Efficiency Analysis And Fault Diagnosis Of Rectifying Column

In modern chemical industry,rectifying column,as a vital equipment in production,plays an important role in production efficiency of the enterprise.It is therefore increasingly important to conduct process monitoring and fault diagnosis on rectifying column.However,in the actual production process,humans cannot accurately and quickly judge equipment failures,calling for research on rectifying column efficiency and fault diagnosis.This study starts out by analyzing the efficiency of rectifying column and conducting correlation analysis on key production variables affecting efficiency,simulates common faults which influence efficiency of rectifying column through HYSYS,and concludes with diagnosis on fault classification that lowers equipment efficiency by improving Particle Swarm Optimization-Support Vector Machine(PSO-SVM).Details of the study are as follows.(1)This study proposes a rapid calculation method of rectifying column efficiency to resolve the low efficiency issue in table look-up to get parameters by combing the actual production condition in the industry and analyzing existing calculation method of rectifying column efficiency.The error calculated through this method is less than those of the table look-up,which not only resolves issues such as low computational efficiency and helps calculate the efficiency of the rectifying column in a real time.(2)Product concentration is subject to other production variables,highlighting the need to identify key production variables to improve equipment efficiency.This study proposes,considering the coupling variables in chemical process,the correlation analysis method based on time delay mutual information.The method calculates the mutual information between variables in consideration of time delay,and selects the variable with higher time delay mutual information from the original variables as the relevant production variables that affect production.The results of actual air separation case analysis show that this method can obtain production variables that are more correlated with product concentration.(3)This study simulates common faults which affect efficiency through HYSYS,analyses the features of the faults and summarizes fault phenomenon.On thebasis of fault classification,this study proposes a fault diagnosis method for rectifying column based on improved PSO-SVM to expel difficulty in indicator selection of support vector.The method combines the idea of simulated annealing algorithm,allows the speed limit value to be changed within a fixed value range,improves the fitness of particles,and makes the support vector machine have better classification effect.Case analysis verifies the applicability of this method.