【摘要】 在传统偏最小二乘法(PLS)建模的基础上,采用滑动窗方法,用递推偏最小二乘法(RPLS)根据在线数据自适应调整PLS混炼模型的结构和参数,建立橡胶混炼质量控制的RPLS在线校正模型。仿真与试验结果显示,该模型能自适应跟踪混炼的过程变化,门尼粘度预测误差为±3,符合生产要求。更多还原

【Abstract】 On the basis of traditional partial least squares(PLS) modeling,the sliding window method and recursive partial least squares(RPLS),which adaptively adjusted the structure and para-meters of PLS model for mixing according to on-line data,were used to build RPLS on-line renewal model for the quality control of rubber mixing.It was verified with experimental data that the model could adaptively track the variation of mixing process,with the predictive error ±3 of Mooney viscosity,which met production requirement.更多还原