Development of accelerated life test criteria

8
Development of
Accelerated Life Test
Criteria
Yung-Li Lee
DaimlerChrysler
Mark E. Barkey
University of Alabama

8.1 INTRODUCTION

Product validation tests are essential at later design stages of product development. In the automotive industry, fatigue testing in a laboratory is an accelerated test that is specifically designed to replicate fatigue damage and failure modes fromproving grounds (PG) testing. Detailed damage analysis is needed to correlate the accelerated test to PG testing. Therefore, accurate representation of PG loading is essential for laboratory durability test development.
PG loading can be measured by driving an instrumented vehicle over the PG with various test drivers. The vehicle is equipped with transducers for component loading histories andsensors for other important vehicle parameters, such as temperature, speed, and displacement. A typical test schedule that was derived on a target customer (usually the 95th percentile customer usage) comprises numerous repeated cycles, usually varying from 800 to 1000. Each cycle then combines several mixed events (e.g., rough road, high-speed laps, city route, and country route). Data acquisitionis usually performed on one test cycle. As illustrated in Figure 8.1, the one-cycle loading measurement for various drivers should be extrapolated to the expected complete loading history. The statistical techniques for cycle extrapolation (Drebler et al., 1996; Roth, 1998; Socie, 2001; Nagode et al., 2001; Socie and Pompetzki, 2003) are briefly described here.
It is believed that vehicleusage, operational conditions (such as weather and road roughness), and driver variability can affect the loading profile. Weather can influence traction force of the tires. For example, wet weather induces a smaller coefficient of friction between a tire and the ground than dry weather does. Roughness of a dirt road varies from time to time and can be washed out abruptly after rains. A study (Socieand Park, 1997) has shown that even for professional drivers, the driving patterns are not repeatable and can significantly alter the measured loading profile. In general, the weather and road roughness conditions have less significant impact to powertrain components than to the chassis and body components. To design components with a very low probability of fatigue failure, it is essential toquantify the severe damage-loading spectrum with a very low probability of occurrence (usually less than or equal to 5%) and the component fatigue resistance with a very high probability of survival. In this chapter, the quantile extrapolation techniques (Drebler et al., 1996; Roth, 1998; Socie and Pompetzki, 2003) in the rainflow cycle domain are reviewed.
This chapter demonstrates how todevelop an accelerated fatigue test criterion to account for test driver variability. An equivalent damage concept is used to correlate the accelerated test to PG loading, meaning that the damage value due to a laboratory test is equivalent to that from a PG loading. Based on the specific percentile damaging loading profile, product validation tests for durability are developed accordingly. This chapterspecifically describes the development procedures of two types of product validation tests: one for powertrain system durability testing (dynamometer testing) and the other for general mechanical component fatigue testing (component life testing).

8.2 DEVELOPMENT OF DYNAMOMETER TESTING

A dynamometer, referred to as a differential or a transmission dynamometer, has been widely used inthe automotive industry to validate durability and mechanical integrity of a gear set. For example, a typical differential dynamometer shown in Figure 8.2 consists of a driving shaft, a differential gear set, and two driven shafts. With a control module, a driving shaft will develop horsepower to turn the differential gears and the driven shafts from which the dynamometer motors absorb power and...