An ABC-BP-ANN algorithm for semi-active control for Magnetorheological damper

Abstract

The Magnetorheological (MR) damper is one of the most popular semi-active devices, which uses MR fluids to produce controllable dampers. In this work, the Back-propagation (BP) Artificial Neural Network (ANN) optimized by the Artificial Bee Colony (ABC) algorithm (ABC-BP-ANN) is proposed to obtain the required voltage for semi-active control of MR damper simulated by Spencer model. It is found that the control-forces of MR damper are close to the results of active control algorithms such as the conventional Linear Quadratic Regulator (LQR) control algorithm. The initial weights and the thresholds of BP-ANN are regarded as solutions; the training errors of BP-ANN are used for the cost function and ABC algorithm is used to optimize the initial weights and the thresholds of BP-ANN. The proposed model uses the Spencer model to calculate the train samples to train proposed ABC-BP-ANN model. The proposed ABC-BP-ANN model predicts the voltage based on the results of control-force calculated by LQR model. Several numerical examples are used to verify the proposed model. The results show that the control-forces of MR damper calculated by proposed model are close to those calculated by LQR algorithm. The proposed ABC-BP-ANN inversion algorithm for obtaining the voltage for MR damper has greater computational efficiency and higher accuracy than the conventional BP-ANN algorithm.