【作者】 姜鹏鹏；

【导师】 吕广明；

【作者基本信息】 哈尔滨工业大学 ， 机械设计及理论， 2006， 硕士

【摘要】 反铲单斗液压挖掘机是一种应用十分广泛的工程机械,在基础设施建设中的作用非常重大。然而由于其控制方式一直没有多大的改进,主要由熟练工人操作控制,因此不仅对操作员来讲劳动强度大,而且在挖掘特定轨迹时精度不够。为了扭转这种局面,引入自动控制机构是势在必行的。本文旨在探讨一种行之有效的自动控制方法,用以对反铲单斗液压挖掘机的挖掘轨迹进行控制,以期能解决上述问题。为了达到这一目标,本文建立了液压挖掘机系统的数学模型,在控制方法上首先采用传统数字PID控制,并利用Matlab7.1对系统的时域和频域特性做了详细的分析,发现加入PID后,在保持系统稳定的基础上,比原系统的截止频带变宽,响应速度明显加快,跟随输入参数变化的能力变强。然后采用了现代智能控制方法中的一种——BP神经网络控制,采用直接逆控制的形式。对引入BP神经网络控制的系统进行计算机的数字仿真。得出,BP神经网络比PID对系统的控制效果更好,表现在阶跃响应时间较短,跟随输入参数变化的能力强,有很强的自适应能力和鲁棒性。为了给实践提供充分的理论指导,本文还利用VC++编程语言对液压挖掘机的工作装置进行了机构仿真,并将两种控制方法对液压挖掘机工作装置的直线挖掘轨迹的控制效果进行了比较。从结果来看,由BP神经网络对挖掘机的控制取得了很好的挖掘效果。在进行精确的理论设计与计算之后,采用PC机做为控制中心,利用VC编程语言编写控制软件,将数字PID控制应用到实验设备中,调试运行。光电编码器采集到的角度信息,经坐标计算得到挖掘机铲斗端点的坐标,将这些连续的坐标描点,即得到一条实际的直线挖掘轨迹。更多还原

【Abstract】 Hydraulic excavator with an adverse bucket is a kind of machinery which is used widely. It is very important in basic construction.But because its method of control is traditional which is controled by a skillful operator, this is hard to the operators and some special effects are not exact enough. The automatic controller is needed.The text is to explore a method of control to control the trace of the hydraulic excavator,and it can also be expected to solve the problem.This article deals with problems as following: It deducted the methematic model of the system.It first selected PID controller which is triditional and analysed the time domain and the frequency domain character of the system. It got that the system held its stability and the time domain and the frequency domain character of the system were better than primary system. Then it used BP neural network which is one of the modern intelligent control methods,and the model of the neural network direct negtive control. It did digital simulation to the controll system, and the results were compared with the PID control system. It got that the result was better which was the response time was shorter, the capability of tracking the input changing and the self adapting was stronger. To supply the sufficiency theory guide to the experience, it emulated the the device of excavator using VC++, and compared the trace. From the results, we can get that the effect controlled by BP neural network is satisfied.After the precise theoretical design and caculation have done, the methods of PID control have applicated in the experimental installation, which used PC as the the central processing unit and compiled the control software by using VC. Then the installation was debuged and run. The angle message from the sensor was conversed through the conversion of coordinates, and the termination of the bucket was obtained. After the continuous points were described in the coordinate system, the practical straight-line excavate tracer came.更多还原