Using CX Proportional Hazard Model for.docVIP

Using COX Proportional Hazard Model for Multivariate Prognostic Analysis of Bearings ? Jinghua Ma and David He ? Intelligent Systems Modeling and Development Laboratory Department of Mechanical Industrial Engineering The University of Illinois at Chicago Chicago, IL 60607 E-mail: davidhe@uic.edu Tel: 312-996-3410 ? Abstract ? The Cox proportional hazard model has been extensively used in biomedical survival analysis.? It can provide a statistically rigorous prediction of system failures based on the time-to-failure data obtained from experimental tests. In this paper, the Cox proportional hazard model is used to analyze the multivariate risk factors influencing the life of the system and then to predict failure rate of the system. Failure rate is a key criterion in estimating the reliability of a system. By employing this method, the failure risk of system which is active at time instantaneous t withinΔt can be obtained, that is, the failure probability of the system between t and (t+Δt).? Prognostic index can be obtained from Cox model which takes into account all significant influencing factors including operational conditions and component materials.? Failure rate can be obtained from prognostic index. Two parameters Weibull distribution is used to represent the baseline hazard function and a Newton-Raphson iterative method of Maximum Likelihood Estimation (MLE) algorithm is used to determine the coefficients. Data collected from real steel and ceramic bearing run-to-failure tests under different operational conditions in a numerical study are used to evaluate the effectiveness of the proposed approach.? The results show that the proposed method can be effective in predicting the real remaining useful life of the bearings. Keywords: Cox proportional hazard model, ? multivariate risk factors, ? failure probability. 1. Introduction The current technique to predict failure can be roughly classified in to two methods. One is physical model (design-based) method in