采煤机作为现代煤炭开采中最核心、重要的设备，是集电力电子、计算机技术、传感器技术为一体的大型复杂系统，保证它的健康稳定运行是煤炭高效生产的关键。加强对采煤机运行状态监测、准确识别采煤机所处健康状态等级以及基于采煤机当前状态制定维护生产计划，对煤炭安全、高效开采具有重要意义。本文通过分析采煤机易发生故障的部位进而选取合适的状态监测参数，在对采煤机监测数据进行异常值处理和构建采煤机健康状态评估体系的基础上，研究基于数据驱动的采煤机健康状态识别方法，并设计开发采煤机健康管理系统。

首先，分析采煤机基本结构，总结常见故障发生位置及故障产生原因，阐述采煤机健康状态识别系统基本框架。从采煤机监测参数的选取、数据预处理方法、健康状态评估理论等方面展开探索，为采煤机健康状态识别工作提供理论基础。

其次，针对采煤机监测数据会不可避免的产生异常值问题，影响状态识别效率，研究基于拉格朗日插值法改进箱型图的异常值处理方法，从而剔除异常值并补充缺失值；针对采煤机监测参数间存在冗余和关联的问题，基于灰色关联度分析方法筛选采煤机状态评估关键指标，构建采煤机健康状态评估指标体系。

然后，针对采煤机健康状态识别方法中存在指标权重确定困难、数据集高维、单一算法抗干扰能力差等问题，建立基于粒子群算法（PSO）优化支持向量分类（SVC）的采煤机健康状态识别模型，并验证算法的实用性；针对采煤机数据集复杂、不平衡等特性导致健康状态识别效率不理想的问题，建立基于卷积神经网络（CNN）和极端梯度提升算法（XGBoost）相结合的模型，首先利用卷积神经网络完成采煤机数据样本特征提取，再利用极端梯度提升算法进行状态识别，经过实验对比，验证算法可行性。

最后，开发采煤机健康管理系统，可实现管理人员登录、数据监测、故障预警、状态评估、信息管理等功能。以采煤机运行状态数据为基础，完成基于PSO-SVC和CNN-XGBoost的采煤机健康状态识别模型的实验分析，并对比模型效果，为采煤机健康状态识别和健康管理提出有效方案。

As the most core and important equipment in modern coal mining, the shearer is a large and complex system integrating power electronics, computer technology and sensor technology. Ensuring its healthy and stable operation is the key to efficient coal production. Strengthening the monitoring of the operating state of the shearer, accurately identifying the health state level of the shearer, and formulating maintenance and production plans based on the current state of the shearer are of great significance to the safe and efficient mining of coal. In this paper, by analyzing the parts that are prone to failure of the shearer and then selecting the appropriate condition monitoring parameters, on the basis of processing the abnormal value of the shearer monitoring data and constructing the shearer health status evaluation system, the data-driven coal mining is studied. The method of identifying the health state of the shearer is designed and developed, and the health management system of the shearer is designed and developed.

First, the basic structure of the shearer is analyzed, the common fault locations and causes are summarized, and the basic framework of the shearer health status recognition system is expounded. From the selection of shearer monitoring parameters, data preprocessing methods, and the theory of health status assessment, it provides a theoretical basis for the work of shearer health status identification.

Secondly, in view of the problem of outliers that will inevitably occur in the monitoring data of the shearer, which will affect the efficiency of state identification, the outlier processing method based on the Lagrangian improved box plot is studied, so as to eliminate the outliers and supplement the missing values; for coal mining Due to the redundancy and correlation between monitoring parameters of the shearer, the research is based on the grey correlation analysis method to screen the key indicators of shearer state evaluation, and build a shearer health state evaluation index system.

Then, in view of the difficulties in determining the index weight, the high dimensionality of the data set, and the poor anti-interference ability of a single algorithm in the identification method of the health state of the shearer, a shearer based on the particle swarm optimization (PSO) optimization support vector classification (SVC) was established. Health status recognition model, and verify the practicability of the algorithm; for the problem of unsatisfactory health status recognition efficiency caused by the complex and unbalanced data set of shearers, convolutional neural network (CNN) and extreme gradient boosting algorithm (XGBoost) Combined, firstly, the convolutional neural network is used to complete the feature extraction of shearer data samples, and then the extreme gradient boosting algorithm is used to identify the state. After experimental comparison, the feasibility of the algorithm is verified.

Finally, a shearer health management system is developed, which can realize the functions of management personnel login, data monitoring, fault warning, status evaluation, and information management. Based on the shearer operating state data, the experimental analysis of the shearer health state recognition model based on PSO-SVC and CNN-XGBoost was completed, and the model effects were compared. Propose an effective scheme for the identification and management of the health status of the shearer.

[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].2021.

[3]袁亮.开展基于人工智能的煤炭精准开采研究,为深地开发提供科技支撑[J].科技导报,2017, 35(14):1-1.

[4]2018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,2018

[5]王金华,黄乐亭,李首滨,黄曾华.综采工作面智能化技术与装备的发展[J].煤炭学报,2014,39(08):1418-1423.

[6]《能源技术革命创新行动计划(2016-2030年)》和《能源技术革命重点创新行动路线图》[J].电力与能源,2016,37(05):609.

[7]王国法,范京道,徐亚军,任怀伟.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(02):5-12.

[8]王国法,王虹,任怀伟,赵国瑞,庞义辉,杜毅博,张金虎,侯刚.智慧煤矿2025情景目标和发展路径[J].煤炭学报,2018,43(02):295-305.

[9]何满潮,马资敏,郭志飚,陈上元.深部中厚煤层切顶留巷关键技术参数研究[J].中国矿业大学学报,2018,47(03):468-477.

[10]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.

[11]刘送永.采煤机滚筒截割性能及截割系统动力学研究[D].中国矿业大学,2009.

[12]邓炳杰.面向不同寿命阶段的机电设备健康状态评价研究[D].重庆大学, 2014.

[13]季瑞.采煤机健康管理关键技术研究[D].中国矿业大学,2014.

[14]张兆连.电牵引双滚筒采煤机[J].煤炭科学技术,1978(08):59-60.

[15]王修义.三井三池400KW滚筒采煤机[J].煤炭技术,1987(02):13-14+7.

[16]王国法.煤矿综采自动化成套技术与装备创新和发展[J].煤炭科学技术,2013,41(11):1-5+9.

[17]吴晓华.CAN-bus系统在DBT EL300型采煤机上的应用[J].中国煤炭,2005(09):43-44.

[18]李国栋,崔旭方,杨旭宏,张鲲鹏.开关磁阻电机调速系统在EL—1000采煤机中的应用[J].煤,2004(05):28-29.

[19]徐国强.久弥坚艾柯夫SL1000型双截割滚筒采煤机[J].矿业装备,2013(08):70-71.

[20]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.

[21]曹秀娟,白跃新.国内切割功率最大采煤机通过鉴定[N].山西日报,2016-01-10.

[22]索智文,赵亦辉,周展.超大采高采煤机智能诊断电气控制系统研究[J].煤炭工程,2020,52(06):15-19.

[23]李国发,王彦博,何佳龙,王继利.机电装备健康状态评估研究进展及发展趋势[J].吉林大学学报(工学版),2022,52(02):267-279.

[24]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.

[25]Merle G, Roussel J M, Lesage J, Perchet V, Vayatis N. Quantitative Analysis of Dynamic Fault Trees Based on the Coupling of Structure Functions and Monte Carlo Simulation [J]. Quality & Reliability Engineering,2016,32(1):7-18.

[26]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.

[27]邱文昊,黄考利,连光耀,张西山,田杰.基于不确定性与重要度的改进DSmT健康状态评估[J].航空动力学报,2017,32(01):96-104.

[28]Tamilselvan P, Wang P F. Failure diagnosis using deep belief learning based health state classification[J]. Reliability Engineering& System Safety, 2020,115:124-135.

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

[30]Wen J, Gao H L. Degradation assessment for the ball screw with variational autoencoder and kernel density estimation[J]. Advances in Mechanical Engineering, 2018,10(9):1-12.

[31]Liu Z L, Zuo M J, Qin Y. Remaining useful life prediction of rolling element bearings based on health state assessment[J]. Proceedings of the Institution of Mechanical Engineers Part C-Journal of Mechanical Engineering Science, 2016, 230(2): 314-330.

[32]李巍华,李静,张绍辉.连续隐半马尔科夫模型在轴承性能退化评估中的应用[J].振动工程学报,2014,27(04):613-620.

[33]Lu C, Li T Y, Liu H M. Online milling tool condition monitoring with a single continuous hidden Markov models approach[J]. Journal of Vibroengineering, 2017, 16(5): 2448-2457.

[34]黄丽琴,赵红红,刘辉,许鹤鹏,孙杰,徐建文,万英杰,陈启明.基于层次故障树的闸门健康状态评估系统研究[J].湖北工业大学学报,2019,34(02):22-27.

[35]Hanachi H, Jie L, Banerjee A, Ying C, Koul A. A physics based modeling approach for performance monitoring in gas turbine engines[J]. IEEE Transactions on Reliability, 2015, 64(1): 197-205.

[36]高千.基于故障树与电气试验的变压器可靠性研究[D].大连理工大学,2019.

[37]曹海斌,郝恒,渠继东.基于失效物理的电子学产品可靠性仿真研究[J].设备管理与维修,2018(13):34-37.

[38]杨光.大型风电机组故障诊断与状态综合评价方法研究[D].华北电力大学,2017.

[39]邓超,孙耀宗,李嵘,王远航,熊尧.基于隐Markov模型的重型数控机床健康状态评估[J].计算机集成制造系统,2013,19(03):552-558.

[40]张金春,张继军,曹彪.基于HMM的机载设备状态健康评估方法研究[J].计算机测量与控制,2015,23(02):491-493.

[41]赵文清,王强,牛东晓.基于贝叶斯网络的电抗器健康诊断[J].电力自动化设备,2013,33(01):40-43.

[42]孙同敏.基于DBN-SVM的航空发动机健康状态评估方法[J].控制工程,2021,28(06):1163-1170.

[43]张寒,刘卫东,潘志敏,栗娟.基于自适应概率神经网络的变压器健康状态评估[J].高压电器,2022,58(02):103-110.

[44]张勇.基于卷积神经网络的谐波减速器健康状态评估方法研究[D].重庆交通大学,2021.

[45]王玉梅,张胤.基于云模型与变权灰色模糊综合评判的真空断路器状态评估[J].制造业自动化,2020,42(02):94-98.

[46]夏湛然.船用电缆绝缘老化状态评估研究[J].舰船科学技术,2020,42(24):112-114.

[47]董雪旺,董煜.基于模糊综合评判的水泵运行状态评估及寿命预测[J].水利技术监督,2021(09):157-161.

[48]王亮,吕卫民,金永川.一种多类型证据的合成评估方法[J].控制与决策,2014,29(11):1973-1978.

[49]陶青.基于综合健康指数的空压机状态评估[D].安徽大学,2019.

[50]司垒.采煤机智能控制关键技术研究[D].中国矿业大学,2015.

[51]何宗政,雷一楠,曹现刚,罗璇,张树楠,宫钰蓉.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.

[52]曹现刚,李彦川,雷卓,雷一楠.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.

[53]李丹宁,郑闯.一种模糊神经网络的采煤机滚筒温度实时故障预警方法[J].煤炭科学技术,2021,49(S1):161-166.

[54]曹现刚,许欣,雷卓,李彦川.基于降噪自编码器与改进卷积神经网络的采煤机健康状态识别[J].信息与控制,2022,51(01):98-106.

[55]丁华,杨亮亮,杨兆建,王义亮.数字孪生与深度学习融合驱动的采煤机健康状态预测[J].中国机械工程,2020,31(07):815-823.

[56]岳加利,郝静,卢海林,余勇,孙舒畅,李培轩.桥梁结构监测数据异常值处理方法[J].武汉工程大学学报,2022,44(01):107-111.

[57]王亮,吕卫民,李伟,等.复杂系统健康状态评估技术现状及发展[J].计算机测量与控制, 2013, 21(4):830-832.

[58]郑锐.三参数威布尔分布参数估计及在可靠性分析中的应用[J].振动与冲击,2015,34(05):78-81.

[59]王盼锋,王寿喜,马钢,全青.基于PCA-PSO-SVM模型的海底多相流管道内腐蚀速率预测[J].安全与环境工程,2020,27(02):183-189.

[60]Bhattacharjee K, Pant M. Hybrid particle swarm optimization-genetic algorithm trained multi-layer perceptron for classification of human glioma from molecular brain neoplasia data [J]. Cognitive Systems Research, 2019, 58: 173-194.

[61]罗巍,卢博,陈菲,马腾.基于PSO-SVM及时序环节的数控刀架故障诊断方法[J].吉林大学学报(工学版),2022,52(02):392-399.

[62]Zeiler M D, Fergus R. Visualizing and understanding convolutional networks[C]//European conference on computer vision. Springer, Cham, 2014: 818-833.

[63]Simonyan K, Zisserman A. Very Deep Convolutional Networks for Large-Scale Image Recognition[J]. Computer Science, 2014.

[64]Szegedy C, Liu W, Jia Y, Sermanet P, Reed S, Anguelov D, Erhan D, Vanhoucke V, Rabinovich A. Going deeper with convolutions[C]//Proceedings ofthe IEEE conference on computer vision and pattern recognition. 2015: 1-9.

[65]He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 770-778.

[66]Howard A G, Zhu M, Chen B, Kalenichenko D, Wang W, Weyand T, Andreetto M, Adam H. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications[J]. arXiv preprint arXiv:1704.04861, 2017.

[67]Hu J, Shen L, Sun G. Squeeze-and-excitation networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2018: 7132-7141.

[68]Huang G, Liu Z, Laurens V, Weinberger K. Densely connected convolutional networks[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2017: 4700-4708.

[69]李全鹏.基于CNN和XGBoost的静态电压稳定裕度预测方法研究[D].郑州大学,2020.

[70]Krizhevsky A，Sutskever I，Hinton G．ImageNet classification with deep convolutional neural networks[J]．Advances in neural information processing systems，2012，25（2）：1097-1105．

[71]车畅畅,王华伟,倪晓梅,蔺瑞管,熊明兰.基于1D-CNN和Bi-LSTM的航空发动机剩余寿命预测[J].机械工程学报,2021,57(14):304-312.

[72]马怀祥,冯旭威,李东升,齐澍椿.基于CNN和XGBoost的滚动轴承故障诊断方法[J].中国工程机械学报,2021,19(03):254-259.

[73]黄颖坤,金炜东,余志斌,吴昀璞.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.

[74]Khoualdia T, Hadjadj A E, Bouacha K, Abdeslam D O. Multi objective optimization of ANN fault diagnosis model for rotating machinery using grey rational analysis in taguchi method[J]. International Journal of Advanced Manufacturing Technology, 2017, 89: 3009-3020.

[75]李劲彬,夏天,黄烁,方华亮,胡傲,韩柳.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.