电站运行优化问题的决策支持系统方案_英文_

电站运行优化问题的决策支持系统_英文_ D S S Sc he me fo r Op eratio n Optimiz atio n Pro ble m o n Power Pla nt s ΞJ iang Hao Xu Zhigao ()Department of Power Engineering , Southeast University , Nanjing 210096 , China Abstract : Based on the target analysi s of the operation optimization for power plant s , a novel system scheme called () operation optimization deci sion support system OODSSi s brought forward. According to the structure and design thinking () of deci sion support system DSS, the overall structure of the OODSS i s studied , and the scheme of the sub2systems in the OODSS such as the user interface system , the problem processing system , the database system , the model base system , ( ) the expert system ESand the data mining system are p ut forward. Key words : power plant , operation optimization , DSS , database system , model base system , data mining , expert system ) 1On the object orientation , they always orient to According to the developing status of the electric the monitoring , but not the deci sion. power industry , the economical operating of power ) 2On the designing orientation , they can p rocess plant s i s a very important p roblem to study. The only structured p roblems , but not semi2structured operation optimization i s a powerf ul method for the p roblems or non2structured p roblems. economical operating management . In thi s paper , a ) 3On the instruction aspect , they are short of the novel operation optimization system scheme called ability to p rocess operation instruction based on () operation optimization DSS OODSSbased on the DSS () artificial intelligence A I. architecture i s studied. The power plant s operation optimization manage2 1 Bac kgroun d ment i s an operator2manager oriented , semi2structured p roblem type deci sion p roblem , and should be The operation optimization system for power plant s designed and implemented with the DSS architecture . ( ) PPOOSi s a software packet who se main object i s to () The operation optimization DSS OODSSmakes make an evaluation or a deci sion on the operating status use of the data p rocessing , model p rocessing and of the power unit with some defined target value , and to 1 analysi s ability of the comp uter system , and help s the make guidance on the unit adjustment. These operators and the managers to make efficiency deci sion include 4 content s : with a powerf ul user interface system. Using the DSS ) 1Provide the operators the performance value to architecture , the OODSS can p rocess the semi struc2 reflect the real2time operating status with the online tured p roblems successf ully. performance calculation. ) 2Provide the operators the optimization target 2 Overall Str ucture value to reflect the optimized operating status with the online optimization value analysi s. 2 . 1 Ba sic str ucture ) 3Provide the operators the online diagno si s and The basic structure of the OODSS i s shown in the adjustment deci sion support with the deviation 2 Fig. 1 . It consi st s of 3 part s : an integrated analysi s. ()component , which includes a user interface U I ) 4Provide the managers the stati stical report ( ) system and a p roblem p rocessing PPsystem , a model about the operating status examination and the () () base MBsystem and a data base DBsystem. The equipment adjustment deci sion support . basic p roblems of the PPOOS are p rocessed by the operation The study and the application of the p roblem p rocessing system together with the user optimization are very active now , but there are al so interface system , the MB system and the DB system. some p roblems in thi s field : Received 2002206218 . Ξ Born in 1976 , male , graduate . Partic ular Str ucture 3 In the OODSS now used , there are several component s , such as the U I system , the PP system , the MB system , the DB system , the DM system and the expert system. Each component has it s own structure . 3 . 1 M B syste m Fig. 1 Basic structure of OODSS The MB system should p rovide model support to 2 . 2 Synthetic str ucture all of the p roblems of PPOOS. As we have di scussed , the operation optimization system has it s own p roblem () The data warehouse DW, the online analysi s structure ; the model system organization structure i s () () p rocessing OLAP, and the data mining DMare something like it . The structure i s shown in Fig. 3 , and new techniques on DSS. They can p rovide novel it i s made up of many sub2model s. The model system schemes for some difficult p roblems on PPOOS , such structure has several layers , and some part s of them are as the optimization value p roblem. shown. ( ) The expert system ESi s used to construct the ( ) intelligent deci sion support system IDSS. ES can help the DSS to p rocess the operation instruction p roblem. The synthetic structure of OODSS shown in 2 Fig. 2i s made up of DW , OLAP , DM , ES and the component s in the basic structure . Thi s structure consi st s of 3 sub2systems : the 1st sub2system i s made up of the MB system and the DB system , the 2nd i s made up of the ES and the DM system , and the 3rd i s Fig. 3 Model system organization structure made up of the DW system and the OLAP. Abbreviation in Fig. 3 : DVM : data validation model ; PCM : performance calculation model ; OVM : optimization value model ; DAM : deviation analysi s model ; DPM : data p rocess model ; SRM : stati stic report model ; A IM : artificial intelligence model ; TM : turbine calculation model ; BM : boiler calculation model ; UM : unit calculation model . The MB system i s made up of the model base () MBand the model base management system () MBMS, and it s structure i s shown in Fig. 4 . The MB consi st s of the model system shown in (Fig. 3 and the model index al so called model ) dictionaryp roviding the index for the model caller . Fig. 2 Synthetic structure of OODSS The MBMS has 2 main f unctions called static managing and dynamic managing. The static managing f unction synthetic Not all of the 3 sub2systems of the includes sub2f unctions such as model creating , model structure are needed to make OODSS. Each sub2system deleting , model querying , and model maintaining. The i s independent , and can be integrated. The OODSS , which will be referred subsequently , i s made up of the dynamic managing2module performs the model s’ running managing and p rovides model s’combination1st and 2nd sub2system , and it i s an IDSS in fact . Ta b. 1 OV analysi s result of a 200 MW power unit Load Item 200 MW 180 MW 150 MW Reference/ Optima l Reference/ Optima l Reference/ Optima l p/ MPa 12. 90/ 12. 96 12. 70/ 13. 06 12. 60/ 12. 93 0 t/ ? 0 535. 00/ 538. 98 535. 00/ 538. 47 535. 00/ 536. 15 t/ ? r 535. 00/ 538. 76 535. 00/ 537. 64 535. 00/ 532. 39 p/ kPa c 5. 20/ 5. 44 4. 80/ 4. 84 4. 20/ 4. 19 t/ ? fw 243. 00/ 238. 06 234. 60/ 234. 61 220. 40/ 225. 96 t/ ? eg 129. 60/ 129. 31 126. 00/ 125. 69 120. 10/ 122. 32 In Tab. 1 pi s the p ressure of main steam ; ti s 0 0 Fig. 4 MB system organization structure the temperature of main steam ; t i s the temperature of r support as well . reheated steam ; pi s the p ressure of turbine exhaust c For example , during the unit performance steam ; ti s the temperature of feed water ; t i s the fw r calculation , the MB system will p rovide running model temperature of boiler exhaust gas. support to the performance calculation module in the The optimal value based on the DM method and p roblem p rocessing system ; the running managing the reference value based on the traditional method are ( module in the MBMS call s the appointed model such both li sted , and with a compare of them , it can be as the turbine calculation model , the boiler calculation concluded that the DM based OV analysi s i s efficient ) model , etc . with the model index to realize thi s and usef ul . support f unction. Since the optimal value and reference value can 3 . 2 D B syste m both p rovide the optimization target value reference to reflect the optimized operating status , certainty factor The DB system should p rovide structured data or weight for each of them must be determined to get support to all of the p roblems of PPOOS. The real2time the universal target value . It i s a multiple model data i s obtained f rom DCS data highway or plant based , semi2structured p roblem in fact ; OODSS can 3 real2time database , and then stored; the manual data connect together with users through the U I system to i s inp utted by user and then stored ; the outp ut data of p rocess thi s type of p roblems , and the users can give the OODSS needs to be stored as well . All of these are the system some extra information about the certainty included in the database of the OODSS. The OODSS factor or weight for each of these value series. DB system i s made up of the database and the database 3 . 4 Expert syste m () management system DBMS, and the DBMS i s a bridge between the OODSS database and the p rocessing ( ) The expert system EShelp s the DSS to p rocess modules in the p roblem p rocessing system. () the operation instruction O Ip roblem. The deviation of controllable operating parameter between it s p hysical 3 . 3 DM syste m value and it s optimization value in a running unit i s The DM system di scovers knowledge needed by caused by some factors of the operation level , the unit OODSS f rom the database with the given method or status , and the automatic control system. The technique . On the one hand , it can p rovide knowledge operation instruction system should have the ability to ( ) to the ES through the knowledge base KBsystem , determine the cause and the effect of the deviation , and and on the other hand , it links the MB system , and give the guidance and the assi stance to the operators. p rovides support to the DVM and OVM in Fig. 3 . In the ES , the deviation cause and it s settling scheme Take the OV module for example : with the can be stored in the knowledge base , and then the stati stical analysi s on the database of the OODSS , the inference engine can make an online diagno si s while DM system can obtain the optimal value of each the unit i s operating and give operators the online performance target s and operating parameters with the guidance or the assi stance . relationship to the load of the unit . In thi s way , the 3 . 5 Integra te d component DM system can help the OV module to f ulfill the optimization value analysi s. An example of The integrated component including the U I system optimization value analysi s result based on the DM and the PP system i s the head component in the method i s li sted in Tab. 1 . OODSS. It works under the monitoring of the users , and communicates with the other component s of the “DA module”and these two modules i s then p ut to theOODSS to p rocess all of the p roblems of PPOOS. The the“O I module”, which perform the deviation analysi s running logic structure of thi s component i s shown in and the operation instruction. The“SR module”f ulfill s 4 (Fig. 5the sub2systems called by the modules are the stati stic report f unction. All the result s of the ) marked in the bracket s, and it i s al so the main modules are fed back to the users through the U I system. running logic of the whole OODSS. Take the“PC module”for an example , when the module i s called , it get s the inp ut data f rom the DB system , and get s the running model support f rom the MB system , and then performs the calculation , and p ut s the calculation result to the DB system in the end. 4 Concl usion In thi s paper , the overall structure of the OODSS and the design scheme of it s sub2systems such as the integrated component , the MB system , the DB system , the DM system and the expert system are p ut forward. Thi s novel scheme for the power plant s’operation optimization system can resolve the p roblems in the PPOOS field di scussed in part 1 , and it i s p roved to be efficient , p reponderant and recommendable . References 1 Hu Huajin. Study on the performance monitoring in fossil Fig. 5 Logic structure of integrated component fueled power plant D . Nanjing : Southeast University , 1998. Abbreviation in Fig. 5 : ()5 7. in Chinese DV i s the data validation ; PC i s the performance 2 Chen Wenwei . Decision support system and its development . calculation ; OV i s the optimization value analysi s ; 2nd ed. M . Beijing : Tsinghua University Press , 2000. 95 DA i s the deviation analysi s ; O I i s the operation ()98 ;324 328. in Chinese instruction ; SR i s the stati stic report . 3 Jiang Hao , Xu Zhigao , Gao Zhengping. Designing and implement of data channels in RT2MIS for power plant J . Automation of The real2time data and manual inp ut data i s obtained by the“data inp ut and DV module”and then () ()Electric Power System , 2002 , 26 2:62 64. in Chinese 4 Jiang Hao , Xu Zhigao. A study on operation optimization DSS for shared by all of the modules in the p roblem p rocessing ( ) a power plant A . In : National Conference on Automation and PPsystem. The “PC module”f ulfill s the perfor2 (mance calculation f unction while the “OV module” 84. in Inf ormation f or Power Plants C . Nanjing , 2002. 82 )Chinese performs the optimization value analysi s. The result of 电站运行优化问题的决策支持系统方案 江 浩 徐治皋 ()东南大学动力系 ,南京 210096 () 摘 要 基于电站运行优化问题的目标分析 ,提出新颖的电站运行优化决策支持系统 OODSS解 () 决方案. 依照决策支持系统 DSS的结构及其思想 ,对 OODSS 的总体结构进行了研究 ,并对其 ( ) 主干子系统 :包括人机交互系统 、问题处理系统 、数据库系统 、模型库系统 、专家系统 ES与数据开 采系统等的设计提出合理的方案. 关键词 发电厂 ; 运行优化 ; 决策支持系统 ; 数据库系统 ; 模型库系统 ; 数据开采 ; 专家系统 中图分类号 TM621 ; TP315