Warranty Cost Forecasting and Prediction Based on Reliability Models

摘要

Warranty cost and policy are directly related to reliability models. Model uncertainty is a critical factor affecting the accuracy of warranty cost forecasting and prediction. The two-parameter Weibull distribution is commonly used in warranty modeling and fatigue S-N data analysis, however, the three-parameter Weibull fits the data shown in this presentation better. The two-parameter and three-parameter Weibull result in different failure rates (percentages), in both short term and long term. The threshold parameter should play an important role in warranty cost forecasting/prediction and policy making, therefore, should not be ignored.