| In modern industrial plants, with the extensive applications of DCS,data acquisitions and storages, process data are identified involving more and more useful operational information. It is convincible that, taking advantage of the rich information, production operations could be extensively improved. In this sense, originated from fields of economy,education and biology, the data mining technology has made inroads into industrial sectors as well.Industrial process data are generally with continuity and target aggregation characteristics. In order to effectively mining the operation information of such data, an APRIORI association rule method based on K-MEANS clustering is proposed and applied to the mining of industrial process rules. The main contents are presented as follows.(1) Based on the analysis and research of the principle and performance of K-MEANS and APRIORI algorithm, an APRIORI association rule method based on K-MEANS clustering is given according to the characteristics of industrial process operation data.(2 ) For industrial processes, the APRIORI association rule algorithm based on K-MEANS clustering is used to process the manipulated variable and the target variable data. Firstly, get the cluster center and class labels. And then the data is segmented according to the 1? rule and translated into boolean codes. Subsequently, the relations between manipulated variables and target variables are extracted using the combinatorial association rule mining method. Finally, the character expressed association rules are transformed into interval expressions so as to achieve the operation rules of high efficiency, reasonable number and easy understanding.(3) The proposed method of combinatorial association rules is applied to the TE process for mining operation rules between the operating variables and the state target variables of the TE reactor. Taking advantage of the process analysis, the validity of the algorithm is verified. |
PDF Full Text Request