基于PLS模型的自适应间歇过程质量预测

间歇生产过程中,很多质量指标不能在线测量,导致过程很难控制。该文应用部分最小二乘(PLS)方法建立软测量模型,通过批次初期在线测量的过程变量对最后的产品质量进行预测。同时,利用过程中得到的中间质量测量值对最后的预测结果进行修正。为了解决过程参数随时间变化的问题,在每个批次结束后利用新数据对原模型进行更新。将该法用于异丁烯酸甲酯(MMA)聚合反应过程,仿真结果显示,该法能够克服过程参数变化的影响,有效地预测最后的产品质量。

Adaptive quality prediction for batch processes based on the PLS model

There are usually no on-line product quality measurements in batch and semi-batch processes, which makes process control very difficult. A model was developed to predict end-product quality from available on-line process variables measured early in the batch using the partial least squares method. Some available mid-course quality measurements were used to correct the final prediction results. Since the process may change with time, the model is updated with new batch data to generate new model parameters after each batch. A simulated batch polymerisation process demonstrates that the method can accurately predict end-product quality.