應用類神經網路於批式蒸餾塔之推論控制

Tunghai University Institutional Repository > 工學院 > 化學工程與材料工程學系所 > 碩博士論文 > Item 310901/4399

Please use this identifier to cite or link to this item: http://140.128.103.80:8080/handle/310901/4399

Title:	應用類神經網路於批式蒸餾塔之推論控制
Other Titles:	Inferential Control of a Batch Distillation Column Using Artificial Neural Networks
Authors:	陳文川
Contributors:	黃琦聰 Huang, Chi-Tsung 東海大學化學工程與材料工程學系
Keywords:	類神經網路;批式蒸餾;推論控制 Artificial Neural Networks;batch distillation;inferential control
Date:	2003
Issue Date:	2011-05-19T06:05:44Z (UTC)
Abstract:	在批式蒸餾塔於操作之時，由於無適當之線上分析儀器可量測其即時組成，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蒸餾塔之組成估測器；此估測器乃利用蒸餾塔塔板溫度與實驗方法來估測蒸餾產物之即時組成。除此外，並利用此估測器所預測之即時組成進一步探討批式蒸餾塔之推論控制。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做組成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組成估測器預測之數據做為批式蒸餾塔之即時組成，並藉此做控制蒸餾塔切換時間之依據是可行的。除此外，於本研究中以此估測器之推論控制方式與傳統之操作方法作比較，在同時考慮產量、產品純度及能源使用的等因素下，由電腦模擬之結果發現在兩種不同之操作狀況下，本研究提出之推論控制方式皆比傳統之方式來的優秀。由此，亦可知以類神經網路應用於批式蒸餾塔之推論控制是可行的。 Industrial interest in batch distillation has increased significantly in recent years as more batch processing is being used for low-volume specialty chemicals. Most studies in the literature have concentrated on the determination of optimum operation strategies, e.g., optimum trajectories of reflux ratio and pressure during the batch and processing of the slop cuts that are produced. Little attention has been paid to problems associated with control. In fact, one of the main difficulties in batch distillation process control is getting reliable and accurate measurement of product compositions. If instantaneous and perfect analyzers were available, the task of batch distillation control would be straightforward once the optimum policies are available. Although composition analyzers, like on-line gas chromatography (on-line GC), have been used in the process industries for a long time, they are normally not so suitable for batch distillation processes. Moreover, the most important drawback for using the on-line GC in controlling a batch distillation process is that it possesses a very large time delay and thereby lowers the achievable control performance. The use of inferential variables, say temperatures, to estimate process compositions in place of direct on-line measurements is usually desired for plant engineers. However, the batch distillation process has been known as a highly nonlinear, dynamic system; it is more complex and difficult to predict real-time compositions through temperature measurements. On the other hand, the artificial neural networks (ANN) with the ability to evolve a good process model from experimental data requires very little or no knowledge of first principles. It has the ability of learning and prediction for nonlinear model, and has therefore been used to identify the process dynamics nonparametrically. The primary object of this investigation, however, is to develop an ANN composition estimator of a batch distillation column from temperature measurements. Then, using this estimator to implement an inferential control on the column. Simulation results in a batch distillation column have demonstrated that the proposed inferential control technique using recurrent and stack ANN can performance better than that of the classic operation.
Appears in Collections:	[化學工程與材料工程學系所] 碩博士論文

Files in This Item:

File	Size	Format
index.html	0Kb	HTML	217	View/Open

Loading...