Application of Operation Guiding Expert System for Thermal Power Units
|Keywords||operation guidance coal-fired power plant coal quality on-line monitoring optimization neural network|
The economic performance of a thermal power unit is one of the most importantworks in the daily production. On the other hand, because of the influence of coalmarket, the coal burnt in the boiler is often very different from the designated one, and itis in urgent need to provide on-line monitoring and technical guidance for the coalblending and coal quality control to improve the safety and economic operation of theunit. Through the information collection of the actual operation parameters of powerplant and the calculation of coal sent into the furnace, and with computer-based quickanalyses, it is of great engineering significance to develop a prediction model of coalquality monitoring when various types of coal are burnt to ensure the optimal economicperformance under secure operation. Based on this consideration, the ShanghaiWaigaoqiao Power Plant purchased and implemented an operation expert guidancesystem on its thermal units.This paper firstly introduced the current status of thermal power unit in ShanghaiWaigaoqiao Power Plant. Based on the literature survey of the application of theoperation expert guide system in the domestic and overseas coal-fired power plants, itconfirmed the necessity and flexibility for its application. Then, it introduced andanalyzes the main components and their functions of the applied expert guidance systemand implemented such a system in the power plant, by combining the existingoptimization system and economic diagnosis system of the unit. According to theanalyses of the input coal properties, using the cross analyses of the internal connectionof the measured components and the neural network prediction method, a model of coalquality monitoring was constructed. With such a model, the proximate and ultimate ofthe coal could be derived based on the field test data of the boiler. In the same time,using the neural network method, the relationship between the coal properties includingthe calculated ratio of hydrocarbon to carbon and the combustion performance of theboiler has been built up. With the model and the constructed relationship, the currentcombustion performance and potential results after adjustment could be forecasted andautomatic optimization could be proposed. The engineering application showed that through the coal quality monitoring andwith active adaption of the operation manner to the frequent coal property changes, theexpert guidance system operation, under the current widespread multi-coal blendingcombustion conditions, could improve the reliability and accuracy of the optimalmanagement system of the unit, optimize the management system, providing therelevant manipulation suggestions for operation personnel and enhancing the safety andthe economy of the power plant.