Using On-Line Coal Analysis to Optimize Combustion in Utility Boilers: a 10-yr Program at Monroe Power Plant

Monroe Power Plant (MPP) of DTE Energy, a 3200 MW generating station fired with blends of Powder River Basin subbituminous coal and Central Appalachian bituminous coal, has developed and advanced the technologies of on-line coal analysis. Technologies involved include both the analytical instrumentation and the communication of results to operators trained in the use of these results. At MPP the x-ray fluorescence (XRF) analyzer provides the full proximate analysis and calorific value of the fuel being transported to the silos; it also provides the complete ash elemental analysis of this fuel. Through additional statistical manipulation the XRF provides ultimate analysis data to the operator. The XRF obtains a data set every 90 seconds. The software used to transmit the data to the power plant operators, developed by Engineering Consultants Group (ECG) of Akron, OH, provides operators, shift supervisors, and engineering staff with data concerning the fuel being burned at the present time, fuel that can be expected in 1 hr, 2 hrs, 4 hrs, and the fuel being loaded into the silos. The information provided to operations includes an assessment of the quality of the fuel from several perspectives: calorific value, moisture content, ash loading, and ash chemistry. Further, using the analyzer and associated PI data, the operator can obtain real time boiler efficiency and flyash resistivity data. Training of personnel provides staff with insights into the meaning of the data, and appropriate responses to the data. For example, the operations group, led by the shift supervisor, periodically adjusts the blend to optimize calorific value, minimize slagging and fouling potential, manage the use of SO3 for flyash conditioning, and manage the use of REF additive for emissions control. These capabilities have been employed by the plant for the past several years, facilitating operation of the four units at MPP. This paper describes the system, its usage, and its technical benefits to the generating station.

1 Using On-Line Coal Analysis to Optimize Combustion in Utility Boilers: a 10-yr Program at Monroe Power Plant David A. Tillman Consultant to DTE Energy Monroe, MI ABSTRACT Monroe Power Plant (MPP) of DTE Energy, a 3200 MW generating station fired with blends of Powder River Basin subbituminous coal and Central Appalachian bituminous coal, has developed and advanced the technologies of on-line coal analysis. Technologies involved include both the analytical instrumentation and the communication of results to operators trained in the use of these results. At MPP the x-ray fluorescence (XRF) analyzer provides the full proximate analysis and calorific value of the fuel being transported to the silos; it also provides the complete ash elemental analysis of this fuel. Through additional statistical manipulation the XRF provides ultimate analysis data to the operator. The XRF obtains a data set every 90 seconds. The software used to transmit the data to the power plant operators, developed by Engineering Consultants Group (ECG) of Akron, OH, provides operators, shift supervisors, and engineering staff with data concerning the fuel being burned at the present time, fuel that can be expected in 1 hr, 2 hrs, 4 hrs, and the fuel being loaded into the silos. The information provided to operations includes an assessment of the quality of the fuel from several perspectives: calorific value, moisture content, ash loading, and ash chemistry. Further, using the analyzer and associated PI data, the operator can obtain real time boiler efficiency and flyash resistivity data. Training of personnel provides staff with insights into the meaning of the data, and appropriate responses to the data. For example, the operations group, led by the shift supervisor, periodically adjusts the blend to optimize calorific value, minimize slagging and fouling potential, manage the use of SO3 for flyash conditioning, and manage the use of REF additive for emissions control. These capabilities have been employed by the plant for the past several years, facilitating operation of the four units at MPP. This paper describes the system, its usage, and its technical benefits to the generating station. INTRODUCTION Monroe Power Plant (MPP) of DTE Energy has operated on blended fuel for many years, using blending to accomplish the use of Powder River Basin (PRB) subbituminous coal. The four units at MPP, each rated above 800 MWe (gross) and constructed in 1970 - 1974, were 2 originally designed to fire Northern Appalachian bituminous coal. With the opening up of the "Joint Line" through the Powder River Basin in Wyoming, this very low sulfur and highly reactive coal became economically available throughout the upper Midwest, and much of the US. Detroit Edison took advantage of this low cost fuel for its flagship generating station- MPP-as a basic energy source. Detroit Edison then switched its bituminous coal from Northern Appalachian higher sulfur slagging coal to low and mid sulfur bituminous coal from Central Appalachia. Optimizing the fuel blending process became an increasingly significant issue as a means for reducing generating coats while maintaining high capacities and efficiencies of the four supercritical MPP boilers. The approach taken was to install an on-line coal analyzer and develop a program around that instrument. The program was envisioned as a 3-legged stool: 1) the analyzer instrument, 2) the software to convey the information to the Shift Supervisor and the Supervising Operators, and 3) training of operations personnel regarding how to understand, interpret, and use the information. In the process the information was also conveyed to the engineering staff and plant management. Further, the program was viewed as a "living effort", one that PROGRAM STRUCTURE Monroe utilizes 2 sets of 3 coal piles under a long tripper gallery (see Figure 1); each pair of piles contains a distinct type of coal. Underpile ploughs are to reclaim the coal and load it on the main conveyor belt (see Figure 2); controls in fuel supply couple plough speed with belt scales to manage the blend composition. The coal analyzer system chosen was supplied by QC, Inc. It utilizes X-ray Fluorescence (XRF) technology. The blended coal passes through the coal sampler and analyzer on its way to the power house. Given the signals, and the extent of fuel blending controls, the desired blend can be supplied to an accuracy level of <1%. Weigh belt scales ensure that the desired blend is achieved. A separate moisture meter is also used in conjunction with the analyzer. Through interpretation of the signals, the instrument (and the moisture meter) can detect and quantify the following: full proximate analysis, calorific value, and full ash elemental analysis. There are numerous competing on-line analyzers on the market. Thermo-Fisher has one as do other vendors. Lehigh University is working on a laser induced spectroscopy breakdown (LIBS) approach, and there are several LIBS designs in the field. 3 Figure 1. Monroe Power Plant showing coal blending facility. Figure 2. Main belt for coal blending. Coal blending facility 4 The fuel tracking software package chosen was AccuTrack™ as developed by ECG Consultants of Akron, OH. This softyware tracks the flow of coal from the coal piles to the plant as is shown in Figure 3. It permits tracking the fuel blend from the coal yard to each of the 28 silos, to the 28 mills, and to the 112 burners supplying coal to the 4 boilers. AccuTrack™ proved to be the most effective fuel tracking software available. It was coupled with the Figure 3. Process schematic of the Digital Fuel Tracking System, AccuTrack™ The third leg of the 3-legged stool has been training. Separate training courses have been developed by Monroe Power Plant staff and ECG Consultants. The training has been designed for shift supervisors, control room (supervising) operators, power plant operators, and plant engineers. It provides all personnel with sufficient information to understand and interpret the data being supplied on operator screens (see Figure 4). These screens use AccuTrack™ to convey the information from the analyzer to operating personnel. They provide guidance concerning how to respond to specific fuel conditions that are being experienced at the present time, and conditions that will occur in 1 hr, 2 hrs, and 4 hrs from the present time. The screens also include fuel quality on the main belt leading to the silos-information required for the shift supervisor to adjust the blend for the coming hours. Digital fuel tracking system by ECG Consultants 5 Figure 4. Operator Advisory Screen for Unit 1; all suchUnit Advisory Screens are Identical. The advisory screens have gone through a significant evolution; the one shown in Figure 4 is the latest configuration designed in the spring of 2012. Note that it has indication of the use of an inorganic additive high in calcium oxide-REF-and it also has more attention to slagging and fouling than previous screens. The attention to slagging and fouling is shown with the Fe2O3/CaO ratio and with other measures and slagging indices. The continued evolution of the on-line analyzer program has led to advisory screens concerning boiler efficiency and controllable losses, temperature profiles of each boiler, and flyash resistivity profiles of each boiler. These screens are based on coupling the on-line coal analysis with PI data regarding operations, certain manual inputs (unburned carbon in ash; air heater in-leakage) and then employing subsequent calculations (e.g., heat transfer calculations for temperature profiling) in order to provide operators with the maximum opportunity to optimize boiler performance. This approach is made possible by the robust PI system of 7500 PI tags per boiler, along with the analyzer program. Figures 5 - 7 are screen shots concerning the efficiency, temperature profile, and resistivity screens. 6 Figure 5. The Boiler Efficiency Screen as Designed and Implemented Figure 6. The Furnace Temperature Profile Screen as Designed and Implemented 7 Figure 7. The Flyash Resistivity Screen as Dessigned and Implemented Operators and shift supervisors have been trained on these screens as well. They permit optimizing the operation of each boiler. Even units that are equipped with scrubbers rely on the flyash resistivity screen to govern the control of sulfur conditioning of flyash. This maximizes the potential for synthetic gypsum sales by maintaining gypsum quality. ANALYZER ACCURACY The entire program rests on the accuracy of the on-line analyzer; without accuracy and some precision, all of the operator screens are simply pretty pictures. Considerable effort has gone into maintaining calibration on the on-line analyzer, such that all of the resulting screens can be used by operations and engineering with confidence. An extensive program was implemented in the spring of 2012, comparing the results of the analyzer with laboratory measurements of coals taken at the same time and from the same belt location as the coals undergoing real-time analysis by the XRF system. Figures 8 - 13 show a representative selection of analyzer-laboratory comparisons. Note that these comparisons have, in fact, been made for all parameters. 8 Figure 8. Fuel Ash Comparison Figure 9. Volatile Matter Comparison Figure 10. Calorific Value Comparison The figures shown in this comparison document the fact that the analyzer is quite accurate for the basic parameters. It shows comparable accuracy for moisture content, fixed carbon content, and alumina content. This accuracy is further shown when comparing the results of firing with and without REF, the inorganic additive used by Monroe Power Plant. During the initial test of REF, on a single boiler, comparisons were made of the coal with and without the additive. The additive was not used in 3 of the 4 boilers. Two such comparisons are shown in Figures 14 and 15. These document the sensitivity of the XRF to changing fuel conditions. Some parameters did not show the level of accuracy and precision identified in Figures 8 - 13. These included calcium oxide, magnesium oxide, sodium oxide, and potassium oxide. While the trends were accurate, there was some spread in the data and in the comparative values. Given the fact that the on-line analyzer is a production instrument-not a laboratory instrument- its results are sufficient for boiler operation, even for those compounds identified above. Significantly, the XRF did pick up the change in calcium concentration resulting from the addition of the REF additive. The overall conclusion is inescapable: the on-line analyzer providing coal compositional data on a real-time basis is sufficiently accurate for operations. 9 Figure 11. Silica Value Comparison Figure 12. Iron Oxide Value Comparison Figure 13. Base/Acid Ratio Value Comparison Because the system has been shown to be accurate, it has become one basis for Unit operators to manage their equipment. It has permitted the use of optimum technical/economic coal blends without causing undue boiler outages as a consequence of coal composition. Future Activities The on-line analyzer, or XRF, has been the focus of continuous improvement over the past 10 years. It has gone from an instrument of little consequence to one that is at the foundation of operations. The operator advisory screens, which include in-silo blends, have gone from simple depictions of the most basic fuel information to substantial advisory devices showing subtle as well as gross trends-and showing efficiency, temperature, and flyash resistivity. They provide information to operators and engineers alike. Future activities will include developing further advisory screens for both operators and engineers alike. More extensive training is planned such that operators and engineers can maximize the use of this information. Other specialized instrumentation may be added as well, making the information available to operators even more complete. 10 Figure 14. Comparison of Calorific Value of Fuel Measured by Analyzer with and Without REF Added to the Fuel. Note that the instrument picked up the drop in Btu/lb caused by REF Figure 15. Calcium Oxide Measurements Before and At the Start of the REF Test. Note that the XRF Analyzer was Sensitive and Accurate Sufficient to Identify the Addition of Calcium Oxide from the Use of the Inorganic Additive.