基于数据挖掘算法的热压罐固化周期预测研究

目前,计划人员只能根据相关工艺文件中固化参数及热压罐固化周期的历史数据对热压罐进行连续排罐,导致计划人员无法制定精细的排产计划,还没有利用数据挖掘算法对热压罐固化周期进行预测的研究。采用支持向量回归和KNN预测两种预测方法,并对两种方法的预测结果进行对比试验。试验表明KNN预测的预测结果中有90%的罐次均优于支持向量回归的预测结果,且有90%罐次的误差小于0.5h。最后对两种方法的预测结果进行了原因分析。

Period Prediction of Autoclave Curing Based on Data Mining Algorithm

At present,the planners can only arrange autoclave continuously according to the curing parameters of relevant process documents and historical data of the curing period,causing planners failure to make a detailed production scheduling plan.Now,data mining algorithm hasn’t been used to predict the curing period of autoclave.To solve the curing time of the autoclave,support vector regression(SVA)method and K–nearest neighbor(KNN)method are used to calculate.The proposed SVR and KNN comparative experiments are performed.Experimental results show that 90%of KNN prediction result is better than support vector regression prediction method,and 90%of the error is less than 0.5 hours.Meanwhile,the reasons for the prediction results of two methods are analyzed.