Modeling of a magnetorheological valve design based on artificial neural networks for heavy equipment cabin application.

Abstract

The artificial neural network (ANN) model can be one of the efficient and accurate models that can be employed in magnetorheological (MR) valve design processes. ANN can be used to assist the process together with finite element method simulation that can predict the desired magnetic field as a function of geometrical sizes quickly and accurately. Therefore, this research aims to develop an MR valve model based on ANN to assist the device application in heavy equipment cabin suspension systems. ANN will predict the value of magnetic flux density as a function of a certain geometrical parameter. Meanwhile, the output data is defined by the value of the magnetic flux density in each zone of the magnetorheological (MR) valve meandering flow path type. ANN training will use the Adam optimization algorithm. The model will be used to calculate the damping force of an MR valve meandering flow path type. The ANN modeling results from the R-squared (R2) value are more than 0.991 for all output zones. Thus, the chosen ANN modeling is considered to be able to accurately predict the value of magnetic flux density in each zone. From the damping force calculation results, there are five variations of the MR valve meandering flow path type design that can be used for heavy equipment cabin suspension systems with a maximum damping force of 5.5 KN.