Optimizing truss members using Machine Learning

Abstract

This paper aims to produce an optimization methodology for a truss for minimal stress on critical sections using Machine learning models. This is done by varying the height and length of the truss span and applying live loads on different joints. The optimization of the truss is done through feed forward with back propagation artificial neural network model, The optimized truss is modelled and analysed using finite element and stiffness software. The paper presents the results and identifies the best dimensions for the truss design with a safety factor.