Analysing Reliability Data
This chapter describes a number of techniques, further to the probability plotting methods described in Chapter 3, which can be used to analyse reliability data derived from development tests and service use, with the objectives of monitoring trends, identifying causes of unreliability, and measuring or demonstrating reliability.
Since most of the methods are based on statistical analysis, the caution given in Section 2.17 must be heeded, and all results obtained must be judged in relation to appropriate engineering and scientific knowledge.
13.2 Pareto Analysis
As a first step in reliability data analysis we can use the Pareto principle of the ‘significant few and the insignificant many’. It is often found that a large proportion of failures in a product are due to a small number of causes. Therefore, if we analyse the failure data, we can determine how to solve the largest proportion of the overall reliability problem with the most economical use of resources. We can often eliminate a number of failure causes from further analysis by creating a Pareto plot of the failure data. For example, Figure 13.1 shows failure data on a domestic washing machine, taken from warranty records. These data indicate that attention paid to the program switch, the outlet pump, the high level switch and leaks would be likely to show the greatest payoff in warranty cost reduction. However, before committing resources it is important to make sure that ...