In the methanol to olefin production process,measuring instrument of the reactor pressure drop may become abnormal,so that the operators were unable to obtain accurate measurement data of the pressure drop.Considering the complexity of actual production process,process data often had dynamic and nonlinear characteristics.A soft sensor technique based on the dynamic kernel ridge regression method was proposed.And a model was established in a coal chemical production enterprise.The results showed that this method could effectively improve the accuracy of the model,compared with the traditional linear regression model,the method had a better predictive effect on the pressure drop in the methanol to olefin reactor.
关键词
DMTO反应器/压降/动态核岭回归算法/最大信息系数/精度
Key words
DMTO reactor/pressure drop/dynamic kernel ridge regression algorithm/maximum information coefficient/accuracy
维普
万方数据