This paper presents a neural network approach to fault diagnosis of dynamic engineering systems based on the classification of surfaces in system output vector space. A simple second-order system is used to illustrate graphically the nature of the diagnosis problem and to develop theory. The approach is then applied to the diagnosis of a laboratory-based hydraulic actuator circuit. Results are presented for networks trained on both simulation and experimental data. An important achievement is the diagnosis of experimental faults using a network trained only on simulation data.
FinkP.LusthJ.Expert systems and diagnostic expertise in the mechanical and electrical domains. IEEE Trans. Systems, Man and Cybernetics, May 1987, SMC-17 (3), 340–349.
2.
WoodC. L.FADES: A tool for automated fault analysis of complex systems. In Research and Development in Expert Systems 89, 1989, pp. 253–262 (Cambridge University Press).
3.
WattonJ.Lucca-NegroO.StewartJ. C.An on-line approach to fault diagnosis of fluid power cylinder drive systems. Proc. Instn. Mech. Engrs, Part I, Journal of Systems and Control Engineering, 1994, 208 (I4), 249–262.
4.
FarB. H.NakamichiM.Qualitative fault diagnosis in systems with nonintermittent concurrent faults: A subjective approach. IEEE Trans. Systems, Man and Cybernetics, January 1993, 23 (1), 14–30.
5.
LinkensD. A.WangH.Fault diagnosis on a qualitative bond graph model, with emphasis on fault localisation. In IEE Conference Publication 389, March 1994, pp. 1329–1334.
6.
LeitchR. R.Engineering diagnosis: Matching problems to solutions. In Proceedings of International Conference on Fault Diagnosis (Tooldiag ‘93), Vol. 3, Toulouse, France, 1993, pp. 837–844.
7.
LeitchR. R.ChantlerM. J.ShenQ.CoghillG. M.A preliminary specification methodology for model based diagnosis. InAnnals of Mathematics and Artificial Intelligence, Special Issue on Current Trends in Research on Model-Based Diagnosis, 1993.
8.
DohertyN. F.KochharA. K.Knowledge engineering for model-based diagnosis: An experience-based approach. Engng Applic. of AI, 1994, (6), 653–663.
9.
DohertyN. F.KochharA. K.MIDAS: An application of model-based reasoning for the diagnosis of hydraulic systems. Knowledge-Based Systems, June 1994, 7 (2), 127–134.
10.
MilneR.Strategies for diagnosis. IEEE Trans. Systems, Man and Cybernetics, May 1987, SMC-17(3), 333–339.
11.
WilskyA.A survey of design methods for failure detection in dynamic systems. Automatica, 1976, 12, 601–611.
12.
PattonR.FrankP.ClarkR.Fault Diagnosis in Dynamic Systems; Theory and Applications, 1989 (Prentice-Hall, Englewood Cliffs, New Jersey).
13.
PattonR.ChenJ.Optimal unknown input distribution matrix selection in robust fault diagnosis. Automatica, 1993, 29 (4), 837–841.
14.
IsermannR.On the applicability of model-based fault detection for technical processes. Control Engng Practice, 1994, 2 (3), 439–450.
15.
YuD.ShieldsD. N.MahtaniJ. L.A non-linear fault detection method for a hydraulic system. In IEE Conference Publication 389, March 1994, pp. 1318–1322.
16.
BernieriA.D'ApuzzoM.SansoneL.SavastanoM.A neural network approach for identification and fault diagnosis on dynamic systems. IEEE Trans. Instrum. Measnt, December 1994, 43 (6).
17.
DaleyS.WangH.On the application of neural networks to the monitoring of a simulated hydraulic rotary drive system. In 6th Bath International Fluid Power Workshop, University of Bath, 1993.
18.
ScaifeM. W.CharltonS. J.MobleyC.A neural network for fault recognition. In SAE International Congress and Exposition, Detroit, Michigan, March 1993, paper 930861.
19.
LeonardJ. A.KramerM. A.Radial basis function networks for classifying process faults. IEEE Control Systems, April 1991, 0272–1708/91, 31–38.
20.
SorsaT.KoivoH.Application of artificial neural networks in process fault diagnosis. Automatica, 1993, 29 (4), 843–849.
21.
PattonR. J.ChenJ.SiewT. M.Fault diagnosis in non-linear dynamic systems via neural networks. In IEE Conference Publication 389, March 1994, pp. 1346–1351.
22.
AdamS.KingR.A neural network-based architecture for on-line fault diagnosis. In Parallel Distributed Computing in Engineering Systems (Eds TzafestasS.BorneP.GrandinettiL.), 1992 (Elsevier Science Publishers BV (North-Holland), Amsterdam).
23.
BecraftW.LeeP.An integrated neural network/expert system approach for fault diagnosis. Computers in Chem. Engng, 1993, 17 (10), 1001–1014.
24.
BishopC.Neural Networks for Pattern Recognition, 1996 (Oxford University Press).
25.
RumelhartD. E.WidrowB.LehrM. A.The basic ideas in neural networks. Commun. ACM, March 1994, 37 (3), 86–92.
26.
MinskyM.PapertS.Perceptrons—An Introduction to Computational Geometry, 1969 (MIT Press, London).
27.
RumelhartD.HintonG.WilliamsJ.Learning internal representations by error propagation. In Parallel Distributed Processing, Explorations in the Microstructure of Cognition, Vol. 1, Foundations, 1987 (MIT Press, London).
28.
BlumE.LiL.Approximation theory and feed-forward networks. Neural Networks, 1991, 4 (4), 511–515.
29.
MastersT.Practical Neural Network Recipes in C++, 1993 (Academic Press, New York).
30.
PollmeierK.Simulation and condition monitoring of fluid power systems. PhD thesis, University of Bath, 1997.
31.
CrowtherW.Fault diagnosis of engineering systems using neural networks: A practical approach. In IEE Colloquium onModelling and Simulation for Fault Diagnosis, University of Leicester, September 1996, paper 1996/260.
32.
LjungL.System Identification Theory for the User, 1987 (Prentice-Hall, Englewood Cliffs, New Jersey).
33.
MilneR.Artificial intelligence for online diagnosis. IEE Proc., July 1987, 134 (D4), 238–244.
