Authors: Tongdan Jin, Yeliz Ozalp
Addresses: Texas A&M International University, 5201 University Blvd, Laredo, TX 78041, USA. ' Teradyne Inc., 300 River Park Drive, SM:3001, North Reading, MA 01864, USA
Abstract: This paper proposes an optimisation procedure to minimise product warranty cost through design for reliability (DFR) in the early development phase. The DFR tool enables the product manufacturers to pro-actively design out potential failure modes due to hardware, software, process and customer usage issues. Reliability prediction for complex hardware–software systems usually is difficult when uncertainties exist in operational environment. This paper proposes a stochastic reliability model to estimate component and non-component failure rates considering uncertainties in operational temperatures and customer usages. Based on the stochastic model, an optimisation program is formulated to minimise the expected product warranty cost subject to the design budget constraint. The trade-off between the design budget and the warranty cost is compared under different design options. Finally, the optimal DFR procedure is demonstrated on a type of automatic semiconductor test equipment. Throughout the paper, systems and products will be used interchangeably.
Keywords: reliability prediction; design for reliability; warranty cost; genetic algorithms; GAs; product development; optimisation; stochastic modelling; product warranty; product design; semiconductor test equipment; design for six sigma; DFSS.
International Journal of Six Sigma and Competitive Advantage, 2009 Vol.5 No.1, pp.42 - 58
Published online: 30 Mar 2009 *Full-text access for editors Access for subscribers Purchase this article Comment on this article