Validation Test Optimization Based on a Statistical Approach for Diesel Engine Cylinder Heads Reliability

This paper describes a statistical method of test sizing optimization, used by RENAULT for diesel engine cylinder heads validation. Based on mission profiles, tests results and damage computations, the paper describes an application of the "Stress Strength" method combined with a sampling plan. Associated with degradation measurements (crack length), it allows to estimate the future field reliability and to optimize the validation test sizing: number of parts to be tested, test duration, confidence level, supplier risk and test acceptance criterion. It permits to obtain less expensive validation tests while demonstrating high level of reliability.