基于神经网络的扣式电池自动化生产线故障诊断系统研究.pdf

河北T业大学
倾士学位论文
基于神经网络的扣式电池白动化生产线故障诊断系统硏究
姓名:闫少伟
申请学位级别:硕士
专业:机械制造及其自动化
指导教师:林树忠
2006110
河北工业大学硕上学位论文
基于神经网络的扣式电池自动化生产线故障诊断系统研究
摘要
自动化和智能化是现代化大生产的显著特点,近年来伴随着生产设备的大型化、高速
化,传统的诊断技术经远远不能适应了。现代化工业生产,一日因故障停机,损失巨大
因此人们对机械设备故障诊断的复杂性、准确性和及时性的要求也逐渐増加,这就迫使人
们在设备诊断方面进行了大量的研究,形成了故障诊断这一新兴的研究领域。
本文首先依据 Li/Mno2扣式电池组装工艺和机械自动化产品设计的系统观点,提出了
一种新型的直线型.i/MnO2扣式电池自动化生产线,并对它的机械系统和控制系统做了详
细地介绍
由于本自动化生产线是一个多输入多输出的复杂系统,作为核心部分,本论文提出了
种基于神经网络的故障检测和诊断方法。其原理是:采用VC++作为前台开发工具,SL
Server2000作为后台运行的数据库,采集生产线上的必要的输入输出信号,在对采集到
的数据进行分析处理的基础上,通过深刻分析维修专家进行设备维修时的思路,采用利用
神经网络推理计算的知识表示方法和启发式推理机制实现故障定位,在很人程度上减少诊
断过程中的工作量,提高诊断的准确性和效率。经实践检验,基本上满足客户的要求。
从上世纪80年代起,诊断技术进入了智能化阶段,经过各个国家和各个领域专家学
者的艰辛探索,神经网络技术在故障诊断方面的应用已经取得了丰硕的成果,本文最后在
对神经网络的研究应用进行总结回顺的基础上展望未来,对神经网络在故障诊断方面的应
用前景作了一个简单介绍。
关键词:自动化生产线,神经网络,故障诊断,数据采集,S?, Server数据库
基于神经网络的扣式电池白动化生产线故障诊断系统研究
RESEARCH ON THE FAULT DIAGNOSIS SYSTEM OF
LINEAR AUTOMATIC ASSEMIBLY LNE FOR BUTTON
BATTERIES BASED ON NEURALNETWORK
ABSTRACT
Nowadays, the prominent characteristic of production Is automatic and Intelligent. No
with the good-sized, high-speed of equipments, traditional diagnose technology is not enough at
all. The expense would be very large once the modern production stopped because of the fault
f machine. So people need complex, exact and timely fault diagnosis system. All these forced
men to investigate fault diagnosis, the rising research domain is formed
First in this paper, according to the system viewpoint of assembly craft and mechanical
automation product design of Li/Mno 2 button batteries, a new linear automatic assembly line
for LI/Mno2 button batteries is proposed. We also Introduce its mechanical and control system
detailedly
Because the automatic product line is a complex system with more inputs and outputs, as
the hard-core, the paper proposed a fault inspect and diagnosis method based on Neural
Network. The principle is as follows: VC++is used as the foreground application and SQL
Server2000 ad the database running backstage. Then collected the necessary In-out signals
based on the analysis and disposal of the data, through deeply analysis of masters thinkin
when they checked or maintained a set of equipments, faults of an fixture can be located b
viable knowledge and heuristic consequence method. The methods can use masters experience
adequately, so it can reduce the work of the procedure and enhance the precision and the
veracity of the result to great extent. By the checkout of practice, it can satisfy the clients
From 80s in last century, the diagnose technology enter into intelligent stage, by hardly
explore of various countries and experts, the Neural Network technology used in fault diagnosis
has obtained plentiful production
In the last, this paper summarized and reviewed the researches and applications of Neural
Network, and give a brief presentation about the application foreground of neural network used
in fault diagnosis
KEY WORDS: automatic product-line, neural network, fault diagnosis, data collection, SQL
Server database