节点文献

基于开放式数控系统的磨料水射流特种精密加工技术

Technology of Abrasive Water Jet Precise Machining Based on Open Numerical Control System

【作者】 贾强

【导师】 雷玉勇;

【作者基本信息】 西华大学 , 机械制造及其自动化, 2006, 硕士

【摘要】 本文在详细研究水射流精密加工工艺特点的基础上,基于人工神经网络理论,建立了水射流精密加工的神经网络模型:应用Visual Basic语言开发界面,通过COM组件将MATAB作为工程计算的核心,实现神经网络的计算、数据分析和仿真。简化了人工神经网络算法的编程,同时为磨料水射流加工模型与水射流精密加工数控系统的集成提供基础。 本文选用PMAC(Programmable Multi-Axis Controller)可编程多轴控制器(PMAC内置PLC功能)作为CNC模块,PC机为系统支撑单元的双CPU开放式数控系统,即以“PC十运动控制卡”方案为硬件平台。采用COM组件技术,用VB语言编制了人机界面、系统管理与调度来实现磨料水射流精密加工数控系统。 本文试图建立一种适合于磨料水射流精密加工的开放式数控系统。论文主要包括以下内容: 1、基于人工神经网络理论,建立磨料水射流精密加工人工神经网络模型,为磨料水流精密加工数控系统提供进给速度预测。 2、应用Visual Basic语言和数据库技术及COM技术,开发人工神经网络运算模块。用MATLAB的COM生成器将训练神经网络的M函数做成COM组件,它作为VB的上层服务器供VB(Visual Basic)调用。通过COM组件将两种语言扬长避短结合起来,实现了MATLAB与VB应用程序

【Abstract】 Abrasive water jet (AWJ) presice machining technology is studied detailedly. A neural network model of abrasive water jet presice machining is established based on artificial neural network (ANN). The visualization interface for computation of neural network is developed. As a calculation unit, the COM component of MATLAB is integrated into the software. Therefor, computation and data analysis as well as simulation of abrasive water jet precise machining can be realized. It simplifies programming the algorithm of artificial neural network. Also, it provides the foundation for integration of ANN model of AWJ with computer numerical control (CNC) system.The open CNC system with double CPU based on PC and PMAC (Programmable Multi-Axis Controller) module is put forward. The system has "PC+PMAC" module. The numerical control system for abrasive water jet machining is developed using Visual Basic and COM technology.The following work is included in the thesis:1. An ANN model for abrasive water jet based on artificial neural network theory was established. The feed speed for the abrasive water jet presice machining can be predicted precisely through this model.2. A module for computation of ANN algorithm based on database, COM component of MATLAB and Visual Basic was developed. M function for training neural network is a component that is generated by COM Builder of MATLAB. Thereafter it can be called as an upper server by Visual Basic.Seamless integration of Visual Basic and MATALB are realized through COM component and M function. Consequently the calculation of artificial neural network is complished.3. The sampling interpolation is employed and studied in the open numerical control system for abrasive water jet machining. An open NC command system based on PMAC is studied. System parameters setting and programming for PMAC are carried out.4. Based on the advanced software developing method, the primary COM-CNC software platform is designed using object-oriented analysis and behavior-based model.

  • 【网络出版投稿人】 西华大学
  • 【网络出版年期】2006年 09期
  • 【分类号】TG596
  • 【被引频次】3
  • 【下载频次】300
节点文献中: 

本文链接的文献网络图示:

本文的引文网络