以數學模型為主要依據的氣路分析法(Model-based Gaspath Analysis Method)，一直是傳統用以監控發動機健康狀態與故障診斷的主要方法。然而，使用此一方法的現有監控診斷軟體卻無法有效的在日常修護中被普及使用。為此，本研究首先利用倒傳遞類神經網路法(Back-propagation Neural Network)來進行故障診斷技術的改善，並證實此網路的確提供了快速且有效的解決方法。其中用於網路訓練、測試的資料乃利用影響係數矩陣(Influence Coefficient Matrix)配合適當的雜訊來產生。研究顯示，當加入合理大小的雜訊時，不論是對於僅有四個量參數的有限型監控系統，或是對於具有八個量測參數的擴展型監控系統來說，此一診斷方法均可達到相當高的診斷正確率。同時，故障隔離的診斷成功率與量測參數的品質良窳有很大的關係，唯有當量測到的參數品質具有一定的可靠性時，才能真正確保其診斷正確率的提升。因此，發展一套於量測參數進入故障診斷神經網路前，先進行訊號驗證與處理的程序，便成為此一研究的重要課題。本文更進一步的針對傳感器訊號前處理系統進行研究。一般其他相關的神經網路訊號處理法在訓練網路時，經常會遭遇到收斂困難的情況，同時此一受過訓練的類神經網路往往無法在傳感器發生故障時正確做出診斷並給予準確的補正。為了克服這些困難，本文提出以類神經網路與遺傳基因演算法(Genetic Algorithms)為基礎的兩步驟法(Two-step Method)，來全面解決此傳感器驗證工作。本法第一步驟為建立一雜訊濾除自相關神經網路(Noise-filtering Auto-associative Neural Network)與一自映射網路(Self-mapping Auto-associative Neural Network)，並利用具快速收斂特性的Levenberg-Marquardt法來訓練網路。第二步驟是以參數識別法求解傳感器偏移(Bias)、飄移(Drifting)或喪失的訊號(Missing Data)。遺傳基因演算法具有全域搜尋的優點，對於搜尋傳感器中的受損參數上有其優勢，因此在本研究中被採用以作為參數識別的工具。研究結果顯示，一套有效的類神經網路與遺傳基因演算法則相互結合，用以進行離線式(Off-line)的氣路量測參數雜訊濾除、受損訊號及傳感器偏移、飄移量偵測與修正的新方法已成功建立。同時，利用此一方法還可及時、無延遲的偵測出發動機性能變化趨勢及無預警式的突跳(Outlier)訊號，是一套具有智能的訊號狀態偵測器。此一方法亦可與倒傳遞故障診斷神經網路相結合而成為一套完整的發動機ECM/FD系統。

　　Gaspath analysis holds a central position in the engine condition monitoring (ECM) and fault diagnostics (FD) technique. However, popularization of this approach has been impeded when practical enforcements were tried in both civil and military sectors. Artificial neural network (ANN) arises as a new approach which avoids the fundamental difficulties associated with the classical model-based methods. The objective of the present work is to develop a reliable ANN-based diagnostic system that can be enforced in the practical applications. Back-propagation, feedforward neural nets are employed for constructing the engine diagnostic networks. Noise-contained training and testing data are generated using an influence coefficient matrix and the data scatters. The results indicate that for situations where sensor scatters are comparable to those of the normal engine operation, the success rates for both 4-input and 8-input ANN diagnoses achieve high scores which satisfy the requirement. The success rate of 4-input diagnosis is almost as good as that of the 8-input. Although the ANN-based method possesses certain capability in resisting the influence of input noise, the success rate of fault diagnosis still depends mainly on the quality of the measurements obtained. A high success rate of diagnosis can only be guaranteed when a correct set of measurement deltas is available. Thus, a design of a preprocessor that can perform sensor data validation is of paramount importance. To this end, the present work proposes a genetic auto-associative neural network algorithm that can perform off-line sensor data validation simultaneously for noise-filtering and bias detection and correction. Neural network-based sensor validation procedure usually suffers from the slow convergence in network training. In addition, the trained network often fails to provide an accurate accommodation when bias error is detected. To remedy these problems, the Levenberg-Marquardt (LM) algorithm is adopted to speed up the network training and a novel two-step approach is proposed for bias accommodation problems. The first step is the construction of a noise-filtering and a self-mapping auto-associative neural network based on the LMBP algorithm. It is shown that the noise can be greatly filtered by the noise-filtering auto-associative neural network. The second step uses an optimization procedure built on top of these noise-filtering and self-mapping nets to perform bias detection and correction. Non-gradient genetic algorithm search is employed as the optimization method. It is shown in the present work that effective sensor data validation can be achieved for noise-filtering, bias correction, and missing sensor data replacement incurred in the gaspath analysis. This newly developed algorithm can also serve as an intelligent trend detector. True performance delta and trend change can be identified with no delay to assure a timely and high-quality engine fault diagnosis.

[1]Algun F., “F110 Engine Monitoring and Maintenance Management Systems for F-16 C/D,” AGARD CP-448, May-June 1988, pp. 11-1－11-12.

[2]Broede J., “Engine Life Consumption Monitoring Program for RB 199 Integrated in the On-Board Life Monitoring System,” AGARD CP-448, May-June 1988, pp. 9-1－9-10.

[3]Carr H. R., Ives G. J., and Jenkins, P., “A Joint Study on the Computer Isolation of In-field Aero Engine Vibration Diagnosis,” AGARD CP-448, May-June 1988, pp. 31-1－31-13.

[4]Cifald, M. L., and Chokani, N., "Engine Monitoring Using Neural Networks," AIAA Paper 98-3548, 1998.

[5]Cue R. W., and Muir D. E., “Engine Performance Monitoring and Troubleshooting Techniques for the CF-18 Aircraft,” ASME Journal of Engineering for Gas Turbines and Power, vol. 113, no. 1, January 1991, pp. 11－19.

[6]DePold H., and Gass, F. D., “The Application of Expert Systems and Neural Networks to Gas Turbine Prognostics and Diagnostics,” ASME Paper 98-GT-101, 1998.

[7]Doel, D. L., “The Role of Expert System in Commercial Gas Turbine Engine Monitoring,” ASME Paper 90-GT-374, 1990.

[8]Doel, D. L., “Gas Path Analysis - Problems and Solutions,” 17th Symposium on Aircraft Integrated Monitoring Systems, Paper A94-24751, Germany, September 21-23, 1993.

[9]Doel, D. L., “TEMPER-A Gas Path Analysis Tool for Commercial Jet Engines,” ASME Journal of Engineering for Gas Turbines and Power, vol.116, no. 1, 1994, pp. 82－89.

[10]Doel, D. L., “An Assessment of Weighted Least-Squares-Based Gas Path Analysis,” ASME Journal of Engineering for Gas Turbines and Power, vol. 116, no. 2, 1994, pp. 366－373.

[11]Doran P. D. “Some Historical Highlights of Turbine Engine Condition Monitoring Using Flight Data,” SAE Paper 700314, 1970.

[12]Eustace, R., and Merrington, G., “Fault Diagnosis of Fleet Engines Using Neural Networks,” International Symposium on Air Breathing Engines, Paper 95-7085, 1995.

[13]Fletcher W. D. M. and Bairsto N. A., “Engine Usage Condition and Maintenance Management Systems in the UK Armed Forces,” AGARD CP-448, May-June 1988, pp. 3-1－3-5.

[14]Fuhs, A. E., “Diagnostics and Engine Condition Monitoring,” AGARD CP-165 Part A, April 1974, pp. 1－12.

[15]Goldberg, D. E., Genetic Algorithm in Search, Optimization, and Machine Learning, Addison-Wesley, Reading, MA, 1989.

[16]Hagan M. T. and Menhaj M. B., “Training Feedforward Networks with the Marquardt Algorithm,” IEEE Transactions on Neural Networks, vol. 5, no. 6, November 1994, pp.989－993.

[17]Hagan M. T., Demuth H. B., and Beale M., “Levenberg-Marquardt Algorithms,” Neural Network Design, PWS Publishing Company, 1997, pp. 12-19－12-27.

[18]Hajela, P., “Nongradient Methods in Multidisciplinary Design Optimization - Status and Potential,” Journal of Aircraft, vol. 36, no. 1, 1999, pp.255－265.

[19]Hall, C.L., Leary, S., Lapierre, L., Hess, A., and Bladen, K., “F/A-18E/F F414 Advanced Inflight Engine Condition Monitoring System (IECMS),” IEEE Proceedings of Aerospace Conference, vol. 6, 2001, pp. 3069－3082.

[20]Haykin, S., Neural Networks, A Comprehensive Foundation, Macmillian College Publishing Co. Inc., New York, 1994.

[21]Hess A. J., “An Overview of US Navy Engine Monitoring System Programs and User Experience,” AGARD CP-448, May-June 1988, pp. 2-1－2-16.

[22]Hines, J. W., Uhrig, R. E., and Wrest, D. J., “Use of Autoassociative Neural Networks for Signal Validation,” Journal of Intelligent and Robotic Systems, vol. 21, no. 2, 1998, pp. 143－145.

[23]Holland, J. H., Adaptation in Natural and Artificial Systems, Univ. of Michigan Press, Ann Arbor, MI, 1974.

[24]Jenkins P. J., “Information Management System for On-Board Monitoring Systems,” AGARD CP-448, May-June 1988, pp. 6-1－6-8.

[25]Kerr, L. J., Nemec T. S., and Gallops G. W., “Real-time Estimation of Gas Turbine Engine Damage Using a Control Based Kalman Filter Algorithm,” ASME Journal of Engineering for Gas Turbine and Power, vol. 114, no. 2, 1992, pp. 187－195.

[26]Kramer, M. A., “Non-linear Principal Component Analysis Using Autoassociative Networks,” AIChE Journal, vol. 37, no. 2, 1991, pp. 233－243.

[27]Kramer, M. A., “Autoassociative Neural Networks,” Computers and Chemical Engineering, vol. 16, no. 4, 1992, pp. 313－328.

[28]Kunz J., and Schulz U., “On-Board Life Monitoring System for Tornado (OLMOS),” AGARD CP-448, May-June 1988, pp. 5-1－5-5.

[29]Luppold, R. H., Roman J. R., Gallops G. W., Kerr L. J., “Estimating In-Flight Engine Performance Variations Using Kalman Filter Concepts,” AIAA Paper 89-2584, 1989.

[30]MacIssac B. D., “Engine Performance and Health Monitoring Models Using Steady State and Transient Prediction Methods,” AGARD Lecture Series 183, May 1992, pp. 9-1－9-21.

[31]Marquardt D., “An Algorithm For Least Squares Estimation of Non-linear Parameters,” Journal of Society of Industry Applied Mathematics, vol. 11, 1963, pp. 431－441.

[32]Mattern, D. L., Jaw, L. C., Guo, T. H., Graham, R., and McCoy, W., “Simulation of an Engine Sensor Validation Scheme Using an Autoassociative Neural Networks,” AIAA Paper 97-2902, 1997.

[33]Mattern, D. L., Jaw, L. C., Guo, T. H., Graham, R., and McCoy, W., “Using Neural Networks for Sensor Validation,” AIAA Paper 98-3547, 1998.

[34]Merrington, G.L., Kwon, O-K., Goodwin, G., Carlsson, B., “Fault Diagnosis in Gas Turbines,” Journal of Engineering for Gas Turbines and Power, Transactions of the ASME, vol. 113,no. 2, April 1991, pp. 276－282.

[35]Merrington, G.L., “Fault Diagnosis in Gas Turbines Using a Model Based Technique,” ASME Paper 93-GT-13, 1993.

[36]Michalewicz, Z., “Genetic Algorithms＋Data Structures＝Evolution Programs,” 3rd Ed., Springer－Verlag Berlin Heidelberg New York, 1999.

[37]Moller J. C., Litt J. S., and Guo T.-H., “Neural Network-Based Sensor Validation for Turboshaft Engines,” AIAA Paper 98-3605, 1998.

[38]Myers D. A., and Hogg G. W., “F100-PW-220 Engine Monitoring System,” AGARD CP-448, May-June 1988, pp. 18-1－18-9.

[39]Napolitano M. R., Neppach C., Casdorph V., and Naylor S., “Sensor Failure Detection, Identification and Accomodation Using On-Line Learning Neural Architectures,” AIAA Paper 94-3598, 1994.

[40]Napolitano, M. R., Neppach, C., Casdorph, V., Naylor, S., Innocenti, M., and Silvestri, G., “Neural-Network-Based Scheme for Sensor Failure Detection, Identification, and Accommodation,” Journal of Guidance, Control, and Dynamics, Vol. 18, No. 6, Nov. -Dec. 1995, pp. 1280－1286.

[41]Napolitano, M. R., Windon, D., Casanova, J., and Innocenti, M., "A Comparison between Kalman Filter and Neural Network Approaches for Sensor Validation," AIAA Paper 96-3894, 1996.

[42]Napolitano M., An Y.-H., Seanor B., Pispitsos S., and Martinelli D., “Application of A Neural Sensor Validation Scheme to Actual Boeing 737 Flight Data,”AIAA Paper 99-4236, 1999.

[43]Neppach D., and Casdorph V. A., “Sensor Failure Detection, Identification and Accommodation in A System Without Sensor Redundancy,” AIAA Paper 95-0011, 1995.

[44]Romer, M., and Atkinson, B., "Real-Time Health Monitoring and Diagnostics for Gas Turbine Engines," ASME Paper 97-GT-30, 1997.

[45]Romer, M., “Testing of a Real-Time Health Monitoring and Diagnostics System for Gas Turbine Engines,” AIAA Paper 98-3603, 1998.

[46]Rumelhart D. E., Hinton G. E., and Williams R. J., “Learning Representations by Back-propagating Errors,” Nature, vol. 323, pp. 533－536, 1986.

[47]Schlueter H., and Schoeddert R., “Engine Condition Monitoring – State-of-the-Art Civil Application,” AGARD CP-448, May-June 1988, pp. 12-1－12-8.

[48]Shephred A. J., “The Levenberg-Marquardt Method,” Second-Order Methods for Neural Networks, Springer, 1997, pp. 69－71.

[49]Stamatis, A., Mathioudakis K., Berios G., and Papailiou K., “Jet Engine Fault Detection with Discrete Operating Points Gas Path Analysis,” Journal of Propulsion and Power, vol. 7, no. 6, 1991, pp. 1041－1048.

[50]Torella G., “Expert Systems for the Trouble Shooting and the Diagnostics of Engines,” AIAA Paper 92-3327, 1992.

[51]Treager I. E., Aircraft Gas Turbine Technology, 3rd Ed., McGraw-Hill Inc., 1997.

[52]Urban, L. A., “Gas Path Analysis Applied to Turbine Engine Condition Monitoring,” AIAA/SAE Paper 72- 1082, 1972.

[53]Urban, L. A., “Parameter Selection for Multiple Fault Diagnostics of Gas Turbine Engine,” Journal of Engineering for Power, Transactions of the ASME, April 1975, pp. 225－230

[54]Urban, L. A., and Volponi, A. J., “Mathematical Methods of Relative Engine Performance Diagnostics,” SAE Technical Paper Series, Aerotech ’92, Anaheim, CA, October 1992, pp. 1－26.

[55]Whitehead, B., Kiech, E., and Ali, M., “Rocket Engine Diagnostics Using Neural Networks,” AIAA Paper 90-1892, 1990.

[56]Yao, X., “Evolving Artificial Neural Networks,” Proceedings of the IEEE, vol. 87, no. 9, September 1999, pp. 1423－1447

[57]Zedda, M., and Singh, R., “Fault Diagnosis of a Turbine Engine Using Neural Networks: A Quantitative Approach,” AIAA Paper 98-3602, 1998.