液压支架群底座位姿及直线度的组合激光检测原理与方法研究.pdf

学校代码10112密级 博 士 学 位 论 文 论文题目液压支架群底座位姿及直线度的组合激光检测 原理与方法研究 英文题目RESEARCH ON PRINCIPLEAND OF COMBINED LASER DETECTION FOR POSEAND STRAIGHTNESS OF BASES OF HYDRAULIC SUPPORT GROUP 作者姓名杨学军. 学号2014310074. 学科矿业工程. 研究方向 煤矿设备机电液一体化 . 指导教师王怀法. 合作导师王然风. 论文提交日期2020 年 6 月 学位论文原创性声明 本人郑重声明所呈交的学位论文，是本人在导师的指导下，独立进 行研究所取得的成果。除文中已经注明引用的内容外，本论文不包含其他 个人或集体已发表或撰写过的科研成果。对本文的研究做出贡献的个人和 集体，均已在文中以明确方式标明。本声明的法律责任由本人承担。 论文作者签名签字日期年月日 学位论文版权使用授权书 本学位论文作者和指导教师完全了解太原理工大学有关保留、使用学 位论文的规定学校有权保留并向国家有关部门或机构送交学位论文的复 印件和电子版；允许本学位论文被查阅和借阅；学校可以将本学位论文的 全部或部分内容编入有关数据库进行检索，可以采用影印、缩印或其他复 制手段保存和汇编本学位论文。 本学位论文属于保密 □在年解密后适用本授权书 不保密□ 论文作者签名导师签名 签字日期年月日签字日期年月日 学位论文答辩信息表 论文题目 液压支架群底座位姿及直线度的组合激光检测原理与方 法研究 课题来源*其他 论文答辩日期2020 年 6 月 11 日答辩秘书付翔 学位论文答辩委员会成员 姓名职称是否博导工作单位 答辩委员 会主席 梁卫国教授博导太原理工大学 答辩委员 1万志军教授博导中国矿业大学 答辩委员 2廉自生教授博导太原理工大学 答辩委员 3田慕琴教授博导太原理工大学 答辩委员 4王怀法教授博导太原理工大学 *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 摘要 I 摘要 液压支架群位姿（位置和姿态）和直线度的检测和控制是智能综采工 作面实现无人化的关键问题之一，是保证无人工作面能够自动连续推进和 实现智能化开采 3.0 和 4.0 的关键因素之一。由于地质条件、控制及其它 原因，在无人工作面正常推进过程中，液压支架经常会偏离其理论位姿， 产生歪斜和不齐现象，迫使工作面停机进行人工调整。液压支架实际位姿 不仅反映整体支架群的直线度和支护工作状态，还部分反映顶板来压情 况。鉴于工作面对液压支架群位姿和直线度检测需求的迫切性，论文在追 踪国内外文献基础上，提出了基于激光组合技术实现液压支架群底座位姿 和直线度检测的新构想，采用该检测方法，有望实现对液压支架群底座六 个自由度（三个位置自由度和三个姿态自由度）位姿和直线度的检测。同 时再配合距离和角度传感器等手段测得支柱、顶梁等部件相对底座位姿信 息，便于掌握液压支架群的实时运行状况，为工作面透明化和下一步自动 调整液压支架位姿提供信息保证。 论文首先分析了液压支架工作环境、位姿偏差产生原因及相关作业标 准，综合考虑累积误差、位姿维度、检测手段和实现过程等问题，提出了 检测思路，并建立了三维和平面检测模型；针对支架底座位姿检测和研究 需求，提出了描述支架底座位姿的运动过程还原法，该法是用从基准坐标 系到待测支架坐标系的三个旋转和三个平移运动过程来描述待测支架底座 的六个自由度位姿θ, φ, ψ, u, v, w；依据检测目标，提出了以首尾架为统 一参考基准、以激光组合装置（即激光接收装置和激光雷达装置的组合） 为检测手段的液压支架群底座六自由度位姿和直线度检测原理，继而明确 了支架群底座位姿和直线度检测流程和计算方法；以斜切进刀割三角煤采 煤工艺为例，设计并展示了支架群底座位姿检测具体过程；提出了分组检 测法和相对位姿检测模式，以应对遇到较差地质条件和激光意外消失时的 特殊情况。 设计了组合激光检测验证系统，对检测系统中的激光接收装置、激光 发射装置、坐标反馈装置和激光雷达装置提出了功能要求；根据六点定位 原理，确定了激光接收装置的三点坐标定位定姿方法；根据光敏电阻φ12 mm对激光亮度敏感性特点，设计了光敏电阻矩阵单元的硬件和软件，并 对其检测精度进行了测试和分析，明确了较为理想的激光光斑检测尺寸 φ18-30 mm；通过对光敏电阻矩阵单元的扩展，完成了激光接收装置的硬 太原理工大学博士学位论文 II 件制作和软件调试；根据激光接收装置的结构和尺寸，研制了能发射三束 激光的激光发射装置和对激光束发射方向进行坐标反馈的坐标反馈装置； 对支架底座的特征进行了提取、简化和分析，在总结激光雷达工作原理的 基础上，研制了能检测支架底座相对位姿的激光雷达装置；以型号 ZZ4000-18-38 液压支架底座为检测方法原型，确定了代表其位姿的坐标系 原点位置和三坐标轴方向，同时为检测基准、目标和检测系统中的各装置 坐标系进行了命名，明确了各坐标系之间的位姿关系；本检测验证系统的 构建完成，为后续组合激光检测算法研究和试验研究奠定了物质基础。 针对激光接收装置位姿检测方法，构建了激光接收装置相对激光发射 装置的位姿检测数学模型，根据激光接收装置上三个点的六个坐标值，利 用坐标法、向量法和几何法等数学方法，探索了激光接收装置位姿的逆运 动解算算法，给出了激光接收装置位姿计算公式；然后，应用正运动解算 法和三维虚拟装配法验证了位姿逆运动解算算法的正确性。 为了验证逆运动解算算法的实际检测效果，设计了六自由度位姿调整 器，对激光接收装置的检测精度进行了验证试验研究；在激光接收装置姿 态单因素变化验证试验中，测得最大角度检测范围为，绕 X 轴14，绕 Y 轴9，绕 Z 轴10；绕 X 轴旋转过程中，姿态最大误差为-0.51，位 置最大误差为 1.08 mm；绕 Y 轴旋转过程中，姿态最大误差为 0.50，位置 最大误差为 3.50 mm；绕 Z 轴旋转过程中，姿态最大误差为-0.40，位置最 大误差为-1.71 mm；在综合位姿精度检测试验中，激光接收装置的最大角 度误差为绕 X 轴旋转时的角度误差，-1.18，最大位置误差为沿 Z 轴平移 时的距离误差，6.25 mm；针对不可避免的检测误差，对激光发射装置增 加了两个激光测距仪后，根据装置对三个点坐标的约束条件和多测出来的 坐标值，提出并设计了一套比例缩放、直线拟合和逐步逼近等算法组合的 优化方法，对检测到的三点坐标进行优化处理，最大角度检测误差由大于 5减小为小于 1.18。对激光接收装置的系统误差和随机误差进行了分 析，并提出了增大光斑-电阻直径比来增加检测精度等的改进措施；通过试 验研究，一方面改进了检测装置结构，增加了检测精度，另一方面针对装 置修改了算法，适应了实际应用场景。 分析了利用激光雷达装置检测液压支架底座相对位姿原理，建立了相 对位姿检测模型，探索了相对位姿检测算法，确定了相对位姿计算公式； 分析了支架底座特征轮廓的特点，探讨了如何确定轮廓特征点和如何检测 轮廓特征点，并分析了有效点的误差影响因素。 摘要 III 针对支架底座位姿检测算法，利用位姿调整器和底座特征物对激光雷 达装置的位姿检测精度进行了验证试验研究，试验结果表明，在单一因素 变化试验中，激光雷达检测装置最大角度误差为 0.44，最大距离误差为- 6.23 mm，综合因素变化试验中，最大角度误差为 0.48，最大距离误差为 6.74 mm，把角度误差折算成距离误差，最大误差为 19.3 mm，根据平均误 差分配原则，满足液压支架50/2 mm 直线度误差检测要求，验证了检测 算法的正确性和有效性。分析总结了激光雷达检测误差、装置误差、计算 误差和随机误差，并提出用增加特征点数量来提高检测精度的改进措施。 在激光接收装置位姿检测算法和支架底座相对位姿检测算法的研究基 础上，对支架群底座位姿和直线度的位姿转换算法进行了研究，推导了支 架群底座位姿和直线度计算公式，为下一步试验验证奠定了理论基础。 针对单个支架位姿检测方法，综合了底座简化模型、激光发射装置、 承运装置模型和激光组合装置对检测系统的综合检测精度进行了验证试验 研究；由于支架底座较大，难以用位姿调整器对其位姿进行预设，所以采 用垫角法对其原始位姿进行了预设。检测结果表明，最大角度误差为 1.37，最大距离误差为 21.48 mm；角度平均误差为 0.44，距离平均误差 为 7.2 mm；满足检测误差要求。 在液支架群底座位姿和直线度检测验证试验中，利用激光发射装置和 激光组合装置对 15 组真实液压支架位姿进行了验证检测；在三维虚拟环 境中，将检测到的六自由度位姿信息还原为支架底座工作时的真实状态， 直观再现了支架底座三维实际工作场景，为实现智能化开采支架透明化提 供了支撑条件；把 15 组真实液压支架底座位姿信息进行了直线度转换计 算，验证试验结果表明所测支架 Y 向坐标值与实际坐标值的最大误差为 26 mm，直线度检测精度满足检测误差要求；真实液压支架群组的验证试 验结果为下一步产品开发奠定了可行的算法与技术基础。 下一步研究工作中，将设计和完善检测装置的承运载体，以实现支架 群位姿和直线度的自动化检测，并针对不同类型支架和采煤工艺对算法进 行验证并完善。 关键词关键词液压支架底座；位姿；直线度；组合激光检测；位姿算法；智能 开采 太原理工大学博士学位论文 IV ABSTRACT The detection and control of poses position and attitude and straightness of hydraulic support group is one of the key issues to realize unmanned intelligent fully mechanized mining face, and it is also one of the key factors to ensure automatic continuous advancement and intelligent mining of 3.0 and 4.0 in unmanned working face. Due to geological conditions, control and other reasons, during the normal advancing process of unmanned working face, hydraulic support often deviates from its theoretical pose, resulting in skew and uneven phenomena, forcing shutdown for manual adjustment. The actual pose of hydraulic support not only reflects the straightness and supporting working state of the whole support group, but also partially reflects the roof weighting of the working face. In view of the urgency of working face to the requirement of detecting the pose and straightness of hydraulic support group, the paper puts forward a new idea of realizing the pose and straightness detection of the base of hydraulic support group based on laser combination technology on the basis of tracking domestic and foreign literatures. Using this detection , it is expected to realize the detection of six degrees of freedom three positional degrees of freedom and three attitude degrees of freedom of the base of the hydraulic support group. At the same time, coordinate with distance and angle sensors and other means to measure the pose ination of the prop, top beam and other components relative to the base, which is convenient to master the real-time operation status of the hydraulic support group and provides ination guarantee for transparency of the working face and automatic adjustment of the pose of the hydraulic support in the next step. Firstly, the paper analyzes the working environment of hydraulic support, the causes of pose deviation and relevant operation standards. Considering the cumulative error, pose dimension, detection and implementation process, the detection idea is proposed, and the three-dimensional and planar detection models are established. According to the requirements of the detection and research of the pose of the support base, a motion process reduction is proposed to describe the pose of the support base. The describes the six degrees of freedom pose θ, φ, ψ, u, v, w of the support base to be tested by ABSTRACT V three rotation and three translation motion processes from the reference coordinate system to the support coordinate system to be tested. According to the detection target, a six-degree-of-freedom pose and straightness detection principle of the hydraulic support group base is proposed, which takes the head and tail frames as a unified reference standard and the laser combination device i.e. the combination of the laser receiving device and the laser radar device as detection means. Then, the pose and straightness detection flow and calculation of the support group base are defined. Taking the mining technology of oblique cutting knife cutting triangular coal as an example, the specific process of detecting the pose of the base of the support group is designed and displayed. A group detection and a relative pose detection mode are proposed to deal with the special situations when poor geological conditions and accidental disappearance of laser are encountered. A combined laser detection and verification system is designed, and functional requirements are put forward for the laser receiving device, laser emitting device, coordinate feedback device and laser radar device in the detection system. According to the six-point positioning principle, the three- point coordinate positioning and attitude determination of the laser receiving device is determined. According to the sensitivity of photoresistor φ12 mmto laser brightness, the hardware and software of photoresistor matrix unit are designed, and its detection accuracy is tested and analyzed, and the ideal laser spot φ18-30 mm detection size is determined. By expanding the photoresistance matrix unit, the hardware and software debugging of the laser receiver are completed. According to the structure and size of the laser receiving device, a laser emitting device capable of emitting three laser beams and a coordinate feedback device for coordinate feedback of the emitting direction of the laser beam are developed. The characteristics of the support base are extracted, simplified and analyzed. On the basis of summarizing the working principle of laser radar, a laser radar device capable of detecting the relative pose of the support base is developed. Taking the model ZZ4000-18-38 hydraulic support base as the prototype of the detection , the origin position of the coordinate system and the direction of the three coordinate axes representing its pose are determined. At the same time, the coordinate systems of the detection benchmark, the target and the devices in the detection system 太原理工大学博士学位论文 VI are named, and the pose relationship between the coordinate systems is defined. The construction of this detection and verification system has laid a material foundation for the subsequent research on combined laser detection algorithms and experiments. In view of the pose detection of the laser receiving device, the pose detection mathematical model of the laser receiving device relative to the laser emitting device is constructed. According to the six coordinate values of three points on the laser receiving device, the inverse motion calculation algorithm of the pose of the laser receiving device is explored by using mathematical s such as coordinate , vector and geometric , and the pose calculation ula of the laser receiving device is deduced. Then, the correctness of the inverse pose motion algorithm is verified by applying the positive motion algorithm and the three-dimensional virtual assembly . In order to verify the actual detection effect of the inverse motion algorithm, a six-degree-of-freedom pose adjuster is designed, and the detection accuracy of the laser receiver is experimentally studied. In the single-factor attitude change verification test of the laser receiving device, the measured maximum angle detection range is 14degrees around the X axis, 9degrees around the Y axis and 10degrees around the Z axis. During the rotation around the X axis, the maximum attitude error is -0.51, and the maximum position error is 1.08 mm; During the rotation around the Y axis, the maximum attitude error is 0.50, and the maximum position error is 3.50 mm; During the rotation around the Z axis, the maximum attitude error is-0.40, and the maximum position error is- 1.71 mm; In the comprehensive pose accuracy test, the maximum angle error of the laser receiving device is the angle error when rotating around the X axis,- 1.18, and the maximum position error is the distance error when translating along the Z axis, 6.25 mm; In view of the inevitable detection error, after adding two laser rangefinders to the laser emitting device, according to the constraint conditions of the device on the coordinates of the three points and the measured coordinate values, a set of optimization s of algorithm combination such as scaling, straight line fitting and step-by-step approximation are proposed and designed, and the detected coordinates of the three points are optimized, the maximum angle detection error is reduced from more than 5to less than 1.18. The systematic error and random error of the laser receiving device are analyzed, ABSTRACT VII and the improvement measures such as increasing the spot-resistance diameter ratio to increase the detection accuracy are put forward. Through experimental research, on the one hand, the structure of the detection device is improved and the detection accuracy is increased; on the other hand, the algorithm is modified for the device, which adapts to the actual application scene. This paper analyzes the principle of using LiDAR device to detect the relative pose of hydraulic support base, establishes the relative pose detection model, studies the relative pose detection algorithm, and deduces the relative pose calculation ula. This paper analyzes the characteristics of the characteristic contour of the support base, probes into how to determine the contour characteristic points and how to detect the contour characteristic points, and analyzes the error influencing factors of the effective points. In view of the algorithm of detecting the pose of the support base, the pose adjuster and the base features are used to carry out experimental research on the pose detection precision of the laser radar device. The results show that the maximum angle error of LiDAR detection device is 0.44and the maximum distance error is -6.23 mm in the single factor change experiment. In the comprehensive factor change experiment, the maximum angle error is 0.48, the maximum distance error is 6.74 mm, the angle error is converted into the distance error, the maximum error is 19.3 mm. According to the principle of average error distribution, it meets the requirements of straightness error detection of hydraulic support 50/2 mm, and verifies the correctness and effectiveness of the detection algorithm. The detection error, device error, calculation error and random error of LiDAR are analyzed and summarized, and the improvement measures to improve the detection accuracy by increasing the number of feature points are proposed. On the basis of the research on the pose detection algorithm of the laser receiving device and the relative pose detection algorithm of the support base, the pose conversion algorithm of the support base pose and straightness is studied, and the calculation ulas of the bracket base pose and straightness are deduced, which lays a theoretical foundation for the next test verification. For a single support pose detection , the comprehensive detection accuracy of the detection system is experimentally studied by combining the simplified base model, laser emitting device, carrier device model and laser 太原理工大学博士学位论文 VIII combination device. Due to the large base of the support, it is difficult to preset its pose with the pose adjuster, so the cushion angle is used to preset its original posture. The experimental results show that the maximum angle error is 1.37 and the maximum distance error is 21.48 mm. The average angle error is 0.44, the average distance error is 7.2 mm. The accuracy of the detection system meet the requirements. In the verification test of the pose and straightness of the base of the hydraulic support group, the poses of 15 hydraulic supports were detected by using the laser emitting device and the laser combination device. In a three- dimensional virtual environment, the