Optimization of Parameters of Electromagnetic System with Magnetorheological Fluid

Magnetorheological (MR) devices, which are easily controllable, fast-acting, and integrable into "smart" systems, are utilized across various industries. Optimizing these devices using machine learning methods is an important and relevant issue. The structure and the operating principle of an electromagnetic brake using MR fluid are described, and a hybrid mathematical model of the system's magnetic circuit is presented. Optimization problems for the parameters of the electromagnetic system using MR fluid have been formulated, aiming to find the maximum values of the electromagnetic force acting on the bridge front of the MR fluid's magnetic particles and the displacement speed of the bridge front. Based on the hybrid mathematical model, the inverse problems of the magnetic circuit have been solved within the value ranges of the optimization variables, determining the magnetomotive force of the control coil for a known magnetic flux value and calculating the electromagnetic force and movement speed. The results of solving these problems have been recorded in a database, which has been trained by using machine-learning algorithms in MATLAB. A simple genetic algorithm has been applied to the trained database. One-objective and bi-objective optimization problems have been solved. The proposed optimization method can be successfully applied to other MR devices.