DESIGN OF OPTIMUM SIMPLE STEP-STRESS ACCELERATED LIFE TESTING PLANS

Abstract

ABSTRACT Accelerated life testing (ALT) is widely conducted to obtain failure-time data in a much shorter time and to make inference about reliability at normal conditions. The accuracy of the reliability prediction is dependent on well-designed ALT plans. A step-stress ALT allows the test condition to change at a given time or upon the occurrence of a specified number of failures. In this paper, we propose a procedure to determine the parameters of the optimum simple step-stress testing plan so that the reliability prediction at normal conditions is accurately determined. We investigate efficient procedures to estimate most, if not all, of the ALT parameters under different operating conditions. The objective function is formulated as a nonlinear programming problem. The resultant optimum plan is demonstrated through numerical example and sensitivity analysis. KEYWORDS Accelerated life test, Proportional hazards, Step-stress, Maximum likelihood estimate 1 INTRODUCTION Accelerated life testing (ALT) is used to quickly obtain reliability-related information on products’ life and/or degradation data, and often promoted as a solution to save test time and costs. Inference about the reliability of products at normal operating conditions can be obtained using data obtained from the accelerated conditions. The accuracy of the inference procedure profoundly affects the reliability estimates at normal conditions and the subsequent decisions regarding system configuration, warranties and preventive maintenance schedules. The accuracy of the reliability estimates mainly depends on two factors: the ALT models and the experimental design of the ALT plans. Optimum plans yield the most accurate reliability estimates of products’ life at the normal conditions, or the design stress. ALT is usually conducted by subjecting the product to severer conditions than normal design conditions (accelerated stress) or by using the product more intensively than in normal use without changing the normal operating conditions (accelerated failure time). Conducting an accelerated life testing requires the development of a proper reliability