Research On Control Technology Of Active Suspension In Underground Vehicle

Trackless rubber tire vehicle is a kind of common instrument for carrying materials and staffs. As it’s flexible, convenient, efficient and safe, it’s being used more and more widely. Trackless rubber tire vehicle with active suspension can adapt the poor road conditions in coal mines better and improve the comfortableness of passengers. The active suspension can generate a suspension control force according to the road condition. The suspension stiffness can be regulated dynamically and the vehicle rideability can be improved. Now many researches have been done on the Control technology of active suspension, but matured product hasn’t been made inland.Two kinds of active suspension controller have been designed. One is the LQG controller which is matured and dominant with lots of research. The other is neural network-PID (NN-PID) controller which hasn’t been used on active suspension. As the road signal is random, it can be denoted by a Gaussian white noise with certain power, the LQG controller has certain advantages at active suspension controlling.As the suspension system is a nonlinear system, and it’s hard to be described. Former active suspension system can’t describe the actual work condition accurately. There is considerable deviation between experimental results and practical applications. A new kind of active suspension controller based on neural network-PID theory has been designed by this thesis. It owns good robustness. Adjusting PID controller parameters is also realizable by online studying. The simulation experiment shows that the active suspension controlled by NN-PID theory has better effects than the other active suspension commanded by a LQG controller.This thesis mainly carried on the following work:(1) Basic knowledge of trackless rubber tire vehicle and detailed description about suspension system including structure, effect and classifications have been introduced. The development history of passive suspension and active suspension are introduced in detail, and the technical specialties, differences are introduced as a main content of the thesis. The1/4vehicle dynamic model based on passive and active suspension has been established respectively. The dissertation introduced the suspension performance estimate manners. At last, the paper introduced the road model and the power spectrum of the road.(2) The dissertation made research into LQG(Linear quadratic Gaussian), the performance of the suspension can be expressed with quadratic functional differential equation. The degrees of parameters’importance can be adjusted by weighting coefficients. The control matrix K is solved by Riccati equation. Finally, it listed the goal and mission of this dissertation.(3) The dissertation analyzed the neural network-PID control theory including its study method. Adaptive neural network-PID controller combines the neural network controller and normal PID controller, and possesses the both advantages of the two controllers. First it possesses the adaptive ability, Second, it has simple structure and it is easier to be modulated. The NN-PID controller is good at expressing no linear structure and setting the parameters automatically. Also, this dissertation designed a NN-PID controller based on1/4vehicle dynamic model.(4) The vehicle model based on active suspension and passive suspension2-DOF model is analyzed with Matlab/Simulink software.Two kinds of active suspension controllers has been built, and simulation experiment about active suspension has been done with Matlab/Simulink software. The body acceleration, the suspension deflection and the tire deflection is been taken as assessment indexes. According to the experimental results, the active suspensions are better than the passive suspension, the active suspension controlled by NN-PID theory is better between active suspensions.