Much Needed Attention to Car Reliability Demonstration Testing and Test Sample Size Determination 2006-01-1616

The automotive industry is suffering from massive recalls, even after many developments have occurred in the fields of Reliability and Quality Engineering. The cost of recalls has been devastating to the industry's profits and product reputation in terms of the bad image to the customer, which is spread to more customers by the dissatisfied customers. Furthermore, the Test Sample Size is often too small to provide useful results concerning the actual designed-in Reliability. The manufacturers do not seem to use sufficient sample sizes in their product Reliability Demonstration Tests, which yield large errors in the test results. In this paper we seek to stress the importance of adequate Testing prior to shipment as a measure to avoid recalls and increase Car Reliability. Millions of recalls every year would have been avoided if adequate Testing had been conducted and the Reliability of the Sample had thus been improved, based on the test results. Here we explore the quantification of the Test Sample Size for the two most important distributions in Reliability Engineering, the Exponential and the Weibull, to assure that the error in determining the Reliability and MTBF (Mean Time Between Failures) is definitely under 10%, and closer to 5%. The goal is to raise the level of alert towards using adequate Test Sample Sizes, and provide the automotive industry with a comprehensive and easy to use guide to determine the required number of units that need to be tested.