發展隱藏式馬可夫模式為基礎的動態影像即時監控以及預測監控於燃燒系統

在本研究中將提出以數位攝影機提供的即時動態影像數據為基礎的監控方法，主要目的在於提早辨識燃燒爐系統異常行為，避免系統效能降低。由於火焰在燃燒中受到其動態行為影響而不停的閃爍，使得火焰具有因閃爍產生的空間性資訊以及受動態行為影響的時間性資訊，因此在本研究提出的方法中包含隱藏式馬可夫模式(Hidden Markov Model, HMM)和多維主成份分析法(Multiway Principal Component Analysis, MPCA)。MPCA方法主要萃取火焰影像中空間性的資訊並降低其變數維度，並以HMM由萃取後的空間性資訊來建立系統的動態關係。由於以此法監控系統不需要此程序相關的背景知識，而此程序的機率分佈由正常影像數據來建立，在系統監控中，線上收集到的序列影像數據可由Viterbi algorithm運算出HMM模式中的最終的隱藏狀態，並以此狀態分佈辨識系統異常，而在本研究中所提出的方法，可像傳統的統計方法中，建立一個簡單的機率管制圖來辨識即時量測到的影像序列是否有異常行為產生，並以此模型即時進行錯誤模式診斷以及預測監控。由於燃燒爐系統中火焰燃燒因受到其動態行為的影響而不停的閃爍，該現象會表現在以MPCA分析法萃取燃燒中火焰的空間性資訊，導致擾動嚴重，由於當燃燒爐系統發生微小的異常時，因火焰閃爍導致的擾動會明顯大於系統變化的趨勢使MPCA-HMM於時間面的分佈不易偵測出，因此本研究將以離散式小波轉換(Discrete Wavelet Transform, DWT)將系統的正常擾動像去除，而異常的擾動保留，以可清楚的還原燃燒爐內火焰變化的趨勢，以提升HMM監控模型的能力，並以此模型即時進行錯誤模式診斷以及預測監控。最後會以一個實際的燃燒爐實驗來驗證本研究所提出的方法。

關鍵字

預測監控；系統監控；燃燒系統；隱藏式馬可夫模式

並列摘要

In this thesis, a novel method of on-line flame detection in video is proposed. Processing the data generated by an ordinary camera monitoring scene, it aims to early detect the current state of the combustion system and prevent the system from further degradation and occurrence of failure. Due to the dynamic change of the combustion system, the turbulent flame flicker produces images with different spatial and high temporal resolutions. The proposed method consists of hidden Markov model (HMM) and multiway principal component analysis (MPCA). MPCA is used to extract the cross-correlations among spatial relationships in the low dimensional space while HMM constructs the temporal behavior of the sequence of the spatial features. Although the prior process knowledge may not be available in the operation process, the probability distribution of the normal status can be trained by the images collected form the normal operation process. Subsequently, monitoring of a new observed image is achieved by a recursive Viterbi algorithm which can find the transition state sequence from series of observed image data. The proposed method, like the philosophy of traditional statistical process control, can generate simple probability monitoring charts to track the progress of the current transition state sequence and monitor the occurrence of the observable upsets. Due to the dynamic behavior of the turbulent flame flicker in the combustion process, the extracted features form images with high perturbation in spatial relationships. It’s hard to detect the minor error of the operation changed which smaller than system perturbation. In this thesis, the perturbations in the normal operation will removed by discrete wavelet transform (DWT) and retain the abnormal system trend. It’s will improve the monitoring capability of the HMM. To demonstrate the power and the advantages of the proposed methods, the real combustion system will be tested.

並列關鍵字

multiway principal component analysis ； statistical process control ； predictive monitoring ； hidden Markov model ； combustor

參考文獻

1. A. R. Jones, “Flame Failure Detection and Modern Boilers,” J. Phys. E: Sci. Instrum., 21 (1988), 921-928.

2. B. C. Juricek, D. E. Seborg, W. E. Larimore, “Predictive monitoring for abnormal situation management,” Process Control, 11, (2001), 111-128.

3. B. Kim and W. S. Kim, “Wavelet Monitoring of Spatial Surface Roughness for Plasma Diagnosis,” Microelectronic Eng., 84 (2007), 2810–2816.

4. B. R. Bakshi, “Multiscale Analysis and Modeling Using Wavelets,” J. of Chemo., 13 (1999), 415-434.

5. B. U. Toreyin, Y. Dedeoglu, A. E. Cetin, “Flame Detection in Video Using Hidden Markov Models,” IEEE Conf. on Image Processing, 2 (2005), 1230-1233.

國際替代計量

發展隱藏式馬可夫模式為基礎的動態影像即時監控以及預測監控於燃燒系統

全文下載

主題瀏覽