The Application Of Failure Warning Method On The Primary Air Fan Of Power Plant

Primary air fan is one of the important auxiliary equipment in power plant. Its running state has always been the attention of the power plant related staff. The main function of the primary air fan is to provide a certain pressure and a certain flow of primary air for boiler combustion system, which will send dry pulverized coal into the burner and provide combustion heat as the pulverized coal volatile part needed. The work environment is relatively poor, so it’s difficult to carry out equipment fault diagnosis and inspection. Since less study is put on failure prognostic system currently, the study on failure prognostic system of a primary air fan in thermal power plants has important theoretical significance and practical value.In this paper, a method is researched deeply to establish the failure prognostic model of a primary air fan. Using Multivariate State Estimation Technique(MSET) to establish the model can make better advancing warning of a fan’s failure state.Firstly, the paper introduces the related theory knowledge of a primary air fan, including the structure and function; secondly, we analyze the historical data taken out of a PI database from a certain power plant SIS system. Preprocess the data from field collection and compress the on-site data through the revolving door algorithm. By applying the method calculus of interpolation, we complete the data among the unequally spaced measuring points in order to establish the failure prognostic data model under MSET; thirdly, data dimension of the memory matrix in the algorithm can be reduced through the correlation analysis and the principal component analysis algorithm to establish memory matrix, respectively. Figure out the estimated value of the primary air fan’s observation vector through MSET core algorithm and work out the residual between estimated value and observed value. Use cluster analysis to figure out the similarity between the two values so that we can choose a better way of data dimension reduction. Finally, the paper makes a brief introduction to the software and hardware of the primary air fan failure prognostic system and analyzes its application effect in power plant. Then systematically introduce and summarize the total effect of the failure prognostic algorithm and point out the further research and development in the future.