COMPLIANCE WITH EPA'S PS-11 CORRELATION REQUIREMENTS

USING THE SICK- FWE200 PM MONITOR

Donald Piening- Sr. Air Quality Scientist - E.ON U .S. Serv ces In c.
Jason Walkerson- Sr. Engineer, Environmental Affairs- E.ON U.S. Services, Inc.
Manfred Strornberg- Head of Product Management- SICK MAIHAK GmbH, Reute
SUMMARY
New performance specification and quality assur-
ance procedures for continuous monitoring of par-
tcula te matter (PM) mass concentrations went into
effect in January 2004 when the United States En-
vironmental Protection Agency (EPA) published
the final versions of its Performance Specification
11 (PS-11) and Procedure 2. Severa! commercially
available PM Continuous Emission Monitoring
Systems (PM CEMS), utilizing various measure-
ment techniques have been tested during the past
few years. The primary requirement of this new
instrumentation is the ability to correlate with gra-
vimetric measurements performed using standard
reference methods such as EPAMethod 17 or EPA
Method5.
-1-
In January 2005, a SICK Maihak FWE200
"scattered-light'' based PM CEMS, was installed at
Louisville Gas and Electric' s (LG&E) Mill Creek
Power Station as part of an evaluation test. The
initial PS-11 correlation test using EPA Reference
Method SB was conducted in April2005. Subse-
quently, in June 2005, a Procedure 2 Response Cor-
relationAudit (RCA) was performed using EPA
Reference Method SB to check the stability of the
established correlation curve. A forced unit outage
prevented completion of the RCA testing in June;
however, the testing was successfully completed in
September 2005 upon the unit's retum to opera-
tion.
The FWE200 demonstrated high reliability and
accuracy with its correlation values and overall
compliance with PS-11 and Procedure 2 require-
ments.
l. Status of Particulate Matter Monitoring in the US
The Federal Standards of Performance for Stationary Sources of Air Pollution has spedfied PM enssion
lints for many years but did not include the requirernent for the c:ontinuous measurement of PM rnass
concentrations. The continuous measurement of stack opacity was required instead.
The EPA first proposed the requirement for continuous PM monitors in 1996 in the New Hazardous Waste
combustor rule. Consequentl}" the EPA started severa! field test programa to evaluate commercially avail-
able PM monitoring systems. In 1995-1996, PM CEMS representing various tec:hnologies were tested at
several waste combustor and cement kiln facilities. Sorne of the tested PM CEMS were manufactured in
Germany where the development of continuous dust monitors started during the 1960s in response to envi-
ronmentallaws. Test results and extensive public discussions helped shape the preliminary proposals into
the regulatory requirements published in J anuary 2004. These regulations were issued in two documents:
40 CFR Part 60 Appendix B Performance Specification 11 - "Specifications and Test Procedures for Particu-
late Matter Continuous Emission Monitoring Systems at Stationary Sources" (PS-11), and 40 CFR Part 60
Appendix F Proc:edure 2 - "Quality Assurance Requirements for Particulate Matter Continuous Emission
Monitoring Systems at Stationary Sources" (Procedure 2).
PS-11 and Procedure 2 apply to any facility that is required to install and operate PM CEMS under provi-
sions of Title 40 of the Code of Federal Regulations. PS-11 and Procedure 2 will help to ensure successful
installation and operation of PM CEMS as well as good quality PM emission data on an ongoing basis.
Performance Specification 11 (PS-11)
Specifications and Test Procedures for Particulate Matter Continuous Emission Monitoring Systems at Sta-
tionary Sources, as amended to Appendix B of Part 60. PS-11 contains guidelines and procedures for selec-
tion, installation and certification of PM CEMS including minirnum performance limits and sample calcula-
tions
Basics of PS-11
... An initial installation and calibration procedure to confirm acceptability of the PM CEMS when it is in-
stalled and put into operation.
... A site-speci:fic correlation of the PM CEMS response vs. manual gravimetric reference method measu-
rements (e.g., EPA Methods 5 or 17). A minimum ofl5 valid test runs must be conducted. Each run
must consist of simultaneous PM CEMS and reference method measurements. Three different levels of
PM mass concentrations should be tested by varying process operation conditions.
... Capability of the PM CEMS to measure across the full range of PM concentrations expected from the
source. The PM monitor range must encompass normal PM levels through the PM emission limit.
Procedure 2
Quality Assurance (QA) Requirements for Particulate Matter Continuous Emission Monitoring System at
Stationary Sources, as amended to Appendix F of Part 60. Procedure 2 contains minimum requirements for
evaluating the effectiveness of quality control (QC) and quality assurance (QA) procedures and the quality
of data produced by the PM CEMS.
-2-
Basics of Procedure 2
~ Assessment of the quality of the PM CEMS data by estimating measurement accuracy.
~ Control and improvement of the quality of the PM CEMS data by implementing QC requirements.
~ Specification of daily drift checks and routine response correlation audits such as
Absolute Correlation Audit (ACA), Sample Volume Audit (SVA) (only if sample volume is required for
determination of PM), Response Correlation Audit (RCA) and Relative Response Audit (RRA) to ensure
the continued validity of the PM CEMS correlation.
PS-11 Correlation Requirement
A fundamental issue of PM CEMS is that commercially available instruments normally do not measure par-
ticulate mass and the flue gas volume directly. Based on their principies of operation, they provide an indi-
rect measure of particulate matter and therefore, each instrument must be calibrated against a direct meas-
urement of actual particulate emissions, e.g. EPAMethod Sor Method 17. PS-11 is based on the principie of
correlating PM CEMS responses relative to emission concentrations as determined by a reference method.
The term "correlation", as used in this paper is defined as follows:
~ Correlation: the primary mathematical relationship for precision tuning the PM CEMS output to a PM
concentration determined by a reference method.
~ Correlation coefficient is a quantitative measure of the association between the PM CEMS output and
the reference measurements. 1his differs from gaseous CEMS that reference to calibration gases of
known concentration. PS-11 allows correlation coefficients 0.85, and for low-emitting sources, 0 7 5
~ PS-11 and Procedure 2 require site-specific correlation of the PM CEM response vs. manual gravimetric
reference method measurements.
2. PM CEMS Technologies
Existing PM CEMS technologies are designed to sample continuously or by batch and use either in-situ or
extractive sampling systems. If entrained water droplets are present in flue gases, the PM CEMS must in-
elude a sample delivery and conditioning system that will heat the sample sufficiently above the dew point
to evaporate water droplets completely. Any droplets remaining in the sample stream will be measured as
particles and cause an error in the measuring result.
Sorne available PM measuring technologies are:
In-situ Technologies (dry gas only)
~ Transmissiometer
- standard double beam
- single beam
~ Scattered light
Extradive Technologies (dry or wet gas)
~ Extractive scattered light
~ Betaray
-3-
In-situ technologies are limited to dry gas applications only. Dry gas meaning the gas temperature is above
the dew point and there are no water droplets, or aerosols present. The presence of water droplets would
interfe.re with the measurement of PM and hence lead to an erroneous result. Extractive technologies can be
used for dry or wet gas applications. Normallft they are used for wet gas applications only.
Light scattering
Light scattering technologies utilize a light beam passing through a measured volume. A detector is
mounted at a predetermined angle to the light beam. PM causes incident light scattering. This scattered
light is detected by the highly sensitive detector and converted into an electrical signal. The intensity of the
measured scattered light is in parallel correlation to the PM. The forward or backward scatter light princi-
pies are used for continuous In-situ PM measurement.
BetaGauge
Beta Ray technologies utilize a partial gas stream extracted from the flue gas and drawn through a filter
tape. The dust quantity on the filter tape is measured by the weakening of the beta radiation passing
through the dust laden filter. To compensate for any time-related reduction of radioactivity and for the
variation of the filter material, the measurement of the absorption is taken before and after dust filtration.
The measured values are then compared.
Light Transmission/Opacity
Light Transmission principie is the classic method for Opacity and PM measurement. A light beam passes
through a defined cross section (e.g. stack) containing a dust laden .flue gas. The light is absorbed by the
particles. The ratio of the received light to the transmitted light is displayed as transmission. The opacity
and extinction (used for PM measurement) is calculated from the transmission.
The formula below shows the relationship between Transmission/ Opacity and Extinction. Instruments ha-
sed on transmission are therefore frequently called opacity monitors.
0-l-T
To get a parameter that is linear in relation to dust concentration we use what we call, Extinction.
1 = received light;
E = extinction or optical density
e = dust concentration
T = Transmission
be
, 1 1 -kcl
Lam rtBeersLaw: = o e
lo = emit.ted light
k = extinction coeffident or optical density coefficent
L = Path length
O=Opacity
The choice of technology used for PM measurement depends on the PM amount and the application condi-
tions.
-4-
3. The FWE200 PM monitoring system
The FWE200 (developed and manufactured by SICK MAIHAK GmbH, Germany) is a particulate matter
mass concentration measuring system that operates according to the scattered-light principie. It was intro-
duced to the market in the year 2000. The FWE200 is specially designed to measure PM concentration in
wet gas streams containing water droplets, typically occurring downstream of wet flue gas desulphuriza-
tion equipment.
Extractive scattered light measurements are used when the gas is below dew point and contains droplets.
This is nota suitable application for In-situ technology because all optical monitors "see" droplets and dust
in the same way and the droplets invalidate any dust measurements. A practica! way to solve the problem
is to extract the gas and to vaporize the droplets.
Operating principie
Figure 1 below shows the principie of operation of the FWE200. A pro be extracts a partial volume from the
flue gas, directing it to a thermocyclone. The thermocyclone improves the energy transfer from heater to
gas and also keeps particles suspended in the gas stream.The gas temperature is controlled to a constant
temperature of typically 320 F. The gas is measured with a standard scattered light instrument in a special
measuring cell, similar to those used for In-situ purposes. An air-driven eductor extracts the gas from the
stack. The eductor eliminates any contad between moving parts and the aggressive gas, and therefore re-
duces maintenance and improves monitor availability. A bypass valve is used in conjunction with the probe
nozzle diameter to adjust the sampling rate to a slightly over-isokinetic sampling condition at full-load unit
operation.
Thermocyclone
Sender/receiver
1- 1
Control unit

Status i g n l s ~ ___j
Output signals -==-.J
Power supply
Figure l. Operating principie of FWE200 (over-isokinetic)
-5-
Regulatory compliance
FWE200 features:
Only a small gas volum.e is need.ed. Therefore the extrac-
tive pump (or the eductor) can be comparatively small and
therefore less expensive.
Th.e smaller volwne requires less heating energy and there-
fore keeps the operatng cost down. System control and
data handling is done by the control unit and the FWE200
MEPA software.
Figu.re 2. FWE200 (installed)
1he FWE200 PM m.onitoring system has been TV (Teclmical Inspection Agency) approved for determina-
ti.on of dust in saturated gases in compliance with the
Directive 2001/80/EC on the limitati.on of emissions of certain pollutants into the air from large com-
bustion plants and the
Directive 2000/76/EC on the incinerati.on of waste.
Additionally, according to EPA's Office of Air Quality Planning and Standards, the FWE200 also complies
with PS-11 equipment crileria (i.e., PS-11 Section 6.0).
System chec:ks
Procedure 2 requires routine system checks on the PM CEMS to ensure proper operation of the system (i.e.,
optics, electronics, light source, detector, etc.). The minimum requirements that pertain to the FWE200 are
(1) th.e zero drift check that ensures stability of the response to the zero check value, (2) the upscale drift
check that ensures stability of the response to the upscale check value, and (3) a check of th.e systems optics
to ensure that the response has not been afiected by the condition of the optical components.
The FWE200 complies with these requirements by utilizing a built-in control cycle. The control cycle (illus-
trated in Figure 3) checks the sender 1 receiver unit automatically and. performs a contamination check as
well as a zero- and span check. The receiver unit is switch.ed to a reference position where it receives the
light beam directly from th.e laser source (100% light intensity). The reading is compared to the factory
value, recorded befare delivery of the system, or altematively to the value that has been recorded after the
last cleaning procedure. After calculating it with the corresponding correction factor (contamination value),
th.e data is stored. Therefore, a contamination of the opticalsurlaces is not influen.cing th.e measuring value.
-6-
-

Mel suEi g,va

,


Zero




1

1

1
1 1


__ .e,.
(',



r ._,
ACA capability
- -
ue
r--
--
r---.-
---
e o trol
1
---
-o



---
--
oint
1
---

..u ff
-
-.-
--
---
---
---
--
r---

















The control point (span check) is then determined by re-
ducing the light intensity of the laser diode to 70%. The
measured Iight intensity is compared to the expected value
and a flag is set in case of a deviatio11 greater than signal
clearly indicates any drift effects of thezero point.2%. The
zero point (zero check) is checked by switching the laser
diode off. The resulting signal clearly indica tes any drift
effects of the zero point.
Figure 3. FWE200 check
P5-11 requires the user to perfonn an "Absolute Correlation Audit" (ACA test) at least once each calendar
quarter to assess the continuous accuracy of the PM CEMS. In the ACA test, the PM CEMS must be chal-
lenged to reproduce measurements using audit standards that cover at least three points in the measuring
range of the instrument. The response of the CEMS to the standards is evaluated. The FWE200 can perform
this test with its "Check Filter Set". This consists of a variety of optical components that can be inserted into
the light path of the instrument when. it is removed from the stack.
As seen in Figure 4, a softening plate, an ab-
sorption filter and a scattering plate are posi-
tioned in the optical measuring volume to
simula te a scattered light value, set to 100%.
Additional reference filter elements are then
inserted to attenuate the nonnalized scat-
tered light value. These filter elements ha ve
various defined transmission values. The
measured responses of the PM CEMS are
compared to the values of the reference fil-
ters.
Filter with a defined
transmission value
Scattering plate
Sender 1 Receiver unit
Figtue 4. Set-up for ACA test
-7-
Absorption filter
Softening plate
Receiver optic in
measuring position
l. Correlaton tests at the LG&E Mili Creek Power Staton
In Noven\ber 2004, LG&E and SICK MAIHAK GmbH agreed to test the appli.cability of the FWE200 PM
CEM for complian.ce with PS-11 requirem.ents in a joint project at LG&E' s Mili Creek. Power Station. Th.e
FWE200 system was delivered in January 2005 and installed in Mili Creek Unit (see Fig. 5 & 6). Full opera-
on started in the sam.e month. First experiences of th.e operation have been very positive from the begin-
ning. Th.e system delivered accurate and reliable measuring data compared to reference method data.
Figure S & 6 Mill Creek Power Station
Correlation Measurements
Unit capacity: 52.5 MW
Mounting: Stack at 200 ft level
:>3 diameter after gas inlet;.
no stratification
Stack diametel:: 19' 6"
Gas remperature: 266 F
Gas is saturatl!d
The capability of the system in view of its compliance with PS-11 requirements was investigated by means
of th.ree site specific correlation test nms ca.rried out on April12-14. 2005 (Test 1) and June 2&-29, 2005 (Test
2) as well as SeptEmber 1, 2005 (Test 3). The relevant reference method measurem.ents were conducted si-
multaneously, using EPA Reference Method SB.
Test 1 (Initial correlati.on test, Aprll
2005)
JI)
Test 1 was carried out at three differ- .,
ent concentration levels of PM as
shown in Figure 7. 'fhe test carne out
80
1
with excellent correlation results. All 540
of the relevant criteria (correlation

coefficient, confidence interval and
tolerance interval) were much better
1!
..,
than the permissible PS-lllimit val-
10
ues.
o
o
PIICBIII--c-v-1

Figure '7. Results of Initial Co.rrelation Test
-8-
Test 2 (RCA test, June 2005)
Test 2 was carried out as a RCA test. Response Correlation Audit Summary
Because of plant operating condi-
tions (i.e., a forced outage), this RCA 80
test was only performed at two (in-
stead of three) different PM concen-
~ 7 0
E
leo -E----
s
tration levels. The results, repre- J
50
sented by the red dots in Figure 8, !
40
-f-------
correspond nicely with those taken
30
during the Test 1 nearly three months
20
previously (see blue dots in Figure
10
o 2 3 4 5 6 7
8). With consideration for the abbre-
viated testing, the data achieves the
RCA criteria of positive test results,
thus proving the excellent perform-
ance and stability of the FWE200 sys-
tem.
PM CEII8 Relponle {lngleenl)
lnitial Cooelaticn Dala
0
RCA Test Data
Figure 8. RCA test data O une 2005)
Test 3 (Completed RCA and low load tests, September 2005)
8 9 10
In September 2005, the third PM concentration level of the RCA test was able to be completed. As seen in
Figure 9, the results of that test corresponded with results of the previous tests. With the completion of that
testing, all criteria of the RCA test were passed with ease, proving once again the excellent performance and
stability of the FWE200 system.
An additional test was performed during this time period to check the effect of over-isokinetic sampling.
PS-11 requires that a PM CEMS operate at (or within 10% of) an isokinetic sampling rate. Currently, the
FWE200 PM monitor operating on Mili Creek Unit 4 samples slightly over-isokinetic at fullload conditions
and maintains that constant sampling rate at allloads. This may not be a substantial issue in maintaining
compliance with emission standards because at lower loads the PM emission rates are typically well below
the PM emission limit. Also, PS-11 does allow for a waiver of the isokinetic sampling requirement if it can
be demonstrated that isokinetic sampling is not necessary. As seen in Figure 9, the results of the additional
"Low Load" test demonstrate that over-isokinetic sampling has not affected the monitors ability to produce
accurate results.
-9-
Figures 8 and 9 show the RCA test
permitted tolerance intervals of
25% compared to the measured and
Response Correlation Audit Summary
90 =-------------------------------------------,
80
70 --- -- --- ------- - -- ---- ------- --- --------- ---------- ------
calculated correlation curve which 60
once more display the outstanding
measuring quality of the system.
50
40 --
30
20
10 ---
~
~ ~ ~ ~ ~ 7 7
o ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
o 2 3 4 5 6 7 8 9 10
PM CEMS Response (mg/acm)
lnilial Correlalion Data RCA Test Data
Low Load (400MW) - Linear (+25 % of Emissi on Limit)
- Linear (-25% of Emi ssion Umit)
Figure 9. Completed RCA and Low Load test data (September 2005)
l. Absolute Correlation Audit (ACA)
As part of the PM CEM's ongoing quality assurance (QA), Procedure 2 of Appendix F of
40 CFR Part 60 requires that an ACAbe performed at least once in each calendar quarter. The ACA checks
three (3) audit points of the PM CEM, all within the monitor' s measurement range.
Utilizing the "Check Filter Set" described prevously in this report, the FWE200 produced the following
data during three ACA tests performed during the evaluation project.
ACA test data:
Customer Louisville Gas & Electric En-
ergy - Mili Creek Generating Station-
Location Stack Unit No. 4
Devce Type:
Devce No.:
Zero point:
Span point:
Contamination:
FWE200
05028734 Ser. no.
4.00mA
15.11 mA
4.00mA
Measuring value: 6.20 mA
Summary of ACA Results on Unit 4 PM CEM
Relcrmce Res]IOIIliC Ablolute
RelereDCe Value Value DHrerence
Date Fter ( %) (%) (%)
January 24, 2005 1 0.00 0.0 0.0
2 37.9 37.8 0.1
3 55.3 56.2 0.9
4 92.6 92.8 0.2
Aprilll, 2005 1 0.00 0.0 0.0
2 37.9 37.1 0.8
3 55.3 56.0 0.7
4 92.6 93.1 0.5
June 28, 2005 1 0.00 0.00 0.0
2 37.9 37.9 0.0
3 55.3 55.6 0.3
4 92.6 93.2 0.6
Figure 10. ACA Results
The data in Figure 10 performed from January 24 through June 28, 2005 are repeatable and produce values
within 1.0 % of the target values .
-10-
2. Conclusion
Within a very short time of installation, the FWE200 demonstrated accurate measuring performance as well
as operational reliability. The most important outcome of the test installation has been the outstanding
compliance with the requirements of EPA's PS-11 and Procedure 2. After only a few months of experience, it
is very likely, that the FWE200 will remain at the Mili Creek Power Station. Proved to be highly accurate
and reliable, the FWE200 will help the plant to comply with regulations involving continuous measurement
of emissions related to particulate matter.
By utilizing the FWE200, LG&E's Mili Creek Station is equipped to comply with it's PM emission limits and
continues the company's tradition of being on the leading edge of environmental technology. We wish to
thank all people and organizations that have supported this evaluation period and contributed to making it
such a successful project.
-11-
-12-