Combustion Optimization of Large Thermal Power Units Based on Semi-supervised Competitive Clustering and Improved Apriori Algorithm
In order to absorb large-scale new energy grid connection,thermal power units need to process higher di-mensional and larger stock data when optimizing combustion through data mining,but the mining efficiency of the exist-ing unsupervised clustering/Apriori algorithm is low and does not meet the requirements of high flexible operation.To solve this problem,the constraint penalty factor was introduced into the unsupervised clustering algorithm to change it into semi-supervised clustering to improve the clustering efficiency.Based on the partition idea,the Apriori algorithm was improved to avoid the generation of redundant rules and improve the mining efficiency and a new data mining algo-rithm based on semi-supervised competitive clustering and partition association rule mining was formed.Taking the 660 MW unit of a power plant as an example,the new algorithm was used for data mining to obtain the optimized values for each operation parameters,and the typical sample database was established to implement combustion optimization and compare with the improved algorithm before improvement.The results show that the new algorithm improves the mining efficiency and storage space utilization,and has a certain reference value for the combustion optimization of large ther-mal power units.
combustion optimizationdata miningtypical sample libraryfuzzy clusteringassociation rulesbig data
万方数据
维普
