Study Of Neural Network Control For Automobile Engine Vibration Isolating System Based On The Magneto-rheological Damper

The engine is one of car vibration and main noise source, whose vibration may produce a number of different complex vibrations with different vibration source and different vibration modes, after mixing so that the engine vibration has the characteristics of the broadband, multi-vibration source and multi-master frequency. At present rubber mount and passive hydraulic mount are mainly used to be vibration isolation and noise reduction. However, it's difficult to fulfill requirements about vibration isolation and noise reduction on the broadband. In a certain condition of the engine vibration it only can play a good vibration isolation effect. But it can't fulfill changes of condition. According to the changes of engine condition, the neural network semi-active control which is based on magnetorheological fluid mount, which could adjust damper force of magnetorheological fluid in engine mount initiatively in accordance with the set of control objectives. To achieve the best vibration isolation and noise reduction effect, it is necessary to research control method based on neural network for the engine vibration isolation.In this paper, on the basis of analyze the structure of magnetorheological (referred to as MR) damper and its mechanical property, the damping force theoretical model of MR damper is established. According to mount's installed technical requirements of diesel engine, magnetorheological fluid controllable suspension based on compression mode is designed. Magnetorheological mount is tested at the national passenger car quality inspection center and the test result is compared with numerical simulation results of MR damping force calculation model. After analyzing the driving force of inline four-cylinder diesel engine, engine vibration isolation control system dynamic model is established and neural network PID control method is brought up. Finally, in different conditions, the engine vibration control model of magnetorheological mount and rubber mount based on PID control and neural network PID control is simulated. And vibration isolation effect of different control methods is analyzed.Through the studies, in the condition of the excited speed is the same as excited frequency, we can obtain that the area surrounded by MR damper indicator curve is increased with increasing coil exciting current intensity and the damping force is gradually increased. But the nonlinear relationship is existed among the cycling dissipated power, excited frequency and exciting current. At the same time, making the displacement and speed of engine vibration as assessment indicators of engine mount vibration isolation effect is reasonable. Finally, on the basis of established MR mount vibration isolation system's dynamics model, the neural network PID controller is designed. And engine vibration isolation simulation system is established by use of MATLAB software and corresponding toolbox, which is based on MR mount. In different conditions, numerical simulation analysis is made. The results show that vibration isolation effect based on neural network PID control is better than MR mount based on PID control and rubber mount.