DOE for product and process optimization, robustness, and tolerance design: 40mm IR illumination grenade

摘要

This paper illustrates how empirical data generated from a systematically designed experiment is used for armament product performance optimization, robustness, and tolerance design, in the context of a 40mm IR flare munition project conducted by the US Army ARDEC. This presentation will introduce modern Robust Design best-practices in a seamless statistical analysis `walkthrough' leaving opportunity for drilldown and audience interaction, and will contrast with Taguchi's approach which many Quality and Reliability professionals are familiar with. This approach uses a number of `Uncertainty Quantification' (UQ) techniques more commonly applied to computer models and simulations. This approach makes use of the validated prediction model built using Response Surface Methodology (RSM), specifically a computer-generated I-optimal experiment, to analytically derive a robust design solution using Propagation of Error (POE), and integrates Monte-Carlo Simulation, Reliability-Based Design Optimization (RBDO), and statistical quality control methods to execute an efficient, cost-effective manufacturing Tolerance Design study.