Study On Optimization Of Operation Parameters Of Thermal Power Unit Based On Multi Clustering Algorithm

With the continuous development of big data and artificial intelligence technology,its application field is expanding.Optimization of operation parameters of thermal power units is one of the main contents of operation optimization.In this paper,combined with mechanism knowledge,a new algorithm of data mining for optimization of operation parameters of power plant boilers and steam turbines is studied,which provides a new exploration and solution for big data and artificial intelligence in optimization of parameters of thermal power units.The main contents and innovations of this paper are as follows:(1)Research on multi clustering algorithm based on K-means clustering and FCM clusteringIn this paper,a multi clustering algorithm based on running data and parameter optimization is proposed.Firstly,the contour coefficient method and the efficiency function partition coefficient are introduced to solve the problem of K-means clustering and FCM clustering algorithm to automatically determine the optimal number of clusters;then,the applicability of K-means clustering and FCM clustering algorithm to the clustering of historical data sets of thermal power units is verified;then,combining the characteristics of these two algorithms,K-means clustering is used for data screening,and FCM clustering is used for operation parameters The multi clustering algorithm is optimized,and the effectiveness of the multi clustering algorithm is verified.(2)Outlier detection based on density DBSCAN clustering algorithmIn the process of clustering,K-means clustering and FCM clustering algorithm are easy to be affected by outliers,resulting in inaccurate clustering results.Therefore,in the process of data preprocessing,this paper introduces DBSCAN clustering algorithm based on density into the historical data of thermal power units for outlier detection.DBSCAN clustering algorithm can find clusters of any shape,and quickly find outliers in each cluster,so the algorithm can effectively eliminate global outliers and local outliers,thus avoiding the impact of outliers on multi clustering.(3)Study on Optimization of operation parameters of thermal power unit based on multi clustering algorithmFirstly,combined with the demand of optimizing the parameters of boiler and steam turbine,using the method of mechanism analysis and grey correlation,the thermal efficiency of boiler,the heat consumption rate of steam turbine and its main related operation parameters are determined,such as exhaust gas temperature,exhaust oxygen content,main steam pressure,main steam temperature,etc.Secondly,by using multi clustering algorithm,the process of data optimization is divided into two stages:high-efficient area selection(K-means)and habitual running area clustering(FCM),which effectively improves the quality of clustering analysis.The result of the final example shows that the result of multi clustering is better than that of single FCM clustering.