Coupling a Dynamic Linear Model with Random Forest Regression to Estimate Engine Wear

##plugins.themes.bootstrap3.article.main##

##plugins.themes.bootstrap3.article.sidebar##

Published Oct 10, 2010
James Schimert Art Wineland

Abstract

Mechanical systems wear or change over time. Data collected over a system’s life can be input to statistical learning models to predict this wear/change. A research effort at The Boeing Company has modeled gas turbine engine exhaust gas temperature (EGT) as a function of other recorded parameters. Our investigation chose gas turbines, but these techniques could also be used for other systems that slowly change (degrade) over time. Previous work trained a flexible empirical regression model at a fixed point of wear, and then applied it independently at time points over the life of an engine to predict wear. However, wear typically occurs slowly and smoothly. This paper describes the benefit of relating wear predictions over time using a dynamic linear model, which is an example of a state space method. The combined model predicts wear with dramatically reduced variability over both our previous effort and a baseline method. The benefit of reduced variability is that engine wear is more evident, and it is possible to detect operational anomalies more quickly. In addition to tracking wear, we also use the model as the basis for a Bayesian approach to monitor for sudden changes and reject outliers, and adapt the model after these events. Experiments compare methods and give some guidance in applying the methodology.

How to Cite

Schimert, J. ., & Wineland, A. . (2010). Coupling a Dynamic Linear Model with Random Forest Regression to Estimate Engine Wear. Annual Conference of the PHM Society, 2(1). https://doi.org/10.36001/phmconf.2010.v2i1.1802
Abstract 226 | PDF Downloads 240

##plugins.themes.bootstrap3.article.details##

Keywords

regression, State Space, Exhaust Gas Temperature, normalization, random forest, dynamic linear model

References
Basu, S. and Farner, S. and Schimert, J. and Wineland, A. (2008a)."Statistical Methods for Modeling and Predicting Maximum Engine Exhaust Gas Temperature (EGT): First Analysis Using Climb Data from a Single Aircraft". Networked Systems Technology Technical Report (NST-08-001)

Basu, S. and Farner, S. and Schimert, J. and Wineland, A. (2008b). "Regression-Based Method for Predicting Engine Wear from Exhaust Gas Temperature", Prognostics and Health Management Conference, Denver, CO.

Breiman, L. and Friedman, J. and Olshen, R. and Stone, C. (1984), Classification and Regression Trees, Wadsworth, Monterey, CA.

Breiman, L. (1996), "Bagging predictors", Machine Learning, volume 26, Pages 123—140.

Breiman, L. (2001a), "Random Forests", Machine Learning, volume 45, number 1, pages 5-32.

Breiman, L. (2001b), "Statistical modeling: The two cultures (with discussion)", Statistical Science, volume 16, number 3, pages 199—231.

Bühlmann, P. and Hothorn, T. (2007). Boosting algorithms: regularization, prediction and model fitting. Statistical Science, 22(4):477-505.

Doucet, A. (Editor) and de Freitas, N. (Editor) and Gordon, N. (Editor), (2001). "Sequential Monte Carlo Methods in Practice, Springer.

Hall, D.L. and McMullen, S.A.H. (2004). “Mathematical Techniques in Multisensor Data Fusion”, Artech House Information Warfare Library.
Friedman, J.H (2001), "Greedy function approximation: A gradient boosting machine", Annals of Statistics, volume 29, pages 1189—1202.

Liaw, A. and Wiener, M. (2002). Classification and Regression by randomForest. R News 2(3), 18--22.

Liggins, M. E., Hall, D. L., Llinas, J. (Editors) (2009) "Handbook of Multisensor Data Fusion: Theory and Practice," 2nd Edition, CRC Press, New York.

Murua, A (2002), "Upper bounds for error rates associated to linear combination of classifiers", IEEE Transactions on Pattern Analysis and Machine Intelligence.

Parikh D. and Polikar R. (2007), “An Ensemble based incremental learning approach to data fusion”, IEEE Transactions on Systems, Man and Cybernetics – Part B: Cybernetics, vol. 37, No. 2.
Giovanni Petris (2010). dlm: Bayesian and Likelihood Analysis of Dynamic Linear Models. R package version 1.1-1.
Petris, G. and Petrone, S. and Campagnoli, P. (2009) "Dynamic Linear Models with R", Springer.

Pole, A. and West, M. and Harrison, J. (1994). Applied Bayesian Forecasting and Time Series Analysis, Chapman & Hall.

R Development Core Team (2009). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3- 900051-07-0, URL http://www.R-project.org.

Schapire, R.E. (2003). The boosting approach to machine learning: An overview. In D. D. Denison, M. H. Hansen, C. Holmes, B. Mallick, B. Yu, editors, Nonlinear Estimation and Classification. Springer.

Schwabacher, M. and Kai Goebel, K. A Survey of Artificial Intelligence for Prognostics. NASA Ames Research Center. MS 269-3. http://ic.arc.nasa.gov/people/schwabac/FS02Schwab acherM.pdf.

Sinhaa, A. and Chenb, H. and Danua, D. G. and Kirubarajana, T. and Farooqc, M. (2008), "Estimation and Decision Fusion: A Survey," Neurocomputing Volume 71, Issues 13-15, pages 2650-2656,

West, M. (1986). Bayesian Model Monitoring”, Journal of the Royal Statistical Society, Series B, volume = 48, pages= 70—78.

West, M. and Harrison, J. (1986). “Monitoring and Adaptation in Bayesian Forecasting Models", Journal of the American Statistical Association, Vol. 81, No. 395, pages 741-750.

West, M. and Harrison, J. (1999). “Bayesian Forecasting and Dynamic Models", Springer.
Section
Technical Research Papers