SCN-SABO-based Parameter Prediction Methods of Calender Rubber Sheet

In the tire manufacturing industry, rubber materials are processed through extruders and calenders to form rubber sheets with specific widths and thicknesses. The dimensional accuracy of these sheets determines the quality of tread production. For the parameter detection of calendered rubber sheets, the existing methods have limited predictive accuracy, which can lead to significant detection errors and the flow of defective products into subsequent processes. Therefore, this paper proposes a predictive model that combines Stochastic Configuration Networks (SCN) with Subtraction-Average-Based Optimization (SABO). Initially, the data of the rubber sheet parameters undergo a correlation analysis to eliminate both highly correlated and lowly correlated features, achieving feature dimensionality reduction and outlier handling. The preprocessed data is then used as input samples, utilizing SABO to search for the optimal process parameters of SCN, thereby outputting the optimal predicted values for the rubber sheet parameters. Experimental results from the test set demonstrate that the SABO significantly enhances the predictive performance of SCN compared to other optimization algorithms, exhibiting strong generalization capabilities. The prediction error can be reduced by more than 40%. This study provides an effective method for predicting the quality of calendered rubber sheets in the rubber tire manufacturing process.