Andre V. Kleyner
Electronics & Safety Division
University of Maryland
College Park, MD 20742
This paper addresses the effect of two-dimensional warranty policies on the procedure for forecasting the cost of unreliability. Automotive warranties are characterized by age or time in service and vehicle mileage. This paper presents a model where the usage time is considered as a primary variable and the mileage accumulation is estimated based on the field return data. At each time interval the probability of exceeding the upper mileage limit on the warranty is calculated based on the mileage distribution and an analytical cumulative distribution function (CDF) of exceeding the upper mileage limit at any point of time is constructed. The CDF function is then used to adjust the forecasted number of units failing during the warranty period. This innovative modeling procedure helps to account for an observed reduction in the number of warranty claims in the second half of the warranty period thus making it a more realistic and accurate evaluation of the cost of unreliability incurred by a manufacturer of products with two-dimensional warranty policies. This paper also presents a case study using an automotive parts example in order to illustrate the methodology.
Complete article is available to CALCE Consortium Members.