Target Value Determination Of Parameter Of Thermal Power Unit Based On Historical Optimal Operation

In the background that automation level and information technology develop rapidly,the operation optimization technology of the power plant is improved continuously.With the popularization of the SIS system and the MIS system,huge amount of historical operating data can be saved.It's a challenge to make further operating optimizing for researchers.Data mining is born in the age of big data.Applying this instrument to analysis the historical databases could overcome the characteristics of these databases such like large number,high dimensions,strong coupling and so on and find some rules that will be influential for decision making.This paper was based on historical operating database,introduced the data mining technology applying in different systems of the power unit and conducted several following aspects of studies:1.Considering the disadvantage that the classic Apriori algorithm usually scans the database for several times and produces many redundant candidate items,we put forward a modified Apriori algorithm.By defining a transaction data boolean vector matrix and corresponding column vector,the number of database scan was reduced.And the efficiency of the algorithm was improved by utilizing the ordering of these items to remove redundant candidate items.2.A historical operating database of a 600 MW power unit about the coal pulverizing system and the air distribution modes was used to build a modified Apriori algorithm mining model which is based on matrix.By mining the parameters of the primary air flow in coal mills,the outlet pressure of coal mills,the speed of coal feeders and the opening of secondary air doors,we got the optimal target reference values of each operating parameters at all operating states.3.A model which is based on PSO-SVR was set up to forecast the vacuum value of the condenser.The circulating water system database is optimized based on the association rules.The PSO method was used for optimization modeling about the parameters of SVR.Finally,the determination about the target value of the condenser vacuum and the evaluation about the operating states of the condenser were achieved.