智能计算机系统提高发电厂的运行能力 中英文对照

1、Intelligent computer systems enhance power plant operationsDouglas J.Smith Without computers and their associated technology, todays power industry would be far less efficient and the maintenance and operation of power plants would be more costly. Computers are used for operating plants, diagnosing

2、operating problems, optimizing the purchase of fuels, and scheduling outages.Artificial intelligence (AI) is one branch of computer science that models knowledge to solve problems in the same way that humans solve problems. The increased speed and power of computers plus their shrinking size has hel

3、ped to expand the use of AI.(Another area in which AI is being used in the power industry is in robotics .See POWER ENGINEERING , July, 1989, page 22.)One of the fastest growing areas of AI is the expert system. Expert systems are programs that incorporate the knowledge of experts. The knowledge con

4、tained in such a system can be used to manipulate data in a manner similar to that used in the human reasoning process.In less than 20years, power engineers have gone from the slide rule to handheld calculators to todays minicomputers .Now, almost all power engineers routinely use PCs.However, quest

5、ions remain about the future use of computers in the power industry:. Where is the technology taking us?. What is the real payoff in terms of greater competitiveness and efficiency of operation?. How is it going to change the way utilities conduct their business?. What is the status of research in a

6、dvanced computer technology and when will this be commercially available to the electric power industry?According to the North American Electric Reliability Council (NERE), the leading cause of forced fossil power plant outages is boiler tube failures. In coal-fired boilers, 3.16 of the forced outag

7、es per year, for each boiler, are attributed to failures of waterwall and superheater tubes. Waterwall tube failures also are the major cause of forced outages in gas-fired and oil-fired units. Superheater tube failures are the third most frequent cause.ESCARTA expert systemOne expert system current

8、ly under evaluation by EPRI is the ESCARTA expert system. This system has been designed to help reduce boiler rube failures (BTF) in fossil steam plants.The systems knowledge base has been developed from the EPRI manual, Investigation and correction of boiler tube failures. In operation, the system

9、first divides the boiler into four main regions: waterwall, superheater, reheater, and economizer. Then it asks the user a series of questions solicit more detailed information should the need arise.ESCARTA has been designed for personnel who are not experts in boiler tube failure analysis. Power pl

10、ant generation and operation managers, maintenance staff, metallurgists, and other plant personnel can easily use the system. Its use requires only two keys on the keyboard, the space bar or arrow keys, and the enter key.According to EPRI, plant personnel are able to use the expert system after less

11、 than thirty minutes of computer training. This is also true for people who have never used a personal computer. The system has been designed for IBM XT, AT, or other compatible computers. It requires 640kbytes of RAM plus a hard disk. The overall structure and functions of ESCARTA are shown in Figu

12、re 1.When a tube failure occurs, the system first determines the failure mechanisms based upon the location of the failed tube, its appearance, and the events leading up to its failure. Diagnosis of the problem is then carried out by obtaining information using “if-then” rules. In this way the exper

13、t system is able to determine one of 22 possible failure mechanisms. From this information it is able to recommend a course of action. Figure 3 shows a partial decision three for finding tube failure locations.Location of failed tubes in waterwalls is referenced relative to a burner level, bends wel

14、ds, and/or weld attachments. Economizer tube failures are located according to their proximity to the feedwater inlet, bends, or flue gas inlet. Superheater tube failures are located in a similar manner.At the same time as the system is asking questions about the location of tube failures, it is als

15、o being asked questions about events leading up to the tube failure. Once information about the tube location and the events leading up to the failure have been determined , the expert system then begins to gather information about the appearance of the failed tube.After determining all the informat

16、ion leading up to the tube failure, and its location, the system tells the plant operator some of the root causes that may have led to the failure.These could include poor quality fuel, malfunction of equipment, and improper operation.To verify a root cause so that corrective action can be taken, th

17、e operator can request the system to supply information on non-destructive evaluation methods, metallurgical tests, and repair procedures.Besides its diagnostic capabilities, ESCARTA creates data bases of the tube failures. Data bases for each of the major sections of the boiler, waterwall economize

18、r, superheater, and reheater are all maintained within the system. Each of the data bases contains:. Outage type, with data and duration. Failure location. Failure appearance. Material and dimensions of the failed tube. The failure mechanism that caused the tube to fail . Details of weld repair, inc

19、luding the contractor usedField experiences with ESCARTAA trial demonstration of ESCARTA has been conducted by five electric utilities: Baltimore Gas & Electric Co., Central Power & Light Co., Ohio Edison, Public Service Co. of Oklahoma, and Delmarva Power Co. During the demonstration phase, each of

20、 the host utilities was given a copy of ESCARTA and trained in its use. After investigating a boiler tube failure, the host utility completed an evaluation form, which was forwarded to EPRI.To become familiar with its use and to determine whether it would be useful to their particular units, the uti

21、lities first used the system to review boiler tube failures that had previously occurred at the plant. After several months of onsite use the utilities reported their use of the system to EPRI. At the same time they indicated areas where the system could be improved.After reviewing and evaluating th

22、e comments and suggestions made by the host utilities, the knowledge base was revised. This revision, which included changes proposed by the utilities, was again distributed to the five utilities. The revised expert system is now being used in the host utilities boiler tube failure investigation and

23、 prevention programs.Results from the field tests revealed the usefulness of ESCARTA. Besides diagnosing boiler tube failures it was also found to be an excellent tool for training power plant personnel about the way boiler tube failures occur. In addition, it is able to show the operators the corre

24、ct procedure for investigating boiler tube failures.Main menuTubespecificationsFailuredatabaseRoot causeNDEWeldingproceduresFailuremechanismContext-sensitive infoDiagnosismoduleCorrectiveactionRepairReferencesMetallurgyOperatingproceduresFigure 1. Overall structure and functions of ESCARTA.One utili

25、ty, after fully evaluating the system, determined that it could save at least three outage days a year by using the expert system. The major saving is due to the reduced time required to locate a damaged tube. EPRI estimates that if a 500-MW power plant, using ESCARTA, were to reduce its outages by

26、three days a year, it could save more than $2 million.By using ESCARTA, utility companies are able to reduce the probability of repeat tube failures. This not only reduces maintenance costs and downtime but also improves plant availability.The system has been designed primarily for use by supervisor

27、y staff. However, it can assist less experienced power plant personnel by helping them to evaluate and diagnose boiler tube problems.CQA evaluates coal quality impactsHouston Lighting & Power Co. (HL&P) and Stone & Webster, in a joint project, have recently completed the development of an expert sys

28、tem for use by utility coal buyers. The coal quality advisor (CQA) runs on a personal computer and is used to assist coal buyers in making detailed assessments of coal costs and the performance impacts of using different coals or coal blends.In this particular application a pulverized coal-fired pla

29、nt with a baghouse and a wet limestone scrubber was the model. HL&P says that the capability exists to replace the baghouse with an electrostatic precipitator, thus allowing the utility to model different pulverized coal-fired units in the HL&P system.Figure 2. Typical solution process used in Houst

30、on Lighting & Powers coal quality advisor. Source: Houston Lighting & Power Co.Slagging index for coal Compare base and actual slagging indexCompare base and actual fouling indexNo. of and usage of deslaggersOperating and maintenance costs of deslagger and sootblower systemsOperator experience1.Soot

31、blowereffectiveness2.Deslagger effectiveness3 FrequendyDisplay actual and limit on paramete-rsClassifycoal qualityimpactsDisplaynetchangeofO&McostsDisplayrulesforeachlimitinglactorDisplayengineer-ingimpactsandadviseNo.of and usage of sootblo-wersFouli-ngindexfor coalDiagnosing heat rate degradationS

32、mart operators aid for power plant optimization (SMOP) is an expert system developed for diagnosing the causes of heat rate degradation on oil- and gas-fired power plants. The project is being developed jointly by Southern California Edison (SCE) and Combustion Engineerings Impell Corp.SMOP is a kno

33、wledged-based system which allows plant operations and performance personnel to optimize a plants heat rate. According to the developers the system is generic and portable.The system is a real-time expert system shell. It can also be embedded and fully integrated into a power plants computer system.

34、 It also is able to utilize data from other systems.SCE has integrated the first prototype system into the PPMS Micro VAX computer at its Huntington Beach GENERATING STATION. The SMOP system has been designed as an operators aid and is used to facilitate the diagnosis of certain operator-controllabl

35、e losses. In addition, it is able to diagnose some system and equipment malfunctions.Of importance to many operators is the systems ability to justify its diagnosis. This is accomplished by highlighting the most probable cause in a logic tree graphic display or by operator messages. These operator m

36、essages enumerate data and logic used by the system in arriving at the conclusion or conclusions.After reaching its conclusion, the system then recommends corrective actions that can be taken. These are prioritized by their ease of execution and relative economic impact in $/shift.The system address

37、es eleven operator controllable losses:1.Condenser system losses 2.Reheater spray flow 3.Reheater temperature4.Main steam temperature 5.Main steam pressure 6.Stack gas exit temperature7.Auxiliary MW losses 8.Auxiliary steam losses 9.Boiler excess air10.Boiler carbon monoxide 11.Turbine cycle lossesE

38、ach of the above parameters has an associated knowledge-base. Losses incurred because of deviations in the reheat and steam temperature are further subdivided according to whether the temperatures are too high or too low.The man-machine interface consists of color graphics that portray portions of t

39、he plant. Data on heat losses and where they are occurring are included in the same color graphics. The screens are touch-sensitive.Currently, the system is undergoing performance testing at the .Huntington Beach station.On-line diagnostic systemWestinghouse Electric Corp. was one of the first compa

40、nies to see the advantages of using artificial intelligence for diagnostic analysis of power plant equipment. Its on-line diagnostic operations center in Orlando, Fla., Figure 5, is linked by redundant telephone lines to various customer plant data centers.The diagnostic center together with the pla

41、nt data centers provide round-the-clock real time diagnostics of turbine-generators. This system considers the turbine-generators as three subsystems. each having its own diagnostic rule base, but with all sharing a common data base of information. The three expert systems are Turbin AID, Gen AID, a

42、nd Chem AID.Thermodynamic and mechanical conditions in the turbine and the turbine cycle are diagnosed by the Turbin AID expert system. This system diagnoses the condition of the turbine, the performance of the high-pressure feed water heaters, the condenser, and the overall heaters, the condenser,

43、and the overall turbine cycle.At the plant data center the system receives information from sensors mounted on the turbine. Using this information, the system calculates various performance indices, including:. Gross heat rate. Efficiencies of the HP, IP, and LP turbines. Terminal temperature differ

44、ences of the heaters. Drain cooler approach temperature differences. Condenser hotwell subcoolingOperators are able to minimize operation costs through key controllable parameters which are displayed on CRT screens together with the heat rate and cost effects of any deviation, Figure 6.A similar scr

45、een displays the heat rate and cost effects of off-target operation of key performance parameters including the HP, IP, and LP turbine efficiencies.SubsystemsGen AID, the generator subsystem, monitors and analyzes more than 250 different inputs of water-cooled generators. The system has been configu

46、red with a combination of sensors and monitors that are segregated into five generator subsystems. These subsystems monitor and analyze the stator winding system, hydrogen auxiliary system, seal oil system excitation system, and mechanical rotating parts system.Controlling a plants chemistry is impo

47、rtant if a power plant is to be efficiently operated. Chemists must have the capability of determining which sources of chemicals are active at any given time. They also must ascertain whether the chemical purification systems are working properly.Westinghouses expert system Chem AID carries out the

48、 above functions by interpreting data supplied by the plant data center monitoring system. The system can diagnose the chemical condition of a power plant in real time.Water chemistry is generally controlled in power plants by manual water sampling and analysis. Complete testing and sampling of a po

49、wer plants water chemistry can take several hours. Unfortunately, significant chemical excursions can occur in 10 minutes or less. Thus manual sampling is not always timely in picking up excursions. With the Westinghouse system, data on the plants water chemistry is analyzed at five-minute intervals

50、.Utilizing this data, the system is able to provide the stations chemists with a variety of tabular and graphical display options. By using the touch-sensitive screens and window-based displays, inexperienced users can find the operating status of the unit and obtain historical data on the plants ch

51、emistry.General Physics and New York State Electric & Gas Co. (NYSEG) have jointly developed a prototype plant performance expert system they call X-TIP (expert thermal information program).The system was developed using an IBM AT with dual monitors-a monochrome monitor for text and a color-graphic

52、monitor for graphics. C language was used as the utility programming language.Importance of a good interfaceDuring development of the system it was found that the most important part, in regard to its acceptance by the operators, was the user interface. In addition, it was determined that high quali

53、ty and easy-to-read color graphics are also essential if users are to accept such systems.Another operator requirement is that the system should provide the most important information first. Secondary information should be available on demand but should not be forced on the user.While not necessaril

54、y important to the operators, the developers found that a system should be designed so that maintenance and/or modifications can be carried out easily. This is important to the systems performance engineers.As a result of the success of the system prototype, General Physics and NYSEG are developing a full-scale model which they will call the fossil thermal performance advisor (FTPA). The scope of the FTPA has been widened beyond that of the original model.Early 1990 will see