{"title":"Fault Diagnosis of Nonlinear Systems Using Dynamic Neural Networks","authors":"E. Sobhani-Tehrani, K. Khorasani, N. Meskin","volume":98,"journal":"International Journal of Computer and Information Engineering","pagesStart":539,"pagesEnd":550,"ISSN":"1307-6892","URL":"https:\/\/publications.waset.org\/pdf\/10001213","abstract":"
This paper presents a novel integrated hybrid
\r\napproach for fault diagnosis (FD) of nonlinear systems. Unlike most
\r\nFD techniques, the proposed solution simultaneously accomplishes
\r\nfault detection, isolation, and identification (FDII) within a unified
\r\ndiagnostic module. At the core of this solution is a bank of adaptive
\r\nneural parameter estimators (NPE) associated with a set of singleparameter
\r\nfault models. The NPEs continuously estimate unknown
\r\nfault parameters (FP) that are indicators of faults in the system. Two
\r\nNPE structures including series-parallel and parallel are developed
\r\nwith their exclusive set of desirable attributes. The parallel scheme is
\r\nextremely robust to measurement noise and possesses a simpler, yet
\r\nmore solid, fault isolation logic. On the contrary, the series-parallel
\r\nscheme displays short FD delays and is robust to closed-loop system
\r\ntransients due to changes in control commands. Finally, a fault
\r\ntolerant observer (FTO) is designed to extend the capability of the
\r\nNPEs to systems with partial-state measurement.<\/p>\r\n","references":"[1] S. Simani, C. Fantuzzi, R. J. Patton, Model-based Fault Diagnosis in\r\nDynamic Systems Using Identification Techniques, Springer 2003.\r\n[2] R. Isermann, Model-based Fault Detection and Diagnosis - Status and\r\nApplications, Annual Reviews in Control 29 (2005) 71-85.\r\n[3] R. Rengaswamy, D. Mylaraswamy, V. Venkatasubramanian, K. E.\r\nArzen, A Comparison of Model-based and Neural Network-based\r\nDiagnostic Methods, Engineering Applications of Artificial Intelligence\r\n14 (2001) 808-818.\r\n[4] P. Frank, Fault Diagnosis in Dynamic Systems using Analytical and\r\nKnowledge-based Redundancy: A Survey and Some New Results,\r\nAutomatica 26 (1990) 459-474.\r\n[5] R. J. Patton, C.J. Lopez-Toribio, F. J. Uppal, Artificial Intelligence\r\nApproaches to Fault Diagnosis,\u201d in: IEE Colloquium on Condition\r\nMonitoring: Machinery, External Structures and Health (Ref. No.\r\n1999\/034), April 1999, pp. 5\/I-5\/18.\r\n[6] N. Vaswani, Adaptive Change Detection in Nonlinear Systems with\r\nUnknown Change Parameters, IEEE Trans. on Signal Processing 55 (3)\r\n(2007) 859-872.\r\n[7] P. Li, V. Kadirkamanathan, Particle Filtering based Likelihood Ratio\r\nApproach to Fault Diagnosis in Nonlinear Stochastic Systems, IEEE\r\nTrans. on Sys, Man Cybernetics 31 (3) (2001) 337-343.\r\n[8] P. Li, V. Kadirkamanathan, Fault Detection and Isolation in Non-linear\r\nStochastic Systems - A Combined Adaptive Monte Carlo Filtering and\r\nLikelihood Ratio Approach, Int. J. of Control 77 (12) (2004) 1101-1114.\r\n[9] L. Guo, L. Yin, H. Wang, T Chai, Entropy Optimization Filtering for\r\nFault Isolation of Nonlinear Non-Gaussian Stochastic Systems, IEEE\r\nTrans. on Automatic Control 54 (4) (2009) 804-810.\r\n[10] E. Sobhani-Tehrani, K. Khorasani, S. Tafazoli, Dynamic Neural\r\nNetwork-based Estimator for Fault Diagnosis in Reaction Wheel\r\nActuator of Satellite Attitude Control System, in: Proceedings of the\r\nIEEE Int. Joint Conf. on Neural Networks (IJCNN), Montreal, Canada,\r\n2005, pp. 2347-2352.\r\n[11] A. Alessandri, Fault Diagnosis for Nonlinear Systems using a Bank of\r\nNeural Estimators, Computers in industry 52 (2003) 271-289.\r\n[12] Z. Xiaodong, M. M. Polycarpou, T. Parisini, A Robust Detection and\r\nIsolation Scheme for Abrupt and Incipient Faults in Nonlinear Systems,\r\nIEEE Trans. on Auto. Control 47 (4) (2002) 576 \u2013 593.\r\n[13] M. M. Polycarpou, A. J. Helmicki, Automated Fault Detection and\r\nAccommodation: A Learning Systems Approach, IEEE Trans. on Sys.\r\nMan Cybernetics 25 (11) (1995) 1447 \u2013 1458.\r\n[14] R. K. Mehra, C. Rago, S. Seereeram, Failure Detection and\r\nIdentification using a Nonlinear Interactive Multiple Model (IMM)\r\nFiltering Approach with Aerospace Applications, in: Proceedings of the\r\n11th IFAC Symposium on System Identification, Fukuoka, Japan, July\r\n1997.\r\n[15] Y. Zhang, Xiao, Detection and Diagnosis of Sensor and Actuator\r\nFailures using Interacting Multiple-Model Estimator, in: Proceedings of\r\nthe 36th IEEE Conference on Decision and Control, San Diego, CA, Dec.\r\n1997, pp. 4475-4480.\r\n[16] T. Jiang, K. Khorasani, S. Tafazoli, Parameter Estimation-Based Fault\r\nDetection, Isolation and Recovery for Nonlinear Satellite Models, IEEE\r\nTransactions on Control Systems Technology 16 (4) (2008) 799-808.\r\n[17] N. Tudoroiu, K. Khorasani, Fault Detection and Diagnosis for Satellite\u2019s\r\nAttitude Control System using An Interactive Multiple Model (IMM)\r\nApproach, in: Proceedings of the 2005 Conference on Control\r\nApplications, Toronto, Canada, August 2005, pp. 1287-1292.\r\n[18] N. Tudoroiu, E. Sobhani-Tehrani, K. Khorasani, Interactive Bank of\r\nUnscented Kalman Filters for Fault Detection and Isolation in Reaction\r\nWheel Actuators of Satellite Attitude Control System, in: Proceedings of\r\nthe 32nd IEEE Conference on Industrial Electronics, Paris, France, 2006,\r\npp. 264-269.\r\n[19] C. Rago, R. Pransanth., R. K. Mehra, R. Fortenbaugh, Failure Detection\r\nand Identification and Fault Tolerant Control using the IMM-KF with\r\nApplications to the Eagle-Eye UAV, in: 37th IEEE Conf. on Decision\r\nand Control, Tampa, FL, December 1998, vol. 4, pp. 4208-4213.\r\n[20] A. Medvedev, State Estimation and Fault Detection by a Bank of\r\nContinuous Finite-Memory Filters, Int. J. of Control 69 (4) (1998) 499-\r\n517.\r\n[21] R. Isermann, Fault Diagnosis of Machines via Parameter Estimation and\r\nKnowledge Processing - A Tutorial Paper, Automatica 29 (4) (1994)\r\n815-835.\r\n[22] A. Houacine, Regularized Fast Recursive-Least Squares Algorithms for\r\nFinite Memory Filtering, IEEE Transactions on Signal Processing 40\r\n(1992) 758-769.\r\n[23] S. Haykin, Kalman filters, in: S. Haykin (Ed.), Kalman Filtering and\r\nNeural Networks, Wiley\/Interscience, 2001, pp. 1-22.\r\n[24] T. Parisini, R. Zoppoli, Neural Networks for Nonlinear State Estimation,\r\nInternational Journal of Robust Control 4 (1994) 231-248.\r\n[25] J. T. Lo, Synthetic Approach to Optimal Filtering, IEEE Transactions on\r\nNeural Networks 5 (5) (1994) 803-811. [26] S. W. Piche, Steepest Descent Algorithms for Neural Network\r\nController and Filters, IEEE Transactions on Neural Networks 5 (2)\r\n(1994) 198-212.\r\n[27] N. Tudoroiu, K. Khorasani, Fault Detection and Diagnosis for Reaction\r\nWheels of Satellite\u2019s Attitude Control System using a Bank of Kalman\r\nFilters, in: Int. Symp. on Signals, Circuits and Systems, Iasi, Romania,\r\nvol. 1, July 2005, pp. 199-202.\r\n[28] H. A. Talebi, R.V. Patel, An Intelligent Fault Detection and Recovery\r\nScheme for Reaction Wheel Actuator of Satellite Attitude Control\r\nSystems, in: IEEE Conf. on Control Applications, Munich, Germany,\r\nOctober 2006, pp. 3282 \u2013 3287.\r\n[29] H. A. Talebi, K. Khorasani, S. Tafazoli, A Recurrent Neural Networkbased\r\nSensor and Actuator Fault Detection, Isolation for Nonlinear\r\nSystems with Application to a Satellite's Attitude Control Subsystem,\r\nIEEE Transactions on Neural Networks 20 (1) (2009) 45-60.\r\n[30] A. Valdes, K. Khorasani, A Pulsed Plasma Thruster Fault Detection and\r\nIsolation Strategy for Formation Flying Satellites, Applied Soft\r\nComputing 10 (3) (2010) 746-758.","publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 98, 2015"}