Failure Rate Prediction

FREE tool to predict MTTF/MTBF and Field Failure Rates and charting to monitor Reliability Growth

How do you make Failure Rate Predictions from Test Data?

This is a Question all reliability Engineers struggle with and often try to use custom software tools, but not necessarily applying the correct principles or assumptions.

Failure Rate Predictions and Reliability Predictions are generally made from Accelerated Life Testing (ALT) where the Mean Time to Failure (MTTF/MTBF) level in hours is predicted from Accelerated Test Data, then converted into Failure Rate in simple steps.

The key factors in any Accelerated Life Test are:

  1. Test Period
  2. No. of units tested
  3. No. of Fails Recorded
  4. Acceleration Factors Applied
  5. Normal Life Duty Cycle or running hours within specific period (month, year, etc.)

The objective of the Life Test is to run minimum sample size for shortest period, but simulate sufficient normal life period to make a Confident Prediction of Field Failure Rate.

The simplest way to set up Accelerated Life Test includes

  1. Define Accelerated Stress Test Environments such as Higher Temperature and Higher Relative Humidity (RH)
  2. Select the appropriate Acceleration models (Arrhenius, Peck’s , Coiffon-Manson, etc) to be used to calculate appropriate Acceleration Factors

    Temperature Acceleration Model

    AF= [exp (-Ea/kT1) / exp (-Ea/kT2)] = exp [ -(Ea/k) (1/T1 – /T2) ]

    T1, T2 = temperature during accelerated test and under normal use temp, respectively
    Ea=activation energy (0.6-0.7 average for Electronic Products),
    k= Boltzmann’s constant (8.6e-5 eV/K), T=absolute temperature

    Humidity Acceleration Model

    AF = (RHtest/RHuse) n

    n = power function , general level often used for product level electronic testing = 1-2

  3. Input Accelerated Life Test environment parameters (Temperature and RH) in to the ‘Reliability Solutions Failure Prediction Model’ against the normal ambient Customer Usage levels to calculate the Acceleration Factors
  4. Set up test with sample quantity provided and run test for selected time periods
  5. Collect Failure Data during test or multiple tests
  6. Input all data to the Reliability Solutions ‘Failure Prediction’ model which calculates Average Hazard Rate Prediction and 60% / 90% Confidence Hazard Rate Predictions
  7. Output is provided in 3 different Failure Rate Predictions
    • a. 12 mth Prediction
    • b. 60% Confidence Prediction
    • c. 90% Confidence prediction

The steps to be followed when using Reliability Solutions MTTF/MTBF Prediction / Field Fail Rate prediction software as below

a) Define Ambient Temperature your product normally operates in (20C, 25C, etc)
b) Specify the High Temperature level to be used in Accelerated Test (50C, 60C, etc)
c) Define Ambient Relative Humidity (RH) level your product normally operates in
d) Specify the Higher RH level to be used in Accelerated Test
e) Input Test Qty in appropriate box
f) Input Test Period in hours (total length of time for the test)
g) Input the 12 mth Normal Usage hrs (could be changed to represent 12 mths, 24, 36, etc)

After inputting all above data, software will auto-calculate all the relevant predictions you require:

Average MTTF/MTBF Prediction

60% Confidence MTTF/MTBF Prediction
(this statement means you can be 60% confident the MTTF/MTBF prediction will be higher than the level you have calculated)

90% Confidence MTTF/MTBF Prediction

Average 12 mth Predicted Field failure Rate

60% Confidence statement on 12 mth Predicted Field failure Rate
(This means you can be 60% confident the Field Failure Rate will be no higher than the level calculated by the software)

90% Confidence statement on 12 mth Predicted Field failure Rate