Warranty Analysis Exercise 1

A reliability engineer wants to predict warranty claims that are caused by defective heat exchangers in air conditioning units. The engineer collects and analyzes annual failure field data. * indicates censored data. The units will have a 10 year warranty. It costs $500 to repair a failed unit in the field.

Use the warranty prediction data in the file FieldData.mwx to forecast future warranty claims.

1. If the failure data is modeled using a Weibull distribution, show the predicted failures and predicted costs plots.
2. Is the Weibull distribution an appropriate reference distribution given the data? Why?
3. What is the range of the predicted number of failures within the next 5 years?
4. What is the range of the predicted cost of failures within the next 5 years?
5. If a warranty was sold for each unit in the field, what should be expected the minimum price of the 10 year warranty?