神经网络算法在电梯导轨校直系统中的应用研究
时间:2018-07-28 14:37 来源:毕业论文 作者:毕业论文 点击:次
| 摘要电梯是现代化高层建筑中必不可少的自动化装置，电梯导轨的直线度是高速电梯运行稳定性的重要影响因素。为了能够使电梯导轨的精度达到一定的要求，直线度检测和校直是导轨出厂前必不可少的环节。为了提高生产效率和导轨的校直精度，把人工神经网络引入到已开发的电梯导轨校直系统软硬件平台中，通过学习大量成功经验的校直压点样本，使系统能够准确地找到压点的弯曲值、支点和压点之间的距离与校直行程的映射关系。26361 本文首先简要概述了已开发的电梯导轨校直系统软硬件平台，接着介绍了神经网络的基础理论知识、LM算法及步骤，然后利用MFC编程实现了神经网络对过压量的计算和预测，重点是在MFC程序中实现样本训练和仿真的过程以及神经网络算法在电梯导轨校直系统中的应用过程，最后又针对特定的情况对神经网络的应用进行了拓展。经过改进的算法有效的提高了导轨的校直精度，同时也提高了导轨的生产效率。 关键词 神经网络 LM算法 导轨校直 MFC编程 毕业论文设计说明书外文摘要 Title Application of neural network algorithm in elevator guide rail straightening system Abstract The elevator is an essential automation equipment of modern high-rise building, and the straightness of the elevator guide rail is an important factor in the stability of the high-speed elevator. In order to make the elevator guide rail achieve the accuracy requirements, the guide rails are generally required to go through the straightness detection and straightening before leaving the factory. In order to improve the production efficiency and the accuracy of guide rail, the artificial neural network has been introduced into the hardware and software platforms of the developed elevator guide rail straightening system. By learning a lot of straightening pressure point samples from successful experiences, the system can accurately find the mapping relation among the bend value of the pressure point, the distance between the supporting points and the pressure point and straightening stroke. This paper begins with a brief overview of the hardware and software platforms of the developed elevator guide rail straightening system, Secondly this paper introduces the basic theory of neural network, LM algorithm and steps. Then this paper describes how to use neural networks based on MFC programming to calculate and predict the overvoltage. This paper focuses on realizing the training and simulation processes of the samples based on MFC program and introduces how to apply neural network algorithm into the elevator guide rail straightening system. Finally this paper introduces the application of neural networks were expanded for a particular case. The improved algorithm can not only effectively improve the guide rail accuracy but also increase the production efficiency. Keywords neural network LM algorithm guide rail straightening MFC programming 目 次 1 绪论 1 1.1 课题研究的目的和意义 1 1.2 相关校直理论的研究现状 1 1.3 课题研究的主要内容 2 2 导轨校直系统整体过程设计 4 2.1 系统总体方案的设计 4 2.2 电梯导轨校直过程的设计 5 2.2.1 电梯导轨直线度误差介绍 5 2.2.2 导轨校直原理介绍 6 2.2.3 导轨校直过程设计 7 2.3 基于数据库的过压量设计 8 2.4 基于神经网络的过压量设计 10 (责任编辑：qin) |