The presenter uses warranty repair data to gain insight into product reliability until the time of first repair. He demonstrates how to examine changes in the reliability of repairable systems over time and the number of failures expected during a particular time period. He determines when warranty claims will peak, the cumulative cost of warranty-related repairs and how to determine resources needed for anticipated repairs.
The presenter shows how to analyze the reliability of systems that can have multiple failures over time, a technique useful for product failures as well as recurrence of an illness. He uses data about systems at two locations to calculate mean time between product failures and compare failures based on location. He interprets Event Plots, Mean Cumulative Failure (MCF) Plots, MCF Differences and Prediction Profilers.
The presenter shows how to model and interpret system reliability changes over time as design improvements are made. He explains that JMP will implement change detection when phase is not selected.