Study of Statistical Failure Rate Analysis for Failure Risk Evaluation of Long-life Equipment

Abstract

In developed countries, the maintenance of ageing electrical power equipment is an important issue, and rational measures are being taken using asset management techniques. Rational maintenance and renewal planning based on risk analysis is required, and it is necessary to accurately estimate the failure rate according to the age of the equipment to be evaluated. One of the methods for understanding ageing characteristics is statistical analysis of equipment operating data such as failures and operating records. However, there are some points to consider when applying this method to equipment that can be operated for a long period of time, has a wide age distribution and is expected to have high reliability through preventive maintenance. Generally, failure statistical analysis involves linear regression, which assumes errors of equal variance in the observed data. However, this approach may have some problem when non-negligible errors are present. It is required regression analysis with appropriate weights and the estimation of appropriate prediction intervals. In this paper, the utilization of variance in binomial distributions and prediction interval estimation by failure occurrence simulation have been proposed, and the safety side estimation method for the failure rate is realized.