基于神经网络的矿井提升机监测与故障诊断系统的研究电力电子与电力传动专业论文.docxVIP

山东科技大学硕士学位论文摘要 山东科技大学硕士学位论文 摘要 摘 要 矿井提升机的故障诊断一直是煤矿安全领域十分必要且有意义的课题。本文对提升 机制动系统故障诊断的实现做了研究和设计。根据提升机制动系统的特点，采用人工神 经网络 BP 算法，利用嵌入式系统进行数据采集，初步实现了提升机制动系统故障诊断。 文中首先简述了目前矿井提升设备的监测方法与系统组成，介绍了人工神经网络概 况，初步介绍了神经网络的学习方法，着重讨论了 BP 神经网络在提升机故障诊断中应用 的可行性和优缺点。 接下来，根据利用嵌入式系统进行硬件设计具有低功耗，集成度高，体积小，成本 低等优点，选取了一款性能优异的 ARM 芯片 S3C44BOX，设计了硬件接口电路。由于 S3C44BOX 不带 MMU，这使得系统没有足够的硬件支持虚拟内存管理技术，因此本系统采 用μClinux 作为嵌入式操作系统。在移植了μClinux 操作系统的基础上设计了部分应用 程序。 最后，通过对提升机制动系统故障特点的研究，建立了基于 BP 神经网络的故障诊断 系统，确定了神经网络每一层的神经元数，并利用 MATLAB 实现了对网络的训练，并对训 练结果进行了仿真，仿真结果表明，该算法达到了预期的要求。 关键词：状态监测，故障诊断，BP 神经网络，嵌入式系统 ABSTRACT Fault diagnosis of Mine hoist is always a quite require and significative task in the field of coal mine security. The text makes research and design for the realization of the fault diagnosis of Mine hoist break system. According to the characteristic of Mine hoist break system, adopting the manual neural network BP arithmetic and using the prevalent embedded system to gather data, the text realizes the fault diagnosis of Mine hoist braking system. At the first, the text simply the monitor measure and the system constitution of Mine hoist equipment at present, and introduces the general situation of the manual Neural Network, nitially introduced the neural network study method and also emphasizes the applying of it in the hoist fault diagnosis; Then, according to use embedded system to carry on the hardware design to have the low power loss, the integration rate high, volume small, the cost is low and so on the superior artillery. we choosing section performance outstanding ARM chip S3C44BOX, designed the hardware connection electric circuit.Because S3C44BOX does not have MMU, this causes the system does not have enough hardware support hypothesized memory management technology, therefore this system uses μClinux embedded operating system. Finally,through research to the elevator braking system, we established fault diagnosis system based on the BP neural network,and makes cert