By Malcolm Boyd, Ph.D., ATCO Power
Deratings due to opacity problems at the Battle River Generating Station led ATCO Power to evaluate and install skewed gas flow technology (SGFT) in one-half of the Unit 5 twin-casing electrostatic precipitator during the August 2000 outage. Preliminary operating results show that the modified casing produces opacity readings at the outlet 40 percent lower than those seen at the outlet of the unmodified casing. The dust loading tests indicate a 27.5 percent improvement in collection efficiency. This article includes a technical review and evaluation of Battle River's SGFT installation, as well as the rationale used to provide the initial economic justification.
 ESP Casing at ATCO Power's Battle River Generating Station |
Battle River Generating Station's 375 MW Unit 5 consists of a tangentially fired Alstom (Combustion Engineering) drum-type boiler burning sub-bituminous coal. The Flakt twin-casing precipitator that serves Unit 5 has 523,561 square feet of collecting surface and a design average gas velocity of 5.12 ft/sec. Unit 5 must be derated whenever the six-minute average opacity level of the stack gases exceeds the environmental license maximum of 20.49 percent. Although the precipitator is normally capable of maintaining opacity below this level, occasional derates/outages occur due to high opacity. Operating records indicated that Unit 5 is derated on average by 38 GWh per year due to high opacity related to precipitator problems and/or coal quality.
Technology Options
Battle River originally considered four options to minimize the opacity derates: (1) precipitator positive isolation, (2) hot suit entry into the precipitator, (3) upgraded controllers, and (4) skewed gas flow technology.
Positive isolation of individual precipitator casings would permit derating to 60 percent maximum continuous rating as opposed to a forced outage in the event that precipitator maintenance was required. This option would require a double guillotine damper and purge system at each of the eight inlet ducts, as well as in the two ducts between the stack and ID fans. The complete installation of a suitable isolation system for each precipitator casing would cost $800,000.
An alternative to positive isolation is "hot suit entry." This option involves isolating and entering the precipitator casings using suitable safety apparel and the appropriate electrical/rapper isolation. Adequately trained personnel can then perform simple maintenance such as the removal of broken wires. A derate to 60 percent is still required because one precipitator casing must be taken off-line. A specialist company quoted $50,000 per visit for this work.
Installing upgraded controllers could provide a maximum reduction in stack opacity levels of 1-2 percent, more likely closer to 1 percent. This would effectively reduce the six-minute opacity average by 0.5 percent if installed on one casing, or 1 percent if installed on both north and south casings. The effect of this reduction in the six-minute opacity average is to decrease the frequency of derates/outages.
Using skewed gas flow technology, Stothert Engineering Ltd. normally estimates a potential performance improvement at between 30 and 55 percent for typical power plant applications. After careful review of this specific precipitator, they advised the use of a conservative performance improvement estimate of 40 percent. SGFT at Battle River, therefore, would result in a stack opacity reduction from 20 percent to 12 percent. Installing SGFT on only one-half of the precipitator, therefore, would provide an improvement of 20 percent; i.e., the stack opacity could be reduced from 20 percent to 16 percent. The effect of this reduction in the six-minute opacity average is to decrease the frequency of derates/outages regardless of the underlying reasons (either equipment failure or variation in coal and ash quality).
The positive isolation and hot suit entry options present safety and investment risks. Also, neither of these two options addresses problems concerned with coal/ash variability, which is a significant issue at Battle River Generating Station. Plant management ultimately chose SGFT over the upgraded controllers for the following reasons:
- Higher net present value
- Excellent internal rate of return
- Higher improvements in efficiency
- The 2000 turnaround represented the only opportunity to implement SGFT until 2002 (the project duration is about 4 weeks)
- Upgraded controllers could be installed at a later date even with Unit 5 on-line
ATCO cautiously chose to modify only one of the twin casings (the north casing) and evaluate the degree of opacity reduction by direct comparison between the performance of the north and south casings.
Skewed Gas Flow Technology
SGFT modifies the gas flow within the precipitator in order to optimize the use of the collecting area by considering the actual dust loading that minimizes the chances for re-entrainment. In a conventional ESP, the combined effect of re-entrainment and gravity can increase ash loading at the bottom of the precipitator relative to the top. By skewing the velocity profile from the bottom left to the top right, SGFT re-averages the re-entrained ash load over the downstream collector plates, thereby improving the overall operating efficiency of the precipitator.
Flue gas flow testing performed in 1998/99 revealed an unevenly distributed gas flow at the precipitator inlet, which translates into a non-uniform ash load. The installation of SGFT and horizontal balancing was intended to result in an equal amount of flue gas (and therefore ash load) flowing through each of the four inlets, and a skewed velocity profile at the inlet and outlet of the precipitator.
Flue gas flow test results also identified problems with "sneakage," whereby a portion of the flue gas flow moves both above and below the high-energy frames, completely missing the collector plates and effectively bypassing the precipitator. SGFT should reduce the level of "'sneakage" because the flow velocities will approach zero at the top and bottom of the precipitator.
 |
The SGFT installation required flow modifications in four areas within the north casing (Figure 1):
- Inlet balancing - Modification of blocker pipes in the common duct upstream of the precipitator to provide an equal amount of gas flow through each of the four inlet ducts.
- Variable porosity inlet screens - These screens were designed to provide the desired skewed velocity profile at the inlet. Modifications required the removal of an existing uniform porosity screen and the installation of a grid of varying porosity 4-foot by 4-foot sheets.
- Additional hopper baffles - These were required to assist in providing the desired skew and to minimize sneakage through the hoppers.
- Variable porosity outlet screens - These were designed to provide the desired skewed velocity profile at the outlet. Modifications required the installation of structural framework and a grid of varying porosity 4-foot by 4-foot sheets.
Technology Evaluation
The success and performance of SGFT is based on an evaluation of the results of air flow, duct opacity, stack opacity and dust loading tests. Air flow tests across several cross-sectional planes within the ESP casing indicated that the mechanical modifications provided the desired skewed flow.
 |
With respect to opacity, Figure 2 illustrates a typical plot showing the relative opacities in the stack, the north (modified) outlet and the south (unmodified) outlet. On average, the converted north casing produces opacities that are 40 percent lower than those produced by the south casing.
Unfortunately, the duct opacity monitors were not installed until after the installation of SGFT in the north casing. Therefore, no baseline data exists for the relative performance of the north and south casing of the precipitator prior to modifications of the north casing. However, the maintenance history does not suggest any major difference in the mechanical or electrical condition of these casings.
A significant reduction has not been observed in the stack opacity since the SGFT installation. It is believed that variable coal quality may be responsible for the lack of a downward trend in opacity. To qualify this belief, a coal quality variable has been suggested that would relate coal analysis data to average opacity. Dr. Ralph Altman of EPRI suggested the following variable as an approximate means of assessing the influence of coal properties on stack opacity:
CoalQualityVariable=(%Ash*Consumption)/(%Sulfur*HeatValue)
 |
Monthly average coal data was used calculate this coal quality variable, which should theoretically follow the same trend as the recorded monthly average opacity. Any improvement in precipitator performance, therefore, should cause the opacity to be lower than that predicted by the normalized coal quality variable. Figure 3 illustrates the average difference between the opacity and normalized coal variable for the period directly before and after the modifications. This shows that the opacity, relative to the coal properties, has been reduced. This relationship will be monitored over the next several months to determine the stability of this trend.
 |
Battle River conducted three dust loading tests over the course of three days using the standard EPA "Method 17" method (Table 1). Each test determined the dust loading for four points in the flue gas system, namely the inlets and outlets of both the north and south precip casings. After further analysis of the results, test 2 was eliminated due to the uncharacteristically high dust loading on the north outlet.
There are several possible explanations for the uncharacteristically high particulate loading found at the north outlet during test 2. High levels of SO2 were observed in the test enclosure of the north-casing outlet on that particular day. This resulted in increased traffic over the roof of the duct because of the frequent gas testing requirements. This traffic may have caused a dust cake on the roof of the duct to dislodge and contaminate the sample stream. Alternatively, operator error may be to blame. It is possible that the probe was inserted too far into the duct, thereby disturbing the dust layer on the floor. Whatever the cause, test 2 measured the north outlet loading to be more than three times higher than that measured for tests 1 and 3.
Assuming the efficiencies of the north and south casings were equal prior to the installation of SGFT, it can be concluded that the modifications have improved the collection efficiency from 97.96 percent to 98.52 percent. This equates to a reduction of particulate emissions of 27.5 percent.
Particulate sizing tests performed on the outlets of the north and south casings indicate that the modified casing is more effective than the unmodified casing for the removal of particles <1µm. This may help explain the difference observed between the 27.5 percent reduction measured by particulate emissions and the 40 percent measured by opacity. Due to the wavelength of the light used in the opacity monitor, a specific density of small particles would have a greater ability to block the light than an equal density of larger particles. This would result in larger reduction in opacity reading.
Limitations and Future Work
A number of limitations make it impossible to categorically state that the performance improvement of the modified casing is a direct result of SGFT. First is the lack of baseline duct opacity and dust emissions measurements. It is possible that the modified north casing has always outperformed the south casing due to some subtle difference in their construction/operation.
Another limitation is the lack of consistent coal quality for the period immediately before and after the installation, which validates the decision to modify only one of the casings for purposes of evaluating SGFT.
Another factor that may explain the difference in performance is temperature difference at the inlets to each casing. Operational factors and the performance of the primary air heaters can cause the precipitator casings to operate at different temperatures. The dust emissions tests involved recording the flue gas temperatures. The average inlet temperature for the modified casing was found to be 13 F less than the unmodified casing, at 297 F and 310 F, respectively. The influence of this temperature difference has not yet been quantified.
Skewing alone might not be responsible for the superior performance of the north casing because upstream balancing was also performed in order to achieve an equal level of flow through each of the four inlets on the north casing. The additional hopper baffles installed on the inlet hoppers may also have helped to reduce sneakage-induced re-entrainment.
Despite these limitations, there is general agreement between those people most closely involved with the project that the SGFT installation has been a success. Further investigation is required to determine the benefit afforded by the SGFT component of the modification as compared with the benefit due to horizontal balancing. Continued evaluation of SGFT will also entail a thorough investigation of the derate history of Unit 5. Monthly, weekly or even daily averages for the stack opacity only tell a small part of the story. The best evidence for the success of SGFT may lie in the derate information. Stack opacity and coal quality will continue to be monitored to determine whether or not the coal quality is really responsible for the lack of a downward shift in stack opacity.
To conclusively evaluate the benefits of SGFT, it would be necessary to proceed with the flow modification of the south casing. Armed with the baseline data concerning duct opacity and dust emissions, it would then be possible to determine accurate values for performance improvement of the south casing. Also, by breaking the flow modification work down into two stages, i.e. horizontal balancing and SGFT, the influence of each could be isolated.
Author-
Dr. Malcolm Boyd is mechanical engineer at ATCO Power's Battle River Generating Station in Alberta, Canada.
Acknowledgements:
The Primary Particulates Control Group of EPRI has sponsored the emissions testing portion of this project. Also, the author would like to thank Dr. Ralph Altman of EPRI for his support and advice, and Ian Clark, Bruce Gowanlock and Jackie Albers (Battle River station's "Precipitator Improvement Team") for their assistance.
Print
Email
Save
