Medição de Particulado - Tests-2011-Report
Medição de Particulado - Tests-2011-Report
Medição de Particulado - Tests-2011-Report
Herewith the finalised report for the particulate emission monitor spot correlations conducted at
Camden Power Station during June 2011.
We thank you for the opportunity to be of service. We trust that your requirements were interpreted
correctly. Should you however have any queries, please contact us at the above numbers, we will
gladly assist.
Stacklabs
PH Pretorius
PURCHASE : 15742
ORDER No.
SUBMITTED BY : PH Pretorius
Stacklabs
Summary................................................................................................................ 1
1. Introduction................................................................................................. 2
3. Results........................................................................................................ 3
4. Discussion................................................................................................... 5
5. Recommendations....................................................................................... 6
6. Acknowledgements...................................................................................... 6
7. References................................................................................................... 6
8. Distribution list................................................................................................ 16
Stacklabs RPHP147
REPORT SUMMARY
ESKOM’s Camden Power Station, situated in the vicinity of Ermelo, Mpumalanga, requested the
spot correlations of three SB100 continuous particulate emission monitors. The SB100 monitors
were supplied and installed by Protea Automated Solutions, who in turn contracted Stacklabs, an
environmental services company, to conduct the required site specific spot correlations. The spot
correlation of the three monitors were completed during June of 2011 and the results thereof have
been presented in this report as follows:
Stack No. 1
3
Average measured particulate concentration = 11.9 mg/Nm dry @ 6% O2
Correlation equation mg/Nm3 dry @ 6% O2 = 2.4722 * mA – 9.8914 @ (0 to 70 SL)
Correlation coefficient of 0.99
The air flow to gas flow correlation for Stack No. 1 remains as follows:
Total Gas Flow = 0.8811 * Total Air Flow + 168.33
With a correlation coefficient of 0.83
Stack No. 2
3
Average measured particulate concentration = 63.5 mg/Nm dry @ 6% O2
Correlation equation mg/Nm3 dry @ 6% O2 = 14.2944 * mA – 57.1242 @ (0 to 500 SL)
Correlation coefficient of 0.99
The air flow to gas flow correlation for Stack No. 2 remains as follows:
Total Gas Flow = 1.812 * Total Air Flow – 369.3
With a correlation coefficient of 0.95
Stack No. 3
3
Average measured particulate concentration = 130.5 mg/Nm dry @ 6% O2
3
Correlation equation mg/Nm dry @ 6% O2 = 17.574 * mA – 70.261 @ (0 to 450 SL)
Correlation coefficient of 0.99
The air flow to gas flow correlation for Stack No. 3 remains as follows:
Total Gas Flow = 1.2532 * Total Air Flow – 100.5
With a correlation coefficient of 0.99
Page | 1
Stacklabs RPHP147
1. INTRODUCTION
In order to meet the requirements of the South African Air Quality Act No. 39 of 2004
(Reference 1), as well as the ESKOM Standard for Emission Monitoring and Reporting GST
36-742( Reference 2), an ESKOM thermal power plant is obligated to continuously monitor
all particulate emissions released into the atmosphere during production. In addition, both
the Air Quality Act and the Eskom GTS 36-742 standards dictate specific requirements
relating to the selected methods and equipment that may be utilised during the continuous
monitoring. Included are requirements for equipment compliance with EN 15267-3 (QAL1)
and correlation procedures incorporating ISO 9096, ISO 10155 and the VDI 2066 part 4.
The particulate emission measurements were carried out with procedures and equipment
that comply with the requirements of ISO 9096 1992 (Reference 3), ISO 12141 (Reference
4) and the VDI 2066 Part 4 (Reference 5). Only quartz filters were used during the spot
correlation and only three measurements were completed on each stack. The spot
correlation procedures complied in part with the requirements of the Eskom standard GST
36-742, Standard for Emission Monitoring and Reporting, with reference to a “correlation
spot check”. The purpose of the standard’s “correlation spot check” could however not be
applied in its entirety due to the following:
· The filter mediums used during the original correlations and the spot correlation were
not the same (see discussion).
· The reporting standard during the original correlation and the spot correlation were
not the same.
· The original particulate emission monitor settings were not the same as the spot
correlation settings.
· The monitor on stack No. 3 had been replaced with a new monitor.
· The maximum range of some of the original correlations was too low for the current
emissions.
Eskom personnel were responsible for the setting of the PJFFs prior to the test period.
Stacklabs were contracted to provide the service of monitor correlation through isokinetic
dust sampling only.
Page | 2
Stacklabs RPHP147
The broad outlines of filter weighing, pre-test preparations, sampling system integrity checks
and sampling procedures are discussed in Appendix A.
3. RESULTS
Stack No. 1
A total of 3 isokinetic measurements were conducted on stack No. 1 during the 14th of June
2011. The test co-ordinates included are as follows:
The particulate emission monitor correlation for stack No. 1 is described by the following
formulae:
3
mg/Nm dry @ 6% O2 = 2.4722 * mA – 9.8914 @ (0 to 70 SL)
correlation coefficient of 0.99
The air flow to gas flow correlation for Stack No. 1 is as follows:
Total Gas Flow = 0.8811 * Total Air Flow + 168.33
With a correlation coefficient of 0.83
From analysis of this stack’s air flow to gas flow correlation data, it is evident that the total air
flow information as presented for Unit No.1 by the station was incorrect. This correlation will
therefore not predict the correct total gas flows for Stack No. 1 and will require correction
ASAP.
Stack No. 2
A total of 3 isokinetic measurements were conducted on stack No. 2 during the 21st of June
2011. The test co-ordinates included are as follows:
The particulate emission monitor correlation for stack No. 1 is described by the following
formulae:
3
mg/Nm dry @ 6% O2 = 14.2944 * mA – 57.1242 @ (0 to 500 SL)
correlation coefficient of 0.99
The air flow to gas flow correlation for Stack No. 2 is as follows:
Total Gas Flow = 1.812 * Total Air Flow – 369.3
With a correlation coefficient of 0.95
Page | 3
Stacklabs RPHP147
Stack No. 3
A total of 3 isokinetic measurements were conducted on stack No. 3 during the 9th of June
2011. The test co-ordinates included are as follows:
The particulate emission monitor correlation for stack No. 1 is described by the following
formulae:
The air flow to gas flow correlation for Stack No. 3 is as follows:
Total Gas Flow = 1.2532 * Total Air Flow – 100.5
With a correlation coefficient of 0.99
The detailed results of the correlation measurements for the north stack have been
presented in the following Tables:
ABBREVIATIONS
The following abbreviations were used in the text, tables and figures:
· ‘Actual’ refers to the measured temperature and pressure conditions of the gases in the
duct
· ‘Normal’ refers to the actual conditions being normalised to 0 °C and 101,325 kPa.
· 'Standard' refers to the actual conditions being converted to 0 °C and 101,325 kPa (This
is the preferred description used by Eskom).
Page | 4
Stacklabs RPHP147
4. DISCUSSION
As mentioned in the introduction of this report, Camden Power Station installed four Sick
SB100 continuous particulate emission monitors during the first quarter of 2010. In addition,
compliant correlations were completed during the first half of 2010 and the resulting
correlation formulae were implemented and utilised until the end of May 2011. During the
second quarter of 2011, some concerns were raised with regards to the continued validity of
the existing correlations and Camden Power Station’s Management requested that the
monitor suppliers, Protea Automated Solutions, investigate and confirm the operational
status of the particulate emission monitors.
After the completion of Protea Automated Solution’s investigation, some of the Camden
Management concerns were eliminated, however the indicated emissions remained high and
it was decided to verify the actual particulate emissions through an isokinetic survey.
Stacklabs was therefore contracted to conduct a series of isokinetic surveys and spot
correlation of the particulate emission monitors.
th
The isokinetic surveys commenced on stack No. 3 on the 9 of June 2011. Here the original
monitor was replaced during the Protea Automated Solution’s investigation. It was
requested that the surveys be conducted utilising quartz filters and that the monitors be set
as described in the ESKOM GST 36-742 standard with reference to the 30 % monitor output
at full load. The monitor’s operating scale was selected to a 0 to 450 Scattered Light (SL)
range, which was significantly higher than the 0 to 30 SL range which was selected during
the original correlation conducted in March of 2010. An increase in the scattered light scale
would indicate an increase in the particulate emissions. It was also noted that the visual
appearance of the stack was significantly darker than during the 2010 correlation period.
The isokinetic measurements were conducted at a 50% MCR loading, as the plant could not
operate at full load. The low plant loading conditions resulted in low gas velocities within the
stack, with some points falling below the lower detection limit of 5m/s. The final results of
the isokinetic measurements on stack No. 3 supported the monitor’s increased indication
3
with measured concentration of 121 to 135 mg/Nm dry @ 6% O2.
The results obtained from the survey which was conducted on stack No. 1 on the 14th of
June 2011 indicated concentrations of 11.5 to 12.3 mg/Nm3 dry @ 6% O2. Here the plant
loading during the survey was as required with an average output of 94% MCR but despite
the high loading, the flow profile remained poor. Even though the average gas velocity was
an adequate 10 m/s, several points fell below 5m/s and it was discovered that the flow angle
at some points was approximately 50 degrees from the vertical. The result of poor flow
profiles may lead to a tendency to under sample during the isokinetic measurements and
therewith, as a result of the momentum of the dust particles, to over report on the particulate
concentrations. This phenomenon will be less evident at lower concentrations and will be
exaggerated at higher concentrations.
The survey that was conducted on stack No. 2 on the 21st of June 2011 was completed at
an average boiler loading off 78% MCR. The flow profiles were significantly better than that
which were experienced on stacks No. 1 & 3. The average velocity of the three
measurements that were conducted during the survey was approximately 9 m/s and here no
velocities below 5m/s were measured. The average particulate concentration measured was
3
63.5 mg/Nm dry @ 6% O2 at an average monitor output of 8.5 mA. The selected monitor
range was 0 to 500 SL. From these results it could be concluded that the emissions on this
stack have also deteriorated since the 2010 correlations but not as significantly as on stack
No. 3.
Page | 5
Stacklabs RPHP147
As discussed under section 2 of this report, the original correlations conducted in 2010 and
the spot correlation of June 2011 cannot be compared directly. Several factors have
changed since the completion of the original correlations of which the most significant may
be the filter materials that were used. In addition, the reporting standard was also changed
from a wet basis to a dry basis with a 6% reference oxygen value. On the monitor side, as
indicated, a new monitor was installed on stack No. 3 and the operating ranges were
changed on both stacks No. 1 & 3. As any one of these factors could invalidate the previous
2010 correlation, there can be no continued confidence in the values determined from the
2010 formulae and therefore the station has no choice but to temporarily implement the June
2011 spot correlation formulae. It must however be stressed that the spot correlation cannot
comply with the requirements of the GST 36-742, VDI 2066, VDI 3950 (Reference 6) or ISO
10155 (Reference 7) standards and that full compliant correlations must be completed on
these stacks as soon as possible.
5. RECOMMENDATION
It is recommended that:
5.1. The particulate emission monitor spot correlation as presented in this report be used to
determine the particulate emissions emitted from stacks No. 1, 2 & 3 at Camden Power
Station as a temporary measure only.
5.2. Full compliant correlations are conducted on Stacks No. 1, 2, 3 & 4 as soon as possible.
6. ACKNOWLEDGEMENTS
The author expresses sincere appreciation for the co-operation of Eskom's & Protea
Automation Solution’s personnel during the test period.
7. REFERENCES
7.2. ESKOM Standard for Emission Monitoring and Reporting GST 36-742 1086.
7.3. ISO 9096 1992 Stationary source emissions – Determination of concentration and mass
flow rate of particulate material in gas- carrying ducts – manual gravimetric method.
7.4. ISO 12141 2002 Stationary source emissions – Determination of mass concentration of
particulate matter (dust) at low concentrations – manual gravimetric method.
7.5. German VDI 2066 part 4, 1989, Determination of dust load by continuous measurement
of optical transmission.
7.6. German VDI 3950 Part 1, 1994, Calibration of automatic emission measuring
instruments.
7.7. ISO 10155 Stationary Source Emissions – Automated Monitoring of Mass Concentration
of Particles - Performance Characteristics, Test Methods and Specification 1995 – 2002.
Page | 6
Stacklabs RPHP147
Table No. 1
Stack No. 1
Particulate Emission Monitor Correlation
Test Results
Page | 7
Stacklabs RPHP147
Table No. 2
Stack No. 2
Particulate Emission Monitor Correlation
Test Results
Page | 8
Stacklabs RPHP147
Table No. 3
Stack No. 3
Particulate Emission Monitor Correlation
Test Results
Page | 9
Stacklabs RPHP147
Plant: Camden PS
Location: Stack No. 1 Operational data:
Monitor information: Operating Range: 0 – 40 [mg/Nm3 dry @ 6% O2]
Make of Monitor: Sick Limits of validity: [as an hourly average]
Model: SB 100 Lower limit: 11.5 [mg/Nm3 dry @ 6% O2]
Serial Number: 10058568 Upper limit: 12.3 [mg/Nm3 dry @ 6% O2]
Monitor setting: 0 – 70 SL Linear function:
Dates: E = 2.4722 * x – 9.8914
Calibration date: 6th June 2011 where: E = Emission [mg/Nm3 dry @ 6% O2]
th
Correlation dates: 14 June 2011 x = Monitor output [mA]
Correlation Coefficient: 0.99
35
30
25
20
15
10 Confidence Intervals
Tolerance Intervals
5
2011 Correlation Data
0
4 6 8 10 12 14 16 18 20
This correlation was produced as described in the German VDI guide with reference to the zero
point hypotheses.
FIGURE 1
Prepared by: Stacklabs report No. RPHP147
Plant: Camden PS
Location: Stack No. 2 Operational data:
Monitor information: Operating Range: 0 – 229 [mg/Nm3 dry @ 6% O2]
Make of Monitor: Sick Limits of validity: [as an hourly average]
Model: SB 100 Lower limit: 60.5 [mg/Nm3 dry @ 6% O2]
Serial Number: 10058565 Upper limit: 66.4 [mg/Nm3 dry @ 6% O2]
Monitor setting: 0 – 500 SL Linear function:
Dates: E = 14.2944 * x – 57.1242
Calibration date: 6th June 2011 where: E = Emission [mg/Nm3 dry @ 6% O2]
st
Correlation dates: 21 June 2011 x = Monitor output [mA]
Correlation Coefficient: 0.99
200
150
100
Confidence Intervals
50
Tolerance Intervals
This correlation was produced as described in the German VDI guide with reference to the zero
point hypotheses.
FIGURE 2
Prepared by: Stacklabs report No. RPHP14
Plant: Camden PS
Location: Stack No. 3 Operational data:
Monitor information: Operating Range: 0 – 281 [mg/Nm3 dry @ 6% O2]
Make of Monitor: Sick Limits of validity: [as an hourly average]
Model: SB 100 Lower limit: 120.9 [mg/Nm3 dry @ 6% O2]
Serial Number: 10038613 Upper limit: 135.7 [mg/Nm3 dry @ 6% O2]
Monitor setting: 0 – 450 SL Linear function:
Dates: E = 17.574 * x – 70.261
Calibration date: 6th June 2011 where: E = Emission [mg/Nm3 dry @ 6% O2]
th
Correlation dates: 9 June 2011 x = Monitor output [mA]
Correlation Coefficient: 0.99
y = 17.574x - 70.261
PARTICULATE EMISSION mg/Nm3 dry @ 6 % O2
250
200
150
100
Confidence Intervals
50 Tolerance Intervals
This correlation was produced as described in the German VDI guide with reference to the zero
point hypotheses.
FIGURE 3
Prepared by: Stacklabs report No. RPHP147
APPENDIX A
Page | 13
Stacklabs RPHP147
1. FILTER WEIGHING
Before the test, the filters are prepared. The required amount of filters are marked with a
unique number and set out for conditioning in a weighing room of which the humidity and
temperature are controlled. After the conditioning period, the filters are weighed and packed
for transportation. Reference filters are treated the same as those earmarked for the tests,
but not used on site.
After the tests, the used and reference filters are set out again in the laboratory and allowed
to condition to the laboratory ambient conditions. The laboratory conditions remain constant,
but a small variation in moisture might be reflected by the reference filters, in which case
adjustment for the change in moisture may be implemented
2. PRE-TEST PREPARATIONS
On site, the equipment is set up at the measuring location. The inside dimensions of the
duct are determined. The number of test points per traverse is determined according to the
standards and the sampling probe marked accordingly
With each change in filter or any other operation, which might influence the integrity of the
vacuum system, a vacuum check is performed. This ensures that only the gas, which
entered the nozzle, will be measured by the gas test meter.
3. SAMPLING PROCEDURE
Gas temperature, pressure and velocity head readings are logged at each sampling point.
Velocity head readings are updated at intervals of 1 minute. During this time, the computer
calculates the orifice flow settings, required for isokineticity and the flow is adjusted
accordingly with each update.
A calibrated orifice flow meter is used to facilitate the adjustment of the sampling flow rate at
one-minute intervals. The relevant parameters for flow calculation are entered into the
computer. The computer is programmed to determine the flow rate through the orifice in
order to achieve isokineticity. A calibrated dry gas test meter is incorporated into the
sampling train. This test meter, measuring the actual volume sampled, is used as a
checking device at the end of each test to determine the percentage isokineticity.
Moisture is separated from the sampled gases during sampling, using a water trap and silica
gel with a blue indicator. The blue indicator turns pink as moisture is absorbed. The
accumulated liquid is used after the test to determine the moisture content on a percentage-
by-volume basis. This value is again incorporated into the volume of dry gas sampled to
determine the concentration of dust in gases at Actual and Normal (sometimes referred to as
Standard) conditions.
The uncertainty before the test, about the moisture content in the gas, fluctuation in the gas
flows and human error contribute towards the final deviation from 100 % isokineticity.
Filters are weighed directly after each test to determine preliminary results on site.
Page | 14
Stacklabs RPHP147
Relevant plant operating parameters may be logged for future reference purposes. It is
usually recommended to take raw product samples during the tests. The content of certain
elements in the raw product has specific bearing on ESP performance and is useful for
future reference.
Page | 15
Stacklabs RPHP147
8. DISTRIBUTION LIST
Page | 16