Neural network modeling of Magnetorheological fluid damper for automobile semi-active suspension

Automotive semi-active suspension based on Magnetorheological damping techniques have received significant attention in recent years, because they offer the adaptability of active suspension without requiring the large power resources. This paper presents a. multi-layer perceptron neural network with 3 input neurons, 1 output neuron and 7 neurons in the hidden layer to simulate the dynamic behavior of automotive Magnetorheological fluid damper. Training of neural network model. have been done by Gauss-Newton based on Levenberg-Marquardt method using data generated from test in laboratory. In comparison with experimental results of Magnetorheological fluid damper, the neural network models are reasonably accurate over a wide range of operating conditions.