Research And Implementation Of Hemodialysis Centrifugal Process Parameters Optimization

In the production of hemodialyzer(dialyzer for short),due to the accuracy of the dialyzer centrifugal processing process parameters is difficult to control,the dialyzer has bubbles,dialysis membrane lack of glue,dialysis membrane hole clogging,and excessive glue penetration in the centrifugation process,which makes the dialyzer yield has dropped.Aiming at the various problems in the processing technology of dialyzer centrifugation because of the imprecision of the process parameters,this thesis designs a method of optimizing the processing parameters of dialyzer centrifuge in enterprise production based on SVM(Support Vector Machine,SVM)algorithm.Firstly,the SVM algorithm involved in the method is improved.Then,orthogonal experiment is designed for the dialyzer centrifuge processing process in actual production.Finally,the software of On-line parameter adjusting system for centrifuge processing of dialyzer based on B/S mode is designed.The main work of this thesis includes the following aspects:1.The production status of dialyzer,the overall structure of dialyzer,the centrifugal processing process of dialyzer,impact factors of the centrifuge processing process of dialyzer are studied.2.The SVM algorithm involved in the optimization method of centrifuge processing technology designed in this thesis is studied.In view of the problems such as slow calculation speed of the original algorithm and the easy over-fitting,the SVM algorithm is improved by a combination of improved kernel function and particle swarm optimization.By comparing the training effect of several improved algorithms on a set of data sets with a random normal distribution sample size of 2000.Compared with frequently used improved SVM algorithm,it has better prediction effect and progress that the improved SVM algorithm is designed in the thesis.3.The optimization model of processing parameters of dialyzer centrifuge is established and orthogonal experiment is designed and the situation after the experiment is analyzed.The data of the orthogonal experiment are analyzed with the analysis of the difference and variance,and the main factors affecting the processing technology of the centrifuge are determined,and the best combination parameters between the factors are found.The optimized SVM algorithm is used to predict the process of centrifuge processing process.The improved SVM algorithm is obviously better than the other models.4.The requirements of optimizing the processing parameters of dialyzer centrifuge are analyzed,and the software of On-line parameter adjusting system for centrifuge processing of dialyzer based on B/S mode is designed to realize the main function of the software.According to the experimental simulation results,the optimized process parameters of the dialyzer centrifuge based on improved SVM algorithm can achieve the expected model prediction effect.