在批式反應器於操作之時，由於無適當之線上分析儀器可量測其即時產物品質，而導致其在操作時往往無法達到最佳之能源效益或是產品純度之要求等之問題。本研究乃以近年來被多方利用之”類神經網路” (artificial neural networks,簡稱ANN)技術來架構一適用於批式反應器之產物品質估測器；此估測器乃利用反應器溫度與實驗方法來估測產物之即時組成。研究結果發現，ANN網路若於訓練時所考慮的變化狀況越完備，則於實際操作時所獲得的結果將會更具準確性；亦即以ANN做產物品質估測時之成功的重要因素之一，乃是訓練組數據收集之完備否。並由研究結果可知，以ANN產物品質估測器預測之數據來預測批式反應器之即時產物品是可行的。 Although composition analyzers, like on-line gas chromatorgraphy (on-line GC), have been used in the process industries for a long time, they are seldom found in the batch reactor process control. The use of inferential variables (like temperatures) to estimate process polymer quality in place of direct measurements is usually desired for plant engineers. The batch reactor has been reported a dynamic, nonlinear system; it is more complex and difficult to predict polymer qualities in a batch emulsion reactor through the use of temperature measurements. The artificial neural networks (ANN) is composed of nets of nonlinear basis functions; it has the ability to evolve a good process model from experimental data which requires very little or no knowledge of first principles. In addition, it has the ability of learning and prediction for nonlinear model, and has therefore been used to identify the process dynamics nonparametrically. The main purpose of this study is using stacked and recurrent ANN to estimate the polymer qualities in a batch emulsion reactor. Several techniques, which include using laboratory data for ANN, have been proposed. Simulation results have demonstrated that the proposed approach can successfully apply to the building of soft sensors for a batch emulsion polymerization reactor.
