An improved case-based reasoning method and its application on fault diagnosis of Tennessee Eastman process

Abstract

In the chemical industry, fault diagnosis is a challenging task due to the complexity and the high demand of product efficiency and consistency in the chemical plant. In this paper, an improved Case-based reasoning method is proposed to predict the status of Tennessee Eastman process (TE process). Firstly, a case reduction method based on selective rules is proposed to maintain the scale of the case base for decreasing the time complexity of the diagnosis process. Afterwards, a case reuse strategy based on trustworthy radius is introduced with the group decision making, this reuse strategy is utilized to evaluate the suggested diagnosis result as trustworthy or untrustworthy by calculating out the trustworthy radius of each category. Finally, the cases that are evaluated as untrustworthy would be adjusted with the group decision making strategy to get the final diagnosis result to accomplish the whole diagnosis process of CBR. Compared with other methods, the diagnosis results obtained from the proposed approach showed higher accuracy and reliability, which demonstrated the improvement of the diagnosis performance of TE process as well as the effectiveness and superiority of the proposed method on fault diagnosis.