Evaluation of a Modified AMSAA (Army Material Systems Analysis Activity) Continuous Reliability Growth Model Using Failure Discounting and Weighting Factors

Abstract

Failure discounting is the practice of removing fractions of failures from test data after corrective actions have been taken and no failures due to the same cause have reoccurred. This thesis examines the affect of discounting failures and weighting test data on the accuracy of an existing reliability growth model, labeled the Modified AMSAA model. Computer simulation is used to evaluate the mean and mean square error of failure rate estimates under the model for a variety of reliability growth patterns each with several discounting and weighting scenarios. Exponential failure times are assumed and testing is truncated at two failures in each test phase. Failure discounting tended to decrease the mean square error slightly for growth patterns with a continual drop in failure rate for each new test phase, but tended to increase the mean square error for other patterns. The Modified AMSAA model is also shown to be superior to the standard AMSAA reliability growth model in bias and mean square error. No discernable benefits due to weighting the data were detected. Keywords: Exponential distribution; Monte Carlo method.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1989
Accession Number
ADA219371

Entities

People

  • Scott L. Negus

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Human Systems

DTIC Thesaurus Topics

  • Accuracy
  • Computer Programming
  • Computer Programs
  • Computers
  • Data Science
  • Equations
  • Failure Mode And Effect Analysis
  • High Reliability
  • Information Science
  • Monte Carlo Method
  • Random Number Generators
  • Random Variables
  • Regression Analysis
  • Reliability
  • Test And Evaluation
  • United States
  • United States Naval Academy

Fields of Study

  • Engineering

Readers

  • Life Cycle Cost Analysis
  • Regression Analysis.
  • Structural Health Monitoring of Composite Structures.