智能化超重型岩巷掘进机动载荷识别系统的开发

发布时间：2018-04-19 13:19

本文选题：岩巷掘进机 + 动载荷识别　；参考：《太原理工大学》2015年硕士论文

【摘要】：本课题来源于国家863计划资源环境技术领域重大项目“煤炭智能化掘采技术与装备（一）”子课题“智能化超重型岩巷掘进机研制”（课题编号：2012AA06A405），是针对岩巷掘进机工作时，工况复杂、负载多变、动载荷实时识别难度大等问题而提出的。岩巷掘进机的动载荷识别是掘进机自动控制的重要组成部分，对提高掘进机智能化水平及使用寿命具有重要意义。近年来，岩巷掘进机在我国开采领域得到了越来越广泛的应用，但智能掘进技术仍处于起步阶段。煤矿井下现在使用的掘进机大多数自动调节水平较低，以司机凭经验手动操作为主。手动操作掘进机不仅劳动强度大，而且因难以及时准确判断截割载荷状态，导致截齿损耗严重。能否根据负载大小自动调节截割速度就显得尤为重要，，而可靠的动载荷识别技术又是自动调节的必备条件。因此，开发智能化超重型岩巷掘进机动载荷识别系统具有非常重要的现实意义。本文在分析岩巷掘进机截割机构动力学特性的基础上，结合先进的信号分析技术、智能识别技术及掘进机实际运行情况，并通过大量动载荷模拟试验，开发了智能化超重型岩巷掘进机动载荷识别系统。本文主要研究内容如下：在查阅大量相关文献的基础上，阐述了智能化超重型岩巷掘进机动载荷识别系统在国内外研究现状及发展趋势。深入分析了不同工况下截割机构的载荷分布，确定了能有效反映截割头动载荷的物理参量，主要包括悬臂振动，截割电动机电流和回转、升降液压缸压力。根据岩巷掘进机截割机构有限的空间范围，选择了适用于井下恶劣环境的多种传感器，完成了监测信息的准确测量。结合项目的功能要求，设计了以数据采集卡及工控机为核心的智能化超重型岩巷掘进机动载荷总体方案。结合动载荷信号为随机信号，频率成分复杂、非平稳的特点，比较了傅里叶变换、小波变换及小波包变换的优缺点，确定了采用小波包变换作为信号处理及特征提取的工具。详细介绍了小波包特征能量提取步骤，并通过分析实测数据得到了振动、电流和压力信号的特征频段范围。结合所选特征频段，选择RBF神经网络及证据组合理论作为动载荷的智能识别方法，并对RBF神经网络和证据理论算法、结构及应用步骤进行简单分析，重点描述了组合神经网络信息识别原理，D-S证据理论动载荷识别原理及基于神经网络和D-S证据理论融合原理。在LabVIEW开发环境下设计了智能化超重型岩巷掘进机动载荷识别的相关应用软件，主要包括LabVIEW平台下MATLAB代码的调用、多传感信号同步采集、小波包特征量提取、神经网络动载荷识别、证据理论融合、数据存储管理以及人机显示界面等程序，并进行了功能调试，调试结果验证了应用软件的有效性。根据掘进机实际运行情况，设计了利用回转液压缸及升降液压缸压力信号先分工况再进行动载荷识别的方案。根据纵向钻进，水平切割，纵向切割三大类工况，系统采用三类工况识别网络。以此为前提，结合截割机构的振动数据、截割电动机的电流及液压缸压力数据，利用数据融合原理，构建了一级和二级RBF神经网络多传感器信息融合动载荷识别模型，提出了将神经网络与证据理论有机结合，优势互补的基于多神经网络与证据理论相融合的掘进机动载荷识别新方法。并用实测数据进行训练、测试及分析。
[Abstract]:This topic derives from the "intelligent mining technology and equipment of coal intelligent mining and mining", a major project of the National 863 plan resources and environmental technology field (one) "intelligent super heavy rock roadway driving machine development" (subject number: 2012AA06A405). It is a problem for the working of Rock Roadheader, which is complex, variable load and difficult to identify the dynamic load in real time. And put forward.
The dynamic load identification of the Rock Roadheader is an important part of the automatic control of the roadheader. It is of great significance to improve the intelligentized level and the service life of the roadheader. In recent years, the rock tunnel boring machine has been more and more widely used in the field of mining in our country, but the intelligent tunneling technology is still in its infancy. The most automatic adjustment level of the machine is low, and the driver is based on the experience manual operation. The manual operation boring machine not only has great labor intensity, but also makes the cutting load serious because it is difficult to judge the cutting load in time and accurately. It is particularly important to adjust the cutting speed automatically according to the load size, and the reliable dynamic load identification technique is very important. As a necessary condition for automatic regulation, it is of great practical significance to develop intelligent load identification system for ultra heavy rock roadway driving.
Based on the analysis of the dynamic characteristics of the cutting mechanism of the Rock Roadheader, this paper combines the advanced signal analysis technology, the intelligent identification technology and the actual operation of the roadheader, and develops a intelligent load identification system for the driving maneuver in the ultra heavy-duty rock roadway through a large number of dynamic load simulation tests. The main contents of this paper are as follows:
On the basis of a large number of relevant documents, the current research status and development trend of the intelligent super heavy rock roadway driving load identification system at home and abroad are expounded. The load distribution of the cutting mechanism under different working conditions is analyzed, and the physical parameters which can effectively reflect the dynamic load of the cutting head are determined, mainly including the cantilever vibration and the cutting motor. Current and rotary, lift hydraulic cylinder pressure.
According to the limited space range of the cutting mechanism of the Rock Roadheader, a variety of sensors suitable for the downhole bad environment have been selected, and the accurate measurement of the monitoring information has been completed. The overall plan of the intelligent and super heavy rock roadway driving load is designed with the core of the data acquisition card and the industrial control machine.
Combining the dynamic load signal as random signal, complex frequency component and non stationary feature, the advantages and disadvantages of Fourier transform, wavelet transform and wavelet packet transform are compared, and the wavelet packet transform is used as a tool for signal processing and feature extraction. The step of wavelet packet characteristic energy extraction is introduced in detail, and the measured data are obtained by analyzing the measured data. According to the frequency range of vibration, current and pressure signal, combining the selected frequency bands, RBF neural network and evidence combination theory are selected as the intelligent recognition method of dynamic load, and the RBF neural network and evidence theory algorithm, structure and application steps are simply analyzed, and the information recognition principle of the combined neural network is described, and the D-S certificate is described. According to the theory of dynamic load identification and the fusion principle based on neural network and D-S evidence theory.
In the LabVIEW development environment, the application software of intelligent super heavy rock roadway driving load identification is designed. It mainly includes the call of MATLAB code under the LabVIEW platform, synchronous acquisition of multi sensing signal, feature extraction of wavelet packet, neural network dynamic load identification, evidence theory fusion, data storage management and human-computer display interface, etc. The debug results verify the validity of the application software.
According to the actual running situation of the roadheader, a scheme is designed to identify the dynamic load by using the pressure signal of the rotary hydraulic cylinder and the lifting hydraulic cylinder first, and then to identify the dynamic load. According to the longitudinal drilling, the horizontal cutting and the longitudinal cutting, the system adopts the three types of working conditions to identify the network. This is the premise, combining the vibration data of the cutting mechanism to cut the electric power. With the current and pressure data of the hydraulic cylinder, using the principle of data fusion, a dynamic load identification model for multi-sensor information fusion of the first and two RBF neural networks is constructed, and a new method is proposed, which combines the neural network and the evidence theory organically, and the advantages are complementary to the multi neural network and the evidence theory. Training, testing and analysis are carried out with measured data.

【学位授予单位】：太原理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TD421.5

【参考文献】