Research On Fault Detection Method Of Wiped Film Evaporation Process And Control Method Of Chaos System

As the most widely used molecular distillation evaporator,the wiped-film molecular distiller is due to its complex physical and chemical reactions and the conversion and transmission of material energy,which makes the entire production system have nonlinearity,strong coupling,large lag,and uncertainty characteristics,so it is impossible to establish an accurate mathematical model.At present,most of the production personnel operate the process parameters in the wiped film evaporation process based on experience,so it is difficult to ensure the continuous and stable operation of the production process.When a fault occurs,the process parameters will fluctuate greatly,affecting the purity and yield of the product.If the fault is not handled in time,it may even cause a major accident.In order to ensure the safe operation of the production process,abnormal operating conditions or failures must be quickly monitored to improve the reliability and safety of the system.In response to the above problems,this article made the following research:First,the Principal Component Analysis(PCA)is used for fault detection.The process data is projected into two subspaces of principal component and residual,and the corresponding T~2 and Q statistics are established to judge the running status of the process.The number of principal elements as the only structural parameter of the monitoring model determines the accuracy of fault detection.The Fault Signal Noise Ratio(Fault SNR)is used to select the optimal number of principal elements and apply it to Fault detection during the wiped film evaporation process.By comparing with the traditional Cumulative Percent Variance(CPV)criterion,simulation results show that the model based on Fault SNR has clearer detection results,higher sensitivity,and better monitoring performance.Secondly,in view of the obvious change in the T~2 statistic and no obvious change in the Q statistic in the PCA test results,it is impossible to accurately identify the changes in operating conditions or failures,introducing process variables that are more relevant to the principal components and calculate their prediction residuals.The new statistics replace the traditional statistics to weaken its conservatism.Through simulation experiments on the fault detection of the wiped film evaporation process,and comparing with the ordinary PCA,the results show that the improved PCA can not only effectively distinguish the change of working condition parameters and the occurrence of faults,but also can detect minor faults,effectively improving the PCA fault detection performance.Finally,the wiped film motor is an important part that affects the separation efficiency of the product.After research,it is found that this type of motor has chaotic behavior under certain parameter conditions.Due to its acyclic and unpredictable nature,it will be considered as an error or a fault and then the motor is damaged.To avoid chaos,a Time-Delay Estimation(TDE)controller based on Supervising Switching Controller(SSC)is designed to suppress chaotic behavior.And a Self-organizing Adaptive Fuzzy Neural Controller(SAFNC)was designed to achieve chaotic synchronization control of the system.Finally,the effectiveness of the algorithm was verified through simulation experiments.