Optimization of Composite Shaft Subjected to Torsional Loading by Using Genetic Algorithm and Particle Swarm Optimization Methods

Optimization of composite shafts subjected to torsional loading has been investigated in the previous studies by considering the constant value of load and so, minimization of shaft mass was defined as the objective function (OF). In the current study, maximization of the torque to mass (T/m) ratio was considered as OF. To do so, the Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) methods were utilized. The number of layers, thickness and angle of each ply as well as the applied torque were considered as the input variables. Moreover, preventing of failure in composite shaft, based on Tsai-Wu failure theory developed in Abaqus finite element software, was defined as the constraint of optimization problem. Also, in order to investigate the effect of OF type, in addition to the T/m, the mass was also defined as OF in a separate optimization problem. The results revealed that despite PSO, GA had suitable convergence in the optimization. Moreover, in spite of the type of OF, using a composite shaft compared to the steel one, had at least 80% mass reduction. Furthermore, although the predicted composite shaft via T/m OF has more mass compared to that predicted via m OF, it can tolerate torsional loading up to 8.5 times more. This point can increase the load carrying capacity of composite shaft.