Dynamic process fault forecasting method based on fuzzy self-adaptive prediction

Abstract

The invention belongs to the technical field of fault diagnosis and forecasting in the chemical industry production process and relates to a dynamic process fault forecasting method based on fuzzy self-adaptive prediction. Fault parameters and state variables are sequentially determined based on a fuzzy self-adaptive unscented Kalman prediction method, and whether future multi-step prediction is needed or not is determined according to fault parameter standard deviations; then, a future multi-step prediction value is worked out; finally, the fault prediction probability is worked out through sigma points of output variables, and fault prediction is carried out according to the fault prediction probability. The whole technical process is simple, principles are reliable, calculating parameters are accurate, fault state estimation precision is high, the application range is wide, logicality is strong, and the method is environmentally friendly.

Description

Technical field: [0001] The invention belongs to the technical field of fault diagnosis and prediction in chemical production process, and relates to a dynamic process fault prediction method based on fuzzy self-adaptive prediction. Background technique: [0002] With the continuous expansion of the production scale of the chemical industry and the increasingly complex production process, once a process failure occurs, it will affect the product quality, and even cause casualties and ecological crisis. In order to ensure the stability of the production process and reduce the loss of failure, automatic production In the process, it is urgent to predict the future operation status of the process, detect the fault as early as possible before the process exceeds the normal operation area, and eliminate the fault in the bud. Fault prediction technology is to use the small abnormal information at the initial stage of the fault to calculate the future predicted value of the process...

AI Technical Summary

Problems solved by technology

At present, research on fault prediction technology mainly focuses on two aspects: one is how to accurately track the fault process and accurately predict future process information; the other is how to use limited future prediction information to judge the future operation status of the process. process and accurately predict future process information, it is urgent to study process state and output prediction methods with strong tracking capabilities. Existing prediction methods include Kalman predictor, particle predictor, strong tracking particle filter, fuzzy adaptive Unscented (unscented) Kalman predictor, support vector machine and prediction method based on correlation vector machine fuzzy model; for the problem of how to use limited future prediction information to judge the future operation status of the process, Juricek and Seborg directly predict the future output Value intervals are compared with control limits; Hu Changhua et al. use strong tracking particle filters to estimate process state variables, and calculate fault prediction probability according to the particles of state variables; fuzzy adaptive unscented Kalman prediction method is published in the paper (Tian Xuemin, Cao Yuping, Chen Sheng , Process Fault Prognosis Using a Fuzzy-Adaptive Unscented Kalman Predictor, International Journal of Adaptive Control and Signal Processing, 2011, 25 (9): 813–830 / Tian Xuemin, Cao Yuping, Chen Sheng, based on Fuzzy Adaptive Unscented Kalman Predictor Process fault prediction, International Journal of Adaptive Control and Signal Processing, 2011, 25 (9): 813–830,); however, as far as the above-mentioned prior art is concerned, for some process models, the state variables have no actual physical It is difficult to determine its normal operating range, and for nonlinear dynamic processes, there is no real-time fault prediction system with complete logic and strong practicability.

Images