3-D Shape optimization of a Sensor Mounting Arm Using MOGA and MLF

In this paper, we present the multi-objective genetic algorithm(MOGA) to find an optimal structure of a sensor mounting arm using the finite element method(FEM). This mounting arm is used to install a weather-station on electrical tower to improve even forecast the transmission capacity through the existing high-voltage overhead lines. The multiple design parameters of this sensor mounting arm structure are determined as variables for the shape optimization. The structural performance of design points with diverse structures are simulated and gathered as the training data for Multi-layer feed-forward neural network(MLF). After training, this MLF model can mimic the mapping from design parameters to structural analysis and replace the time-consuming simulation. Given design constraints with the implementation of the algorithm MOGA, one tradeoff optimal structure is found. In comparison with the optimal structure obtained from Standard Response Surface (2
^nd
order polynomial) and Kriging Response Surface, the solution structure from MLF and MOGA is more light and shows a better static mechanical performance. This work is implemented within the FEM program Ansys without any interfaces with other programs. Additionally, the material selection and analyse of fatigue life for sensor mounting arm is also presented in this work.