Study of discharge modeling method using support vector machine for rubber mixing process

Rubber mixing is an important production process in the tire plant, and the control of discharge point, which is the most important action point, is seriously relevant with the final tire quality. This paper presents a SVM (support vector machine) discharge modeling method based on statistical learning theory (SLT) to treat the limitation of conventional methods. Abnormal modeling samples are first divided into three different types according to the input and output distribution properties, and corresponding approaches are developed to eliminate the outliers, respectively. Then SVM technique is applied to build the discharge model to establish the rubber discharge condition. The obtained model was applied to a pilot plant of the Hangzhou Zhongce Rubber Co. Ltd., China and the result shows that the proposed method suffices the demand of data-driven modeling in finite-sample, complexity control and robustness of actual mixing process, and the discharge modeling method is valid.