In response to the 1990 Clean Air Act Amendments, Pennsylvania Power & Light Company is switching to low sulfur coals to meet Title IV requirements for limiting sulfur dioxide emissions. In addition, low NOx burners were also installed as part of PP&L's overall compliance plan. The Precipitator Performance Improvement Program (PIP) had the goal of producing consistent particulate emissions at levels lower than that prior to program implementation, while coping with higher resistivity fly ashes coupled with the potential for increased unburnt carbon. The precipitator PIP focused on Unit 2 of the Brunner Island Station, Units I and 2 of the Martins Creek Station, and Units 3 and 4 of the Sunbury Station. A tailored PIP was developed for each Unit, addressing its particular needs. The PIP blended theoretical concepts and practical considerations in developing a least-cost plan. The PIP encompassed installation of new rappers and controls, new automatic voltage controllers, increased electrical sectionalization, addition of collecting surface area, and the use of single and dual agent flue gas conditioning systems.

The precipitator PIP was undertaken on a fast-track schedule approach, moving from the study and specification phase, through installation, in slightly more than one year. The precipitator PIP's results are impressive, and suggest that the performance capabilities of many existing precipitators were under estimated.

In response to the 1990 Clean Air Act Amendments, Pennsylvania Power & Light Company embarked on system-wide studies to identify the least-cost methods to control emissions of sulfur dioxide and oxides of nitrogen. The studies examined both Phase I and Phase 11 requirements to assure that the compliance plan provided for the least cost options while maintaining flexibility to cost-effectively respond to an uncertain future. The compliance plan identified that three units at the Brunner Island Station, two units at the Martins Creek Station and two units at the Sunbury Station will be switching to lower sulfur, eastern bituminous coals and would have low NOx burners installed. Each unit's precipitator was evaluated in terms of assessing the impact of low sulfur coal in light of on-going maintenance, repair and upgrading programs. It was determined that one unit at the Brunner Island Station and four units at the Martins Creek and Sunbury Stations required remedial attention to accommodate low sulfur coal firing.

The Martins Creek Steam Electric Station is located on the Delaware River near Martins Creek, Northampton County, Pennsylvania. Each unit has a nominal rating of 150 MW. Each unit bums pulverized bituminous coal in a Foster Wheeler front wall-fired balanced-draft boiler and fly ash is collected in a Buell weighted-wire design precipitator. The precipitator consists of a single casing with two chambers having a total of 102 gas passages. The collecting plates are 30 feet in height with 27 feet of treatment length producing an aspect ratio of 0.9. The fields are electrically configured with the first four fields having a six foot depth and the fifth field having a three foot depth. There is a total of 165,200 ft² of collecting plate area with an SCA of 275 ft²/1000 ACFM.  The T-R sets in the first four fields each serve 30,000 ft² of collecting plate area, while T-R's in the last field serve 15,000 ft².

The Sunbury Steam Electric Station is located on the Susquehanna River, in the town of Shamokin Dam, Snyder County, Pennsylvania. Units 3 and 4 have nominal ratings of 100 MW and 140 MW, respectively. Each unit bums a blend of pulverized bituminous coal in a Foster Wheeler front wall fired balanced draft boiler. Fly ash collection is accomplished by the original Research-Cottrell precipitator in parallel with a retrofitted Buell precipitator, both of which are flue gas conditioned with sulfur trioxide.

The Unit 4 Research-Cottrell precipitator consists of four chambers having a total of 84 gas passages. The collecting plates are 17-1/2 feet in height with 18 feet of treatment length producing an aspect ratio of 1.03. There are two electrical/mechanical fields, each nine feet in depth. The collecting area is 52,920 ft² resulting in an SCA of 207 ft²/1000 ACFM. Each T-R set serves 13,230 ft². This precipitator was originally equipped with an empty third field, six feet in depth, which if internals were installed, would increase the SCA to 276 ft²/1000 ACFM.

The compliance plan focused on providing flexibility in coal purchasing. As such, evaluation of precipitator performance improvement options involved the ability to fire coals which have the following characteristics:

Ultimate Analysis                                                  Medium Sulfur                 Low Sulfur
  (% as received)                       Range                      Central PA                     Central PA

Ash                                              5-15                              12.0                                 11.4
Carbon                                       60-80                              70.7                                 69.7
Hydrogen                                  3.5-5.75                             4.2                                   4.2
Moisture                                      4-10                                6.1                                   7.2
Nitrogen                                    0.5-2.0                               1.0                                  1.2
Oxygen                                         2-8                                  4.0                                  4.9
Sulfur                                        0.5-3.0                               2.0                                  1.4
Volatile                                      18-40                              23.9                                24.8
Fixed Carbon                              45-65                             58.0                                56.6
Btu/lb.                                  12,000-13,600                    12,050                           12,310
Hardgrove Grind, Index              45-10                               80                                   80

The precipitator PIP focused on maximizing the effectiveness of the existing equipment rather than
wholesale replacement with new precipitators or fabric filters. On-going maintenance and
inspection programs determined that the precipitator casings ductwork and the supporting structures
were in good condition, which permitted attention to be aimed at internal component upgrades to
reliably increase precipitator performance. Coupled with the switch to lower sulfur coals, the
affected units would also undergo the installation of low NOx burners which could increase the
unburnt carbon content of the fly ash. It was determined that the precipitators needed to be in
good mechanical and electrical condition in order to cope with the changes.

Mechanical Repair

Mechanical repair is a broad category which includes replacement of broken emitting electrodes and insulators; electrode realignment; casing and ductwork repair of perforations or other air infiltration sources; repair or strengthening of electrode hanger attachments; replacement of rapper transmission components broken due to long term rapping cycles; and replacement of internals degraded due to long term corrosion. Such mechanical repairs, to a cost-effective degree, are generally considered to be a mandatory step to restoring a precipitator to a near-original condition which allows the best utilization of available SCA, irrespective of any other performance enhancement approaches applied.

Although less understood on a theoretical basis than other precipitator performance factors, effective electrode cleaning and minimization of last-field rapping re-entrainment puffs is considered to be an important factor in maintaining performance capability over the long term. Testing has proven that for high efficiency cold side precipitators, rapping re-entrainment can represent 30% of the total emission. There is a general consensus that rapping performance (exclusive of controls) by single large impact rapper designs are more effective than older style vibrators. In addition, it is highly desirable to have rappers with adjustable impact when trying to optimize performance of weighted-wire designs. Highly sectionalized designs with larger numbers of rappers and consequently shorter horizontal transmission paths through rapper anvils and beams, is desirable for reducing the stress on the maximum impact portion of a section, providing the most flexibility in optimizing and avoiding detrimental over-rapping.

The general concept on electrical energization (excluding controls) is to match the T-R set energy supply ratings and capabilities to the precipitator load. Secondary voltage ratings of 45 kV de average for plate-to-plate spacings of nine to ten inches are adequate to satisfy the ultimate applied electric field strength limitation (spark over) for most cold-side precipitator applications.

Although a seemingly safe, conservative approach of oversizing T-R set current ratings would appear to be appropriate, several problems can be introduced. Aside from the cost of the medium and low voltage power supplies, which are designed to electrical equipment ratings and not expectations, T-R sets operating at less than 50% of their ratings ma be subject to problems with control stability and arc suppression. In addition, for control of an oversized T-R set, the conduction angle for silicon controlled rectifier (SCR) controllers is low and the wave shape is not optimal for best collection performance. Such performance is dependent on half cycle peak times average voltage, and not just the average. A special consideration arises with older electrical designs, which tended to be much oversized in terms of current ratings employing saturable core reactor voltage control, and which were later converted to more modem control designs. Since the power supply circuit requires a current limiting reactor (CLR) to protect the components (SCRs and rectifiers) from high instantaneous current flow during SCR firing and sparking, it was common practice to convert the saturable core reactor to a passive CLR by disconnecting the de control voltage function. This practice usually leads to CLR sizing which is too small for optimum wave shaping. In this case, the practice has been to derate the T-R set by replacement of old reactors with larger CLRs typically sized at greater than 30% of the precipitator design impedance.

Consistent with general criteria of matching power supplies to precipitator loads, the last important performance enhancement consideration is increasing electrical sectionalization with more lower current rated T-R sets even though the total precipitator load remains essentially the same. For variations of electrical load requirements due to changes in ash resistivity, gas temperature and particulate loading, greater sectionalization provides for a higher degree of localized control optimization. Greater sectionalization also minimizes the size of sections which could be otherwise performance limited due to misalignment, excessive electrode ash deposits or outage due to broken wires. Increasing sectionalization by the addition of more T-R sets, with derating of existing sets, is considered to be of substantial benefit for older precipitators which may exhibit substantial flue gas temperature stratification due to air heater and ductwork designs.

Depending on the level of sophistication of the existing controls, upgrading of both rapper and automatic voltage controls (AVC) to modem microprocessors is one of the most cost-effective approaches to performance enhancement. Microprocessor AVCs have the advantage of response speed and sophistication to achieve on-going optimums rather just maintaining preconceived optimum set points. Microprocessor AVCs also have the ability to communicate through a dedicated supervisory system or Distributed Control System (DCS) with other controllers (T-Rs and rappers) as well as the stack continuous emission monitoring system (CEMS) for more elaborate control schemes and data acquisition.

For rapper controls, modem microprocessor designs generally provide a more reliable approach to anti-coincident rapping (simultaneous same lane rapping in multiple fields or simultaneous multiple lane rapping in the last field). Modem microprocessor designs also allow the anti-coincident function to extend to two precipitators on the same generating unit via communication capabilities. As for AVCS, communication capability provides the potential for elaboration of control schemes to include some type of feedback or external intelligence which is usually lacking, such as, briefly pausing the last field rapping sequence in anticipation of soot blowing or when opacity is otherwise high. Although the software and strategy for rapper optimization has not been developed to the same extent as for AVCS, the potential for incorporating such advances in the future is generally sufficient to justify modernization at the present time.

Most analyses of precipitator performance fundamentals indicate that when the flue gas characteristics are not uniform with respect to velocity, temperature, dust loading and conditioning agent concentrations, there is typically an effect of a lower collection efficiency. The general approach to gas flow distribution is the use of inlet duct turning vanes, inlet nozzle ladder vanes or splinters, and inlet and outlet multiple perforated plates with variable porosity's. The design approach for gas flow distribution improvement on existing units involves off-line baseline data acquisition with in situ air flow velocity measurements, analysis of data to provide a preliminary design for distribution devices for a three dimensional scale flow model, fine tuning by flow testing in the model, and installation of new devices during an outage. This approach has been successfully applied many times in the industry. However, the duration of the process is long-, and it is difficult to predict before the data acquisition and model test activities, whether the effort can achieve significant performance enhancements. Some progress has been made with the development of numerical analysis methods as a substitute for physical model testing. Confidence in such a technique appears to be limited to those arrangements which have a comparatively simple inlet duct configuration or for similar arrangements for which a before/after database exists.

For flue gas temperature maldistribution resulting from air heater rotation, the situation is more complex, especially for a gas conditioned unit where the gas conditioning injectors are normally located in the space required for the internal vanes and duct compartments required for substantial cross-duct mixing. Although temperature stratification corrections have been successfully applied on a limited number of units (e.g., at the Keystone and Conemaugh Stations), generalized predictions on performance enhancement have not been fully developed.

When precipitator performance is pushed to ideal limits through the application of all of the previously described technologies, then further improvements can be achieved by increasing the SCA, which is also equivalent to increasing treatment time. From the short term perspective, this may be achieved by decreasing gas flow (i.e., derating the generating unit). Practical, long term solutions involve the addition of more collecting area. This may be accomplished by increasing precipitator height with new, taller internals, the addition of another field (if space permits) or the addition of another precipitator, preferably in a parallel arrangement to decrease gas velocity through the existing precipitator. Although larger SCAs are frequently used for new units as a conservative approach to compensating for all other non-ideal conditions which may develop, SCA increases for existing units are generally a last resort, because of the substantial cost involved compared to the other technologies described.

When lower sulfur coals are fired, fly ash resistivity generally increases and may be detrimental to precipitator performance. It has been well demonstrated that the injection of sulfur trioxide into the flue gases can maintain fly ash resistivity at an optimum level of about 2 x 10¹⁰ ohm-cm for good precipitator performance. Thus, existing precipitators are able to meet required emission levels with low sulfur coals. In addition, some units have found that the effects of high unburned carbon levels in the fly ash (a condition expected when low NOx burners became operational) can be mitigated with the injection of ammonia into the flue gases ahead of the precipitators. Flue gas conditioning, either single or dual agent, is best applied to precipitators which are mechanically and electrically in good condition and properly operating.

Each of the performance improvement options were evaluated for the five units. The evaluation considered both the theoretical and practical impacts on precipitator performance and reliability. The evaluation produced the following remedial action plans:

The original Research-Cottrell precipitator would be refurbished with new internals including a higher degree of electrical sectionalization and reusing the existing T-R sets to the maximum degree possible. New microprocessor based AVC's would also be installed. The number of rappers for collecting plates and discharge electrodes would be increased to reflect modern practices as well as adding a new rapper microprocessor control system. A new penthouse heated air pressurization system would also be installed as well as replacing access doors in the precipitator casing, penthouse and ductwork. For the retrofitted Buell precipitator, the rapper control system would be replaced with a microprocessor based system.

The Buell precipitator would have its discharge electrode vibrators replaced with gravity impact rappers and new microprocessor based controls. The collecting plate vibrators would also be replaced with twice as many gravity impact rappers and associated controls. The number of T-R sets would be doubled and the CLR's on the existing T-R sets would be replaced while providing new AVC's for all sets.

Sunbury Station, Units 3 and 4

The original Research-Cottrell precipitators would have internals and associated accessories installed in the empty third mechanical field position. The AVC's and CLR's for the existing T-R sets would be replaced. The retrofitted Buell precipitators would have the AVC's and rapper controllers replaced with microprocessor based equipment.

Considering the potential affects of lower sulfur content coals and possible increases in fly ash unburned carbon levels due to low NOX burner installation, considerable attention was directed at assessing the risks associated with not moving forward with dual flue gas conditioning systems at the three stations. From a classical standpoint, lower sulfur coals would produce higher resistivity fly ashes thereby limiting the amount of useful power which can be input to the precipitators. The basic question was "at what point will precipitator performance be significantly degraded when considering the repairs, replacements and upgrades associated with the balance of the PIP?" Compounding this question was the realization that coals for each station may be derived from multiple sources. In addition, it was uncertain as to what level the fly ash unburned carbon would be with the LNB's and how would the carbon affect resistivity. Faced with these uncertainties, the cost of the gas conditioning system and projected unit loading regimes in the short ten-n, it was decided to suspend flue gas conditioning system purchase pending assessment the precipitator performance after the implementation of the PIP.

The precipitator PIP began with the award of an engineering services contract in March 1993 to prepare project study plans, write specifications, contractor evaluate proposals, perform detail engineering and provide construction support. The precipitator PIP concluded with equipment installation for Sunbury Unit 3 by March 6, 1994; Brunner Island Unit 2 by June 12, 1994; Martins Creek Unit 2 by July 24, 1994; Martins Creek Unit I by September 11, 1994; and Sunbury Unit 4 by February 27, 1995. In order to meet schedule requirements, the project team adopted a parallel path task approach integrating the resources of PP&L, Raytheon, Research-Cottrell and General Electric Environmental Services.

Interface information requirements for detailed balance of plant design were identified early and aggressively pursued to support the program schedule. Significant effort was expended to expedite design drawings and release final engineering within one month of vendor release of final drawings. Critical milestones dealt with release of preliminary and final engineering to PP&L Construction. Despite several unforeseen challenges (which could have forced delays), the program team was able to maintain completion and issuance of all engineering within PP&L's standard milestone targets, i.e., three months before the start of the outage. Engineering for pre-outage work, primarily for current limiting reactor installation and cable tray runs, was issued first and in sufficient time to support PP&L's construction effort to complete this work before the outage. The success of meeting the Program's schedule requirements is due to the early pursuit of interface design information, the focusing of work efforts and close cooperation of all members of the program team.

The implementation of the Precipitator PIP has produced particulate emission levels which are significantly less than originally anticipated. The Commonwealth of Pennsylvania limits particulate emissions to a maximum of 20% opacity on a one-minute average basis. When firing low sulfur coals and with the operation of low NOx burners at the Brunner Island, Martins Creek and Sunbury Stations, the precipitators are consistently producing opacity levels at the 5% level which certainly more than meets the objectives of the PIP.

As electric utilities are faced with emerging emission regulations in an increasingly competitive environment, existing electrostatic precipitators may have unrealized capacity to cost-effectively reduce particulate emissions. In order to tap this resource, utilities need to move beyond the theoretical and demonstrate precipitator capabilities by fine tuning and upgrading the existing equipment. PP&L's Precipitator Performance Improvement Program points the way to maximizing performance economically.