Title: Reliability in automotive design

Authors: R.G. Bertodo

Addresses: Director, Product Engineering, Rover Group, Coventry, UK

Abstract: A survey of upper-medium sector cars showed that, for a specific definition, reliability levels of the order of 0.78 were required for customer satisfaction under European operating conditions. Low variability, however, appeared to be at least as important an indicator of market acceptability. A mathematical model of the field reactions was derived from conventional statistical theory, and was used to demonstrate a high level of sensitivity to safety margin and |loading roughness|. This inherent sensitivity led to the conclusion that reliability could not be calculated with adequate accuracy at the design stage or satisfactorily demonstrated by simulated testing. This finding did not align with practical experience, since many examples of successful design existed, as demonstrated by the survey. Analysis of the design and development process used for a motor car recognised by the market as |reliable| allowed the evolution of a ten-step design rule. This is presented with examples from published work and supported by a case study. The implications of the design rule are discussed.

Keywords: vehicle design; failure prediction; reliability; automotive components; failure rate; defect rate; mathematical modelling.

DOI: 10.1504/IJVD.1988.061478

International Journal of Vehicle Design, 1988 Vol.9 No.2, pp.140 - 158

Published online: 26 May 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article