Research Of Data Mining Technology In Optimization Of Power Plant Operation

White paper of the development of large data in the electric power industry marks the Chinese power industry usher in the age of big data and it brights the intelligent mode of production.As is known, a power plant thermal system is very complex and the parameter coupling is serious.The existing system of optimization for operation is difficult to get an ideal solution, so we use data mining technology to analysis the huge operation data and find a new control scheme to keep the unit with the best operating condition, thereby improving the safety and economy of unit.This paper relies on the Guizhou Datang Power Generation Co. Ltd. simulation project.We get the running data needed by the research from the SIS server.We also participate in upgrading the SIS system and study on the key technologies of data storage and transmission, so ensuring the accuracy of data derived and making a good foundation for the following research.Then giving the combined forecasting model based on Lagrange interpolation and the GARCH model in order to complete the data preprocessing which is very important in data mining and taking up a lot of time.Neural network can use GARCH model which is good response to data volatility to overcome the Runge phenomenon from Lagrange interpolation, at the same time using fast operation of Lagrange interpolation to enhance the speed.Coal consumption for power supply protection correlation based on sequential patterns(CCPC)algorithm is the core content of this paper.The design idea of algorithm comes from Ma Jin, Jin Maojing et who come from Shanghai Jiaotong University and propose proposed privacy preserving multi step attack alarm correlation(PPMAC)algorithm.in We use he coal consumption of power supply which fully reflect unit’s economy as optimization objectives and based on the division of time series.The coal consumption for power supply caused by different operation parameters change in the time series of the rise or fall as attack or to protect the power supply coal consumption.Through sequential pattern mining,learning attack scene to reduce the coal consumption of electricity supply and improve the management level of the actual running of the unit.Finally,using statistical analysis field of the prestigious SPSS Clementine to complete the simulation experiment and getting the predicted result required to verify the feasibility of the CCPC algorithm.