proves most helpful in situations
where, due to transmission path attenuation, the casing vibration signal has a
signal-to-noise ratio insufficient to detect the problem (i.e., the noise obscures the
signal). Torsional vibration is especially effective in situations where unsteady
forces excite the resonance of the structure or housing. Measure torque by using
pairs of matched sensors spaced at a sonic interval to take advantage of the
phase difference in the signals.

Multi-Channel Vibration Analysis

Multi-channel vibration analysis offers several extremely powerful methods for

machinery analysis such as force-response analysis, cross-coupling phase analy-
sis, analysis of resonance mode characteristics, and multi-plane balancing. Addi-
tionally, coherence functions offered by multi-channel analyzers allow for check-
ing the quality and linearity of data collected with typical dataloggers.

Shock Pulse Analysis

This type of analysis is used to detect impacts caused by contact between the
surfaces of the ball or roller and the raceway during rotation of anti-friction
bearings. The magnitude of these pulses depend on the surface condition and the
angular velocity of the bearing (RPM and diameter). Spike energy is similar in
theory to shock pulse.

Vibration Sensor Mounting (Permanent Installations)

Identify each monitoring point and epoxy magnetic corrosion resistant steel discs
(sound discs) at every location. Sound discs can be purchased for approximately
$26.00 each or the maintenance machinist can cut them from 1-in. alloy 410 or
416 magnetic stainless steel bar stock using a fly cutter. Use a tested epoxy such
as Hysol Gray Epoxy Patch available from Structural Adhesives to adhere the
sound disc to the equipment. Prepare the surface for the epoxy by grinding and
wiping down with a solvent. The use of super-glue type adhesives is not recom-
mended as the mounting pads tend to fall off.

As an alternative, the monitoring locations may be machined on the machines

surface at the locations listed in the next section. The machined surface should
be between 32 and 125 micro-inches (1 micro-inch [mic] = 0.001 in.) and within
one degree of perpendicularly with respect to the shaft center line.
54 USACERL TR 99/41

The use of a low mass accelerometer (100 mV/g) and a rare earth super magnet
to attach the accelerometer to the sound disc is recommended. This technique
has been proven by the U.S. Navy to provide a usable upper frequency limit of 5
kHz. By using special purpose accelerometers and a couplant between the accel-
erometer, magnet, and mounting disc, accurate measurements to 20 kHz are
possible.

Laser Shaft Alignment

Laser shaft alignment is a natural compliment to vibration analysis. Properly

aligning shafts eliminates one of the major causes of vibration in rotating ma-
chines and also drastically extends bearing life. For the minimal amount of
work involved, the payback is great. Figure 9 illustrates how drastically cylin-
drical roller bearing life is shortened by minimal misalignments of a few minutes
(i.e., a fraction of a one-degree misalignment).

Limitations

The effectiveness of vibration monitoring depends on sensor mounting, resolu-

tion, machine complexity, data collection techniques, and the ability of the ana-
lyst. This last factor, the ability of the analyst, is probably the most important
aspect of establishing an effective vibration monitoring program. The analyst
must be someone who possesses a thorough understanding of vibration theory
and the extensive field experience necessary to make the correct diagnosis of the
vibration spikes that may appear in the data acquired.

Complex, low speed (<120 RPM), variable speed, and reciprocating machinery
are extremely difficult to monitor effectively. Additionally, single channel analy-
sis cannot always accurately determine the source of the vibration on complex
machines.

Logistics

Equipment Required

For permanent data collection, vibration analysis systems include microproces-

sor-based data collectors, vibration transducers, equipment-mounted sound
discs, and a host personal computer with software for analyzing trends, estab-
lishing alert and alarm points, and assisting in diagnostics. Portable hand-held
USACERL TR 99/41 55

data collectors consist of a hand-held data collection device (about the size of a
palm-top computer) and a magnetized sensing device.

Operators
Requires personnel who have the ability to understand the basics of vibration
theory and who have a basic knowledge of machinery and failure modes.
Manning requirements are site specific.

Available Training
Training is provided by equipment vendors, as well as by the Vibration Insti-
tute, 6262 South Kingery Highway, Willowbrook, IL 60514 (708) 654-2254.

The Vibration Institute has published certification guidelines for vibration

analysts. Passing a written examination is required for certification.

Cost

Vibration analysis systems cost $20,000 to $120,000 for narrow band systems,
software, and primary training. The $20,000 will buy an adequate hand-held
system (about $13,000) and the training (about $7000) required for the techni-
cian to use that device and interpret the data acquired with it. The high end of
this price range is for a basic multi-channel, permanently installed system that
can be expanded for an additional cost of $500 to $1,000 per additional sensor.
In integrating vibration-monitoring into a CM program, it is often advisable to
contract out-of-house for a year or two to companies specializing in vibration
analysis. During that time, MEDCOM facility maintenance personnel can be-
come familiar with the technology and analysis process, make sure the correct
equipment is being analyzed, and make sure vibration analysis has an accept-
able payback for the facility. If only a few pieces are being analyzed at a facility,
it may make sense to contract the service out on a regular basis.
56 USACERL TR 99/41

6 Thermography
Infrared Thermography (IRT) is the application of infrared detection instru-
ments to identify temperature differences. The test instruments used are non-
contact, line-of-sight, thermal measurement and imaging systems. Because IRT
is a noncontact technique, it is especially attractive for identifying hot/cold spots
in energized electrical equipment, large surface areas such as boilers and build-
ing walls, and other areas where stand off temperature measurement is neces-
sary. Instruments that perform this function detect electromagnetic energy in
the short wave (3 to 5 microns) and long wave (8 to 15 microns) bands of the elec-
tromagnetic spectrum. The short wave instrument is the best choice for facilities
inspections due to the varied inspections (electrical, mechanical, and structural)
encountered. However, the short wave instrument is more sensitive than long
wave to solar reflections. The maintenance technician will need to be aware of
this when performing outdoor inspections in areas such as transformers, motor
control centers, switchgear, substations, switchyards, or power lines. In such
cases, sunlight reflected from shiny surfaces may make those surfaces appear to
be hotter than the adjacent surfaces when they really are not. To be effective
in facilities applications, IRT instruments must be portable, sensitive to within
0.20 C over a range of temperatures from -100 to +3000 C, and accurate within
+/-3 percent. In addition, the instrument must be capable of storing an image of
the thermogram for later analysis.

IRT inspections are identified as either qualitative or quantitative. The quantita-

tive inspection attempts the accurate measurement of the temperature of the
item of interest. To perform a quantitative inspection requires detailed knowl-
edge and understanding of the relationship of temperature and radiant power,
reflection, emittance, and environmental factors, as well as the limitations of the
detection instrument. This knowledge and understanding must be applied in a
methodical fashion to control the imaging system properly and to obtain accurate
temperature measurements. Quantitative measurements of temperature are ex-
tremely time-consuming, and are rarely needed in facilities applications.

The qualitative inspection is interested in relative differences, hot and cold spots,
and deviations from normal or expected temperature ranges. The knowledge and
understanding discussed above is needed to perform a meaningful qualitative
inspection. However, qualitative inspections are significantly less time-
consuming because the thermographer is not concerned with highly accurate
USACERL TR 99/41 57

temperature measurement. In qualitative inspections the thermographer obtains

accurate temperature differences (T) between like components. For example, a
typical motor control center will supply three-phase power, through a circuit
breaker and controller to a motor. Ideally, current flow through the three-phase
circuit should be uniform so the components within the circuit should have
similar temperatures. Any uneven heating, perhaps due to dirty or loose connec-
tions, would quickly be identified with the IRT imaging system. Because the
many variables that influence the quantitative inspection (reflection, emittance,
etc.) are the same between like components, the thermographer can quickly focus
on the temperature differences. The factors so important to a highly accurate
quantitative temperature measurement have very little influence on the tem-
perature differences between like components.

Theory and Applications

IRT can be used to identify degrading conditions in facilities electrical systems

such as transformers, motor control centers, switchgear, substations, switch-
yards, or power lines. In mechanical systems, IRT can identify blocked flow con-
ditions in heat exchanges, condensers, transformer cooling radiators, and pipes.
IRT can also be used to verify fluid level in large containers such as fuel storage
tanks. IRT can identify insulation system degradation in building walls and
roof, as well as refractory in boilers and furnaces. Temperature monitoring, in-
frared thermography in particular, is a reliable technique for finding the mois-
ture-induced temperature effects that characterize roof leaks, and for determin-
ing the thermal efficiency of heat exchangers, boilers, building envelopes, etc.

Deep-probe temperature analysis can detect buried pipe energy loss and leakage
by examining the temperature of the surrounding soil. This technique can be
used to quantify ground energy losses of pipes. IRT can also be used as a dam-
age control tool to locate mishaps such as fires and leaks. In soliciting consult-
ants to perform thermography, one should remember that (unless requested oth-
erwise) the thermographer will normally provide only an exception report that
identifies finds/faults (i.e., his analysis will be of qualitative temperature differ-
ences).

In summary, IRT can assess the in-service condition of electrical and mechanical
systems. Once this is done, the maintenance supervisor/manager can prioritize
work based on the temperature difference criteria. The greater the T, the more
urgent the problem.
58 USACERL TR 99/41

Limitations

Thermography is limited to line of sight. Errors can be introduced due to color of

material, material geometry, and by environmental factors such as solar heating
and wind effects.

Logistics

Equipment Required
Equipment ranges from simple, contact devices such as thermometers and
crayons to full color imaging, computer-based systems that can store, recall,
and print the thermal images.
The deep-probe temperature technique requires temperature probes, analy-
sis software and equipment to determine the location of piping systems.

Operators
Operators and mechanics can perform temperature measurements and
analysis using contact-type devices with minimal training on how and where
to take the temperature readings.
Because thermographic images are complex and difficult to measure and
analyze, training is required to obtain and interpret accurate and repeatable
thermal data and to interpret the data. With adequate training and certifica-
tion, electrical/mechanical technicians and/or engineers can performed this
technique.
Maintenance personnel can apply deep-probe temperature monitoring after
being trained, although this service is often contracted.

Training Available
Training is available through infrared imaging system manufacturers and
vendors.
The American Society of Non-destructive Testing (ASNT) has established
guidelines for nondestructive testing (NDT) thermographer certification.
These guidelines, intended for use in non-destructive testing, may be used as
guidelines for thermography in CM if appropriately applied. Certification is
not easily obtained. When deciding which maintenance technician should be
certified as a thermographer, the maintenance manager should consider gen-
eral background, work experience, and any previous thermographic experi-
ence or thermographic training.
USACERL TR 99/41 59

Cost
Prices of noncontact infrared thermometers/scanners start at approximately
$1,000. Full color microprocessor imaging systems with data storage and
print capability range from approximately $25,000 to $70,000.
Average thermographic system rental is approximately $1,500 per week. Op-
erator training costs approximately $1,250 per week of training.
Thermographic contractor services cost approximately $3,000 per day. Con-
tract services for deep-probe temperature analysis cost $5,000 to $6,000 for
the first day of service, and $1,500 to $2,000 for each additional day.
60 USACERL TR 99/41

7 Passive (Airborne) Ultrasonics

Airborne ultrasonic devices are highly sensitive listening guns (similar in size
to the radar speed guns used by police at speed traps). The maintenance techni-
cian should use a piece of equipment to obtain a translation of the noise (into the
range audible to human ears) produced by that piece of equipment. It provides a
convenient, nonintrusive means of testing equipment. It is especially easy and
useful in testing remote (i.e., pole-mounted) electrical equipment, as well as
shielded electrical equipment (i.e., connections inside switchgear and panels.) In
the case of high voltage insulator failures, airborne ultrasonic devices can often
detect faults earlier than infrared thermography can. Except for severe cases
where a current path to ground was established, infrared thermography would
not detect high-voltage insulation failures because the corona or tracking typi-
cally produces little or no heat. Airborne ultrasonic devices can also detect the
noise caused by loose connections as they vibrate inside of panels.

Airborne ultrasonic devices operate in the frequency range from 20 to 100kHz

and translate the high frequency signal to a signal within the audible human
range. This allows the operator to hear changes in noise levels associated with
leaks, corona discharges, and other high frequency events. For example, a main-
tenance technician could use ultrasonic equipment to hear a bearing ring and
surrounding housing resonating at the resonant frequency. Once detected, a
maintenance technician could then proceed to find the cause of the problem. (In-
sufficient lubrication and minor bearing material defects would be the likely
cause of this malfunction.)

Theory, Applications, and Techniques

Some of the most common plant applications of ultrasound detection are: leak
detection in pressure and vacuum systems (i.e., boiler, heat exchanger, condens-
ers, chillers, distillation columns, vacuum furnaces, specialty gas systems),
bearing inspection, steam trap inspection, valve blow-by, pump cavitation, detec-
tion of corona in switch gear, compressor valve analysis, and the integrity of
seals and gaskets in tanks, pipe systems, and large walk-in boxes. The informa-
tion in the following sections is applicable to all airborne ultrasonic devices. Any
specific procedures listed are those followed when using an UltraProbe 2000
(manufactured by UE Systems). While specific to a particular ultrasonic probe,
USACERL TR 99/41 61

these procedures illustrate the effort involved when using any ultrasonic sensing
equipment.

Basic Theory of Ultrasonic Detection

All operating equipment and most leakage problems produce a broad range of
sound. The high frequency ultrasonic components of these sounds are extremely
short wave in nature, and a short wave signal tends to be fairly directional. It is
therefore easy to isolate these signals from background noises and detect their
exact location. In addition, as subtle changes begin to occur in mechanical
equipment, the nature of ultrasound allows these potential warning signals to be
detected early before actual failure.

Airborne ultrasound instruments, often referred to as ultrasonic translators,

provide information two ways: qualitatively, due to the ability to hear ultra-
sounds through a noise isolating headphone, and quantitatively, via incremental
readings on a meter. This is accomplished in most ultrasonic translators by an
electronic process called heterodyning, which accurately converts the ultra-
sounds sensed by the instrument into the audible range where users can hear
and recognize them through headphones.

Although the ability to gauge intensity and view sonic patterns is important, it is
equally important to be able to hear the ultrasounds produced by various
equipment. That is precisely what makes these instruments so useful; they allow
inspectors to confirm a diagnosis on the spot by being able to discriminate among
various equipment sounds.

The reason users can accurately pinpoint the location of a particular ultrasonic
signal in a machine or from a leak is due to its high frequency short wave. Most
of the sounds sensed by the human ear range between 20 Hz and 20 kHz (20 cy-
cles per second to 20,000 cycles per second). They tend to be relatively gross
when compared with the sound waves sensed by ultrasonic translators. Low fre-
quency sound in the audible range are approximately 1.9 cm to 17 m in length,
whereas ultrasounds sensed by ultrasonic translators are only 0.3 to 1.6 cm long.
Since ultrasound wave lengths are magnitudes smaller, the ultrasonic environ-
ment is much more conducive to locating and isolating the source of problems in
loud plant environments.

Airborne ultrasound translators are relatively simple to use. They consist of a

basic handheld unit with headphones, a meter, a sensitivity adjustment, and
(most often) interchangeable modules that are used in either a scanning or a
contact mode. Some instruments have the ability to adjust the frequency re-
62 USACERL TR 99/41

sponse from between 20 to 100 kHz. An ultrasonic transmitter called a tone gen-
erator is often included.

Many of these features are useful in helping a user adapt to a specific test situa-
tion. For example, if an ultrasound source is too difficult to locate due to an in-
tense signal, a user can focus on the exact site by adjusting the sensitivity
downward. In another instance, if a low level leak occurs in a water valve, the
frequency tuning can be adjusted to help a user hear the trickle of the water
leak.

Interchangeable modules allow users to adjust for different types of inspection

problems. The scanning mode is used to detect ultrasounds that travel in the
atmosphere such as a pressure leak or a corona discharge, while the contact
mode is used to detect ultrasounds generated within a casing such as in a bear-
ing, pump, valve, or steam trap housing.

Leak Detection

This category covers a wide area of plant operations. It can be viewed as a way of
keeping a system running more efficiently. Some plants include it as part of an
energy conservation program, while others refer to it as fugitive emissions. Re-
gardless, leaks cost money, affect product quality, and can wreak havoc with the
environment. Ultrasonic detection can often locate the problem, whether the
leakage occurred in a liquid or a gas system.

The reason ultrasound is so versatile is that it detects the sound of a leak. When
a fluid (liquid or gas) leaks, it moves from the high pressure side through the
leak site to the low pressure side, where it expands rapidly and produces a tur-
bulent flow. This turbulence has strong ultrasonic components. The intensity of
the ultrasonic signal falls off rapidly from the source, allowing the exact spot of a
leak to be located.

Generalized gas leak detection is also very easy. An area should be scanned while
listening for a distinct rushing sound. With continued sensitivity adjustments,
the leak area is scanned until the loudest point is heard.

Some instruments include a rubber focusing probe that narrows the area of re-
ception so that a small emission can be pinpointed. The rubber focusing probe is
also an excellent tool for confirming the location of a leak. This is done by press-
ing it against the surface of the suspected area to determine if the sound of the
leak remains consistent. If it decreases in volume, the leak is elsewhere.
USACERL TR 99/41 63

Vacuum leaks may be located in the same manner; the only difference being that
the turbulence will occur within the vacuum chamber. For this reason, the inten-
sity of the sound will be less than that of a pressurized leak. Though it is most
effective with low-mid to gross leaks, the ease of ultrasound detection makes it
useful for most vacuum leak problems.

Liquid leaks are usually determined through valves and steam traps, although
some successes have been reported in locating water leaks from pressurized
pipes buried underground. A product can be checked for leakage if it produces
some turbulence as it leaks.

Valves are usually checked for leakage with the contact probe on the downstream
side. This is accomplished by first touching the upstream side and adjusting the
sensitivity to read about 50 percent of scale. The downstream side is then
touched and the sound intensity is compared. If the signal is lower than up-
stream, the valve is considered closed; if it is louder than upstream and is ac-
companied by a typical rushing sound, it is considered to be leaking.

Steam traps are also inspected easily with ultrasonic translators. During the
steam trap operation and while observing the meter, trap condition can be inter-
preted. The speed and simplicity of this type of test allow every trap in a plant to
be routinely inspected. Each type of steam trap produces a distinct sound as
briefly described below.

1. Intermittent Traps - When the trap is operating properly, the operator will hear
an open and closing sound. The trap normally fails in the open position, produc-
ing a continuous, rushing sound.

2. Inverted Bucket - A normal trap sounds as if it is floating; a failed trap sinks,

producing a continuous flow noise.

3. Thermostatic - Ultrasonic testing results of thermostatic traps vary. The noise

produced by these traps can be continuous or intermittent and will produce dif-
ferent sounds accordingly.

4. Float and Thermostatic (Continuous Load) - Flow in these traps is usually

modulated. Failed traps are normally cold and silent.

5. Continuous Flow - This type of trap, when operating normally, produces the in-
termittent sound of condensate flow only. If it has failed in the open position, a
continuous flow sound should be heard.
64 USACERL TR 99/41

Leaking tubes in heat exchangers and condensers as well as boiler casing leaks
are detectable with ultrasonic translators. In most power plants, the problem of
condenser in-leakage is a major concern. Condenser fittings are often routinely
inspected using the leak detection method previously described. If a leak is sus-
pected in a condenser tube bundle, it is possible to locate the leak by putting a
condenser at partial load and opening up a water box of a suspected tube bundle.
After the tube sheet is cleared of debris, the tube sheet is scanned.

How To Locate Leaks

Select the Log setting on the meter selection dial. Use fixed band position on
the Frequency selection dial. If too much background noise is present, try some
of the shielding methods. Start off with the sensitivity selection a 10 (maximum).
Begin to scan by pointing the module towards the test area. The procedure is to
go from the gross to the fine with more subtle adjustments made as the leak
is approached.

If there is too much ultrasound in the area, reduce the sensitivity setting and
continue to scan. If it is difficult to isolate the leak due to competing ultrasound,
place the rubber focusing probe over the scanning module and scan the test area.
Listen for a rushing sound while observing the meter. Follow the sound to the
loudest point. The meter will show a higher reading as the leak is approached. To
focus in on the leak, keep reducing the sensitivity setting and move the instru-
ment closer to the suspected leak site until you can confirm a leak.

To confirm a leak, position the rubber focusing probe (if it is on the scanning
module) close to the suspect leak site and move it slightly back and forth in all
directions. If the leak is at this location, the sound will increase and decrease in
intensity as you sweep over it. In some instances, it is useful to position the rub-
ber focusing probe directly over the suspect leak site and push down to seal it
from surrounding sounds. If the probe is over the leak, the rushing sound will
continue. If it is not the leak site, the sound will drop off.

Overcoming Competing Ultrasounds

If competing ultrasounds make it difficult to isolate a leak, there are two options:
manipulate the environment, i.e., when possible, turn off the equipment that is
producing the competing ultrasound or isolate the area by closing a door or win-
dow. Manipulate the instrument and use shielding techniques.

If environmental manipulation is not possible, try to get as close as possible to

the test site and manipulate the instrument so that it is pointing away from the
USACERL TR 99/41 65

competing ultrasound. Isolate the leak area by reducing the sensitivity of the
unit and by pushing the tip of the focusing probe up to the test area, checking a
small section at a time.

In some extreme instances, when the leak check is difficult in the fixed band
mode of the frequency selection dial, try to tune in to the leak sound by tuning
out the problem sound. In this instance, adjust the frequency selection dial until
the background sound is minimized and then proceed to listen for the leak.

Since ultrasound is a high frequency, short wave signal, it can usually be blocked
or shielded.

Note: when using any method, be sure to follow your plant or company safety
guidelines.

Electrical Problems

Three types of high voltage electrical problems detectable with ultrasound are:

1. Arcing: An arc occurs when electricity flows through space. Lightning is a good
example.

2. Corona: When voltage on an electrical conductor, such as an antenna or high

voltage transmission line, exceeds threshold value, the air around it begins to
ionize to form a blue or purple glow.

3. Tracking: Often referred to as baby arcing, electricity follows the path of dam-
aged insulation, using surrounding dirt, debris, and moisture as the conductive
medium.

Although theoretically, ultrasonic detection can be used in low, medium, and high
voltage systems, applications normally use medium and high voltage systems.
When electricity escapes in high voltage lines or when it jumps across a gap in
an electrical connection, it disturbs the air molecules around it and generates
ultrasound. Often this sound will be perceived as a crackling or frying sound; in
other situations, it will be heard as a buzzing sound. Applications include: in-
sulators, cable, switchgear, bus bars, relays, contractors, and junction boxes. In
substations, components such as insulators, transformers, and bushings may be
tested. Ultrasonic testing is often used for evaluation at voltages exceeding
2,000 volts, especially in enclosed switchgear. This is especially useful in identi-
fying corona problems. In enclosed switchgear, the frequency of detection of co-
rona greatly exceeds the frequency of serious faults identified by infrared. It is
66 USACERL TR 99/41

recommended that both tests be used with enclosed switchgear. When testing
electric equipment, follow plant or company safety procedures. When in doubt,
ask your supervisor. Never touch live electrical apparatus with the system.

The method for detecting electric arc and corona leakage is similar to the proce-
dure outlined in leak detection. Instead of listening for a rushing sound, a user
will listen for a crackling or buzzing sound. In some instances, as in trying to lo-
cate the source of radio/TV interference or in substations, the general area of dis-
turbance may be located with a gross detector such as a transistor radio or a
wide-band interference locator. Once the general area has been located, the
scanning module is used with a general scan of the area. The sensitivity is re-
duced if the signal is too strong to follow on the meter until the loudest point is
located. Determining whether a problem exists is relatively simple. By compar-
ing sound quality and sound levels among similar equipment, the problem will
become easy to identify, even though the sound itself will differ somewhat as it
resonates through various types and sizes of equipment.

On lower voltage systems, a quick scan of bus bars will often pick up a loose con-
nection. Checking junction boxes can reveal arcing. As with leak detection, the
closer one gets to the leak site, the louder the signal. If power lines are to be in-
spected and the signal does not appear to be intense enough to be detectable
from the ground, you can use an ultrasonic waveform concentrator (a parabolic
reflector), which will double the detection distance of the system and provide
pinpoint detection.

Mechanical Inspection

Ultrasonic inspection and monitoring of bearings is a reliable Method for de-

tecting incipient bearing failure. The ultrasonic warning appears prior to a rise
in temperature or an increase in driving torque. Ultrasonic inspection of bearing
is useful in recognizing the beginning of fatigue failure, brinnelling of bearing
surfaces, flooding of (or lack) of lubricant.

In ball bearings, as the metal in the raceway, roller, or bearing balls begins to
fatigue, a subtle deformation begins to occur. This deforming of the metal will
produce an increase in the emission of ultrasonic sound waves. When testing,
changes in amplitude of from 12 to 50 times the original reading is indication of
incipient bearing failure. When a reading exceeds any previous reading by 12
dB, it can be assumed that the bearing has entered the beginning of the failure
mode. This information was originally discovered through experimentation per-
formed by NASA on ball bearings. In tests performed while monitoring bearings
at frequencies ranging from 24 through 50 kHz, the changes in amplitude indi-
USACERL TR 99/41 67

cated the onset of, or incipient, bearing failure before other indicators, including
heat and vibration changes. (An ultrasonic system based on detection and
analysis of modulations of bearing resonance frequencies can provide subtle de-
tection capability, whereas conventional methods have difficulty detecting very
slight faults.) As a ball passes over a pit or fault in the race surface, it produces
an impact. A structural resonance of one of the bearing components vibrates or
rings by this repetitive impact. The sound produced is observed as an increase
in amplitude in the monitored ultrasonic frequencies of the bearing.

Brinelling of bearing surfaces will produce a similar increase in amplitude due to

the flattening process as the balls get out of round. These flat spots also produce
a repetitive ringing that is detected as an increase in amplitude of monitored
frequencies. The ultrasonic frequencies detected by the system are reproduced as
audible sound. This signal can greatly assist a user in determining bearing
problems. When listening, it is recommended that a user become familiar with
the sounds of a good bearing; often heard as a rushing or hissing noise. Crack-
ling or rough sounds indicate a bearing in the failure stage. In certain cases a
damaged ball can be heard as a clicking sound, whereas a high intensity may
indicate uniform ball damage. Loud rushing sound of a good bearing only
slightly rougher can indicate lack or lubrication. There are two basic procedures
of testing for bearing problems: comparative and historical. The comparative
method involves testing two or more similar bearings and comparing potential
differences. Historical testing requires monitoring a specific bearing over a pe-
riod of time to establish its history. By analyzing bearing history, wear patterns
at particular ultrasonic frequencies become obvious, allowing for early detection
and correction of bearing problems.

Comparative Test

Use the contact (stethoscope) module. Select LIN on the meter selection dial.
Select the desired frequency on the frequency selection dial. Select a test spot on
the bearing housing and mark it for future reference. Touch that spot with the
contact module. In ultrasonic sensing, the more media or materials ultrasound
has to travel through, the less accurate the reading will be. Therefore, be sure
the contact probe is actually touching the bearing housing. If this is difficult,
touch a grease fitting or touch as close to the bearing as possible. For consis-
tency, always approach the test spot at the same angle. Reduce sensitivity until
the meter reads 20. Listen to the bearing sound through headphones to hear the
quality of the signal for proper interpretation. Select same type bearings under
similar load conditions and same rotational speed. Approach the bearings at the
same angle, touching approximately the same area on the bearing housing.
Compare differences of meter reading and sound quality.
68 USACERL TR 99/41

Historical Bearing Test

There are two methods to historically trend a bearing. The first is a very com-
mon, field proven method called the simple method. The other provides greater
flexibility in terms of decibel selection and trending analysis, and is referred to
as the attenuator transfer curve method. The attenuator transfer curve
method is used in the Bearing Trac software, which provides trending, graphs,
and historical analysis. Before starting with either of the two historical methods
for monitoring bearings, the comparative method must be used to determine a
baseline.

Simple Method

Use the basic procedure as outlined above in the comparative test. Note fre-
quency, meter reading, and sensitivity selection on your bearing history chart.
Compare this reading with previous or future readings. On all future readings,
adjust frequency and sensitivity level to the original level recorded in the bear-
ing history chart. If the meter reading has moved from the original 20 mark up
to or past 100, there has been a 12 dB increase. (Increments of 20 on the meter
in the linear mode is about 3 decibels; e.g., 20 40 = 3 dB, etc.) Note: Increase of
12 dB or greater indicated the bearing has entered a failure mode. Lack of lubri-
cation is usually indicated by an 8 dB increase over baseline. It is usually heard
as a loud rushing sound. If lack of lubrication is suspected, after lubricating, re-
test. If readings do not go back to original levels and remain high, consider that
the bearing may fail soon and recheck frequently.

Lack of Lubrication

To avoid lack of lubrication, note the following: as the lubricant film reduces, the
sound level will increase. A rise of about 8 dB over baseline accompanied by a
uniform rushing sound will indicate lack of lubrication. When lubricating, add
just enough to return the reading to baseline. Use caution. Some lubricants will
need time to run to uniformly cover the bearing surfaces. Lubricate a little at a
time. Do not over lubricate.

Overlubrication

One of the most frequent causes for bearing failure is overlubrication. The ex-
cess stress of lubricant often breaks bearing seals or causes a buildup of heat,
which can create stress and deformity. To avoid overlubrication do not lubricate
if the baseline reading and baseline sound quality is maintained. When lubri-
cating, use just enough lubricant to bring the ultrasonic reading to baseline. As
USACERL TR 99/41 69

mentioned above, use caution. Some lubricants will need time to uniformly cover
the bearing surfaces.

Ultrasonic Translators

It is advisable to have instruments that are sensitive enough to detect the type of
problems you will encounter in the plant. A wide dynamic range in an instru-
ment enables you to look for small leaks on one end and locate gross mechanical
problems on the other. Since sound quality is an important consideration, make
sure the instrument heterodynes the ultrasonic signal. This will assure users
that they are getting an accurate reproduction of the ultrasonic signal, for signal
clarity and interpretation of the headphone sound. It is essential to have noise-
attenuating headphones with good sound quality. If the sound quality is not
clear, it will be difficult to understand what is being sensed. It is advisable to get
over-the-ear headphones that will block out ambient plant sounds during inspec-
tions. Without proper radio frequency radio frequency (RF) shielding, stray elec-
tronic signals will interfere with test results. In some instances, a radio pro-
grams have been heard, a phenomenon that totally confused operators.

Since every plant is different, special accessories may be needed to assist in some
situations. For example, compressor valve analysis might be easier with a mag-
netically mounted probe and an oscilloscope interface. If you are going to inspect
a variety of equipment or have fluids of different viscosities, it would be useful to
have the ability to change frequencies. For leak detection of potentially explo-
sive or flammable gases, it is advisable to use equipment that is rated intrinsi-
cally safe.

Limitations

Airborne ultrasonics are subjective and depend on perceived differences in

noises. To maximize the usefulness of this technology, care should be taken when
setting test equipment controls for frequency ranges, sensitivity, and scale. Addi-
tionally, the operator should realize that piping bends and the presence of mois-
ture and solids may dissipate and/or block the ultrasonic signal.

Logistics

If a vibration program already exists for bearing analysis, an ultrasonic bearing

monitoring program can help. Ultrasound translators can be used to aid a diag-
nosis. The high frequency, short wave characteristic of ultrasound allows the
70 USACERL TR 99/41

signal to be isolated so that a user can determine if a bearing has been correctly
diagnosed as failing.

At times equipment connected to a particular bearing can generate false signals.

By adjusting the sensitivity and frequency, and by listening to the sound, one can
determine whether it is the bearing, a rotor, or something else that is the root of
the problem. The ability to hear what is going on can prove very important. Ul-
trasound detectors work well on slow speed bearings. In some extreme cases,
just being able to hear some movement of a bearing through a well greased cas-
ing could provide information about potential failure. The sound might not have
enough energy to stimulate classic vibration accelerometers, but will be heard
via ultrasonic translators, especially those with frequency tuning. Sometimes
there are so many bearings in a plant that not every piece of equipment can be
checked routinely by a limited staff of trained technicians. Since ultrasound
technicians require little training, a technician or the machine operator can de-
termine potential bearing problems and notify the vibration technician for fol-
low-up.

Equipment Required

The equipment required for performing ultrasonic detection of faulty equipment

is an ultrasonic monitoring scanner for airborne sound or ultrasonic detector for
contact mode. (With the contact mode detector, contact to the equipment in
question is made with a metal rod. This is the high-tech version of the mainte-
nance technician putting a screwdriver on a piece of equipment and putting his
ear to the other end.)

Operators

Maintenance technicians and maintenance engineers are easily taught ultra-

sonic detection.

Training Available/Required

Minimal training is required with the exception of that required for use of multi-
channel Acoustic Valve Leak Detectors (AVLDs).

Cost

The price of scanners and accessories ranges from $1,000 to approximately

$8,000. Complex acoustic valve leak detection systems are approximately
$100,000.
USACERL TR 99/41 71

8 Lubricant and Wear Particle Analysis

Purpose

Lubricating oil analysis is performed for three reasons:

1. To determine the machine mechanical wear condition

2. To determine the lubricant condition

3. To determine if the lubricant has become contaminated.

A wide variety of tests can provide information regarding one or more of these
areas. The test used will depend on the test results sensitivity and accuracy, the
cost, and the machine construction and application. The three areas are not un-
related; changes in lubricant condition and contamination, if not corrected, will
lead to machine wear. Because of the important relationships, commercial analy-
sis laboratories will often group several tests in cost effective test packages that
provide information about all three areas.

Machine Mechanical Wear Condition

The criteria for analyzing the lubricating oil to determine the machines condi-
tion are generally the same as for performing vibration analysis. This analysis is
applicable to all machines with motors 7.5 HP or larger, critical machines, or
high cost machines. Generally the routine sampling and analysis periodicity will
be the same as the vibration analysis periodicity (when using a portable vibra-
tion data collector). For machines with a condition history (a year or more of
data), this is typically performed quarterly.

Lubricant Condition

Lubricating oil is either discarded or reconditioned through filtering and/or re-

placing additives. Analyzing the oil to determine the lubricant condition is,
therefore, driven by costs. Small machines, those with oil reservoirs 1 gal or
less, have the oil changed on an operating time basis. An automobile is the most
common example of time-based lubricating oil maintenance. In this example,
72 USACERL TR 99/41

the costs to replace the automobile oil (the replacement oil, labor to change the
oil, and disposal costs) are lower than the cost to analyze the oil (i.e., the cost of
sample materials, labor to collect the sample, and the analysis). In the case of
automobile oil, time-based replacement is cheaper than analysis due to competi-
tion and the economies of scale that have been created to meet the consumer
need for replacing automobile oil.

In the case of lubricating oil used in facility equipment, simply replace and dis-
card the machine lubricating oil if it is cheaper than analyzing it. When making
this decision, the maintenance manager must have firm prices for materials used
to take samples and the labor hours it will take to collect, package, and send the
samples out for analysis. Remember, though, that one oil sample is sufficient for
many tests.

Lubricant Contamination

Lubricating oil can become contaminated due to the machines operating envi-
ronment, improper filling procedures, or through the mixing of different lubri-
cants in the same machine. If a machine is topped off with oil frequently, the
maintenance technician should send the oil out for analysis periodically to check
the machine for any serious problems.

Standard Analytical Tests

Lubricating oil and hydraulic fluid analysis should proceed from simple, subjec-
tive techniques such as visual and odor examination through more sophisticated
techniques. The more sophisticated (and expensive) techniques should be used
when conditions indicate the need for additional information and the equipment
cost or criticality justifies the cost.

Visual and Odor

Simple inspections can be performed weekly by the equipment operator to look

at and smell the lubricating oil. A visual inspection looks for changes in color,
haziness or cloudiness, and particles. This test is very subjective, but can be an
indicator of recent water or dirt contamination and advancing oxidation. A small
sample of fresh lubricating oil in a sealed, clear bottle, can be kept on hand for
visual comparison. A burned smell may indicate oxidation of the oil. Other
odors could indicate contamination. Odor is more subjective than the visual in-
spection because peoples sensitivity to smell varies, and there is no effective way
USACERL TR 99/41 73

to compare the odor between samples. The operator must be careful not to intro-
duce dirt into the system when taking a sample.

Viscosity

Viscosity is a measure of oil flow rate at a specified temperature. A change (in-

crease or decrease) in viscosity over time indicates changes in the lubricant con-
dition, or it may indicate lubricant contamination. Viscosity can be tested using
portable equipment, or it can be tested more accurately in a laboratory using the
ASTM D445 procedure. Viscosity is measured in centistoke (cSt), and minimum
and maximum values are identified by the ISO grade. Testing oil viscosity is
usually part of a commercial laboratory standard test package.

Water

Water in lubricating oil and hydraulic fluid contributes to corrosion and forma-
tion of acids. Small amounts of water (less than 0.1 percent) can be dissolved in
oil and can be detected using the crackle test or infrared spectroscopy (minimum
detectable is 0.05 percent or approximately 500 ppm by both methods), the
ASTM D95 distillation method (minimum detectable is 0.01 percent/100 ppm)
ppm), the ASTM D1744 Karl Fischer method (minimum detectable is 0.002 per-
cent/100 ppm). If greater than 0.1 percent water is suspended or emulsified in
the oil, the oil will appear cloudy or hazy. Free water in oil collects in the bottom
of oil reservoirs and can be found by draining them from the bottom.

Percent Solids/Water

A simple, inexpensive test is used to provide a gross estimate of solids and/or wa-
ter in the oil. A sample is centrifuged in a calibrated tube and the resulting vol-
ume is measured. The test is effective for amounts in the range of 0.1 to 20 per-
cent of volume and is usually part of a commercial laboratory standard test
package.

Total Acid Number (TAN)

Total acid is an indicator of the lubricating oil condition and is monitored relative
to the TAN of new oil. In some systems, the TAN will also be used to indicate
acid contamination. TAN is measured in milligrams of potassium hydroxide
(KOH) per gram of oil (mg KOH/g). KOH is used in a titration process and the
end point is indicated by color change (ASTM D974) or electrical conductivity
change (ASTM D664).
74 USACERL TR 99/41

Total Base Number (TBN)

Similar to the TAN test method, the TBN test measures alkalinity (ability to
neutralize acid) of oil sample. This test is used on oil with high detergent addi-
tives such as diesel and gasoline engines. KOH is used in a titration process and
the end point is indicated by electrical conductivity change (per ASTM D664 or
ASTM D2896). When comparing test results from your oil against baseline data
from the oil supplier, make sure that the same test method was used for your oil
as was used in generating the baseline data. Results can vary significantly be-
tween test methods.

Spectrometric Metals

Also known as emission spectroscopy, this technique examines the light (spec-
trum) emitted from the sample during testing, and identifies up to 21 metals.
Metals are categorized as wear, contaminate, or additive metals. The procedure
identifies both soluble metal and metal particles up to 5 to 10 microns (5-10 m).
The test cost is moderate, and is usually part of a commercial laboratory stan-
dard test package. Other techniques (e.g., absorption spectroscopy and X-ray
spectroscopy) are used by some laboratories to identify metals.

Infrared Spectroscopy

This technique is also known as infrared analysis, infrared absorption spectros-

copy or spectrophotometry, and Fourier Transform Infrared (FTIR) spectroscopy.
The technique examines the infrared wavelength that is absorbed by the oil
sample. The test is used to identify nonmetallic contamination (see Water, dis-
cussed above in this chapter) and lubricant conditions (e.g., oxidation, anti-
oxidant, other additive depletion). In the future, it may become possible to cou-
ple computer expert system analysis with known oil spectrums, in an effort to
produce highly accurate diagnosis of small changes in the oil condition. Costs
vary, depending on the level of sophistication required. Infrared spectroscopy is
usually part of a commercial laboratory standard test package.

Analytical Ferrography

More detailed than Direct Reading (DR) ferrography, analytical ferrography is

often initiated based on changes in DR, spectrometric metal increases, or in-
creased particle count. The analysis is sometimes performed on a regular basis
on expensive or critical machines. The test process is labor intensive and in-
volves the preparation of sample and examination under magnification. Results
vary with the analysts capability, but the procedure can provide detailed infor-
USACERL TR 99/41 75

mation regarding wear: e.g., wear type (rubbing, sliding, cutting), color, particle
types (oxide, corrosive, crystalline), and other nonferrous particles. This detailed
information can be critical in finding the root cause of wear problems. Costs are
moderately high; the test is performed on a fixed price basis (per sample) from a
commercial laboratory.

Special Tests

Special tests are sometimes needed to monitor lubricant conditions on some ex-
pensive or critical Systems. Usually the special test is used to monitor a lubri-
cant contaminate, a characteristic, or additive depletion. This section identifies
some of the special tests available. Special tests are rarely needed for routine
monitoring of lubricants. The list of special test presented here is not meant to
be all-inclusive only a list of samples. Test procedures are constantly being
developed and refined. The annual ASTM Standards provides a description of
current test methods,

Glycol Antifreeze

Glycol contamination can be detected using infrared spectroscopy (see Infrared

Spectroscopy, discussed earlier) at levels greater than 0.1 percent (1,000 ppm),
which is usually adequate for condition monitoring. However, additional tests
can be specified to identify if small amounts of glycol are present. ASTM D2982
will indicate if trace amounts are present. ASTM D4291 uses gas chromatogra-
phy to quantify small amounts of glycol.

Karl Fischer Water

Water contamination can be detected using infrared spectroscopy (see Infrared

Spectroscopy, p 74) at levels greater than 0.05 percent (500 ppm), which is usu-
ally adequate for condition monitoring. Using a titration process with a Karl
Fischer reagent, low levels of water can be detected and quantified. The test,
ASTM D1744, is useful when accepting new oil or evaluating clean up efforts.
Cost of the test is moderate.

Foamlug

Some oil may have anti-foam agents added to improve the lubrication capability
in specific applications such as gear boxes or mixers. ASTM test D892 can be
used to test the oils foam characteristics, The test blows air through a sample of
the oil and measures the foam volume. Cost of the test is moderately high.
76 USACERL TR 99/41

Rust Prevention

Some systems are susceptible to water contamination due to equipment location

or the system operating environment. In those cases, the lubricating oil or hy-
draulic fluid may be fortified with an inhibitor to prevent rust. The effectiveness
of rust prevention can be tested using ASTM D665 (or ASTM D3603). Results
are pass/fail and the cost of the test is high.

Rotating Bomb Oxidation Test (RBOT)

Also known as the Rotary Bomb Oxidation Test, ASTM D 2272 is used to esti-
mate oxidation stability and the remaining useful life of oil. The test simulates
aging, identifying when rapid oxidation takes place and indicating that anti-
oxidants have been depleted. The test is not a one time test; it must be per-
formed over time, starting with a baseline test of the new oil. Subsequent tests
are necessary to develop the trend line. Because of the high cost and the multiple
tests required, this test is usually only performed on large volume reservoirs or
expensive oil.

Application

Typically, lubricating oil analysis should be performed on a quarterly basis on all

machines with motors 7.5 HP or larger, and on all critical or expensive machines.
The analysis schedule should be adjusted in the same way that the vibration
analysis schedule is adjusted. Analyze more frequently for machines that are
indicating emerging problems; less frequently for machines that operate under
the same conditions and are not run on a continuous basis. A new baseline
analysis will be needed following machine repair or oil change out. All hydraulic
systems, except mobile systems, should be analyzed on a quarterly basis. Mobile
systems should be considered for analysis based on the machine size and the cost
effectiveness of performing the analysis. Generally, it is more cost effective in
mobile equipment to maintain the hydraulic fluid based on the fluid condition.
However, for small systems, the cost to flush and replace the hydraulic fluid on a
time basis may be lower than the cost to analyze the fluid on a routine basis.
Grease is usually not analyzed on a regular basis. Although most of the testing
that is done on oil can also be done on grease, there is a problem getting a repre-
sentative sample. To get a representative sample that is a homogeneous mixture
of the grease, contaminants, and wear, the machine must usually be disassem-
bled. Once a machine has failed and must be disassembled, analysis of the
grease to diagnose the failure can sometimes be useful.
USACERL TR 99/41 77

A concern common to all machines with lubricating oil systems is keeping dirt
and moisture out of the system. Common components of dirt, such as silica, are
abrasive and naturally promote wear of contact surfaces. In hydraulic systems,
particles can block and abrade the close tolerances of moving parts. Water in oil
promotes oxidation and reacts with additives to degrade the performance of the
lubrication system. Ideally, there would be no dirt or moisture in the lubricant;
this, of course, is not possible. The lubricant analysis program must therefore
monitor and control contaminants. Large systems with filters will have steady-
state levels of contaminates. Increases in contaminates indicate breakdown in
the systems integrity (leaks in seals, doors, heat exchangers, etc.) or degradation
of the filter. Unfiltered systems can exhibit steady increases during operation.
Operators can perform a weekly visual and odor check of lubricating systems
and provide a first alert of contamination. Some bearing lubricating systems
have such a small amount of oil that a weekly check may be impractical.

Motors, Generators, Pumps, Blowers, Fan

For machines with less than 5 gal in the lubrication system, the analyst is
mostly concerned with machine condition. Lubricant condition and contamina-
tion are of interest because they provide some indication of machine condition.
Routinely monitor viscosity, percent solids/water, and spectrometric metals.
Monitor trends and discard or refresh the oil when viscosity changes 10 percent
from the baseline. Viscosity normally increases above the baseline with the oil
service time. If the viscosity decreases below the baseline, it usually means that
the oil is contaminated, probably from adding the wrong type of makeup oil.
There should be no water present (minimum detectable water is 0.1 percent). If
there is water, the source of the water needs to be identified and corrected. For
machines with more than 5 gal of oil in the system, add infrared spectroscopy
(minimum amount of water detectable is 0.05 percent) and particle counting.
Changes in particle count can indicate increased contamination or increased
wear. Correlate particle count with spectrometric metals. The rate of particle
count change indicates how quickly the lubricant is degrading. Visual particle
counting can be used to identify the source of the contamination. In addition,
perform DR ferrography for expensive or critical machines. In all machines,
changes in spectrometric metals or DR should be investigated further using
analytical ferrography and correlated with vibration analysis.

Gearboxes

Same as above, except for gearboxes with less than 5 gal of oil, add particle
counting. Implement DR ferrography for high cost or critical gearboxes. Monitor
trends and correlate with vibration readings.
78 USACERL TR 99/41

Chillers

In addition to the items identified above, add Total Acid Number (TAN) and DR
ferrography.

Diesel Engines

Use the same procedure as for chillers except substitute Total Base Number
(TBN) for TAN when oil has high detergent additives. A decrease in viscosity
below the baseline may indicate fuel contamination. Coolant leakage (glycol and
other characteristics) is identified from the infrared spectroscopy analysis.

Compressors

Centrifugal compressors should be treated the same as chillers. Reciprocating

compressors should be treated the same as diesel engines.

Hydraulic Systems

Perform the same oil analysis as that performed on gearboxes. Monitor particle
count by ISO category. Each hydraulic system will have limiting clearances that
will determine critical particle sizes. Note that some hydraulic systems use flu-
ids other than oil (water or glycol). For these systems, oil analysis does not ap-
ply; however, perform particle control the same as for oil-filled hydraulic sys-
tems.

Large Reservoirs

For reservoirs over 500 gal, consider performing an Rotating Bomb Oxidation
Test (RBOT) to assess the oxygen stability. Cost is usually the deciding factor.
At least three tests are needed to develop a trend. Once the trend has been es-
tablished, additional retesting should be performed at least once a year. Mainte-
nance dollar are saved when replacement or refreshing of a large volume of oil
(or smaller volume of expensive oil) can be deferred.

Lubrication Analysis

As one can see from reading the above, there are numerous lubrication tests.
Commercial laboratories performing the tests have charts available that sum-
marize the various lubricant tests, monitoring interval, and application.
USACERL TR 99/41 79

Sampling

Oil samples must be collected safely and in a manner that will not introduce dirt
and other contaminates into the machine/system, or into the sample. It may be
necessary to install permanent sample valves in some lubricating systems. The
oil sample should be representative of the oil seen in the machine. The sample
should, therefore, be collected from a mid-point in reservoirs and upstream of the
filter in circulating systems. Sample collection bottles and tubing can be pro-
cured through testing laboratories. The testing laboratory can also provide
guidance as regards to the cleanliness level needed. Oil sample pumps for ex-
tracting oil from reservoirs must be used properly to avoid contamination. Sam-
ples must be collected from the same point in the system to ensure consistency in
the test analysis; therefore, the maintenance procedure must provide detailed
direction on where and how to collect samples. The equipment operators can
collect samples. Each sample is marked with the system/machine name, sample
location point (the system may have multiple sample points), date, elapsed oper-
ating time for the system/machine, and other comments such as last topping off
or filtering operation. The analyst will also need to know the amount of oil in the
reservoir to make recommendations to correct abnormalities.
80 USACERL TR 99/41

9 Electrical Condition Monitoring

Electrical condition monitoring encompasses several technologies and techniques
used to provide a comprehensive system evaluation. Electrical equipment repre-
sents a major portion of a facilitys capital investment. From the power distribu-
tion system to electric motors, efficient operation of the electrical systems is cru-
cial to maintaining operational capability of a facility.

Monitoring key electrical parameters provides the information to detect and cor-
rect electrical faults such as high resistance connections, phase imbalance, and
insulation breakdown. Since faults in electrical systems are seldom visible,
these faults are costly (increased electrical usage), present safety concerns (fires),
and involve life cycle cost issues (premature replacement of equipment). Ac-
cording to the Electric Power Research Institute, voltage imbalances of as little
as 5 percent in motor power circuits result in a 50 percent reduction in motor life
expectancy and efficiency in three phase AC motors. A 2.5 percent increase in
motor temperatures can be generated by the same 5 percent voltage imbalance
accelerating insulation degradation.

Techniques
infrared thermography

airborne ultrasonics
transformer oil analysis
megohmmeter testing
high potential testing (HiPot)
surge testing
conductor complex impedance

time domain reflectometry (TDR)

motor current spectrum analysis
RF monitoring

power factor and harmonic distortion

starting current and time
USACERL TR 99/41 81

motor circuit analysis (MCA)

motor current readings

turns ratio

(Note: HiPot and surge testing should be performed with caution. The high volt-
age applied during these tests may induce premature failure of the units being
tested. For that reason these tests normally are performed only for acceptance
testing, not for condition monitoring.)

Megohmmeter Testing

A hand-held generator (battery powered or hand cranked) is used to measure the

insulation resistance phase-to-phase or phase-to-ground of an electric circuit.
Readings must be temperature-corrected to trend the information. Winding
temperatures affect test results. An enhanced technique compares the ratio of
the Megohmmeter readings after 1 minute, and then again compare the readings
after 10 minutes. This ratio is referred to as the polarization index.

High Potential Testing (HiPot)

HiPot testing applies a voltage equal to twice the operating voltage plus 1000
volts to cables and motor windings to test the insulation system. This is typi-
cally a go/no-go test. Industry practice calls for HiPot tests on new and re-
wound motors. This test stresses the insulation systems and can induce prema-
ture failures in marginal motors. Due to this possibility, HiPot is not
recommended as a routinely repeated condition monitoring technique, but as an
acceptance test. An alternative use of the equipment is to start with lower volt-
age and increase the applied voltage in steps and measure the change in insula-
tion resistance readings.

Surge Testing

Surge Testing uses equipment based on two capacitors and an oscilloscope to de-
termine the condition of motor windings. This is a comparative test evaluating
the difference in readings of identical voltage pulses applied to two windings si-
multaneously. Like HiPot testing, the applied voltage equals two times operating
voltage plus 1000 volts. This test also in primarily an acceptance, go/no-go test.
Data are provided as a comparison of waveforms between two phases indicating
the relative condition of the two phases with regard to short circuits. The read-
ings for a particular motor can be trended, but the repeated stress of the insula-
tion system is not recommended.
82 USACERL TR 99/41

Conductor Complex Impedance

The total resistance of a conductor is the sum of its resistance, capacitive imped-
ance, and inductive impedance. Accurate measurement of the conductor imped-
ance allows minor degradations in a motor to be detected and addressed prior to
motor failure. The condition of the insulation system can be determined by
measuring the capacitance between each phase and ground. The presence of
moisture or other conducting substance will form a capacitor with the conductor
being one plate, the insulation the dielectric, and the contaminate forming the
second plate. Maintaining proper phase balance is imperative to efficient opera-
tion and toward realizing the full lifetime of electrical equipment.

Time Domain Reflectometry

In this test, a voltage spike is sent through a conductor. Each discontinuity in

the conductor path generates a reflected pulse. The reflected pulse and time dif-
ference between initial pulse and reception of the reflected pulse indicate the lo-
cation of the discontinuity.

Motor Current Spectrum Analysis (MCSA)

MCSA is a method of detecting the presence of broken or cracked rotor bars or

high resistance connections in end rings. Motor current spectrums in both time
and frequency domains are collected with a clamp-on ammeter and Fast Fourier
Transform (FFT) analyzer. Rotor bar problems will appear as side-bands around
the power supply line frequency. MCSA evaluates the amplitude of the side
bands that occur about the line frequency.

Radio Frequency (RF) Monitoring

RF monitoring was developed to detect arcs caused by broken windings in gen-

erators. It consists of establishing RF background levels and the amplitude trend
over a narrow frequency band.

Power Factor and Harmonic Distortion

Maintaining optimum power factor maximizes the efficient use of electrical

power. Power factor is the ratio of real power to reactive power usage. Dual
channel data-loggers are used to determine the phase relationship between volt-
age and current, then calculate the power factor. If this detects a low power fac-
tor, subsequent engineering analysis will be required to devise a means of im-
proving power system power factor.
USACERL TR 99/41 83

Motor Current Readings

Clamp-on ammeter attachments provide the capability to take actual current

draw information while the equipment is operating. On three-phase equipment,
comparison of current draws can reveal phase imbalance conditions.

Airborne (Passive) Ultrasonics

Refer to Chapter 7.

Transformer Oil Analysis

Refer to Chapter 8.

Applications

Equipment to be Monitored

Specific equipment that can be monitored by electrical condition monitoring

techniques are:

1. Electrical Distribution Cabling - Megohmmeter, Time Domain Reflectometry,

HiPot, IRT (if visible), and Airbourne Ultrasonics

2. Electrical Distribution Switchgear and Controllers - Timing, Visual Inspection,

IRT, and Airborne Ultrasonics

3. Electrical Distribution Transformer - Oil Analysis, Turns Ratio, Power Factor,

and Harmonic Distortion

4. Electrical Motors - Current Draw, Motor Current Spectrum Analysis, Motor Cir-
cuit Analysis, Megohmmeter. HiPot, Surge Test, Conductor Complex Impedance,
Starting Current, and Coast-Down Time

5. Generators - Megohmmeter, RF, and Coast-Down Time

6. Distribution System - HiPot, Ultrasonics, Power Factor, and Harmonic Distortion.

84 USACERL TR 99/41

Conditions Monitored

Conditions to be monitored are: voltage, current, resistance, complex impedance,

capacitance, insulation integrity, phase imbalance, mechanical binding and pres-
ence of arcing.

Detection Interval

Monitoring intervals of several weeks to several months for various technologies

will provide sufficient condition information to warn of degrading equipment
condition. Specific expectations of the length of warning provided should be fac-
tored into developing monitoring intervals for specific technologies. Several of
the technologies outlined are also effective when used for acceptance testing and
certification.

Accuracy

Accuracy depends on the technique applied and rating of instrument.

Limitations

The technologies presented can be divided into two categories:

Energized. Those technologies that can safely provide information on ener-
gized systems and require the system be energized and operational. These
technologies include IRT, Ultrasonics, Motor Current Readings, Starting
Current, Motor Current Spectrum Analysis RF, Power Factor, and Harmonic
Distortion.
De-Energized. Technologies that require the circuit to be de-energized for
safe usage include Surge Testing, HiPot Testing, Time Domain Reflectometry
(TDR), Megohmmeter, Motor Circuit Analysis, Transformer Oil Analysis,
Turns Ratio, and Conductor Complex Impedance.

Each technology will require specific initial conditions to be set prior to con-
ducting the test. For instance, prior to an IRT survey, typical equipment pow-
ered through the switch board should be running to bring the distribution
equipment to normal operating temperatures. Higher load accentuates problem
areas. Conducting the survey at low load conditions may allow a problem to re-
main undetected.
USACERL TR 99/41 85

Logistics

Equipment Required

A comprehensive electrical testing program includes: multi-meters/volt-

ohmmeters, current clamps, TDR, motor current spectrum analysis software,
and integrated motor circuit analysis testers.

Operations

Electricians, electrical technicians, and engineers should be trained in electrical

CM techniques such as motor current signature analysis, motor circuit analysis,
complex phase impedance, and insulation resistance readings/analysis.

Training Available

Equipment manufacturers and consultants specializing in electrical testing

techniques provide classroom training and seminars to teach these techniques.

Cost
Equipment. Equipment costs vary from $20 for a simple multi-meter to ap-
proximately $40,000 for integrated motor current analysis (MCA) testers. A
full inventory of electrical testing equipment should cost from about $30,000
to $75,000.
Training. Training averages between $750 and $1,000 per week per person
trained.
86 USACERL TR 99/41

10 Non-Destructive Testing
Non-Destructive Testing (NDT) evaluates material properties and quality of
manufacture for expensive components or assemblies without damaging the
product or its function. Instead of statistical sampling techniques that use only
surface measurements or require the destructive testing of selected components
from a production lot, NDT is used when these testing techniques are cost pro-
hibitive or ineffective. Typically, NDT has been associated with the welding of
large high stress components such as pressure vessels and structural supports.
Process plants such as refineries or chemical plants use NDT techniques to en-
sure integrity of pressure boundaries for systems processing volatile substances.

Techniques

The following section discussed various NDT techniques.

Radiography

Radiography is performed to detect sub-surface defects. Radiography or X-ray is

one of the most powerful NDT technique available in industry. Depending on the
strength of the radiation source, radiography can provide a clear representation
(radiograph) of discontinuities or inclusions in material several inches thick. X-
ray or gamma ray sensitive film is placed on one surface of the material to be ex-
amined. The radiation source is positioned on the opposite side of the piece. The
source may be either a natural gamma emitter or a powered X-ray emitter. The
source is accurately aligned to ensure the proper exposure angle through the ma-
terial. When all preparations and safety precautions are complete, the radiation
source is energized or unshielded.

Gamma or X-rays pass through a material and expose film placed under the ma-
terial. By developing the film in a manner similar to photographic film, an im-
age of defects or inclusions in the material is produced. More advanced radiolu-
minescent film does not require photographic processing. Multiple shots from
varying angles provide a complete picture of the thickness of the material. Dual
angles are required to determine the size and orientation of an inclusion.
USACERL TR 99/41 87

Once the type, size, and orientation of each inclusion are defined, these can be
classified as either acceptable inclusions or unacceptable defects.

Defects in the material must be accurately located to facilitate minimal material

removal, yet ensure the defect has been completely eliminated. Minimizing ma-
terial removal also minimizes repair cost and reduces the likelihood of additional
defects created by the repair. The repair is then re-evaluated to ensure the de-
fect removal and subsequent repair were conducted properly.

Radiography, though a versatile tool, is limited by the potential health risks.

Use of radiography usually requires the piece be moved to a special shielded
area, or that personnel be evacuated from the vicinity to avoid exposure to the
powerful radiation source required to penetrate several inches of dense material.
Temporary shielding may also be installed, but the installation and removal of
thousands of pounds of lead is labor intensive and rarely worth the expense.
Radiography technicians are trained in radiation health physics and material
properties. These technicians can visually distinguish between welding slag in-
clusions, porosity, cracking, and fatigue when analyzing radiographic images.

Ultrasonic Testing (Imaging)

Ultrasonic testing provides detection of deep sub-surface defects. Ultrasonic (UT)

inspection of welds and base material is often an alternative or complementary
NDT technique to radiography. Though more dependent on the skill of the op-
erator, UT does not produce the harmful radiation entailed with radiography.
UT inspection is based on the difference in the wave reflecting properties of de-
fects and the surrounding material. An ultrasonic signal is applied through a
transducer into the material being inspected. The speed and intensity with
which the signal is transmitted or reflected to a transducer provides a graphic
representation of defects or discontinuities within the material. A couplant fluid
is often used to provide a uniform transmission path between the transducer, re-
ceiver and the material of interest. Transducer configurations differ depending
on the type of system used. Some systems use a single transducer to transmit
and receive the test signal. Others use a transmit transducer in conjunction
with a separate receive transducer. Dual transducer systems may be configured
with both transducers on the same surface of the material or with transducers
on the opposite surfaces of the material.

Three scan types are most commonly used: A Scan, B Scan and C Scan. A
Scan systems analyze signal amplitude along with return time or phase shifts
the signals travel between a specific surface and discontinuities. B Scan systems
88 USACERL TR 99/41

add signal intensity modulation and capability to retain video images. C Scan
Systems include depth gating to eliminate unwanted returns.

UT inspection is a deliberate process covering a small area (4 to 8 sq in.) at each

sampling. Consistency in test method and interpretation of results is critical to
the reliable test results. Surface preparation is also critical to reliable UT re-
sults. Any surface defects such as cracks, corrosion, or gouges will adversely af-
fect the reliability of UT results.

Due to the time and effort involved in surface preparation and testing, UT in-
spections are often conducted on representative samples of materials subjected
to high stress levels, high corrosion areas and large welds. By evaluating the
same sites at regular intervals, one can monitor the condition of the material.
One hundred percent UT inspection is typically reserved for original construction
of high stress components such as nuclear reactor vessels or chemical process
vessels.

Magnetic Particle Testing

The NDT technique uses magnetic particle detection of shallow sub-surface de-
fects. Magnetic Particle Testing (MT) techniques are useful during localized in-
spections of weld areas and specific areas of high stress or fatigue loading. MT
provides the ability to locate shallow sub-surface defects. Two electrodes are
placed several inches apart on the surface of the material to be inspected. An
electric current is passed between the electrodes producing magnetic lines.
While the current is applied, iron ink or powder is sprinkled in the area of inter-
est. The iron aligns with the lines of flux. Any defect in the area of interest will
cause distortions in the lines of magnetic flux, which will be visible through the
alignment of the powder. Surface preparation is important since the powder is
sprinkled directly onto the metal surface and major surface defects will interfere
with sub-surface defect indications. Also, good electrode contact and placement
is important to ensure consistent strength in the lines of magnetic flux.

A major advantage for MT is its portability and speed of testing. The hand-held
electrodes allow the orientation of the test to be changed in seconds. This allows
for inspection of defects in multiple axes of orientation. Multiple sites can be in-
spected quickly without interrupting work in the vicinity The equipment is port-
able and is preferred for on-site or in-place applications. The results of MT in-
spections are recordable with a high quality photograph or transfer to tape.
Fixing compounds are available to glue the particle pattern in-place on the test
specimen. Interpretation of results depends on the experience of the operator.
USACERL TR 99/41 89

Dye Penetrant

Dye Penetrant is used to detect surface defects. Dye penetrant (DP) inspections
provide a simple method for detecting surface defects in nonporous materials.
DP allows large areas to be quickly inspected. Once the surface has been
cleaned, a penetrating dye (magenta or fluorescent color) is sprayed liberally on
the entire surface. The dye is allowed to penetrate for several minutes. The ex-
cess dye is then wiped from the surface leaving only the dye that has been drawn
into surface defects. A developer (usually white) is sprayed on the entire surface
(same area as the dye application). The developer draws the dye from the de-
fects, producing a visual indication of the presence of surface defects. The defec-
tive areas are then identified for repair and the remaining dye and developer are
removed.

Hydrostatic Testing

Hydrostatic Testing (Hydro) is an NDT method for detecting defects that com-
pletely penetrate pressure boundaries. Hydros are typically conducted prior to
the delivery or operation of completed systems or subsystems that act as pres-
sure boundaries. As the name implies, hydrostatic tests fill the system to be
tested with water or the operating fluid. The system is then sealed and the pres-
sure is increased to approximately 1.5 times operating pressure.

This pressure is held for a defined period. During the test, inspections are con-
ducted to find visible leaks to well as monitor pressure drop and make-up water
additions. If the pressure drop is out of specification, the leak(s) must be located
and repaired. The principle of hydrostatic testing can also be used with com-
pressed gases. This type of test is typically called an air drop test and is often
used to test the integrity of high pressure air or gas systems.

Eddy Current Testing

Eddy current testing is used to detect surface and shallow subsurface defects.
Also known as electromagnetic induction testing, eddy current testing provides a
portable and consistent method for detecting surface and shallow subsurface de-
fects. This technique provides the capability to inspect metal components
quickly for defects or homogeneity. By applying rapidly varying AC signals
through coils near the surface of the test material, eddy currents are induced
into conducting materials. Any discontinuity that affects the materials electrical
conductivity or magnetic permeability will influence the results of this test.
Component geometry must also be taken into account when analyzing results
from this test.
90 USACERL TR 99/41

A set of magnetizing coils are used to induce electrical currents (eddy currents)
into the component being tested. The induced currents produce magnetic fields,
that are then detected by a set of sensing coils. Typically, the two sets of coils are
combined into a single probe. In some systems, Hall effect devices are used in-
stead of sensing coils. The frequency of the AC signal used (5 to 10 MHz) deter-
mines the depth of penetration through the material for the eddy currents.
Lower excitation frequencies increase the penetration depth and improve effec-
tiveness in detecting deeper defects. Higher frequencies are used to enhance de-
tection of surface defects. Analysis equipment senses several parameters in-
cluding magnitude time lag, phase angles, and flow patterns of the resulting
magnetic fields. Automated analysis methods reduce reliance on operator expe-
rience for consistent results.

Location and Intervals

Before implementing an NDT program, a formal plan should be developed de-

tailing the type of technique to be used, location, frequency, number and orienta-
tion of samples, information to be gained from each sample, and failure mode
that each sample addresses. This plan will point out excessive testing as well as
omissions from the NDT Program. Two of the more difficult variables to address
are location and interval (time period) between inspections.

Intervals

When establishing sample intervals or frequency, several factors must be

weighed: operating cycle of the system, historical failure rate, type of container
material, type of contained substance, chemistry control, major corrosion mecha-
nisms, expected corrosion rate, erosion mechanisms, expected erosion rate, con-
sequences of system breach, and type of NDT techniques applicable to the situa-
tion. All these will affect the inspection interval (Table 2).

Table 2. Recommended maximum inspection intervals (API 570).

Piping Circuit Thickness External Visual

Classification Measurements Inspection
Class 1 5 years 5 years
Class 2 10 years 5 years
Class 3 10 years 10 years
Injection points 3 years By class
Soil-to-air interfaces N/A By class
USACERL TR 99/41 91

American Petroleum Institute (API 570) recommends the following criteria for
establishing intervals for NDT inspection: a piping service classification where
Class 1 has the highest potential of resulting in an immediate emergency if a
leak were to occur, any Class 3 has the lowest potential of resulting in an imme-
diate emergency.

Relevant regulatory requirements must be taken into account when determining

NDT inspection intervals. The multitude of professional codes and government
regulations present this section from attempting to cover specific regulations.
Many government regulations provide sufficient leeway for the experts within an
organization to set intervals in accordance with technically sound methods.
Some regulations simply require a technically sound plan that the organization
follows.

Before accepting what seems like an unreasonable interval simply in the name of
regulatory compliance, investigate the document that originates the require-
ment. In industry, many regulatory requirements have been needlessly made
more stringent by the philosophy, If a little is good, a lot must be better. In an
effort to avoid falling afoul of regulatory inspectors, inspection costs were signifi-
cantly increased without a corresponding increase in plant safety or reliability.
Investigate basic requirements and, if these are unclear, ask the originating
agency for clarification of their expectations.

After the base inspection intervals have been established based on corrosion
rate, class, and regulatory requirements, specific system intervals can be modi-
fied based on actual conditions, historical data, and operating parameters.
Evaluate intervals based on operating conditions, previous history, current in-
spection results, or other indications of other than normal conditions. By con-
ducting statistical analysis on historical NDT results and failure rates, intervals
can be refined with a higher level of confidence. Pareto and Weibull analysis
techniques can be applied to indicate systems where unusual failure rates are
occurring. Corrosion coupons can be used to provide specific information on the
corrosion rate of systems, allowing further refinement of inspection intervals.
Process parameters can be used as triggers for specific NDT inspections. As
thermodynamic properties change, they can provide indications of increased cor-
rosion product deposits. Analysis of fluids transported within the system can also
indicate changes in corrosion activity, allowing NDT inspection schedules to be
appropriately adjusted. Procedures for addressing adverse events such as over-
pressurization and out-of-specification temperatures should include the require-
ment for more frequent or immediate NDT inspections. Details of type, location,
parameter of concern, and acceptable value should be indicated to facilitate a
safe and expeditious recovery from the incident.
92 USACERL TR 99/41

Locations

The following should act as guidelines for locating NDT sampling points:
Presence of Dead-Heads that can create turbulence or stagnate areas where
material may accumulate and set up corrosion cells,
Junctions of dissimilar metals. Galvanic corrosion is prevalent in these areas
unless specific steps are taken to prevent it.
Abrupt changes in direction of flow (elbows) and changes in pipe diameter will
cause turbulence that may accelerate many corrosion mechanics.
Stressed areas, welds, high stress fasteners and areas that undergo cyclic
temperature/pressure changes, or flow changes.
Some applications may warrant specifying top, middle or bottom of pipe or
areas of where more than one phase of a substance is present.
Areas where accelerated corrosion/erosion mechanisms have been identified
Areas susceptible to cavitation.

Applications

1. Radiography. Radiographic techniques are readily applicable to metal compo-

nents, including weld deposits. Specialized applications for plastics or composite
materials are possible, though typically these materials are not most economi-
cally inspected with radiography. For thick cross-sections, radiography is often
the only reliable method for inspection.

2. Ultrasonics. UT techniques are applicable to metal components including weld

deposits. Specialized applications for plastics or composite materials are com-
mon. When possible, UT is a preferred method over radiography for in-place ap-
plications, due to expense and safety precautions required by radiography. UT is
especially useful since it only requires access to one surface of the material. Ul-
trasonic techniques provide excellent penetrating power for thick cross-sections.

3. Magnetic Particle. MT techniques are applicable only to materials that conduct

electric current and magnetic lines of flux. Only shallow defects are detectable
with MT inspection. Typically, these techniques are most effective on welded ar-
eas. The speed of testing allows multiple inspections to be conducted along dif-
ferent axes to detect defects in different orientation planes.
USACERL TR 99/41 93

4. Dye Penetrant. DP inspections are applicable for any nonporous material that is
chemically compatible with the dye and developer. This is the simplest NDT
technique in which to gain proficiency.

5. Hydrostatic Testing. Hydros test the integrity of pressure boundaries for compo-
nents and completely assembled systems that contain pressurized fluids or gases.
Identification of defects that penetrate the entire pressure boundary is the pri-
mary application for hydrostatic testing.

6. Eddy Current. Eddy current techniques are used to detect internal defects such
as cracks, seams, holes, or lamination separation (on both flat sheets and more
complex cross-sections ), as well as monitoring the thickness of metallic sheets,
plates and tube walls. Portable systems arc used extensively in the condition
monitoring of installed heat exchanger and chiller tube wall thickness. Where
coating thickness is an important factor, there is sufficient difference in electrical
or magnetic properties between the base material, and the coating. Eddy Cur-
rent Testing can determine the actual coating thickness. In more production ori-
ented applications, installed systems can determine material composition, uni-
formity, and thickness of materials being produced.

Limitations

1. Magnetic Particle. MT techniques are applicable only to materials that conduct

electrical current and influence magnetic lines of flux. The difference in the in-
fluence of the lines of flux between base material and the defect is the basis for
MT inspection. Only small areas (30 sq in.) between the two electrodes can be in-
spected. Surface preparation is important, though not as critical as with UT.
Consistent electrode contact is critical. Loose contact will weaken the magnetic
lines of flux to the point where the influence of a defect may not be visible in the
filing pattern. Operator skill is important, though this is a relatively simple
technique. No historical record is produced for each test, unless specific steps are
taken to photograph the result of each test.

2. Ultrasonics. UT techniques are one dimensional. Unless special techniques are

applied., defects that parallel the axis of the test will not be apparent. Compo-
nents constructed using laminate techniques or layered construction present spe-
cial problems for UT techniques, since the boundary between each layer may be
interpreted as a defect. The thicker the layers of base material, the more likely
UT will provide usable results.

3. Radiography. Effective use of radiography mandates expensive equipment, ex-

tensive safety precautions and skilled technicians to interpret the images. Ex-
94 USACERL TR 99/41

pensive tracking and security for radiation sources is mandatory. Safety precau-
tions often demand evacuation of areas adjacent to the piece being examined or
installation of extensive shielding. Even with these limitations, radiography is
often the most effective method of assuring integrity of critical welds, structural
members, and pressure boundaries. As material thickness increases, radiogra-
phy is often the only acceptable method to achieve a 100 percent penetration.

4. Dye Penetrant. Minute surface discontinuities such as machining marks will be-
come readily apparent. The inspector must be trained to distinguish between
normal surface discontinuities and defects that must be repaired. The dye and
developer are usually sprayed or painted on the piece to be inspected, so over-
spray and protection of internal surfaces are prime concerns for systems with
stringent chemistry and cleanliness control. Product cleanliness standards may
prohibit the use of DP inspection.

5. Hydrostatic Testing. Cleanliness and chemistry control of the fluid must be con-
sistent with the operating standards of the system. Close attention should be
given to controlling system thermodynamic parameters during the test to pre-
vent overpressurization of the system. Overpressurization could lead to unin-
tended damage to the system. Individual component hydros do not ensure sys-
tem integrity. A final hydro of the completed system is used to ensure the
integrity of the assembled systems pressure boundary.

6. Hydros will not identify defects that are present, but have not completely pene-
trated a pressure boundary. The pressure applied to the system is generally not
sufficient to enlarge existing defects to the point of detection by the test. Hydro-
static testing requires a pressure source capable of expeditiously filling and pres-
surizing the system, extensive instrumentation and monitoring equipment, along
with a sufficient quantity of fluid to fill the system. A method of isolating pres-
sure relief devices and connecting the pressure source to the system must be
provided.

7. Eddy Current - Eddy Currents tend to flow parallel to tile surface to which the
exciting field is applied. Some orientations of laminar discontinuities parallel to
this surfaces tend to remain undetected by this method. Eddy Current Testing
will not penetrate deeply into the material of interest, and so is limited to shallow
subsurface and surface defects.
USACERL TR 99/41 95

11 Conclusions
Management of maintenance activities at facilities on military installations is a
complex and expensive task. This report presents a variety of techniques that
can monitor equipment condition and anticipate failure. For some noncritical,
inexpensive, and easily replaced components, run-to-failure method may be an
acceptable practice. For large, complicated, expensive, mission-critical items,
run-to-failure may be unacceptable. Maintenance to maximize service life of
equipment or components and surveillance of performance degradation can allow
repairs/replacement without interruption of mission-critical activities. For cer-
tain installations, it may be more economical to use contract services to maintain
infrequent, complex, and expensive equipment and processes.
96 USACERL TR 99/41

References

American Society of Non-Destructive Testing, Recommended Practice Number SNT-TC-1A (Colum-

bus, Ohio, 1992).

American Society for Testing and Materials (ASTM), Annual Book of ASTM Standards, Section 5
(Philadelphia, PA, DATE).

API670, Vibration, Axial-Position Bearing-Temperature Monitoring System (American Petroleum

Institute, New York, NY, DATE).

Design and Manage Life Cycle Cost (M.A. Prace, Farmer Grove, OR, 1978).

General Motors Specification A 1.0, Laser Alignment Specification for New and Rebuilt Machinery
and Equipment (1993).

Harris, Tedric A., Rolling Bearing Analysis Second Addition (John Wiley & Sons, New York, NY,
DATE).

Infraspection Institute, Guideline for Infrared Inspection of Building Envelopes and Insulated
Roofs (Shelburne, VT, DATE).

Infraspection Institute, Guideline for Infrared Inspection of Electrical and Mechanical Systems
(Shelburne, VT, DATE).

Infraspection Institute, Guideline for Measuring and Compensating for Reflected Temperature,
Emittance and Transmittance (Shelburne, VT, DATE).

Lundberg, G., and A. Palmgren, Dynamic Capacity of Rolling Bearings, Acra Polytech, Mechanical
Engineering Series 1, R.S.A.E.E., No. 3, 7 (1947).

National Aeronautical and Space Administration (NASA), Reliability Centered Maintenance Guide
for Facilities and Collateral Equipment (December, 1996).

Nowlan, F.S. and H.F. Heap, Reliability-Centered Maintenance (Dolby Access Press, San Francisco,
CA 1978).
USACERL TR 99/41 97

USACERL DISTRIBUTION

Madigan Army Medical Center 98433-4000

ATTN: (2)

Chief of Engineers
ATTN: CEHEC-IM-LH (2)
ATTN: CEHEC-IM-LP (2)
ATTN: CECC-R
ATTN: CERD-L
ATTN: CERD-M

Defense Tech Info Center 22304

ATTN: DTIC-O (2)

11
11/96