A Quality Integrated Fuzzy Inference System for the Reliability Estimating of Fluorochemical Engineering Processes

• Main Authors: Feng Xue, Xintong Li, Kun Zhou, Xiaoxia Ge, Weiping Deng, Xu Chen, Kai Song
• Format: Article
• Language: English
• Published: MDPI AG 2021-02-01
• Series: Processes
• Subjects: process reliability estimating; fluorochemical engineering process; fuzzy inference system; quality prediction; prognostics and health management
• Online Access: https://www.mdpi.com/2227-9717/9/2/292

Hypertoxic materials make it critical to ensure the safety of the fluorochemical engineering processes. This mainly depends on the over maintenance or the manual operations due to the lack of precise models and mechanism knowledge. To quantify the deviations of the operating variables and the product quality from their target values at the same time and to overcome the measurement delay of the product quality, a novel quality integrated fuzzy inference system (QFIS) was proposed to estimate the reliability of the operation status as well as the product quality to enhance the performance of the safety monitoring system. To this end, a novel quality-weighted multivariate inverted normal loss function was proposed to quantify the deviation of the product quality from the target value to overcome the measurement delay. Vital safety process variables were identified according to the expert knowledge. Afterward, the quality loss and the vital variables were inputs to an elaborate fuzzy inference system to estimate the process reliability of the fluorochemical engineering processes. By integrating the abundant expert knowledge and a data-driven quality prediction model to design the fuzzy rules of QFIS, not only the operation reliability but also the product quality can be monitored on-line. Its superiority in estimating system reliability has been strongly proved by the application of a real fluorochemical engineering process located in East China. Moreover, the application of the Tennessee Eastman process also confirmed its generalization performance for other complicated black-box chemical processes.