Research On Performance Diagnosis Of Steam Turbine Unit Under Data Driven

At present,the traditional thermal power generation industry is affected by new energy power generation,and the market share has been declining year by year.The thermal power generation in the new environment pays more attention to the flexibility peak-adjusting capability of the unit,which also puts higher requirements on equipment performance monitoring and operation optimization in the field of thermal power generation.With the development of information technology,the value mining of massive operational data has gradually been widely recognized.How to obtain the correct information through data mining is also an urgent problem to be solved in the field of thermal power.In this paper,the three main components of the thermal power turbine group:the through-flow part,the regenerative heater part and the condenser part,the characteristic parameters that can characterize the thermal economy of the operation are selected,and the traditional thermal calculation research ideas are used for analysis.At the same time,according to the characteristics of its massive actual operational data,the relevant processing methods of data mining technology are used to perform quasi-steady-state screening based on sliding window method and abnormal data detection based on BP neural network,and use steady state data after screening to build.Modular analysis is mutually validated with the results of traditional thermodynamic calculations.Based on the analysis of the characteristic flow area,heater end difference and condenser performance status monitoring factors,and the information mining of massive power station operation data,the rationality of evaluating the thermal economy of steam turbines is also verified.A complete steam turbine performance analysis program has a certain guiding effect on the efficient and safe operation of thermal power generating units.