Neural Network Control Simulation On Semi-active Air Spring Suspension

Along with the development of high way construction in china, the demand for thevehicle's handling stability and ride comfort is increasing. At the same time, the upkeep costs ofthe road has become the economic burden of the government, so, how to decrease thedamage to the road surface becomes a serious problem for the automobile industry.The unchangeable parameters of passive suspension limit the improvement of the automobileperformance. The semi-active suspension, which can adjust the parameters ofsuspension actively according to the conditions of road surface and the vehicle,helps to improve the ride comfort and control stability, and also remarkablyminimize the damage to road surface.Because of the low natural frequency, the adjusted stiffness, the little vibration,the low noise and the long serviceable life, air spring suspension is a popularizationin passenger cars and heavy trucks. The function of air spring, especially the designof air spring suspension and control arithmetic and the matching with vehicle, isbeing researched in China. Electronic control air suspension system which makethe best of the adjusted stiffness can improve the automotive performances to agreat extent and decrease environmental damage.The thesis is the continuation of the research for the project of JiLin Science andtechnology's development "The research of Automobile electronic control system for air springsuspension". The main work is to research into control strategy and control system design forair spring suspension in order to improve the vehicle's handling stability and ride comfort to thehighest extent.1. Characteristic analysis of air spring suspension systemThe air spring suspension system consists of air spring,guided framework,electro-magnetic valve ,air compressor,damper,sensor,gas storage and so on.The air spring is composed of air spring bellow,piston and top plate. Air springshave strong nonlinear stiffness characteristic due to different pistons and air springbellow, which is a complicated vulcanized material of rubber and lining cloth. Airsprings can have perfect stiffness characteristic by using it's nonlinear stiffnesscharacteristic, which will improve the vehicle's performance. In addition, thestiffness characteristic of air spring has a long time-delay in response toelectro-magnetic valve. All these factors have influence on the control effect.The nonlinear stiffness characteristic of the air spring is due to rubber materialnonlinear,composite of rubber and lining cloth nonlinear,contact nonlinear andgeometry nonlinear. So air spring suspension is considered a nonlinear suspensionwith adjustable stiffness. When the load of vehicle changes, the stiffness values ofair spring can been changed by electro-magnetic valve. But the frequency ofautomobile body and the height of the car body will not been changed basically.This will improve the vehicle's ride comfort greatly. When the vehicle is in theconditions of sharp turn,acceleration and deceleration,the state of the car body canbeen controlled with the adjusting stiffness of air spring. This will also improve thevehicle's handling stability. In addition the weight of air suspension is light and thestiffness value of air spring is adjustable. When the vehicle is in the condition ofhigh speed, the friction between the tire and the ground will be increased and thebrake distance will be shortened. This will also improve the vehicle's handlingstability.When the pressure of air spring can varies with on/off time of electro-magneticvalve on the test bed, the relation curve between stiffness of air spring and on/offtime of electro-magnetic valve can be established through measuring the load anddeformation of air spring and the formula has been drawn up. Therefore thepreparation has been made for designing of controller.2. Neural network control of air suspensionThe published researches on the control algorithm of semi-active suspensionand active suspension almost have been concerned with the all branches of moderncontrol theories. Some control algorithms are applied in the suspension controlsystem, but most of them still rest on the academic research stage. neural networkcontrol method is original in the latter of the 20th century and the is quicklydeveloping. For its excellent control performance (nonlinear mapping, self-studyadjustability, and so on), it gives a new idea on the mathematics model and controlof nonlinear system, and comes to many specialists and scholars' notice, who putforward lots of successful theories and plays and perfect the theory of neuralnetwork control.This paper designs a neural network controller by use of Matlab/simulink andapplies the controller to the quarter-vehicle air suspension system.Neural network control regards the control system as a mapping between inputand output, and realize the mapping characteristic through the learning ability andapplicability of neural network, and then it can achieve very outstanding result ofthe control processes and accurate mathematics model.3. Verify the effect of neural network control algorithmTo verify the effect of neural network control algorithm used in this paper, a testsystem for the quarter-vehicle air suspension model has been designed according tothe performance requirements to the test of air spring suspension and practicalconditions. The electro-hydraulic servo actuator system of Test Center in JilinUniversity are applied to the design of the test system. The dynamic simulation andstiffness test of air suspension can be carried out in this test system. Many differentcontrol algorithms can be applied to the stiffness control of air suspension in thistest system.The results show the neural network control system used in this paper has agood control effect on the attenuation of perpendicular vibration of the sprung massand dynamic tyre load. Therefore it can improve the vehicle's ride comfort andhandling stability greatly.