Experimental 'Reliability Growth' or 'Reliability Verification' Related with the Amount of Innovation in a New Car Model

Reliability growth testing, i.e. the ‘growth’ of reliability, is a consolidated systematic method used by Fiat Auto to increase and ‘demonstrate’ reliability while testing vehicles. If a new car model is not very different from the previous one (being used as reference), a single ‘verification’ experimental phase on the vehicles which is consistent with the end product may be sufficient for a reliability statistical demonstration. In this case, the resultant confidence interval may be very large, especially when we want to predict the in field failure frequency after applying corrective actions. Conversely, if the new product shows too many problems, we could obviously define, at any moment, a specific (and more demanding) reliability growth plan and restart the experiments. This paper suggests the cases in which we should choose a reliability growth test or a reliability verification test, whilst considering the total amount of design variations with respect to the reference model: completely new model, restyling, face lifting, etc. Copyright © 2004 John Wiley & Sons, Ltd.