Stochastic – based optimisation methods for finding optimum part orientation in additive manufacturing

Abstract

Additive Manufacturing has meant a revolution in the industrial world and its applications to different fields are increasing. However, within this technique, there is a problem that engineers have to face and which should be highlighted: part orientation. A component can be built up with different configurations and each one can offer different production performance, for example, one orientation can improve the cost of the part while others can worsen it. The problem resides in choosing the best part orientation that satisfies certain manufacturing parameters which are compromised each other. The objective of this project is to find an adequate part orientation in a problem proposed by Airbus Group Innovations (AGI). To achieve it, optimisation methods were employed, namely, a stochastic optimisation technique based on Stochastic Annealing (SA) which was implemented in a Matlab™ programme. Before facing the Airbus problem, a study on simpler examples was carried out to characterise the behaviour of the method. The results obtained were satisfactory and it was proven that the chosen method solves the problem in an efficient way. Initially, a basic method was developed and later it was updated with the incorporation of two improvements which noticeably enhanced the searching performance of the developed optimisation tool. The best outcomes were acquired with the improvement that involved adding a characteristic of an optimisation method called Particle Swarm Optimisation (PSO).