Systemwide power plant performance monitoring using microcomputers

  • Published on

  • View

  • Download


684 IEEE Transactions on Power Systems, Vol. 3, No. 2, May 1988 SYSTWIDE POWER PLANT PERFORMANCE MONITORING USING MICROCOMPUTERS M. M. Hillhouse Gulf States Utilities Company P. 0. Box 2951 Beaumont, Texas 77704 Abstract - Gulf States Utilities recently installed microcomputer-based monitoring systems in its power plants to accurately compute fuel consumption. electrical generation, and unit efficiencies. Gulf States engineers programmed the microcobnputers using BASIC language to perform instantaneous and totalized energy calculations from continuous metering. Electrical generation, boiler and supplier BTU fuel flows, and unit heat rates are used by the plant operators to make adjustments to improve unit efficiency. The information is also automatically transmitted via a SCADA system to the Energy Control Center. This information is used to meet generation requirements with the most efficient generating units, while minimizing fuel costs. INTRODUCTION Gulf States Utilities (GSU) recently completed installation of a system-wide energy management system. Previous papers have described the central Energy Control Center (ECC), the five Division Control Centers, and the Remote Terminal Units (RTUs). See references attached. This paper describes the power plant microcomputers and associated equipment. !lhenty-nine generating units at eight plant sites were instrumented. Microcomputers were installed at six plants to accurately calculate fuel consumption and unit heat rates. Computers were not added at two plants, as fuel calculations were not required. Figure 1 shows the power plant system overview. Plant personnel use the information for historical reports and to optimize generating units efficiency. The fnformation is also used at the ECC to minimize systemwide fuel and electrical generation costs. Prior to this project volumetric gas and oil metering accuracy was 5% or worse on each generating unit. Fuel consumption was manually calculated by Operators and phoned into the ECC. Instantaneous fuel flows and unit heat rates were not availdble. Goals of the new system were to: (1) provide accurate continuous fuel, electrical, and heat rate monitoring to better control natural gas suppliers and to dispatch units more efficiently; and (2) provide an in-plant tool for Operators to improve unit performance. The following plant information was provided to the ECC in this project: 1) Instantaneous and totalized .'Unit gas consumption . Unit oil consumption,(No. 2 oil) . Unit oil consumption (No. 6 oil) . Unit coal consumption ' . Unit gross generation This paper was sponsored by the IEEE Power Engineering society for presentation at the IEEE Power Industry Computer Applica- tion Conference, Montreal, Canada, May 18-21,1987. Manuscript was published in the 1987 PICA Conference Record. . Unit station service electrical load . Gas supplier flow 2) Instantaneous unit heat rate 3) Status . Computer power loss . Transmitter in calibration mode . Gas supplier on pressure (versus flow) control Fuel flows transmitted to the ECC are in British Thermal Units (BTUs). Fuel flows provided at the plant level are in both BTU and volumetric units. Accuracy goals were determined based on American Gas Association (AGA) and American Petroleum Institute (API) standards: . Gas - 1% volumetric; 2% BTU . Oil - 1% volumetric; 3% BTU . Coal - 3% weight; 5% BTU Oil and coal BTU accuracies were worse than gas since constants were used instead of continuous heating value instruments. These instruments were not commercially available for oil, and were unreliable and cost prohibitive for coal. Electrical signal accuracy goals for gross generation and station service usage were 1%. The turndown goal for flow signals (ratio of maximum flow to minimum flow that is accurately measured) is 10 to 1. No preprogrammed device was available to handle the variety of inputs, outputs, and special functions required for this application. Design goals were to minimize equipment complexity and cost, yet maintain flexibility to make future changes as functional requirements evolved. The computer equipment had to be easily programmable in a high level language to perform 4 variety of complex calculations and operational functions. To minimize field wiring costs and the number of units affected by a computer or RTU failure, the decision was made to provide a computer and RTU for each control room. In several cases two or three units share a control room and computer. The individual system architecture is shown in figure 2. FIELD INPUTS As much as 90% of GSU's fossil plant generation is provided by natural gas fired units. Because of this and the complexity of our gas supplier contracts, it is important that gas is accurately measured. Gas flow calculations are defined by the American Gas kssociatioq Report No. 3 (AGA3). Metering installations were upgraded by replacing orifice plates, adding flange taps, and adding straightening vanes. New resistance temperature detectors (RTDs) and pressure transmitters were added. Both low and high range differential pressure transmitters were installed across each orifice. The low range transmitter accurately measures differential pressures from 1% to 10%. which corresponds to flows of 10% to 32%. The high range transmitter accurately measures differential pressures from 10% to loo%, which corresponds to flows of 32% to 100%. Since the square root of differential pressure is used to calculate flow, the dual transmitter method improved total flow turndown from "3 to 1" to "10 to 1". Gas chromatographs were added to measure heating value and specific gravity. The computer automatically selects the proper differential pressure transmitter, uses the six inputs for each gas 0885-8950/88/0504Wl .WO 1988 IEEE 685 stream, and calculates volumetric and BTU gas flow. The computer is also programmed with a low range cut off (1% of span) feature to eliminate flow errors caused by transmitter drift. Several of our units can burn No. 2 or No. 6 fuel oil in addition to natural gas. These were instrumented with oil metering per American Petroleum Institute standards. Turbine meters and RTDs were added to No. 2 oil piping. Signal adapters were added to existing No. 6 oil positive displacement meters along with RTDs. In order to detect and compensate for meter inaccuracy, a portable trailer-mounted ballistic oil prover was purchased. It has a 1000 to 1 turndown with a 0.01% accuracy. By inserting it in series with a meter, a K-factor (flow correction multiplier) is determined over the full flow range. Oil heating value is determined based on samples analyzed on a laboratory bomb calorimeter. The meter K-factor and oil heating value are manually entered into the computer by the Operator. From the oil inputs and constants, volumetric and BTU oil flows are calculated. Only one of GSU's units burns coal, and it was also instrumented. Boiler coal flow signals are received from gravimetric feeders. Weighing scales also monitor the conveyors feeding coal silos, and provide a more accurate indication of daily flow (about 1%). Heating value is obtained from laboratory samples measured bomb calorimeter. The heating value is manua ly entered into the computer by the Operator. The computer calculates weight and BTU flows. 7" Electrical wattfvar transducers and watt-hour meter pulse initiators were added. Test switches were installed to allow on-line maintenance. The electrical signals are paralleled to both the computer and the RTU. Thus a microcomputer failure will not affect the ECC's Economic Dispatch program. The microcomputer uses the electrical values in power factor and heat rate calculations. The number of inputs to and outputs from each computer system was intentionally limited to a few key parameters. This minimized maintenance requirements. The small number of field instruments are closely monitored to ensure they remain in calibration. COMPUTERS GSU decided to purchase a general purpose programmable computer and write our own pragrams. We could not predict how much memory would be required for our application programs on any manufacturer's computer. For this reason we leased our first computer, with the lease cost applying towards purchase of the first computer. The leased computer had to pass an on-site performance test with our application program running before actual purchase of it and twelve additional computers were made. This allowed GSU to verify that the computer selected would meet our present requirements as well as allow for future expansion. After installation, each computer was required to pass an on-site 90 day acceptance test with a 99% or better availability. Progradng The Acurex Autodata 10150 microcomputer was selected for our application. It has the following general capabilities: . Detachable ASCII programming keyboard . Front panel operator and programing interface . 5 inch diagonal CRT, 21 characterlline alphanumeric printer . Fill in the blank "Menu" programming for inputs and outputs BASIC programming for calculations and log sheets 128K system PROM, 64K programmable RAM memory On board cassette (300K) for program save or load Math package, input scaling, alarms, keylock, report formatting Battery backed clock GSU engineers programmed the microcomputers using an extended BASIC language. The program is recorded on data cassette tape, and will automatically load and begin executing following a power interruption. Descriptive variable names have been used to make the program easier to understand by a user. The program calculates fuel flows, power factor, and heat rates; formats and prints reports; displays information on the CRT; manipulates and stores data; and provides an easy to use interface for entering or viewing system constants and variables. The program also calculates a target heat rate as a function of unit net generation. The target heat rate is developed from performance test results and recent unit operating characteristics. The Operator uses target heat rate as a tool to improve unit efficiency. It is a guide by which he can adjust controllable parameters, such as excess air flow, steam pressure, and steam temperature. Additionally, calculations are made for heat rate variance, and the dollar value of this variance based on unit fuel costs. Variance is the difference between actual and target heat rate. Thus the Operator can see how much better or worse his unit is performing in comparison to the target. Operator Interface The system is designed to operate with a minimum of Operator interaction. Normally the machine will scan all field inputs, update calculated outputs, and print reports. Routine reports are printed each hour, each operator shift (8 or 12 hour), and each day. Figure 3 shows a typical hourly report. Scan rates vary between 10 and 45 seconds, depending on the number of inputs, outputs, and calculations required of each of the 13 microcomputers. Pushbuttons on the front panel were programmed to perform the following functions: Display and update on the CRT each scan cycle gross generation, net generation, actual heat rate, and target heat rate. Print a report that shows the instantaneous value of all continuous inputs, constants, and calculated quantities. It is a useful check of instrument calibration and calculation accuracy. See Figure 4. Enter a constant value for any transmitter that is an input to the system. Maintenance can then be performed without impairing data sent to the ECC. Plant printouts and the ECC are flagged automatically that calculated results may be in question. Enter unit inputloutput curve coefficients. The program will calculate a unit target heat rate as a function of net generation. The equation used is of the form Y = AfX + B + CX, where A, E, and C are the inputloutput curve coefficients as determined by performance tests, Y is the target heat rate, and X is the unit net load in megawatts. Enter unit fuel cost (in dollars per million BTU), which is used to calculate the dollar value of variances from target heat rate. 686 Enter orifice plate sizes for those gas suppliers who frequently change orifice plates. Enter the heating value of No. 2 oil, No. 6 oil, and coal, based on the latest bomb calorimeter test results. Enter the average K-factor for each oil meter based on oil prover runs over the full flow range. The computer corrects the volumetric flow with this factor. Calculate an average heat rate over a time period determined by the Operator. This is useful during performance tests. Validation The AGA3 calculations are relatively complex and time consuming to manually perform. To quickly verify that each computer is correctly calculating gas flow from a given set of field inputs, two tools have been provided. First, GSU wrote an AGA3 program in FORTRAN on the corporate computer, accessible at each plant by a mainframe terminal. Second, a demand listing can be printed by each microcomputer's printer (Figure 4). The list includes all inputs, constants, and calculated outputs for a single point in ti=. By entering the demand listing input values into the corporate computer, calculated outputs can be verified. ECC - Each plant microcomputer sends output signals to an RTU. The interface is 4-20 ma for instantaneous values. Contact closures are used for status and accumulated values. Accumulated information is totaled in the RTU. The ECC computer polls the RTUs, gathering analog values every 10 seconds and accumulated values every ways. 30 minutes. The ECC uses the data in several Each generating unit's real time fuel consumption and corresponding electrical generation are used to automatically update off-line heat rate curves to current conditions. The on-line heat rate curves used in the Economic Dispatch program are based on unit performance tests conducted approximately once a year. These tests address seasonal changes, and rated pressure versus sliding pressure operation. An Analyst or Operator can compare the on-line versus the off-line heat rate curves. If there are significant differences, such as increases due to equipment failure, actual heat rates can be substituted. System generation can then be incrementally added or reduced more efficiently. Electrical generation is used to automatically prepare load and generation forecasting reports for the next day, week, etc. Fuel flows are used for historical reports (Instantaneous, hourly, daily, monthly), and for on-line graphic displays. Daily reports are used by GSU Management to make operational decisions in terms of contract bum rates and allocation among gas suppliers. Graphic displays provide another means to prevent overburns. Since the Economic Dispatch program has no fuel constraints, the Operator can manually restrict units based on display information. The Performance Test Department also uses this data for plant statistical analysis of unit heat rates and startup fuel requirements. The startup fuel costs are factored into the unit commitment program. . Fuel flow analog values are integrated and compared to totalized values each hour. When these deviate more than a selected amount, an alarm occurs, identifying possible equipment failure. This is a useful maintenance tool. CONCLUSION This paper described the power plant energy metering microcomputers and associated field inputs installed across the GSU system. They are part of an overall energy management system, and provide information for two means of efficiency improvement. First, accurate fuel flow and heat rate information are provided as a performance tool for plant operators to use in controlling their unit. Actual gas flow results have been repeatable in the 0.5% accuracy range, which is better than the 2% goal. Actual and target heat rates have proven valuable to both plant operators and plant management. Second, the continuous signals at the ECC are used to: (1) make better, more timely gas supplier operating decisions to minimize fuel costs and to stay within plant target gas bums; and (2) update heat rate curves to current conditions for use in the economic dispatch program, thus minimizing megawatt generation costs within system constraints. Several unique features associated with this project included: . Accurate liquid and gaseous continuous flow calculations . Accurate continuous heat rates used in economic dispatch . Actual and target heat rate as an operator tool over a broad megawatt range . General purpose microcomputer programmed by GSU . Testing the computer to ensure its adequacy for the application using a leasing agreement . Oil prover for meter accuracy compensation over broad flow range . Corporate AGA3 program as maintenance tool Since the computer systems were installed in 1983 and 1984, GSU has improved its system heat rate by about 3%. The use of these microcomputer systems is one I. 2. 3. 4. 5. of several reasons for this improvement. REFERENCES L . J. Bodemann, R. W. Griffith, C. H. Lammons, and R. E. Smith, "The Pre-Implementation Phase of the Gulf States Utilities Energy Management System", IEEE COPS Conference, March 1982. C. L. Byars, R. W. Griffith, C. H. Lammons. G. S . Martirie, and R. E. Smith, "Distributed Energy Control System Installation Challenge", IEEE PICA Conference, 1984. G. L. Viviani, C. E. Lin, and M. G. Webb, "A Real Time Approach For Revising Generation Unit Performance Characteris tics", IEEE Transactions on Power Apparatus 6 Systems, February 1985, Vol. PAS-104, No. 2. R. W. Griffith, "Implementation of a Remote Fault Isolation System", Presented at the EEI Engineering and Operating Computer Forum, Engineering and Operating Division Computer Committee, Washington, DC, September 9, 1986. R. W. Griffith, "Gulf States Utilities Company's Energy Management System - A Distributed Approach", Presented at the 158th meeting of the EEI Electrical Systems and Equipment Committee, Baton Rouge, La., February 19, 1986. NO. 2 O I L INPUTS NO. 6 O I L I N P U T S . GAS INPUTS e d - 687 POWER PLANT OVERVIEW - -1 I I I I -4 I r- - - - - -* I I I I I NECHES L- r------ S A B I N E 1 I I M I CROWAVE r-------- 1 I A; RTU t rA-7 I I I I I I L-----J U N I T I NELSON COAL U N I T I U N I T S G S - GAS SUPPLIERS FIGURE I INDIVIDUAL SYSTEM ARCHITECTURE I r 1 STATUS CONTACTS - R T U - - NO. 2 O I L OUTPUTS NO. 6 O I L OUTPUTS SCADA r GAS OUTPUTS MICRO COMPUTER 4 MICROWAVE F IGURE HEAT RATE OUTPUTS E L E C T R I C A L INPUTS INTERNAL CRT, CASSETTE TAPE OPERATOR KEY BOARD 2 TYPICAL HOURLY REPORT . ._ 2 I 118.0 7.7 0.0 110.3 1131 1167 10585110433 UNITED 1053 1075 -. 1;a.- - _ _ _ 9 B . J SUGARBOWL 1 285 1340 FIGURE 3 TYPICAL DEMAND LISTING 4 -~ SPEC IF I C - B R A V f T V H E I T l N G VALUE lBTU/CF) 1020 1009 1040 STATIC PRESSURE ( P S I G l 441.4 549.5 126. k -5- .5i64--- .5989 D I F F E R E N T i A L - P R E 9 9 ~ ~ . Z V." ., ". C - 4.?------8.-44.0 ?&.t----iP. + GAS FLOW (MCFIHR) 1223 0 0 0 353 0 1026 1054 1149 1092 GAS FLOW (MMBTUIHR) 1247 0 0 0 0 1067 1096 1195 1136 356 --1 _ _ _ ~ ~ _ _ _ _ _ _ ~ UNIT DATA, UNIT 1 UNIT 2 196.0 I --S.? __ _____-_- ~ - - _ _ _ _ - ~ - GROSS GENERATION I M W l 150.8 NORMAL ST+lTlBKSERVfE-ttlUi 4-13 RESERVE STATIqN SERVICE IHW) 0.0 0.0 I 192.3 NET GENERATION (MU) 146.0 POWER FAL!T:TOR- 6:9855 P I P E DIAMETER (INCHES) 12.090 13.250 ORIF ICE DIAMETER (INCHES) 6. 160 8.990 BASE PRESSURE-tpstet 14: 696-4 r696 -.__ -2 - map--- _- 1 ___ _ _ _ _ ~ BASE TEMPERATURE (DEGREES F) 60 60 I GAS TEMPERATURE (DEOREES F) 45.2 46. I SPECIFIC O R A V I W .SE67 .se79 -_. HEATING VALUE (BTU/SCF) 1043 1045 STATIC PRESSURE (PSIG) 56.4 99.1 DP-LOW RANQE - tf"ES1Wfk, tare fS;(t DP-HIGH RANGE (INCHES WATER) 67.0 79.0 GAS FLOW (flCFIHR) 1403 less OAS FLOW tflMBTU/ttRt 1464 19313 HEAT RATE (ACTUAL BTUIKWH) 10027 100713 HEAT RATE VARIANCE FROM TAROET fBTU/KWWl 40 -152 _____ ---.--.--A -__- _ _ ~ - ~ - _ _ _ _ _ - - HEAT RATE fTARGEI-BMfKYH) 9987-------------10230 - ~ - ._ __ __ -____ - - __ MOLE PERCENT NITROGEN- 0.53 MOLE PERCENT-lZW-- $.36- - __-- - - _ _ _ _ _ _ ~ - -__ _.--____ ______ t OP!ZRATOR E N T E R E D - W W " -~ 0 % CALCULflTION BASED ON OPERATOR ENTERED F I X E D V ILUE FIGURE 4


View more >