在本研究中我們利用模糊推論來鑑識線上數據的變化趨勢並應用在失誤診斷上面。此方法需以兩階段執行：首先在離線準備階段，我們依照動態特徵將變化趨勢分為十一類，接著根據轉換函數以及有向圖可以推論出每一變數的動態行為及失誤傳遞路徑，並據以建構出三種不同的IF-THEN法則，接下來的線上診斷是依三層式推論架構執行，在第一層架構中，主要將線上數據依上述十一種變化趨勢分類，而第二層架構則是鑑識出每一變數的響應型式，再利用第三層架構來判斷出不同變數間的失誤傳遞路徑是否與有向圖模式相符，最後以案例研究之模擬結果來驗證此方法的可行性。

In this study, fault diagnosis based on trend recognition exhibited in the sensors measuring the process is considered. A two-staged strategy is employed: (1) in the off -line preparation stage, the characteristic trend relate to possible observed variables are classified as eleven kind of trend language of primitives. According to the primitives, dynamic behavior with trend information in each measured variable can be inferred by the transfer function on a specified fault. In addition, fault propagation paths are also inferred by SDG model. Finally, the inferred dynamic behavior and propagation paths will be coded into three different IF-THEN rules with trend language of primitives. (2) On-line implementation stage, a three layer fuzzy inference system is adopted. In first layer, identifying the most likely trend based on the similarity between observed trend and the dynamic behaviors. Second layer is used to evidence the validity the dynamic sequences relate to each variable. Third layer consider the fault propagation path between different variables. Finally, simulation studies have been carried out to demonstrate the feasibility of the proposed approach.

Chang, S. Y., Chang, C. T. A-fuzzy-logic based fault diagnosis strategy for process control loops. Chemical Engineering Science 58, 3395-3411, 2003.
Chang, S. Y., Lin, C. R., Chang, C. T. A fuzzy diagnosis approach using dynamic fault tree. Chemical Engineering Science 57, 2971-2985, 2002.
Chen, J. Y., Chang, C. T., Fuzzy diagnosis method for control systems with coupled feed forward and feedback loops. Chemical Engineering Science 61, 3105-3128, 2006.
Chen, J. Y., Chang, C. T., Systematic enumeration of fuzzy diagnosis rules for identifying multiple faults in chemical processes. Industrial & Engineering Chemistry Research 46, 11, 3635-3655, 2007.
Dash, S., Rengaswamy, R., Venkasubramanian, V. Fuzzy-logic based trend classification for fault diagnosis of chemical process. Computer and Chemical Engineering 27, 347-362, 2003.
Dash, S., Venkasubramanian, V. Challenges in the industrial applications of fault diagnostic systems. Computer and Chemical Engineering 24 785-791, 2000.
Henley, E. J., Kumamoto, H. Reliability engineering and risk assessment. Prentice-Hall, Inc., Englewood Cliff, New Jersey, U.S.A., 1981.
Hoskins, J. C., Kalivur, K. M., Himmeblau, D. M. Fault diagnosis in complex chemical-plants using artificial neural networks. A.I.Ch.E. Journal 37, 137, 1991.
Iri, M., Aoki, K., O’Shima, E., Matsuyama, H. An algorithm for diagnosis of system failure in the chemical process. Computers Chemical Engineering 3, 489, 1979.
Ju, S. N., Chen, C. L., Chang, C. T. Constructing fault trees for advanced process control systems-application to cascade control loops. IEEE Transactions on Reliability 53, 43-60, 2004.
Mathworks, 2002a. Fuzzy logic toolbox-user guid. The Mathworks Inc.
Mathworks, 2002b. SIMULINK-dynamic system simulation for MATLAB-using simulink. The Mathworks Inc.
Petti, T. F., Klein, J., Dhurjati, P. S. Diagnostic model processor: Using deep knowledge for process fault diagnosis. A.I.Ch.E. Journal 36, 565, 1990.
Sugeno, M. Industrial Applications of Fuzzy Control. Elsevier Science Pub. Co, 1985.
Ulerich, N. H., Powers, G. J. Online hazard aversion and fault diagnosis in chemical prosces: The digraph + fault-tree method. IEEE Transactions on Reliability 37, 171-177, 1988.
Venkasubramanian, V., Rengaswamy, R., Kavuri, S. N., Yin, K., A review of process fault detection and diagnosis Part III: Process history based methods. Computer and Chemical Engineering 27, 327-346, 2003.