Engineered Software

Reliability Test Design

Overview

Bogey Testing

Bayesian Testing

Engineered Software Home Page

 A reliability test plan requires:
  1. a sample size,
  2. a test duration,
  3. a criteria for acceptable performance, and
  4. the number of failures allowed.

For example, given a Weibull time to fail distribution with a shape parameter of 2, 95% reliability can be demonstrated at 100 hours with a confidence level of 90% by any of the test plans shown in the table below.

Test
Plan

Number of
Units Tested

Number of
Failures

Test Duration
(Hours)

A

45

0

100

B

76

1

100

C

192

5

100

D

318

10

100

E

45

1

131

The most efficient reliability test plans assume there will be no failures. Of course this is an unrealistic assumption, but if failure occurs the item under consideration has not necessarily failed to meet the reliability requirements. The test may be continued until the reliability requirements are met, or until it is evident that the reliability requirements will not be met. This defies traditional statistical thinking which requires test parameters and pass fail requirements to be determined prior to testing.

BackHomeNext