基于神经网络的轮式装载机防滑差速器的控制研究

发布时间：2018-01-18 09:59

本文关键词：基于神经网络的轮式装载机防滑差速器的控制研究　出处：《浙江理工大学》2012年硕士论文　论文类型：学位论文

【摘要】：轮式装载机工作环境比较恶劣，经常由于车轮打滑引起车辆驱动性能和通过性能下降，影响作业进度和效率。国内工程车辆差速器多采用普通差速器或摩擦片式被动限滑差速器。主动限滑差速器在汽车上应用较多，工程车辆上没有应用。本文结合国内外限滑差速器的发展状况，分析当前常用限滑差速器的结构和缺陷。以ZL50装载机为研究对象，对比常用的普通差速器和摩擦片式被动限滑差速器的结构和工作原理，提出结构改进方案，设计了一种新型的电磁式主动限滑差速器，完成结构选型，传力过程分析。 ZL50装载机行驶路况复杂，针对电磁式限滑差速器控制器的输入和输出呈非线性，设计BP神经网络模型，确定了网络拓扑结构，建立带反馈的神经网络。基于BP神经网络的PID控制器对电磁式限滑差速器的控制机构电磁阀进行控制。确定初始权值和阈值，BP神经网络的输出值作为PID控制器的三个参数实现控制，将误差信号从输出端回送，反向传播，修正加权系数。根据路面输入情况调整，反复迭代，自适应得到控制参数。对ZL50装载机动力传动系统进行分析和建模，对驱动系统组成部件分别建模。建立发动机模型，液力变矩器模型，探讨发动机和液力变矩器的共同工作输入特性、输出特性，，建立变速器模型，驱动桥模型和轮胎模型。分析直线行驶和转向行驶的驱动过程。将差速器从驱动系统中独立出来，选用基于BP神经网络的PID控制算法，控制电磁阀推杆作用于液压油缸的压力，最终实现对限滑力矩的控制。在Matlab中设计程序，三种对分路面上直线行驶情况进行仿真，对比分析有无装用限滑差速器的仿真结果。进行直线行驶和转向行驶工况的仿真，对比分析装用普通差速器、摩擦片被动限滑差速器、电磁式主动限滑差速器三种差速器的仿真结果。仿真结果可知：基于BP神经网络的PID控制的电磁式限滑差速器，克服了普通差速器平均分配扭矩的缺点，对附着系数对分的路面具有良好的自适应性，两侧路面附着系数相差越大，限滑效果越明显。低附着区域能够实现较好的限滑作用，控制轮胎的滑转率，使装载机具有较高的驱动力和侧向附着力，保证车辆的动力性能。仿真结果证明所选用的电磁式限滑差速器的结构设计、控制方法的合理性和适用性。
[Abstract]:The working environment of the wheel loader is very bad, and the driving performance and passing performance of the vehicle are often decreased because of the skid of the wheel. The common differential or friction plate passive slip limiting differential is mostly used in the domestic engineering vehicle differential. The active limiting slip differential is widely used in the automobile, but it is not used in the engineering vehicle. This paper analyzes the structure and defects of the current limited slip differential based on the development of the limited slip differential at home and abroad. The ZL50 loader is taken as the research object. Comparing the structure and working principle of common differential and friction plate passive limited slip differential, the structure improvement scheme is put forward, and a new type electromagnetic active limiting slip differential is designed to complete the structure selection. Analysis of force transfer process. The driving condition of ZL50 loader is complex. Aiming at the nonlinear input and output of electromagnetic limited slip differential controller, the BP neural network model is designed and the network topology is determined. The PID controller based on BP neural network is used to control the solenoid valve of the electromagnetic limited slip differential. The initial weight and threshold are determined. The output value of BP neural network is used as the three parameters of PID controller to realize the control. The error signal is sent back from the output end, and the weighting coefficient is revised. Adaptive control parameters are obtained. The power transmission system of ZL50 loader is analyzed and modeled. The components of the drive system are modeled respectively. The engine model and the torque converter model are established. The working input and output characteristics of engine and hydraulic torque converter are discussed. The transmission model, drive axle model and tire model are established, and the driving process of straight line driving and steering driving is analyzed. The differential is independent from the drive system and the PID control algorithm based on BP neural network is selected to control the pressure of the solenoid valve push rod acting on the hydraulic cylinder. Finally, the control of the limit slip torque is realized. In Matlab, three kinds of simulation programs are designed to simulate the straight line driving on the sub-pavement. Compare and analyze the simulation results of the limited slip differential with or without the installation. Carry on the simulation of the straight line driving and steering driving condition, compare and analyze the ordinary differential and the friction plate passive limit slip differential. Simulation results of three kinds of electromagnetic active limiting slip differential. The simulation results show that the electromagnetic slip limiting differential based on BP neural network PID control overcomes the shortcomings of average torque distribution of the ordinary differential and has good adaptability to the road surface with the adhesion coefficient. The higher the adhesion coefficient of both sides of the pavement, the more obvious the slip limiting effect. The lower the adhesion area is, the better the slip limiting effect can be, and the higher the wheel slip rate can be, the higher the driving force and the lateral adhesion of the loader can be. The simulation results show the structure design of the electromagnetic limited slip differential, the rationality and applicability of the control method.
【学位授予单位】：浙江理工大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：TH243;TH132.46

【相似文献】