Study On Vibration Damping Performance Of Automobile Suspension Under Intelligent Control

As an important part of automobile chassis, suspension system has an significant effect on ride comfort and handling stability. According to road shocks and motion signals from vehicle body,semi-active suspension system can adjust suspension damping and other parameters to improve itself cushion performance, which shows a great advantage over passive suspension with unchangeable parameters. Structure of semi-active suspension is more complex correspondingly. It’s difficult to achieve satisfied control effect by conventional control methods. Thus,it attaches great importance to studying on intelligent control of semi-active suspension.This paper put forward evaluation indexes of automobile suspension performance firstly. And road input model of random filtering white noise was derivated and established in the time domain. Based on a few reasonable assumptions about some properties of automobile driving, Each suspension(passive or semi-active) is respectively corresponding to a half vehicle dynamics model with 4 DOFs. Secondly, the research of intelligent control methods was conducted. In combination with fuzzy logic theory, it takes several steps to construct fuzzy inference system, including fuzzifying, fuzzy inference and defuzzifying. Hence a closed loop fuzzy control system of semi-active suspension and its simulation model were built. Then, in order to improve effect of fuzzy control, two kinds of neuro-fuzzy inference system which was based on Mamdani model or TSK model were designed as core controllers. Each controller was trained to gain fuzzy rules through BP algorithm. Respectively with S-function in Simulink or ANFIS toolkit from MATLAB software itself, two types of semi-active suspension system model under fuzzy neural network control were achieved. Finally, this paper has completed the simulation experiment in Simulink for the passive suspension system and semi-active suspension systems under fuzzy control and fuzzy neural network control, obtaining the simulation results.Through analysis of RMS of body acceleration, working space and dynamic tyre load from the results, it can be inferred that comprehensive performance of semi-active suspension system under fuzzy control is better than that of passive suspension system; comprehensive performance of semi-active suspension system under fuzzy neural network control is far superior to that of semi-active suspension system under fuzzy control; comprehensive performance of semi-active suspension system based on TS model under fuzzy neural network control is the best. Therefore, the intelligent control methods used in this thesis can improve performance of semi-active suspension system effectively.