Machine Learning in Design of Optimization of MR Actuators: A New View of Solution

In this study, a new progress of optimization of smart structure is carried out. The application of machine learning through neural networks brings a new view in design. The data using for optimization is collected from the simulation results, the real values of the closed structure with the proposed model. The combination of these data and the outstanding property of the neural networks helps improve the obtained variables, and then performance of the proposed structure is also improved. The proposed structure is built based on cantilever beam model. Smart material is covered outside of the beam based on magnetorheological elastomer (MRE in short). This model is used for evaluate the performance of the neural networks in optimization. Results of the progress show that this method can use in case of no specialize software, and satisfy the initial requirements.