采煤机是煤炭开采过程中最基本、核心的大型复杂设备之一，加强对采煤机的状态监测、准确评估采煤机健康状态、及时进行预前维护，对煤炭安全开采具有不可忽视的工程意义。本文针对采煤机运行可靠性低、监测参数冗余、健康状态评估效率不高等问题，以电牵引采煤机为研究对象，分析采煤机基本结构及易发故障部位以选取采煤机监测指标参数、研究采煤机状态参数异常值检测方法、构建采煤机健康状态评估体系、研究数据驱动的采煤机状态评估方法、开发健康状态评估系统，提高采煤机运行可靠性和提升采煤机健康管理水平。

首先，分析采煤机主要结构、总结常见故障发生位置和发生原因，阐述数据驱动的采煤机健康状态评估技术基本框架。通过对采煤机监测参数选取、数据预处理、评估模型构建等主要理论进行分析，为采煤机健康状态评估工作的展开提供理论基础。

其次，针对采煤机运行过程中产生的数据异常值影响评估效率等问题，分别提出基于LOF和iForest算法的采煤机状态参数异常值检测方法，并对两种算法的参数进行优化，对比异常值检测算法对采煤机状态数据异常值检测的适用性；针对采煤机监测参数存在冗余性和相关性的问题，通过采煤机监测参数相关性分析筛选状态评估关键指标，构建采煤机健康状态评估指标体系。

然后，针对基于BP神经网络的状态评估方法易陷入局部极小值、评估效率不高等问题，引入遗传算法来优化BP神经网络模型，并对该采煤机状态评估方法进行实验验证；针对神经网络可解释性差、采煤机数据类别不平衡等问题，建立基于改进XGBoost的采煤机状态评估模型，通过对比分类错误率选择XGBoost模型关键参数的最优值，并计算改进后的XGBoost模型的平均准确率，验证评估方法的可行性。

最后，以状态评估算法研究为基础，基于LabVIEW和MATLAB脚本节点开发采煤机健康状态监测及评估系统。该系统可实现运行状态监测、状态评估、管理人员登录、数据库存储等功能，并利用实时运行数据对评估系统进行验证；选取采煤机历史运行数据对评估算法模型进行对比，为采煤机健康状态评估和健康管理工作提供有效的解决方法。

The shearer is one of the most basic and core large-scale complex equipment in

the coal mining process. Strengthening the state monitoring of the shearer, accurately

assessing the health status of the shearer and timely pre-maintenance,it has engineering

significance that cannot be ignored for safe coal mining.Aiming at the problems of

low operating reliability of the shearer, redundant monitoring parameters and low effic-iency of health evaluation, the electric traction shearer is taken as the research object, and the basic structure and fault-prone parts of the shearer are analyzed to select the

shearer for monitoring index parameters, studying the detection method of the abnormalvalue of the shearer status parameter, constructing the shearer health evaluation system,researching the data-driven shearer status evaluation method, developing the health

evaluation system,improving the operation reliability of the shearer and enhancing the

health management level of coal mining machinery.

Firstly,this paper analyzes the main structure of the shearer, summarizes the locat-

ions and causes of common failures and describes the basic framework of data-driven shearer health assessment technology. Through the analysis of the main theories such

as the selection of the monitoring parameters of the shearer, data preprocessing, and

the construction of the evaluation model, it provides a theoretical basis for the evaluat-ion of the health status of the shearer.

Secondly, in view of the problem of data abnormal value generated during the

operation of the shearer affecting the evaluation efficiency, etc., the shearer state

parameter anomaly detection methods based on LOF and iForest algorithms are

respectively proposed, and the parameters of the two algorithms are optimized and the anomaly detection is compared. The applicability of the algorithm to the abnormal

detection of the state data of the shearer; in view of the redundancy and correlation

between the monitoring parameters of the shearer, the key indicators of the status

evaluation are screened through the correlation analysis of the monitoring parameters ofthe shearer, and the health status of the shearer is constructed Evaluation index system.

Then, in view of the problems that the state evaluation method based on BP

neural network is easy to fall into local minimum and the evaluation efficiency is not high, genetic algorithm is introduced to optimize the BP neural network model, and

the state evaluation method of the shearer is experimentally verified;Aiming at the pro-blems of poor interpretability of neural network and unbalanced data types of shearer,

this paper establishs a state evaluation model for the shearer based on the improved

XGBoost,selects the optimal value of the key parameters of the XGBoost model by

comparing the classification error rate, and calculates the average accuracy of the impr-oved XGBoost model Rate to verify the feasibility of the evaluation method.

Finally, based on the research of the state evaluation algorithm, the shearer health

state monitoring and evaluation system is developed based on LabVIEW and MATLABscript nodes.The system can realize several functions such as operation status monitorin-g, status evaluation, manager login, database storage, etc. It uses real-time operation

data of the shearer subsystem to verify the evaluation system with examples; By com-paring the evaluation algorithm model with the historical data of shearer, it provides aneffective solution for shearer health status evaluation and health management.

[1]赵敏, 杨伟红, 王国平, 等. 我国煤炭资源的战略储备研究[J]. 中国矿业, 2017, 26(10): 90-92.

[2]中华人民共和国国家统计局. 原煤生产总量[R]. 2020.

[3]2020年中国煤炭工业发展研究报告[M]. 北京: 中国经济出版社, 中国煤炭工业协会, 2020

[4]王金华, 黄乐亭, 李首滨, 等. 综采工作面智能化技术与装备的发展[J]. 煤炭学报, 2014, 39(8): 1418-1423.

[5]袁亮. 煤炭精准开采科学构想[J]. 煤炭学报, 2017, 42(01): 1-7.

[6]王国法, 范京道, 徐亚军, 等. 煤炭智能化开采关键技术创新进展与展望[J]. 工矿自动化, 2018, 44(2): 5-12．

[7]邱锦波. 滚筒采煤机自动化与智能化控制技术发展及应用[J]. 煤炭科学技术, 2013, 41(11): 10-13.

[8]Sean Kilcarr. Aiming at condition-based maintenance[J]. Fleet Owner, 2015.

[9]李晋, 朱强强, 范旭峰, 等. 大型机电设备健康状态评估方法研究[J]. 工矿自动化, 2015,41(01): 6-9.

[10]季瑞. 采煤机健康管理关键技术研究[D]. 徐州：中国矿业大学, 2014.

[11]谢贵君. 电牵引采煤机的现状与发展趋势[J]. 煤矿机械, 2009, 30(02): 1-3.

[12]芮冰, 黄钦宗. 我国采煤机30年发展回顾和展望[J]. 煤矿机电, 2000(05): 36-40.

[13]谭超. 电牵引采煤机远程参数化控制关键技术研究[D]. 徐州：中国矿业大学, 2009.

[14]戴绍成. 高产高效综合机械化采煤技术与装备. 煤炭工业出版社. P215-218.

[15]张世洪. 我国综采采煤机技术的创新研究[J]. 煤炭学报, 2010, 35(11): 1898-1902.

[16]崔妮. 基于无线传感器网络的采煤机状态监测系统设计与实现[D]. 太原：中北大学, 2016.

[17]陈相丞. 采煤机摇臂系统维修状态评估方法研究[D]. 徐州：中国矿业大学, 2018.

[18]段蛟龙. 基于物联网的采煤机状态监测及寿命管理系统的开发[D]. 太原：太原理工大学, 2016.

[19]卫薇, 龙玉江, 钟掖. 基于概率统计模型的电力IT监控对象特征异常检测[J]. 山东农业大学学报(自然科学版), 2019, 50(04): 612-618.

[20]罗峰, 胡强, 侯硕, 等. 基于支持向量机的CAN-FD网络异常入侵检测[J]. 同济大学学报(自然科学版), 2020, 48(12): 1790-1796.

[21]JIA H J, DING S F, SHI Z Z. Approximate Weighted Kernel k-means for Large-Scale Spectral Clustering[J]. Journal of Software, 2015, 23(1): 12-17.

[22]刘芳, 齐建鹏, 于彦伟, 等. 基于密度的Top-n局部异常点快速检测算法[J]. 自动化学报, 2019, 45(09): 1756-1771.

[23]李春生, 于澍, 刘小刚. 基于改进距离和的异常点检测算法研究[J]. 计算机技术与发展, 2019, 29(03): 97-100.

[24]Bandaragoda, T.R., Ting, K.M., Albrecht, D., Liu, F.T., Wells, J.R.: Efficient anomaly detection by isolation using nearest neighbour ensemble. In: 2014 IEEE International Conference on Data Mining Workshop, pp. 698–705 (2014)

[25]刘明. 虚拟变量回归模型的应用: 方差分析及异常值检验[J]. 统计与决策, 2014(16): 10-14.

[26]沈小军, 付雪姣, 周冲成, 等. 风电机组风速-功率异常运行数据特征及清洗方法[J]. 电工技术学报, 2018, 33(14): 3353-3361.

[27]寇元宝, 汪崇建, 贠瑞光. 采煤机运行数据预处理算法研究[J]. 煤矿机械, 2020, 41(10): 41-43.

[28]杨全纬, 唐向红, 郑阳, 等. 电机状态异常数据检测算法研究[J]. 组合机床与自动化加工技术, 2016(08): 34-38.

[29]倪景峰, 刘丽华, 顾煜炯. 基于最小二乘支持向量机算法的测量数据时序异常检测方法[J]. 华北电力大学学报(自然科学版), 2008(03) : 62-66.

[30]孙栓柱, 宋蓓, 李春岩, 等. 一种基于贝叶斯后验的异常值在线检测及置信度评估算法[J]. 中国科学技术大学学报, 2017, 47(08): 644-652.

[31]周林, 赵杰, 冯广飞. 装备故障预测与健康管理技术[M]. 北京：国防工业出版社，2015.

[32]Tobon-Mejia D A, Medjaher K, Zerhouni N, et al. A Data-Driven Failure Prognostics Method Based on Mixture of Gaussians Hidden Markov Models[J]. IEEE Transactions on Reliability, 2012, 61(2): 491-503.

[33]陈发广, 周步祥, 曾澜钰. 基于多信息融合的变压器运行状态评估模型[J]. 电力系统及其自动化学报, 2013, 25(04): 140-144.

[34]Ali J B, Chebel-Morello B, Saidi L, et al. Accurate bearing remaining useful life prediction based on Weibull distribution and artificial neural network[J]. Mechanical Systems & Signal Processing, 2015, 56-57: 150-172.

[35]李鹏, 李丹, 李喜旺, 等. 基于模糊神经网络的电力调度自动化设备健康评估[J]. 计算机系统应用, 2019, 28(02): 207-212.

[36]徐岩, 陈昕. 基于合作博弈和云模型的变压器状态评估方法[J]. 电力自动化设备, 2015, 35(3): 88-93.

[37]Mandal S, Maiti J. Risk analysis using FMEA: Fuzzy similarity value and possibility theory based approach [J]. Expert Systems with Applications. 2014, 41, 3527-3537.

[38]Merle G, Roussel M, Lesage J, et al. Quantitative Analysis of Dynamic Fault Trees Based on the Coupling of Structure Functions and Monte Carlo Simulation [J]. Quality & Reliability Engineering, 2014, 32(1): 7-18.

[39]J. Schmidhuber, Deep learning in neural networks: An overview. Neural Netw, vol. 61, pp. 85–117, Jan. 2015.

[40]Ge D, Yang Y. Reliability analysis of non-repairable systems modeled by dynamic fault trees with priority AND gates [J]. Applied Stochastic Models in Business & Industry, 2015, 31(6): 809-822.

[41]Chen Yanglong, Ma Yanheng, Yin Yuanwei. Equipment State Assessment Based on Convolutional Neural Network, Journal of Physics: Conference Series. 2019;40:48-52.

[42]LAI Yue-hua, DONG Hai-ping, CHENG Tao, et al. Reliability Estimation of Aviation Pyrotechnics System Based On Mixed Beta Distribution [J]. Journal of Donghua University (English Edition), 2014, 31(06): 785-788.

[43]罗毅飞, 王磊, 黄永乐, 等. 焊接型电力电子器件失效机理及量化评估方法[J]. 中国电力, 2019, 52(09): 38-47.

[44]Thaduri A, Verma A K, Kumar U. Comparison of failure characteristics of different electronic technologies by using modified physics-of-failure approach[J]. International Journal of System Assurance Engineering & Management, 2015, 6(2):1-8.

[45]周宁, 马建伟, 胡博, 等. 基于故障树分析的电力变压器可靠性跟踪方法[J]. 电力系统保护与控制, 2012(19): 72-77.

[46]黄晓明, 凌万水, 吴栋萁. 基于故障树分析法的配电自动化实用化运维指标研究[J]. 电力系统保护与控制, 2017(24): 98-104.

[47]董兴辉, 张鑫淼, 郑凯, 等. 基于组合赋权和云模型的风电机组健康状态评估[J].太阳能学报, 2018, 39(08): 2139-2146.

[48]韩洋, 刘混举. 基于模糊层次分析法的煤机装备齿轮健康评价[J]. 煤炭技术, 2018, 37(05): 279-282.

[49]彭道刚, 陈跃伟, 范俊辉, 等. 基于层次分析法和粗糙集的变压器状态评估研究[J]. 高压电器, 2019, 55(07): 150-157.

[50]耿苏杰, 王秀利. 基于模糊贝叶斯网络的电力设备故障诊断和状态评估[J]. 计算机集成制造系统, 2016, 96(03): 775-783.

[51]赵东明, 程焱明, 曹明. 基于卷积神经网络的柴油发电机健康评估[J]. 计算机科学, 2018, 45(S2): 152-154.

[52]Yang Zhichun, Shen Yu, Yang Fan, Wan Zilin.A comprehensive real‐time state evaluation strategy for distribution equipment based on MI-PSO-BP algorithm. 2019;9(2): 1345-1353.

[53]曹惠玲, 黄乐腾, 李志伟, 等. 基于SOM神经网络的航空发动机滑油系统健康评估[J]. 中国民航大学学报, 2014, 32(06): 19-22.

[54]张春, 舒敏. 基于支持向量机的健康状态评估方法[J]. 计算机系统应用, 2018, 27(03): 18-26.

[55]翟文睿. 采煤机性能退化评估研究[D]. 徐州：中国矿业大学, 2020.

[56]雷一楠. 采煤机健康状态监测与识别方法研究[D]. 西安：西安科技大学, 2020.

[57]司垒. 采煤机智能控制关键技术研究[D]. 徐州：中国矿业大学, 2015.

[58]龚思丞, 黄文焘, 邰能灵, 等. 基于复杂网络的电热微网拓扑综合评估方法[J]. 电力系统自动化, 2019, 43(23): 173-189.

[59]徐敬友, 陈冲, 罗纯坚, 等. 基于改进复杂网络模型的电网关键环节辨识[J]. 电力系统自动化, 2016, 40(10): 53-61.

[60]Zhu Weixing, Zhi xue, Wang Guolin, et al. Capability-based context ontology modeling and reasoning for C~4ISR communication[J]. Journal of Systems Engineering and Electronics, 2016, 27(04): 845-857.

[61]Wei K C, Kyritsis V, Katsaros K V, et al．Resilience of interdependent communication and power distribution networks against cascading failures[C]//Ifip Networking Conference. Vienna, Austria：IEEE, 2016：37-45．

[62]Sen A, Mazumder A, Banerjee J, et al. Identification of K most vulnerable nodes in multi-layered network using a new model of interdependency[C]//Computer Communications Workshops .Toronto, Canada：IEEE, 2014：831-836．

[63]Tian Meng, Wang Xianpei, Dong Zhengcheng, et al. Cascading failures of interdependent modular scale-free networks with different coupling preferences[J]. Europhysics Letters, 2015, 111(1)：18007．

[64]何宗政, 雷一楠, 曹现刚, 等. 基于劣化度的采煤机健康状态评价方法研究[J]. 煤矿机械, 2019,40(12): 54-57.

[65]董泽,贾昊.基于EWT-LOF的热工过程数据异常值检测方法[J].仪器仪表学报,2020,41(02):126-134.

[66]Liu F T，Kai M T，Zhou Z H . Isolation Forest［C］//Eighth IEEE International Conference on Data Mining. IEEE, 2008.

[67]徐岩, 刘婧妍, 张诗杭, 等. 基于遗传算法的直流配电网线路故障定位方法[J]. 太阳能学报, 2020, 41(12): 1-8.

[68]李彬, 张云, 王立平, 等. 基于遗传算法优化小波神经网络数控机床热误差建模[J]. 机械工程学报, 2019, 55(21): 215-220.

[69]曹现刚, 李彦川, 雷卓, 等. 采煤机健康状态智能评估方法研究[J]. 工矿自动化, 2020, 46(06): 41-47.

[70]文楠, 严波, 林翔, 等. 基于BP神经网络的导线脱冰跳跃高度预测模型[J]. 振动与冲击, 2021, 40(01):199-204.

[71]何华锋, 何耀民, 徐永壮. 基于改进型BP神经网络的导引头测高性能评估[J]. 系统工程与电子技术, 2019, 41(07): 1544-1550.

[72]高玉明, 张仁津. 基于遗传算法和BP神经网络的房价预测分析[J]. 计算机工程, 2014, 40(04): 187-191.

[73]邓伟锋, 李振璧. 基于GA优化BP神经网络的微电网蓄电池健康状态评估[J]. 电测与仪表, 2018,55(21): 56-60+85.

[74]兰建义, 乔美英. 费用成本下煤矿综采设备检修周期的优化分析[J]. 煤矿机械, 2011, 32(04): 218-220.

[75]廖伟涵. 基于机器学习与自然语言处理的变压器故障诊断与状态评价研究[D]. 杭州：浙江大学, 2020.

[76]黄颖坤, 金炜东, 余志斌, 等. 基于深度学习和集成学习的辐射源信号识别[J]. 系统工程与电子技术, 2018, 40(11): 2420-2425.

[77]CHEN T, GUESTRIN C. XGBoost:A scalable tree boosting system [C]. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco,CA, United states, 2016: 785-794.

[78]李劲彬, 夏天, 黄烁, 等. 基于XGBoost的集成式隔离断路器状态评估[J]. 高电压技术, 2020, 46(05): 1800-1806.

[79]王志杰. 基于LabVIEW的高速列车轴承故障监测诊断系统研究[D]. 大连：大连交通大学, 2019.

[80]李达. 基于LabVIEW的武器综合测试与评估系统软件设计[D]. 哈尔滨：哈尔滨工业大学, 2016.