發展隱藏式馬可夫樹模式之即時批次監控系統

本研究是發展以小波轉換為基礎的隱藏式馬可夫樹模式（Hidden Markov tree Model, HMT）之即時批次製程監控系統。與大部分現存只在時間面上分析的批次製程監控方法不同，HMT不只能在時間與頻率的多尺度下分析量測訊號，還可以捕捉真實量測數據的群聚性與延展性兩個統計特徵。這個方法不僅提供更好的過濾雜訊效果，並且減少訊號扭曲。為了不在批次製程完成後再控管品質變數，本研究修正數據展開的結構以利即時批次監控的建立，當然也提出批次監控系統模式的訓練過程。在擷取過去操作的資訊建立模式之後，該監控方法的原理與傳統的SPC相似，建立簡單的監控管制圖，可以很容易追蹤每一個批次製程的進行，和監控製程錯誤的發生。本研究將利用一個模擬的數學例子，fed-batch的盤尼西林（penicillin）發酵製程，以及杜邦公司聚合反應製程的數據，並與一些傳統的批次監控方法做比較，以說明所提出的批次製程監控系統的優點。

關鍵字

小波轉換；隱藏式馬可夫樹模式；批次監控；主成分分析法

並列摘要

A novel technique of on-line batch processes monitoring based on wavelet-based hidden Markov tree (HMT) is developed. Unlike most of the existing batch process monitoring methods for time scale only, HMT can not only analyze the measurements at multiple scales in time and frequency but also capture the clustering and persistence of the statistical characteristics for practical measured data. This approach provides better noise reduction and less signal distortion. In order to conduct the on-line batch monitoring without real-time quality measurement at the end of batch run, a simple modification of the unfolding structure is applied. The training procedures for setting up the batch monitoring model are also included. After extracting the past operating information, the proposed method, like the philosophy of traditional statistical process control, can generate simple monitoring charts, easily track the process of each batch run, and monitor the occurrence of process faults. The applications are discussed through one simulation case and two sets of benchmark data, the fed-batch penicillin production and the DuPont polymerization process, to illustrate the advantages of the proposed method in comparison to some conventional methods.

並列關鍵字

wavelet ； MPCA ； monitor ； fault detection ； HMT

參考文獻

1.Aradhye H. B.; Bakshi B. R.; Strauss R. A. and Davis J. F., “Multiscale SPC Using Wavelets Theoretical Analysis and Properties,” AIChE J., 49, 939 (2003)

2.Bajpai R. K. and Reuss M., “A Mechanistic Model for Penicillin Production,” J. Chem. Tech. Biotechnol., 30, 332 (1980)

3.Bakhtazad A.; Palazoglu A. and Romagnoli J. A., “Detection and Classification of Abnormal Process Situation Using Multidimensional Wavelet Domain Hidden Markov Trees,” Comp. Chem. Eng., 24, 769 (2000)

4.Bakshi B. R., “Multiscale PCA with Application to Multivariate Statistical Process Monitoring,” AIChE J., 44, 1596 (1998)

5.Birol G.; Ündey C. and Cinar A., “A Modular Simulation Package for Fed-batch Fermentation: Penicillin Production,” Comp. Chem. Eng., 26, 1553 (2002)

國際替代計量

發展隱藏式馬可夫樹模式之即時批次監控系統

未授權

主題瀏覽