Laser Peening for Reliable Fatigue Life. Delivery Order 0025: Volume 1 - Simulation and Optimization of a Laser Peening Process

摘要

We develop a finite element simulation of the 3D Laser Peening (LP) process and validate it with experimental results available from the literature. The simulation model is then used for parametric studies of the LP process including pressure pulse properties, spot properties, shot number, shot locations, shot sequencing, shot overlap configurations, and various geometries. The models were optimized via Mixed-variable Niche Particle Swarm Optimization (MNPSO), a tool developed for this work. MNPSO was implemented within the framework of a Progressive Multi-Fidelity Optimization Strategy (PMOS), which initially conducts simulations using low-fidelity models to identify a reduced design space. Two stages of low-fidelity design space reduction are conducted before high-fidelity models are implemented in a third stage that finalizes the optimization. This method has been demonstrated on several models ranging from a parametric quarter model to a basic bending coupon model. The effects of varying key parameters upon the residual stress field were identified for Ti-6Al-4V. The niching algorithm was implemented, and the results showed multiple 'optimal' results for a given Laser Peened surface. This method allows the designers to utilize their own preferences in determining which 'optimal' peening condition to use.