矿井煤流数字孪生激光扫描质量评价方法研究
Research on the quality evaluation method for digital twin laser scanning of mine coal flow
李晓雅 1李猛钢 2胡而已 1裴文良3
作者信息
- 1. 应急管理部信息研究院,北京市 朝阳区,100029
- 2. 中国矿业大学机电工程学院,江苏省徐州市,221116
- 3. 中信重工开诚智能装备有限公司,河北省唐山市,063020
- 折叠
摘要
激光扫描技术是矿井数字孪生建模的基础手段,其模型点云质量好坏直接影响孪生建模和智能分析的精度.介绍了煤流激光扫描数字孪生建模的原理,通过分析煤流点云几何参数的信息熵,明确了点云信息熵与噪音强度之间的关系,并可将其作为建立激光扫描煤流点云模型质量评价方法的实验依据.阐明了信息熵的影响因素,开展了多因素方差实验进行显著性效应检验,结果表明信息熵的变化主要与煤流量、噪音强度有关.结合二次曲面方程、高斯牛顿迭代法,对噪音强度的预测模型进行了优化训练;通过构建模糊关系矩阵,实现对点云滤波参数的模糊推理,进而得到高保真、高可靠以及高精度的煤流点云模型.针对煤矿综放工作面实地采集的煤流点云样本进行实验,结果表明点云质量评价算法具备更好的一致性、稳定性和单调性,更适用于煤矿井下复杂工况条件.
Abstract
Laser scanning technology is the basic method of mine digital twin modeling,and the quality of the model point cloud directly affects the accuracy of twin modeling and intelligent analysis.The principle of digital twin modeling for coal flow laser scanning was introduced.The analysis of information entropy of the geometric parameters of the coal flow point cloud clarified the relationship between the point cloud information entropy and the noise intensity,which can be used as an experimental basis for establishing the quality evaluation method of the laser scanning coal flow point cloud model.The influencing factors of information entropy were ascertained,and a multi-factor variance experiment was carried out to test the significant effect,and the experiment results showed that the change of information entropy was mainly related to coal flow and noise intensity.Combining the quadric surface equation and the Gauss Newton iteration method,the prediction model of noise intensity was optimally trained.By constructing a fuzzy relationship matrix,the fuzzy inference of point cloud filter parameters was achieved,resulting in a high-fidelity,high-reliability and high-accuracy coal flow point cloud model.Based on the experiment with coal flow point cloud samples collected on the coal mine fully mechanized caving face,the results indicated that the point cloud quality evaluation algorithm had better consistency,stability and monotonicity,and was more suitable for complex working conditions in underground coal mines.
关键词
数字孪生/智能煤流/点云数据/模型质量/信息熵Key words
digital twin/intelligent coal flow/point cloud data/model quality/information entropy引用本文复制引用
基金项目
国家自然科学基金资助项目(52274159)
出版年
2024
NETL
NSTL
万方数据