30 May 2005


Estimation of the mean time between failures (hereinafter termed MTBF) for random failure is a
well-known part of equipment engineering, but no discussion of MTBF presently exists in connection
with building equipment. In this study, an investigation of MTBF estimation for random failure is
made using field data of building equipment operations, and a comparison of results is conducted by
applying two methods of interval estimation. The study touches upon the relationship of the results
achieved by the Bayesian approach with the assessed values of practical operations. The author
employed simulation of a maximum-likelihood method with the Bayesian method in estimating the
failure rate and then analyzed it by using the data of building equipment and devices for a period of
random failure. Especially it estimated the comparison between heat pump system and other
building system. These results suggest the most suitable and practical way of MTBF estimation for
actual maintenance sites.