THE QUANTIFICATION OF HUMAN RELIABILITY. A FEASIBILITY DEMONSTRATION
Abstract
The purpose of this study was to demonstrate a method of calculating the probability of human errors during prelaunch testing. Completion of studies described herein will permit the following statements: (a) The probability that the test can be completed without human error, (b) The probability that human errors will remain undiscovered, (c) The probability that undiscovered errors will result in failures. In addition to a reliability number for the total test, such studies will provide a breakdown of reliabilities for all human tasks in the test. Areas of greatest risk can then be pinpointed and corrective efforts focused on them. A proposed method for calculating system reliability is described. Development and application of this method will permit assessment of the contribution of each preflight test to over-all system reliability, and will also point out the need for adding or removing tests. This method takes into account the probabilities: (a) that the hardware was manufactured correctly, (b) that the hardware will not be damaged by human handling, (c) that the tests administered to the hardware will reveal all malfunctions, and (d) the inherent reliability of the hardware. Reliability thus determined would be a function of the following: R = f (CI) where R = reliability, C = confidence, and I = inherent reliability.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1965
- Accession Number
- AD0614134
Entities
People
- Allan M. Stave
Organizations
- General Electric