Abstract: In response to insufficient training samples under multiple operating conditions of petrochemical plants, which makes it difficult to meet the data requirements of traditional state evaluation methods, a state evaluation method based on improved TOPSIS (the technique for order preference by similarity to ideal solution, TOPSIS) was proposed to avoid dependence on historical data. First, the positive and negative ideal solutions were determined by expert experience and process index control limits of process cards in the improved TOPSIS. Then, the two types of indicators, such as fixed value index and fixed interval index, were established according to the characteristics of operating parameters. Moreover, the evaluation standards were determined. The proposed method was employed to deal with the abnormal data of a reaction and regeneration system in a fluid catalytic cracking unit from a refining company. The results showed that the proposed method could quantify the real-time operating status of petrochemical plants.

Key words: TOPSIS, petrochemical plant, state evaluation, health index, fluid catalytic cracking unit