岩巷掘进机截割头振动与岩石硬度映射关系研究.pdf
学校代码10112 密 级 硕 士 学 位 论 文 (专(专 业业 学学 位)位) 论文题目 英文题目 作者姓名 王 茜 学 号 2017520191 专业领域 电气工程 研究方向 电力电子与电力传动 指导教师 吝伶艳 副教授 校外导师 陈荣海 高级工程师 论文提交日期 2020 年 5 月 岩巷掘进机截割头振动与岩石硬度 映射关系研究 Study on the Mapping Relationship Between Vibration of Cutting Head of Rock Roadheader and Rock Hardness 万方数据 万方数据 硕士学术学位、硕士非工程类专业学位 学位论文答辩信息表 论文题目 岩巷掘进机截割头振动与岩石硬度映射关系研究 课题来源* 国家自然科学基金(项目编号U1510112) 论文答辩日期 2020 年 5 月 17 号 答辩秘书 梁定康 学位论文答辩委员会成员 姓名 职称 博导/硕导 工作单位 答辩委员 会主席 李保来 高高工 山西汾西重工有 限责任公司 答辩委员 1 杜欣慧 教授 硕导 太原理工大学 答辩委员 2 郑丽君 副教授 硕导 太原理工大学 *课题来源可填国家重点研发计划项目、国家自然科学基金项目、国家 社科基金项目、教育部人文社科项目、国家其他部委项目、省科技厅项 目、省教育厅项目、企事业单位委托项目、其他 万方数据 摘 要 I 摘 要 本课题研究内容来源于国家自然科学基金“基于多参量的超重型岩巷 掘进机截割动载荷智能识别方法的研究” (项目编号 U1510112) , 主要是针 对井下环境恶劣, 操作工人人身安全存在极大隐患、 动载荷无法准确识别等 问题提出来的。 实际工作中, 操作工人由于无法准确判断掘进机截割头所接 触岩壁状况, 及时调整掘进机各项参数,可能会导致掘进机零件损坏、使用 寿命变短。 因此, 本文提出了一种适用于井下岩壁开采的超重型岩巷掘进机 截割头动载荷识别方法, 为掘进机智能化发展奠定了基础, 具有极其重要的 意义。 本文通过查阅大量国内外文献,在深入了解掘进机及动载荷识别方法 研究现状的基础上发现 现有的掘进机智能化水平较低, 动载荷识别技术在 掘进机煤岩切割领域方面应用较少。 因此, 本文以超重型岩巷掘进机为研究 对象, 提出了一种基于多参数的掘进机截割头动载荷识别方法, 分析掘进机 截割头振动信号与岩石硬度之间的映射关系,具体研究内容如下 深入了解掘进机结构组成及其工作原理,对不同参数下的掘进机截割 机构进行动力学分析。根据掘进机截割头载荷计算公式,分析、 确定影响截 割头振动信号变化的各个参数,包括岩壁硬度、截深、摆速、转速等。确定 掘进机各项参数值,在 MATLAB 软件中编写了截割头载荷模拟程序,得到 了不同参数变量下掘进机截割头受力曲线,并深入分析了各参数对载荷的 影响,为掘进机动载荷识别参考依据。 以某型号的超重型岩巷掘进机为例搭建了虚拟样机三维模型。在 Pro/E 软件中搭建掘进机截割部、 本体部、 行走部等部件的三维模型并进行组装, 建立掘进机三维模型。将模型导入 ADAMS 软件中进行柔性化处理并添加 各类约束以及驱动条件构成掘进机虚拟样机模型。将不同参数下的掘进机 截割头受力曲线导入 ADAMS 软件内并施加于截割头质心处,启动程序在 ADAMS 内进行掘进机切割煤岩仿真模拟得到不同参数下的截割头振动信 号,为截割头载荷特征量提取提供依据。 根据掘进机动载荷识别要求,提出了一种基于多源数据融合的掘进机 动载荷识别方法。 首先, 提出了一种将相关性阈值去噪法和经验小波变换相 结合的特征量提取方法。利用掘进机信号时频图的极大值点构建自适应滤 波器组, 选择经验小波变换对信号进行分解, 利用相关性阈值去噪法对各信 号分量进行去噪处理并计算去噪后各分量与初始信号之间的相关系数,选 万方数据 太原理工大学硕士学位论文 II 择其标准偏差作为阈值, 对信号进行删选构成特征量。 最后, 将特征量作为 卷积神经网络的输入信号对掘进机截割头信号进行模式识别。 设计了掘进机截割头动载荷识别方案流程。利用经验小波变换与相关 性阈值去噪方法相结合对不同参数下的掘进机截割头信号进行特征量提取, 此后设置合适的神经网络参数构建神经网络模型,利用一维卷积神经网络 对输入信号进行模式识别,实验证明,该方案具有较高的识别准确率。 关键词关键词超重型岩巷掘进机;动载荷识别;改进经验小波变换;虚拟样机; 卷积神经网络 万方数据 ABSTRACT III ABSTRACT The subject is an important part of national science fund projects “Research on Intelligent Recognition of Cutting Dynamic Load of the Super-heavy Rock Roadheader Based on Multi Parameters” NO U1510112, which is mainly aimed at solving the problem of difficulty in recognizing the hardness of cutting rock wall because of the poor underground environment and the hidden danger to the personal safety of operators. During actual coal and rock mining, the operating workers cannot accurately judge the condition of the rock wall contacted by the cutting head of the roadheader and adjust the parameters of the roadheader in time, which may cause damage to the roadheader parts and shorten it’s life. Therefore, a for identifying the dynamic load of the cutting head of the super-heavy rock roadheader suitable for actual coal and rock mining is proposed, which provides a basis for the intelligent development of the roadheader and this research has extremely important significance. In this paper, through consulting a large number of domestic and foreign literatures, on the basis of in-depth understanding of the development process of the roadheader and the dynamic load identification , it is found that the existing roadheader has a low level of intelligence, and the dynamic load identification technology is rarely applied in the field of coal and rock cutting of roadheader. Therefore, taking super-heavy rock roadheader as the research object, a dynamic parameter recognition for the cutting head of the roadheader based on multi-parameters is proposed to analyze the mapping relationship between the vibration signal of the head of roadheader and the hardness of the rock. The specific research contents are as follows Through the deep understanding of the structure and working principle of the roadheader, the dynamic analysis of the cutting mechanism is carried out. According to the load calculation ula of the cutting head of the roadheader, the main parameters affecting the cutting head of the roadheader including the cutting rock hardness, the cutting head depth, the swing speed, the rotation speed , etc. According to the design parameters of the cutting head of the roadheader, the cutting head load simulation program is written in MATLAB software, the stress curve of the cutting head of roadheader under different parameter variables is 万方数据 太原理工大学硕士学位论文 IV gotten, and the influence of each parameter on the cutting head load is analyzed in depth, which provides the reference basis for the identification of the driving load of the cutting head load of the roadheader. According to the example of a super-heavy rock roadheader, a three- dimensional virtual prototype model was built. The three-dimensional models of the roadheader’s cutting part, body part, walking part and other components are built and assembled in Pro / E software to establish the three-dimensional model of the roadheader. In the ADAMS software, the three-dimensional model is flexibly processed, and various constraints and driving conditions are added to a virtual prototype model of the roadheader. The force curve of the cutting head of the roadheader under different parameters is imported into the ADAMS software and applied to the center of mass of the cutting head. In the ADAMS , the coal and rock cutting of the roadheader is simulated to obtain the vibration signal of the cutting head under different parameters, which provides a basis for the extraction of the cutting head load feature quantity. According to the requirements of roadheader maneuver load identification, a of roadheader maneuver load recognition based on multi-source data fusion is proposed. First, a feature extraction combining correlation threshold denoising and empirical wavelet trans is proposed. According to the maximum value point of the time-frequency map of the roadheader signal, an adaptive filter bank is constructed, the empirical wavelet trans is used to decompose the signal, and the correlation threshold denoising is used to denoise each signal component, and the standard deviation of the correlation coefficient between each component after denoising and the initial signal is used as a threshold to per signal deletion to constitute a feature quantity. Finally, the feature quantity is used as the signal of the convolutional neural network to per pattern recognition of the cutting head signal of the roadheader. The process flow of the dynamic load identification scheme for the cutting head of the roadheader is designed. The empirical wavelet trans and correlation threshold denoising are used to extract the feature quantity of the cutting head signal of the roadheader under different parameters, then the neural network model is constructed by setting appropriate neural network parameters, and then the one-dimensional convolution neural network is used to 万方数据 ABSTRACT V per patten recognition on the signal, experiments show that this scheme has a high recognition accuracy. KEY WORDS Super Heavy Rock Roadheader; Dynamic Load Identification; Improved Empirical wavelet Trans; Virtual Prototype; Convolutional Neural Network 万方数据 太原理工大学硕士学位论文 VI 万方数据 目 录 VII 目 录 摘 要 ......................................................................................................................................... I ABSTRACT ............................................................................................................................ III 第一章 绪论 ............................................................................................................................ 1 1.1 课题研究背景和意义 ................................................................................................ 1 1.2 课题国内外现状 ........................................................................................................ 2 1.2.1 掘进机国内外研究现状 .................................................................................. 2 1.2.2 动载荷识别技术国内外研究现状 .................................................................. 3 1.3 本文研究目标及主要内容 ........................................................................................ 5 第二章 岩巷掘进机截割机构动力学分析 ............................................................................ 7 2.1 掘进机结构及工作原理 ............................................................................................ 7 2.1.1 岩巷掘进机结构 .............................................................................................. 7 2.1.2 掘进机工作原理 .............................................................................................. 8 2.1.3 掘进机分类 ...................................................................................................... 9 2.2 掘进机截割头动力学分析 ...................................................................................... 10 2.2.1 掘进机截割头运动 ........................................................................................ 10 2.2.2 掘进机截割头受力分析 ................................................................................ 12 2.3 掘进机截割头载荷模拟 .......................................................................................... 15 2.3.1 确定截割头截割参数 .................................................................................... 15 2.3.2 载荷模拟程序 ................................................................................................ 16 2.3.3 截割头载荷模拟分析 .................................................................................... 19 2.4 本章小结 .................................................................................................................. 27 第三章 岩巷掘进机虚拟样机模型 ...................................................................................... 29 3.1 掘进机三维模型搭建 .............................................................................................. 29 3.1.1 掘进机部件搭建 ............................................................................................ 29 3.1.2 掘进机整机模型搭建 .................................................................................... 31 3.2 掘进机虚拟样机模型 .............................................................................................. 32 万方数据 太原理工大学硕士学位论文 VIII 3.2.1 掘进机虚拟样机建立 .................................................................................... 33 3.2.2 掘进机虚拟样机的运行 ................................................................................ 35 3.3 本章小结 .................................................................................................................. 38 第四章 岩巷掘进机动载荷识别方法的研究 ...................................................................... 39 4.1 动载荷识别基本要求 .............................................................................................. 39 4.2 信号特征量提取 ...................................................................................................... 39 4.2.1 信号分析方法的对比 .................................................................................... 40 4.2.2 奇异值分解 .................................................................................................... 41 4.2.3 经验小波变换 ................................................................................................ 43 4.2.4 改进经验小波变换 ........................................................................................ 46 4.3 信号特征量提取 ...................................................................................................... 48 4.4 信号动载荷识别 ...................................................................................................... 49 4.4.1 神经网络理论应用 ........................................................................................ 49 4.4.2 动载荷识别方案 ............................................................................................ 50 4.4.3 改进 RBF 神经网络 ...................................................................................... 51 4.4.4 卷积神经网络 ................................................................................................ 52 4.5 本章小结 .................................................................................................................. 53 第五章 岩巷掘进机动载荷识别实验结果分析 .................................................................. 55 5.1 信号采集 .................................................................................................................. 55 5.2 奇异值分解信号处理 .............................................................................................. 56 5.2.1 最优小波包参数确定 .................................................................................... 56 5.2.2 奇异值分解特征量提取 ................................................................................ 58 5.3 改进经验小波变换特征量提取 .............................................................................. 59 5.4 基于神经网络的模式识别 ...................................................................................... 66 5.4.1 RBF 神经网络模式识别 ............................................................................... 66 5.4.2 卷积神经网络模式识别 ................................................................................ 66 5.5 本章总结 .................................................................................................................. 69 第六章 结论与展望 .............................................................................................................. 71 6.1 工作总结 .................................................................................................................. 71 6.2 工作展望 .................................................................................................................. 72 参考文献 .................................................................................................................................. 73 攻读学位期间取得的科研成果 .............................................................................................. 77 万方数据 目 录 IX 致 谢 ...................................................................................................................................... 79 万方数据 太原理工大学硕士学位论文 X 万方数据 绪 论 1 第一章 绪论 1.1 课题研究背景和意义 我国煤炭资源十分丰富,其生产和消费总量均居于世界前列,根据研究表明我国 将在很长一段时间内坚持以电力为中心,煤炭作为主体,油气和各类新能源全面共同发 展的战略政策[1]。我国地质构造复杂,煤炭开采难度大,绝大部分以井下开采为主,因 此,实现煤炭开采的智能化、自动化以及信息化变得尤为重要。 煤炭开采主要包括露天开采以及矿井开采两大类。由于我国煤炭工业基础较为薄弱, 露天开采总含量占煤炭总量的不足 10%,因此煤炭开采以矿井开采为主[2]。巷道掘进技 术是提高煤炭产量的关键,其中煤矿井下爆破技术不仅效率低下,不规范使用还易引发 人员伤亡、 岩石破碎分撒、 瓦斯爆炸、 煤岩突出坍塌等问题[3], 严重影响煤矿安全生产, 因此我们应尽量减少该方法的使用次数。与之相对,岩巷掘进机由于其开采效率高、安 全性能好、巷道环境优良等特点,逐渐成为煤矿开采的主要使用设备[4]。 岩巷掘进机的机械装备水平及其制造工艺与煤矿安全生产息息相关,因此要想实现 采煤工业快速、经济、安全生产,就必须不断提高我国岩巷掘进机制造水平[5]。传统的 岩巷掘进机破岩效率低下、应用范围小、可靠性低且智能化水平低,因此如何快速提高 掘进机智能化发展,实现多功能于一体的高效开采成为提高我国煤岩开采质量和效率的 关键[6]。目前,超重型岩巷掘进机凭借其稳定性好、截割能力强等特点被广泛使用,然 而,工人在操作掘进机时所处环境恶劣,井下飞尘严重,导致其视野受限无法准确判断 掘进机所切割岩壁的状态,及时调整掘进机工作时截割头的各项参数,最