Probabilistic Reliability vs. Failure Mode Avoidance Methodologies Within the Automotive Industry

Reliability from a probabilistic definition has been defined as the probability that a system will meet its intended function under stated conditions for a specified period of time. Therefore, in order to predict reliability, one must know three things: 1) function, 2) stated conditions, and 3) the specified time for the prediction. Manufacturers within the automotive industry have begun to challenge the probabilistic definition-stating reliability is too difficult to measure. Their reasoning is the fact that one cannot accurately define the time to measure reliability nor the stated conditions in the field. An emerging concept is not to measure reliability using conventional tools and methodologies, but to measure the failure modes that occur. If one can then distance the product from the failure mode, by default product reliability will increase. Therefore an emerging definition for reliability is simply failure mode avoidance. It is the intent of this paper to review probabilistic methods of reliability measurement, modeling and prediction as applicable to the automotive industry. Probability distribution functions and their applications in reliability evaluation will be explored. Failure mode avoidance methodologies, including noise management strategies will be explored as well. In conclusion, contrasting viewpoints of probabilistic reliability and failure mode avoidance will be presented.