If failure time data from several systems of the same kind are available it is often desirable to analyze these data simultaneously. Two concepts for doing simultaneous trend testing in data from more than one system are presented. These two concepts are compared in a simulation study, and a general strategy for trend testing in data from several systems that draw beneets from both concepts is proposed. 1 INTRODUCTION For repairable systems it is important to detect possible changes in the pattern of failures, which for instance can be caused by various aging effects or reliability growth. We say that there is a trend in the pattern of failures if the inter-failure times tend to alter in some systematic way, which means that the inter-failure times are not identically distributed. By using statistical trend tests it is possible to decide whether such an alteration is statistically signiicant or not. If we have failure time data from several repairable systems, we can either test for trend in each separate system, or do a simultaneous analysis. Doing a simultaneous analysis of data from more than one system requires that all the systems are judged to be suuciently equal for a simultaneous analysis to make sense. This decision must primarily be based on knowledge of the systems and their operating conditions. Graphical techniques for exploring the pattern of failures in each system can also be useful in this stage of the analysis. If doing a simultaneous analysis can be justi-ed, this will in general be much more powerful than analyzing each system separately. Changes in the pattern of failures which is impossible to
[1]
J. Lawless.
Regression Methods for Poisson Process Data
,
1987
.
[2]
Harold B. Sackrowitz,et al.
Evaluating tests for increasing intensity of a Poisson process
,
1993
.
[3]
Richard E Barlow,et al.
Analysis of Time Between Failures for Repairable Components.
,
1977
.
[4]
Lee J. Bain,et al.
Tests for an Increasing Trend in the Intensity of a Poisson Process: A Power Study
,
1985
.
[5]
P. C. Sander,et al.
Repairable systems reliability: Modeling, inference, misconceptions and their causes
,
1986
.