Abstract
A study on automotive reliability in real conditions of operation can be performed based on data gathered from service records for a lot of vehicles. Because the mileage of vehicles are different, it is required to process the data considering as truncated time the minimum mileage of the group of vehicles studied. In this case, we work with concepts specific to incomplete tests of truncated type. In this situation it occurs the next opportunity: the verisimilitude of the achieved reliability model can be checked by extrapolation of reliability beyond the time of truncation and the number of expected defects based on theoretical model for any moment shall not be lower than the one found in the service records for the vehicles running more than the truncation time considered. This paper presents a batch of 200 vehicles with mileage over 225,000 km, at which we made a reliability study for the oxygen sensor based on the Weibull model, thus demonstrating the viability of the proposed algorithm.
Keywords: Oxygen Sensor, Reliability Modeling, Incomplete Tests, Truncated Tests, Weibull Law