THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS

345 E. 47th St., New York, N.Y. 10017

The Society shall not be responsible for statements or opinions advanced In
papers or discussion at meetings of the Society or of its Divisions or Sections,
947G 1.452
or printed In its publications. Discussion Is printed only if the paper is pub-
lished in an ASME Journal. Papers are available from ASME for 15 months
after the meeting.
Printed in U.S.A. Copyright © 1994 by ASME

ON-LINE DETERGENT FLUID EVALUATION

ON A TF4OB GAS TURBINE ENGINE

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

Howard Harris
Naval Surface Warfare Center
Philadelphia, Pennsylvania

c-
Marc Calabrese
Naval Surface Warfare Center 111 1 M BREAKRU 01111
Philadelphia, Pennsylvania

ABSTRACT

This paper presents results, evaluation, and comparisons

between eleven on-line water wash detergents and water that were
tested on a TF4OB gas turbine engine. Comparisons were based
on test cell data, field use data, environmental requirements, cost
and U.S. Navy specification MIL-C-85704 Rev B, "Cleaning
Compound, Turbine Engine Gas Path." Detergents tested were:
Turbotect 920 and 950, B&B TC-100, ZOK-27, R-MC,
Fyrewash WB and SB. Turco 6783-50, Turboclean 2, EDC R-
2701 and TEC 20. Fyrewash SB was the only solvent based on-
line detergent tested. All other detergents were aqueous based.

INTRODUCTION

TF4OB Gas Turbine Compressor Description

The compressor assembly consists of seven axial stages and a

single centrifugal stage, mounted on a hollow shaft supported by
the numbers one and two bearing assemblies. The compressor
housing completely encloses the compressor rotor, and contains
seven stator assemblies which serve as diffusers and guide vanes
for each axial rotor stage, figure I.

The compressor blade material is a martensitic high-chromium Figure 1 TF4OB COMPRESSOR

stainless steel similar to AM 350 except for a slightly higher
carbon content.

Detergents Tested

Tables 1 and 2 lists the detergents and data summary for each
detergent that were tested on the TF4OB gas turbine at our land
based test site.

Presented at the International Gas Turbine and Aeroengine Congress and Exposition
The Hague, Netherlands — June 13-16, 1994
 TABLE 1
DETERGENTS TESTED

PRODUCT MANUFACTURER BASE DILUTION COST COST OF COST MIL-C- NSN

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

IDENTIFICATION RATIO CONC. SOLUTION PER 85704
PER WASH QPL
GAL[2] 3 GALS. 13]
$ $ $

ZOK-27 AIRWORTHY LTD. WATER 1:4 21.20 5.30 15.90 NO 7930-00-

ELSTEAD 340-
ENGLAND 2176

R-MC ECT INC. WATER 1:6 27.20 6.80 20.40 NO NO

KING OF
PRUSSIA, PA

FYREWASH WB ROCHEM, INC. WATER 1:4 22.00 5.50 16.50 NO NO

KETTERING, OH

FYREWASH SB ROCHEM, INC. SOLVENT 1:4 22.00 5.50 16.50 NO NO

KETTERING, OH

TURBOTECT 920 SPRAYTEC, INC. WATER 1:4 15.00 3.75 11.25 NO NO

BROOKFIELD, CT

TURBOTECT 950 SPRAYTEC, INC. WATER 1:4 15.60 3.90 11.73 NO NO

[1] BROOKFIELD, CT

TURCO 6783-50 TURCO, INC. WATER 1:4 10.50 2.63 7.88 YES NO
WESTMINSTER, Type II
CA

TC-100 B&B TRITEC, INC WATER PRE-MIX n/a 6.50 19.50 YES NO
MIAMI, FL Type
IIA

TURBOCLEAN 2 TRAFFIC AIR WATER 1:4 11.90 3.00 9.00 NO NO

ALPHARETA, GA

EDC-R-2701 EDC WATER PRE-MIX n/a 5.73 17.19 NO NO

SYOSSET, NY

TEC-20 NAWC WATER 1:4 n/a n/a n/a YES • NO

WARMINSTER, Type
PA HA

Notes:
[1). Same as Octagon SW 2-43-C
[2]. Cost based on one 55- gallon drum
[3]. Qualified Products List (QPL) approved detergents as of August 2, 1993.

2
 TABLE 2. DATA SUMMARY FOR EACH DETERGENT TESTED

PRODUCT PERCENT PERCENT % NON- VOC % WEIGHT VAPOR SPECIFIC pH FLASH

IDENTIFICATION WATER SOLVENT VOLATILE CONCENTRATE PRESSURE GRAVITY POINT
Irnm Hg) • °F

ZOK-27 77.5 5.2 17.3 5.4 0.6 1.018 7.7 >160

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

R-MC 98 none 2.0 neg.111 24 . 1.017 7.1 none

FYREWASH WB 47 <1 52 0 >5 0.97 9 >200

FYREWASH SB 16.7 <50 21.62 0 >1 0.951 8.0 >140

TURBOTECT 920 70 none 30 <5 not known 1.02 7.1 >212

TURBOTECT 950 61.5 TBD 23.2 9.89/1121 1.8 1.01 7.2 none

TURCO 6783-50 53.2 <15 31.5 5.75 15 1.016 8.5 none

TC-100 94 none 3.62 <5 0.01 1.00 7.97 >200

TURBOCLEAN 2 86 none 22.0 0 17.5 1.023 7.8 none

EDC R-2701 TED TED TBD <5 770 1.1 TBD none

TEC-20 95.2 <5 1.17 <5 18 1.0 7.5 >212

Notes:
111. Volatile organic compound (VOC) level negligible: Although the product contains organic material it is not volatile at
110 'C called out for determining volatile material content.
121. Used amended rule 1122 of the South Coast Air Quality Management District.

How Detergents Work A sixteen gallon holding tank was used to mix the detergent and
deionized water. A low pressure air line in the test cell was used
All aqueous detergents work in a similar manner. Detergent to pressurize the holding tank to 65 psig and was monitored
(soap) has two parts: one is a long, straight hydrocarbon chain locally by a pressure gauge. The system was activated by
that is oil soluble and the other is the water soluble sodium or depressing a latching push button installed in the engine control
potassium salt of an acid. console. When the push button was depressed, the solenoid valve
opened, allowing the detergent solution to flow from the tank
Oily or greasy dirt adheres firmly to the substrate and is through a 15 micron strainer and branching into the two nozzle
undisturbed by washing in water alone. Soap acts as an blocks. The solution was atomized by the four on-line nozzles.
emulsifying agent and brings the dirt into colloidal dispersion. The duration of each wash was three minutes.
Without the emulsifying agents the droplets would not combine
and the two liquids would separate. The hydrocarbon end of the The on-line detergent wash nozzles were manufactured by
soap molecule is soluble in the oil or grease. The anionic end is Spraying System Co.(pan number H-1/4-VV-SS-1502). The
not soluble in the oil or grease, but protrudes from the colloidal nozzles were modified by adding a sleeve providing a slot for the
droplet, making them water soluble and repelling other similar set screw to hold the nozzle in the nozzle block. These nozzles
droplets, thus preventing coalescence, figure 2. emit a flat shape pattern with an orifice size of 0.036 inches, and
a flow capacity of 0.25 gallons per minute at 60 psig. The flat
During an on-line detergent wash, the water acts as a carrier spray pattern distributes the liquid as a flat fan. The atomized
which moves the small droplets of oily dirt from the compressor spray droplets at 60 psig have a droplet mean volume diameter
blades to the combustor. Once in the combustor, the water of 800 microns.
flashes and the droplets of oily dirt burn with the fuel.
This nozzle block was designed to provide placement for the two
on-line detergent wash nozzles and two detergent crank-wash
Test Cell On-Line Detergent Wash System nozzles. The location of the nozzle blocks on the TF4OB gas
turbine is shown in figure 3. There are two blocks per engine.
This system was designed to spray detergent solution at a
constant pressure into a running TF4OB gas turbine engine by To verify that the on-line spray nozzles completely covered each
remote operation. compressor blade from root to tip, the compressor case was split
open and water soluble green lines were drawn on all compressor
•
3
 At each power level, engine performance data was recorded by
0 OIL • SOLIIBLE END
the data acquisition system. Engine performance data was
corrected to standard day conditions (80 ° F, 29.92 inches of
WATER SOLUBLE
ANIONIC PIM mercury). This data was used in comparing engine performance
(a) A MOLECULE OF bras= IC DETERGENT when the engine was "clean" and when the engine completed five
on-line detergent washes.

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

TABLE 3. BASELINE CALIBRATION RUN PROFILE

% Ni % N2 TIME

60 25 4 MINUTES
SUBSTRATE
(o) ACTION OF SYNTH:NM =ENGEM 70 35 4 MINUTES
ON OILY DI=
co.; 80 45 4 MINUTES

85 80 4 MINUTES

90 92 4 MINUTES
CO2
95 84 4 MINUTES
Cc) Liatue tattEasznarar GLOBUI-E IS
DRONER DM SMALL DROPLETS (MICELLES)
100 91 4 MINUTES

Ni- Gas producer speed

N2- Power turbine speed

Video Borescope Inspection

(a) KICELLW ARE DISPERSED 24 soma
A video borescope inspection was done to record the condition
of the engine components before and after each fluid tested.
Figure 2 DETERGENT ANALYSIS

blades from tip to root. A borescope inspection showed the green Test Run For Each Detergent Tested
lines were removed by on-line detergent washing.
The engine was operated continuously at 90% NI and 80% N2
Test Operation with salt and carbon ingestion. When a 4 percent drop in
compressor pressure ratio occurred, an on-line detergent wash
To simulate the operating conditions by the TF4OB in fleet use, was performed. Each on-line wash was for three minutes, with
this engine was subjected to various contaminates. Salt water a total of three gallons of solution being ingested into the engine.
was fed into the running TF4OB engine by a nebulizer. Salt There were a total of five on-line detergent washes per fluid.
water was made to specification MIL-E-17341C. Since there was Prior to each fluid tested, the TF4OB engine was detergent crank-
no inlet filtration system in the test cell, airborne contaminates washed with Eldorado ED-563 detergent (MIL-C-85704, Type I
were ingested into the engine during operation. In addition, a approved solvent detergent).
peripheral gap was present where the exhaust collector duct
exited the test cell. This caused exhaust gases to back-flow into The on-line detergents were tested in the following order:
the test cell, resulting in exhaust carbon re-ingestion. The I. Fyrewash WB
amount of contaminates entering the engine was not a concern, 2. Turbotect 920
because the purpose of this test was to enhance compressor 3. R-MC
fouling as quickly as possible, and to verify if the detergents 4. ZOK-27
could clean the engine. 5. B&B TC-100
6. EDC R-2701
Baseline Calibration Run Profile 7. Turboclean 2
8. Deionized water
A calibration run at seven power levels was done at the 9. Fyrewash SB
beginning and end of each fluid test. Table 3 list the sequence of 10. Turbotect 950
the baseline calibration run. 11. Turco 6783-50
12. NAWC TEC 20

4
Test Criteria

Two operating conditions would terminate

the test These were:

I. High exhaust gas temperature (EGT),

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

greater than 1225° F
2. Compressor surge

The minimum performance criteria used to

determine if the fluids were successful in
cleaning the engine were:

I. Compressor discharge pressure (P 3) had

to increase after each on-line detergent wash.
2. LOT had to decrease after each on-line
detergent wash.
3. Shaft horsepower (SHP) had to increase
or remain the same after each on-line
detergent wash.
4. A 4 percent compressor pressure
degradation from initial to final baseline runs.
5. Contaminates could not redeposit in the
latter stages of the compressor, nor in the hot
section of the engine (as determined by
borescope inspection.)

RESULTS
Figure 3 NOZZLE BLOCK LOCATION

Two methods were used to determine how

effective the fluids were in cleaning the TF4OB gas turbine cleaning salt and carbon from the compressor blades.
engine. These methods were:
I. Video borescope inspection 3. R-MC: The fluid was successful in cleaning salt on the
2. Engine dam analysis compressor blades, however, light carbon deposits were
observed.
Video Borescope Inspection
4. ZOK-27: The fluid was not completely successful in cleaning
The inspection included examination of the following salt and carbon from the compressor blades.
components:
I. Compressor assembly- Stages 1,2,3,4,5 (upstream 5. TC-100: The fluid was successful in cleaning salt from the
side), 6 (downstream side) and 7 (upstream side). compressor blades. Medium to heavy carbon deposits were
2. Gas producer assembly- First stage gas producer found on the compressor blades. Heavy deposits of detergent
nozzle vanes and first stage gas producer disc blades. were found on the leading edges on the 7th stage blades.
3. Combustor assembly- Liner, curl, fuel nozzles and
swirler assemblies. 6. EDC R-270I: The fluid was successful in cleaning salt from
the compressor blades. Traces of carbon deposits were found on
An evaluation between the baseline and follow up borescope several stages of compressor blades. Traces of detergent deposits
inspections for the eleven detergents and water showed the were found throughout the compressor assembly.
following:
7. Turboclean 2: The fluid was successful in cleaning salt from
I. Fyrewash WB: The fluid was successful in cleaning carbon the compressor blades. Traces of carbon deposits were found in
and salt from the compressor blades. The blades were coated several stages of compressor blades. Traces of detergent deposits
with a light film of the solution. were found throughout the compressor assembly.

2. Turbotect 920: The fluid was not completely successful in 8. Water: Light to medium carbon deposits found throughout

5
the compressor assembly. Heavy salt deposits were found degrade from initial to final baseline runs. Figure 6 shows that
throughout the compressor assembly. SFC dropped after each on-line wash.

9. Fyrewash SB: The fluid was successful in cleaning salt from

the compressor blades. Medium to heavy carbon deposits were Field Data
found throughout the compressor assembly. Traces of detergent

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

deposits were found on several stages of compressor blades. Landing Craft Air Cushion (LCAC) 012, 018 and 001 have all
been equipped with an on-line detergent wash system. The LCAC
10. Turbotect 950: The fluid was successful in cleaning salt is powered by four TF408 gas turbines. As of December 1993,
from the compressor blades. Traces of carbon deposits were ZOK-27 was the only detergent used on all three craft. The
found in several stages of compressor blades. Traces of detergent results of each craft are:
deposits were found on several stages of compressor blades.
LCAC 012
11. Turco 6783-50: The fluid was successful in cleaning salt
from the compressor blades. Medium to heavy carbon deposits Over 700 operating hours without a detergent crank wash
were found in several stages of compressor blades. Traces of Two hot end inspection showed no evidence of sulfidation
detergent deposits were found on several stages of compressor Borescope inspection showed only light deposit of carbon and
blades. salt
Crew response was positive
12. Naval Air Warfare Center TEC-20: Medium to heavy
carbon deposits were found in several stages of compressor
blades. Medium salt deposits were found on several stages of LCAC 018
compressor blades.
Over 150 operating hours without a detergent crank wash
Borescope inspection showed only traces of carbon
Engine Data Analysis Crew response was positive

Compressor fouling, and hence engine performance LCAC 001

deterioration can be detected by changes in compressor shell
pressure. When the compressor is fouled the inlet mass flow is Over 150 operating hours without a detergent crank wash
reduced. As a result the engine operates at a lower pressure ratio. Borescope inspection revealed only traces of carbon
Therefore, the amount of compressor fouling can be deduced at Crew response was positive
any time by comparing the actual site combustor shell pressure to
the expected pressure if the engine were clean flhor, 19911. Safety and Environmental Requirements
During this test, the compressor shell pressure ratio was Pa o.
P3 is the compressor discharge pressure, and P r, is the ambient All detergents are equally safe when used as specified. The
pressure. Both parameters were measured and recorded during most serious concerns are about possible dermatitis or minor
this test. respiratory discomfort. These should not be serious problems if
the detergents are diluted to the ratio of water specified.
EGT, SHP, and fuel consumption were the other engine
parameters that were measured and recorded. San Diego Air Pollution Control District (SDAPCD) monitors
and issues environmental permits for "gas path cleaners" used in
Table 4 summarizes the results of each fluid by comparing final San Diego County. This district includes Camp Pendelton, CA
baseline performance to initial baseline performances and by where Assault Unit Craft Five (ACU 5) is located. ACU 5 is the
comparing before-wash performance to performance two minutes station for the LCACs on the west coast. SDAPCD uses "Rule
after each on-line wash. 67.6- Solvent Cleaning Operations" to determine if a solvent gas
path cleaner requires a permit.
Three graphs were used to compare and analyze how each fluid
effected engine performance by on-line detergent washing. The Rule 67.6 states, "a gas path cleaner (corrosion control carts)
graphs were: means equipment which applies solvent to the interiors of gas
1. Corrected EGT and Corrected SHP vs Corrected NI turbines or jet engines for removal of corrosion or combustion
2. Compressor Pressure Ratio (P3/Pc,) vs Corrected Ni deposit," and a "solvent is any liquid containing more than ten
3. Specific Fuel Consumption (SFC) vs Wash percent by weight of organic compounds and which is used to
dissolve, clean, strip, or remove impurities, coatings, stains, or
Figures 4, 5, and 6 show the performance of EDC-R-270I film from surfaces."
detergent. Figure 4 shows that corrected EGT and SHP versus
corrected Ni decreased slightly from initial to final baseline runs. A permit is required of any solvent gas path cleaner if the volatile
Figure 5 shows that the compressor pressure ratio (1 33/1%) did not organic compounds (VOC) exceeds 10 percent by weight as used.

6
 Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024
TABLE 4. TF4OB GAS TURBINE ENGINE DATA ANALYSIS

FLUID AVG % AVG % AVG % AVG AVG % AVG AVG % AVG %

CHANGE IN CHANGE IN CHANGE IN INCREASE IN CHANGE IN EGT CHANGE IN CHANGE IN CHANGE IN
P3/1% FROM P31P0 2 SHP FROM SHP 2 MINS. (°F) FROM EGT (°F) 2 SFC FROM SEC 2 MINS.
INITIAL TO MINS. INITIAL TO AFTER EACH INITIAL TO MINS. AFTER INITIAL TO AFTER
FINAL AFTER FINAL ON-LINE FINAL EACH ON- FINAL EACH ON-
BASELINE EACH ON- BASELINE WASH BASELINE LINE WASH BASELINE LINE WASH
LINE WASH

+ Optimum + Optimum + Optimum + Optimum - Optimum - Optimum - Optimum

- Optimum

TURCO 6783-50 -2.63 +0.97 -7.30 +22 +3.37 +1.8 12) +5.14 -0.38

FYREWASH SB -1.51 +1.19 -5.64 +45.2 +3.32 +8.2 121 +5.85 -0.84

FYREWASH WB -0.88 +0.03 +1.00 +6.1 +3.22 +5.9 121 0.00 +0.28

EDC R-2701 0.00 +1.91 -2.26 +92.1 +1.04 -3.85 +1.34 -2.64

R-MC -1.69 +0.95 -3.26 +1.5 +0.68 -1.8 -1.87 0.00

TC-100 -3.04 +1.60 -9.38 +38.5 +3.60 -9.3 +4.99 -1.73

TEC-20 -3.43 +1.38 -5.49 +36.6 +3.45 -0.8 +9.36 -0.94

TURBOCLEAN 2 -1.42 +1.66 -2.30 +69.8 -2.25 -4.4 -1.03 -2.48

TURBOTECT 920 0.00 +0.43 0.00 +4.6 +0.55 -4.6 -0.47 -0.02

TURBOTECT 950 -2.95 +1.14 -7.06 +42.7 +3.35 -1.3 +6.65 -1.20

WATER -3.65 111 +0.22 -7.36 +22.8 +2.27 +4.9 121 . +4.24 -0.58

ZOK-27 -0.46 +0.07 -0.68 +3.7 +2.07 -14.3 +2.28 0.00

Note:
111 Performed only three on-line detergent washes, due to a lack of CDP recovery.
12) Failed test criteria.
 TFACR DATA
112701 DEMME
CORRECTED COT • CORRECTED OW vic. • BASELINE RON
CORSEC1133141
7 NAL BASELINE

ma IIIflIIfl i

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

6
IMO Yi
umuunumi _
1003
OA
. -1.=.71151FA .
. i MIR . 4

. iuuiaaiu • .
. EMS OM 3

MO
1 W1 := LJ 1
65 CO 65 70 7b DO 95
CORRECTED 1.11 (IC RPM)
PO 95 103 IC

CORRECTED NI IR RHO DIM conriernanoso 0 F M Er OF It

FINAL 11A-SBJ WE OCOSED SIER I AMMO
DATACallia10 One F. MP N. SOW
Illiamain =MD An Seam

FIGURES
TROB BASELINE DATA
Figure 4 Corrected EGT & Corrected SHP EDC FISTO1 DETERGENT
vs Corrected NI P3/P0 vs CORRECTED NI

1ST WASH 2ND WASH 3RD WASH 4TH WASH 5TH WASH

7T4013 °FERRERO AT • 1 I4R BEORE WASH al 1 MINS AFTER WASH SEPT 23. 1993
" PO% NI, 9031N2
0 JUST BEFORE WASH Eg 1 HR AFTER WASH

JUST AFTER WASH

FIGURES
CORRECTED SFC vs. WASH
Er1C R-2701 DETERGENT

8
SDAPCD was provided with the necessary information to on-line detergent cleaning.
determine the VOC for each detergent tested. All eleven
detergents have a VOC limit below 10 percent as used. II. On-line detergent washing retains engine power.

MIL C 85704
- - 12. On-line detergent washing does prolong detergent crank
washing frequency.

Downloaded from http://asmedigitalcollection.asme.org/GT/proceedings-pdf/GT1994/78842/V002T03A014/2404135/v002t03a014-94-gt-452.pdf by guest on 12 March 2024

MIL-C-85704 Rev B is the military specification for "Cleaning
Compound, Turbine Engine Gas Path'. This specification 13. ZOK-27 detergent has worked well in field use.
establishes the requirements for two types of turbine engine
cleaning compounds used for starter cranked engine cleaning 14. The sequence of fluids tested did not have an effect on
operations. Type I is for solvent emulsion cleaner, Type II is for engine performance. That is, the fluids tested first did not have
aqueous cleaners and Type IIA is for ready-for-use aqueous an advantage to the fluids tested last.
cleaner. The military specification test for chemical composition
content, toxicity, corrosion, effect on materials, rinseability, FUTURE U.S. NAVY TEST
storage, cleaning efficiency, stability, and workmanship.
The U.S. Navy will be performing limited on-line detergent
As of August 2, 1993 only one Type II detergent (Turco 6783-50) wash testing on an Allison 501-KI7 gas turbine engine at our
had passed all MIL-C-85704 specifications. Two Type HA land-based test site. The test will compare the results obtained
detergents, Turco 6783-10 and B&B TC-100, have passed all from TF4OB testing to the Allison 501-K17 engine. The test
MIL-C-85704 specifications. There is currently no category for criteria and procedure will be similar to those of the TF40B.
on-line detergents in the MIL-C-85704.
Future plans also include a long-term shipboard evaluation on a
DD-963 Class destroyer. Two of the ships' four GE LM2500
engines and one of the three Allison 501-K17 engines will be
CONCLUSIONS outfitted with an on-line wash system. Based on the results of
previous tests, several detergents will be used. We will
frequently monitor the test with borescope inspections and engine
I. The on-line detergent wash nozzles provided complete performance data evaluation. The long-term performance
compressor blade coverage. Borescope inspection showed the degradation of the on-line washed engines will be compared to
green line was removed from each compressor blade. the engines that used only detergent crank wash.

2. Borescope inspection revealed TC-I00, TEC-20 and water

were not acceptable in cleaning carbon from the latter stages of REFERENCE
the compressor.
(II DIOR S. DIAICUNCHAIC, "Performance Deterioration in
3. Borescope inspection revealed all detergents were effective in Industrial Gas Turbines", ASME paper no. 9I-GT-228. June 91
preventing salt from accumulating an the compressor blades, with
the exception of water and TEC-20.

4. Turco 6783-50 and Fyrewash SB & WE failed the test criteria

to lower EGT after each on-line detergent wash.

5. Water failed two test criteria: it did not increase P3 after each
on-line detergent wash nor did it decrease EGT after each on-line
detergent wash.

6. EDC R-2701 detergent provided the best results in cleaning

the engine components.

7. Fyrewash SB (solvent based) did not perform any better than

any aqueous based detergent.

8. All detergents are safe when used as specified.

9. All detergents have a VOC content below 10 percent as used

10, MIL-C-85704 should to be updated to include detergents for