Regression models for magnetic flux density using DoE techniques and geometric optimization of MR valve

Abstract

Magnetorheological (MR) fluids have found wide application in engineering devices, especially, in MR dampers. Calculation of magnetic flux density in the MR valve plays a critical role in the performance analysis of MR dampers. In the present study, regression models have been fitted for magnetic flux densities in the active and core regions of the MR valve for a commercially available MR fluid. The fitted model is presented as a function of the MR valve’s geometric parameters and applied current. For fitting the regression model, simulation experiments have been performed in ANSYS-APDL software using different design of experiment (DoE) techniques. For each technique, the model is improved upon by p-value test and the best model among different DoE techniques is selected by analyzing the coefficients of determination. The best-improved models are then used for geometric optimization of MR valve in the MATLAB® environment and the optimal results are compared to the one obtained from the approximated magnetic circuit with constant relative magnetic permeability.