Research On The Spool Correction System Based On Support Vector Machine

In order to realize the sustainable development of economy,the remanufacturing technology has been widely concerned.The current remanufacturing is mainly focused on surface repairing of parts,and there are few researches on automatic correction theory in view of deformed parts.The research of automatic correction theory mainly focuses on the determination of correction stroke.Due to the influence of non-linear factors such as material,size,deformation degree and deformation position,the traditional method is difficult to predict the correction stroke of the spool,but artificial intelligence provides an effective method.This paper takes the spool as the research object,the spool's correction stroke prediction model are studied.First of all,when the spool is plastically deformed,in order to solve the problem that it is difficult to predict the correction stroke of spool by analytical model,a method of modeling using support vector machine is proposed.Based on the research of support vector machine and least squares support vector machine,the least squares support vector machine is selected to establish the spool correction stroke prediction model.In order to improve the prediction accuracy of the model,the parameters of the model are optimized by the combination of genetic algorithm and cross validation.Secondly,in order to make the prediction accuracy of the least squares support vector machine increase gradually with the continuous accumulation of sample sets in the correction process,an incremental learning algorithm of least squares support vector machine based on pruning is proposed.This algorithm uses genetic algorithm to prune the sample set,sets pruning threshold according to the performance of the regression machine,and uses this threshold as the basis for screening important samples,so as to realize the incremental learning process of the least squares support vector machine.Thirdly,according to the correction principle of the spool,the spool correction device is designed,and the device is introduced from three aspects:structure design,hardware selection and software design,among them,the software design part uses VB and MATLAB mixed programming technology.The correction experiment of the spool is achieved by using the device,and the experimental data are obtained.The correct experiment of spool was carried out with the device,and the experimental data were obtained.The experimental data are processed by interpolation method to obtain the sample data needed to establish the model.Finally,BP neural network,improved BP neural network,the least Squares support vector machine based on genetic algorithm optimization and the least squares support vector machine incremental learning algorithm based on pruning processing are used to predict the correction stroke of the spool,and the predicted results are validated with the experimental results.The results show that the incremental learning algorithm proposed in this paper has a smaller predictive error,which proves the feasibility and superiority of the method in the prediction of the spool correction stroke.It also provides a good method for predicting the corrected stroke of other parts.