Research On Vehicle Semi-Active Suspension System Based On Road Estimation

The concept of controllable suspension system has been of great concern in recent decades. As an important part of controllable suspension system, semi-active suspension system has been successfully applied in high end passenger vehicle for its remarkable merits, e.g. low energy consumption, easy to be built-in, and similar ride comfort performance compared with active suspension.Essentially, by taking ride comfort and road handling as objective functions, and rattle space as constraint, semi-active suspension system can be regarded as multi-objectives optimization problem. In this case, the chosen of control objective function for different semi-active suspension control strategies recently attracts more attention. To solve this problem, this thesis sums puzzles related to objective function up in 4 questions, and performs road estimation based semi-active suspension control strategies studying. Answers to these questions are finially given. The detailed content can be concluded as:To solve questions related to the essential contradiction of suspension system and performance comparison, five most representative control strategies are compared. Corrected delay mechanism of controllable damper is then proposed, and system performance with delay time of 0ms, 5ms, 10 ms and 15 ms is compared. The concept and principle of road estimation based semi-active suspension system are proposed.Two new road estimation methods are then proposed to estimate road profile in time domain and in statistical way to remedy the problems existing in current methods. The method of road estimation in the time domain transforms the problem into inverse system identification, realizing precise estimation based on system response. To reduce the influence of unavoidable noise, a new method called ANFIS-GMDH is then applied. As for road estimation in statistical way, according to the difference in both time and frequency domain, a classifier based on ANFIS and wavelet transform is proposed, and simulation and experiments prove its accuracy. Proposed methods provide a solid foundation for application of semi-active control strategies.To solve the question about how to combine road estimation and semi-active suspension, road estimation based hybrid control and model predictive control for semi-active system are proposed. The performance of hybrid suspension system is firstly transformed into analytical expression and solved with NSGA-II, and control parameters are selected according to road statistic information. With consideration of state depended constraints in semi-active suspension system and road profile information, hybrid MPC is implemented in both quarter and half vehicle model. This thesis realizes a combination of road estimation and semi-active control strategies.