Magnetorheological Semi-active Suspension Control Strategy With Grey Prediction And Fuzzy For D Class Vehicle

In recent years, the rapid development of electronic control technologies makes itpossible to meet people’s requirements on vehicle riding comfort and handling stability.With its simple structure, cheap price, being able to improve vehicle ride comfort andhandling stability, force adjustable damping semi-active suspension has gained extensiveattentions among researchers. At the same time, researches on control strategies and how toeliminate the control lagging have become hot issues in the discipline of electroniccontrolled suspension.This paper is based on a research and development project of adaptive electroniccontrolled suspension. The purpose is to study motion prediction method and damping forcecontrol strategy for magneto-rheological semi-active suspension. The front axle sprung massmotions prediction algorithm based upon grey prediction, rear sprung mass motionsprediction algorithm based on the pre aim between axles, pitch and roll motions predictionalgorithm based on neural network are designed respectively. On this basis, an adaptivevariable domain fuzzy damping force control algorithm for magneto-rheological semi-activesuspension is designed. A CarSim/Simulink simulation platform for off-line simulation and adSPACE hardware-in-the-loop test platform for real time test of are set up respectively toverify the control algorithm. At last, offline simulation and hardware-in-the-loop test dataunder various typical working conditions are analyzed to evaluate the effectiveness of thecontrol algorithm. The specific content includes:1) A Simulink model of road surface incentive is built on the basis of the road elevationmathematical model and discrete pulse road surface elevation incentives mathematicalmodel constructed through filtered random white noise method. It generates thecorresponding road surface pavement excitation data. The magneto-rheological damperdeveloped by our research group is test on the test bench based on MTS850to obtain itspower characteristics and speed characteristics. And the magneto-rheological damperpolynomial model and its inverse model are established according to the test data. Theaccuracy of the models can reach more than99.9%.2) The grey prediction algorithm aiming at predicting motions of the front axle sprung,the pre aim algorithm between two axles to predict rear sprung mass motions, the neuralnetwork algorithm to predict pitch and roll motions are designed. Besides, an adaptive variable domain fuzzy damping force control algorithm is proposed. A Simulink semi-activesuspension control model is developed. It is divided into riding comfort mode and handlingstability mode combined with the prediction algorithm and the adaptive variable domainfuzzy control algorithm.3) A simulation platform including the CarSim vehicle model, the road surfaceexcitation model, the semi-active suspension state of grey prediction fuzzy control model isbuilt. The semi-active suspension state of grey prediction fuzzy control model is test throughrandom roughness, vertical impulse input, emergency braking, the step steering and thesteer-braking five typical working conditions. Simulation results show that semi-activesuspensions with state of the grey prediction fuzzy controller obtain better riding comfortand handling stability than passive suspension.4) Proper hardware equipment is selected and the corresponding software interface isestablished according to the electrical characteristics. A software model is built and adSPACE real time simulation system is set up accordingly as the hardware-in-the-loop (HIL)test bench. The fuzzy controller based-on state of grey prediction of the semi-activesuspension is tested under five typical working conditions. Data from HIL tests shows thatcontrol algorithm is able to improve vehicle riding comfort and stability obviously.