Research On Neural Network Control Strategy Of Hydro-pneumatic Suspension System

At present,nine-axle all-terrain cranes are increasingly used in super-large industrial applications.Compared with other cranes,their advantages lie in chassis design technology.In particular,the representative research of interconnected oil-air suspension systems,the improvement of the suspension system itself has improved the performance of all-terrain cranes.However,only relying on passive suspension to improve the performance of the vehicle is limited.How to further improve the role of the overall crane hydro-pneumatic suspension has become the key to future research.The main purpose of this paper is to design an active hydro-pneumatic suspension that enables the all-terrain crane to adapt itself to various roads by autonomous adjustment.This article studies an active hydro-pneumatic suspension based on RBF neural network PID control strategy.The dynamic suspension can adjust the network output according to the road excitation to achieve learning,adapt to various road conditions,and improve the stability of the operation of the all-terrain crane.Sex and ride comfort.Picking up the all-terrain crane hydro-pneumatic suspension as the research object,the composition,installation,and working principle of the hydro-pneumatic suspension are analyzed.The mathematical model of the hydro-pneumatic suspension cylinder was formed.Through the MATLAB/simulink simulation software,the impact of road surface excitation on the hydro-pneumatic suspension was simulated.The detailed description is achieved through the changes of the gas volume and pressure of the accumulator,the suspension chamber and the stiffness and damping characteristics of the suspension.Based on RBF neural network PID control theory,an operational hydro-pneumatic suspension is designed.The RBF neural network is a local approximation network that can approximate any continuous function with arbitrary precision.And learning speed,there is not any local minimum problem.It is built on the gradient descent method to calculate the weight,node center and width of the output layer in the RBF neural network.Online adjustment of K_P,K_I,and K_D parameters to achieve dynamic and steady-state stability of all-terrain cranes.The interconnected hydro-pneumatic suspension was used as the controlled object.The MATLAB/simulink simulation software was used to simulate the vehicle body displacement,vehicle acceleration,and suspension dynamic travel with the RBF neural network PID control and the traditional PID control.The results prove that the controller integrated by the RBF neural network PID can attenuate the vibration of the all-terrain crane in a shorter time and reach a steady state.The vibration amplitude is also significantly lower than the conventional PID-controlled oil-air suspension.The content of the above research has evident reference value for the research of active hydro-pneumatic suspension.According to the vibration model and a mathematical model established by the hydro-pneumatic suspension,it provides an effective solution to the dynamic performance analysis of the all-terrain crane hydro-pneumatic suspension system.