An innovative approach to the design of multi-bolted joint assemblies in the uncertain early stages

• Published in: Transactions on the built environment Vol. 106; pp. 55 - 68
• Main Authors: DERICQUEBOURG, T, LENOIR, D, DEBARLE, F, JEZEQUEL, L
• Format: Conference Proceeding Journal Article
• Language: English
• Published: Southampton WIT 2009; W I T Press
• Subjects: Algorithms; Assemblies; Automobile industry; Automotive engineering; Bolted joints; Clusters; Configurations; Design; Design optimization; Exact sciences and technology; Hypercubes; Mathematics; Preliminary designs; Probability and statistics; Robustness (mathematics); Sampling theory, sample surveys; Sciences and techniques of general use; Statistics; Bolted joint; Preliminary design; Design process; Motor car; Uncertain system; Latin hypercube sampling; Process planning; Automobile industry; Duration; Risk management; Optimization
• ISBN: 9781845641856; 184564185X
• ISSN: 1746-4498; 1743-3509
• DOI: 10.2495/OP090061

In the automotive industry, the optimization of the preliminary stages is a major asset for reducing the time of the design process. So, a new approach for setting up the optimal configuration of a bolted joints assembly has been developed. The aim is to efficiently fix the main characteristics of bolted joint assemblies in a particularly uncertain context. Indeed, in preliminary design, choices are made with little information and many indeterminations, but these should be relevant whatever the following steps of the design process may be. The methodology developed is based on simplified models included in a strategy consisting of testing configurations to choose the best one. In order to scan the design space well, an optimized Latin hypercube strategy is used. Then, algorithms have been developed to find clusters of good items and of unacceptable ones. These clusters are then used to specify a robust solution or to avoid incorrect configurations. An application was conducted on an automotive mounting. The reduced time needed to obtain useful results really contributes to shortening the design process duration and the approach allows one to control the risks of the process uncertainties.