截割部是采煤机截割煤层的核心部件，其寿命是影响采煤机整体可靠性的关键，对截割部的状态监测与剩余寿命预测至关重要。截割部的运行工况复杂，在以往对截割部剩余寿命预测中，大多数学者通过建立截割部齿轮和轴承动力学仿真模型来预测其疲劳寿命，但是这样不能直接反应截割部的实时状态退化信息。因此本文通过分析截割部的传动系统和运行工况，安装传感器，采集运行状态数据，提取状态数据特征，融合数据信息，从数据驱动方向入手采用深度学习方法对截割部进行剩余寿命预测。本文主要工作包括：

首先，对截割部传动系统特点和运行工况加以说明，针对截割部的传动系统特点和实际工况条件部署传感器测点位置和种类，采集监测信号，分析信号中噪声原因，采用基于3sigma准则剔除异常值方法与研究一种基于改进小波阈值降噪方法实现信号去噪。

其次，针对目前对截割部剩余寿命预测时退化指标提取困难与预测精度低的问题，研究一种基于PCA-GRU的截割部剩余寿命预测方法。分析能够直接反应截割部运行状态的振动信号，提取振动信号的时域、频域特征。通过PCA对振动信号的多域原始特征集进行融合得到能够反映截割部衰退趋势的指标。通过GRU能够学习时间序列数据之间的相关性，实现对截割部退化状态的跟踪和剩余寿命的预测。

然后，针对采煤机截割部剩余使用寿命预测过程中状态数据维度高、数量大以及时间序列相关信息难以充分考虑的问题，研究一种基于MSCNN-GRU融合的采煤机截割部剩余寿命预测方法。构建剩余寿命预测标签，优化模型结构。通过MSCNN不同尺度的卷积核学习不同尺度的特征，提取数据空间详细特征，进行特征融合。结合GRU提取时间相关性特征，预测剩余使用寿命。

最后，通过搭建采煤机截割部仿真实验平台，获取实验平台全寿命周期数据，对截割部进行剩余寿命预测。采集实验数据，对数据进行筛选，特征提取。构建截割部剩余寿命预测模型。

综上所述本文所提的基于深度学习的采煤机截割部剩余寿命预测方法，能够有效解决截割部寿命预测所面临的退化指标提取困难、数据量大、维度高、时序性数据关联度难以挖掘等问题，对于采煤机的预测性维护提供一定的理论指导。

The cutting part is the core component of the shearer cutting the coal seam, and its life is the key to the overall reliability of the shearer, and it is very important for the condition monitoring and remaining life prediction of the cutting part. The operating conditions of the cutting part are complex. In the previous prediction of the remaining life of the cutting part, most scholars established the dynamic simulation model of the cutting part gear and bearing to predict its fatigue life, but this cannot directly reflect the cutting part. Real-time state degradation information. Therefore, this paper analyzes the transmission system and operating conditions of the cutting part, installs sensors, collects operating state data, extracts the characteristics of the state data, fuses the data information, and uses the deep learning method to predict the remaining life of the cutting part from the data-driven direction. The main work of this paper includes:

Firstly, the characteristics and operating conditions of the transmission system of the cutting part are explained. According to the characteristics of the transmission system of the cutting part and the actual working conditions, the positions and types of sensor measuring points are deployed, the monitoring signals are collected, and the reasons for the noise in the signals are analyzed. Criterion outlier removal method and research A method based on improved wavelet threshold denoising to achieve signal denoising.

Secondly, in order to solve the problems of difficulty in extracting the degradation index and low prediction accuracy when predicting the remaining life of the cutting part, a PCA-GRU-based residual life prediction method for the cutting part is studied. Analyze the vibration signal that can directly reflect the operating state of the cutting part, and extract the time domain and frequency domain features of the vibration signal. The multi-domain original feature set of the vibration signal is fused by PCA to obtain an index that can reflect the decline trend of the cutting part. Through the GRU, the correlation between time series data can be learned, and the tracking of the degradation state of the cutting part and the prediction of the remaining life can be realized.

Then, aiming at the problems of high latitude and large quantity of state data and the difficulty of fully considering the relevant information of time series in the process of predicting the remaining service life of shearer cutting part, a kind of remaining life of shearer cutting part based on MSCNN-GRU fusion is studied. method of prediction. Build remaining life prediction labels and optimize model structure. The features of different scales are learned through the convolution kernels of different scales of MSCNN, and the detailed features of the data space are extracted for feature fusion. Combined with GRU to extract time-dependent features, the remaining service life is predicted.

Finally, by building a simulation experimental platform for the cutting part of the shearer, the whole life cycle data of the experimental platform is obtained, and the remaining life of the cutting part is predicted. Collect experimental data, filter the data, and extract features. Build the remaining life prediction model of the cutting section.

In summary, the deep learning-based residual life prediction method of shearer cutting section proposed in this paper can effectively solve the difficulties in extracting degradation indicators, large amount of data, high dimension, and time-series data correlation faced in life prediction of cutting section. It provides certain theoretical guidance for the predictive maintenance of shearers.

[1]张宏. 煤炭行业经济运行形势分析与展望[J]. 中国煤炭工业，2020 (01): 10-13.

[2]本刊记者. 煤炭行业改革发展成效显著——中国煤炭工业协会在京召开2020煤炭行业年度新闻发布会[J]. 中国煤炭工业，2021 (03): 28-29.

[3]熊妍婷. 煤炭行业市场结构与企业竞争策略选择[J]. 煤炭工程，2020, 52(07): 196-200.

[4]刘睿卿. 基于深度学习的采煤机状态预测预警研究[D]. 西安科技大学.

[5]张强,张晓宇. 不同工况下采煤机滚筒截割性能研究[J]. 应用力学学报，2021, 38(06): 2360-2368.

[6]程泽银，丁华，杨亮亮. 采煤机摇臂剩余寿命预测系统设计与开发[J]. 煤矿机械，2021, 42(01): 16-18.

[7]王步康. 煤矿巷道掘进技术与装备的现状及趋势分析[J]. 煤炭科学技术，2020, 48(11): 1-11.

[8]牟凯，孙琳，尚文利，等. 基于传感器网络的烟机设备在线监控系统研究[J]. 微计算机信息，2010, 26(16): 120-121.

[9]李帅帅，任怀伟. 综采工作面“三机”设备位姿测量技术研究现状与展望[J]. 煤炭科学技术，2020, 48(09): 218-226.

[10]廖雯竹，潘尔顺，王莹，等. 统计模式识别和自回归滑动平均模型在设备剩余寿命预测中的应用[J]. 上海交通大学学报，2011, 45(07): 1000-1005.

[11]毛君，朱煜，陈洪月，等. 采煤机摇臂传动系统寿命预测与优化分析[J]. 机械设计，2018, 35(12): 52-58.

[12]周久华. 神东矿区采煤机故障统计与原因分析[J]. 煤炭科学技术，2015, 43(S2): 139-143.

[13]陈渊，马宏伟. 基于PSO-SVM的采煤机摇臂齿轮箱故障诊断研究[J]. 煤矿机械，2015, 36(10): 303-306.

[14]周久华. 采煤机摇臂齿轮箱故障诊断技术研究[D]. 重庆理工大学，2013.

[15]Yurek O E, Birant D. Remaining Useful Life Estimation for Predictive Maintenance Using Feature Engineering[C].2019 Innovations in Intelligent Systems and Applications Conference(ASYU), 2019, 1-5.

[16]藏义明，杨明波. 工业互联网条件下智能维修的预测性维护策略[J]. 设备管理与维修，2017, 19: 62-63

[17]Poor P,Basl J, Zenisek D. Predictive Maintenance 4.0 as next evolution step in industial maintenance development[C]. 2019 Inteernational Research Conference on Smart Computing and Systems Engineering(SCES), 2019 ,245-253.

[18]年夫顺. 关于故障预测与健康管理技术的几点认识[J]. 仪器仪表学报，2018, 39(08): 1-14.

[19]ZIO E, MAIO F D.A data-driven fuzzy approach for predicting the remaining useful life in dynamic failure scenarios of a nuclear system[J]. Reliability Engineering & System Safety, 2013,95(1): 49-57.

[20]魏建华，程起元，李广义. 医疗设备预知性维护技术的探究与实践[J]. 中国医疗设备，2019, 34(08): 128-130.

[21]Sutharssan T, Stoyanov S, Bailey C,et al. Prognostic and health management for engineering systems: a review of the data-driven approach and algorithms[J]. The Journal of Engineering, 2015, 2015(7): 215-222.

[22]FAN J J, YUNG K C , PECHT M. Physics-of-failure-based prognostics and health management for high-power white light-emitting diode lighting[J]. IEEE Transactions on Device and Materials Reliability, 2011, 11(3): 407-416.

[23]刘帮俊，陈荣刚，吴斌. 火炮身管失效机理和寿命预测[J]. 兵器装备工程学报，2016, 37(12): 121-125+149.

[24]Maio F D, Tsui K L, Zio E. Combining Relevance Vector Machines and exponential regression for bearing residual life estimation[J]. Mechanical Systems and Signal Processing, 2012,31.

[25]李文华，关欣，周露露，等. 环境应力下铁路电器触点失效机理分析及寿命建模[J]. 机车电传动，2017, 255(02): 14-17+23.

[26]梁宁，张志刚，刘翠翠，等. 电磁脉冲环境下HVDC换流阀晶闸管的失效机理及寿命模型分析[J].电力电容器与无功补偿，2019, 182(02): 142-146+171.

[27]白梁军，黄萌，饶臻，等. 基于GARCH模型的IGBT寿命预测[J]. 中国电机工程学报，2020, 653(18): 5787-5796.

[28]吴超勇. 大数据驱动的轴承寿命预测关键技术研究[D]. 华中科技大学，2017.

[29]Doksum k A,Hoyland A.Models for variable stress accelerated life testing experiments based on Wiener processrs and the inverse Gaussian distribution[J]. Technometrics, 1992, 34(1): 74-82.

[30]Huang Z Y, Xu Z G, Wang W H, Sun Y X. Remaininguseful life prediction for a nonlinear heterogeneous wienerprocess model with an adaptive drift.IEEE Transactionson Reliability, 2015, 64(2): 687−700.

[31]朱磊，左洪福，蔡景. 基于Wiener过程的民用航空发动机性能可靠性预测[J]. 航空动力学报，2013, 28(28): 1006-1012.

[32]赵广社,吴思思,荣海军.多源统计数据驱动的航空发动机剩余寿命预测方法[J].西安交通大学学报，2017, 51(11): 150-155+172.

[33]SIMON H.Neural network:A comprehensive foundation[M]. New Jersey:Prentice Hall PTR,1994.

[34]WANG L X,WU Z H,FU Y D,et al.Remaining life predictions of fan based on time series analysis and BP neural networks[C]//Information Technology,Networking, Electronic and Automation Control Conference, IEEE. 2016: 607-611.

[35]SANTHOSH T V, GOPIKA V，GHOSH A K，et al. An approach for reliability prediction of instrumentation and control cables by artificial neural networks and weibull theory for probabilistic safety assessment of NPPs[J]. Reliability Engineering and System Safety, 2018, 170: 31-44.

[36]李志刚，刘伯颖，李玲玲，等. 基于小波包变换及RBF神经网络的继电器寿命预测[J]. 电工技术学报，2015, 30(14): 233-240.

[37]王烨，左洪福，蔡景，等. 基于Bayesian推断和LS-SVM的发动机在翼寿命预测模型[J]. 南京理工大学学报，2013, 37(6): 955-959.

[38]禹鑫燚，施甜峰，唐权瑞，等. 面向预测性维护的工业设备管理系统[J]. 计算机科学，2020, 47(S2): 667-672+677.

[39]程梦瑶. 大数据驱动下的PHM技术[J]. 软件和集成电路，2016 (05): 22-25.

[40]李天梅,司小胜,刘翔,裴洪.大数据下数模联动的随机退化设备剩余寿命预测技术[J/OL].自动化学报:1-23[2022-04-15].

[41]刘惠，刘振宇，郏维强，等. 深度学习在装备剩余使用寿命预测技术中的研究现状与挑战[J]. 计算机集成制造系统，2021, 27(01): 34-52.

[42]周福娜，高育林，王佳瑜，等. 基于深度学习的缓变故障早期诊断及寿命预测[J]. 山东大学学报，2017, 47(5): 30-37.

[43]REN L, SUN Y, WANG H,et al. Prediction of bearing remaining useful life with deep convolution neural network[J]. IEEE Access, 2018, 6(99): 13041-13049.

[44]MALHI A, YAN R, GAO R X. Prognosis of defect propagation based on recurrent neural networks[J]. IEEE Transactions on Instrumentation & Measurement, 2011, 60(3): 703-711.

[45]ZHANG Y Z, XIONG R, HE H W,et al.Long short-term memory recurrent neural network for remaining useful life prediction of lithium-ion batteries[J]. IEEE Transactions on Vehicular Technology, 2018, 67(7): 5695-5705.

[46]姚德臣，李博阳，刘恒畅，等. 基于注意力GRU算法的滚动轴承剩余寿命预测[J]. 振动与冲击，2021, 40(17): 116-123.

[47]赵丽娟，杨程程. 薄煤层采煤机截割部可靠性分析[J]. 制造业自动化，2012, 34(02): 4-6.

[48]毛君，姜鹏，谢苗. 斜切状态下滚筒采煤机液压拉杠力学分析与寿命预测[J]. 工程设计学报，2015, 22(01): 95-100.

[49]张啸，刘混举，郭生龙，等. 采煤机截割部行星架疲劳寿命的仿真分析[J]. 矿山机械，2013, 41(06): 18-21.

[50]陈洪月，焦思维，张坤，等. 实际工况下采煤机行走机构的啮合应力与疲劳寿命研究[J]. 机械强度，2018, 40(06): 1456-1462.

[51]罗璇. 数据驱动的采煤机摇臂寿命预测方法研究[D]. 西安科技大学, 2020.

[52]鲁翅，赵强，潘小兵. 深度探讨采煤机应用与维修技术[J]. 科技资讯，2011(13): 114.

[53]张露, 丁艳红, 赵津, 等. 基于FTA和FMEA滚筒采煤机截割头潜在故障预测[J]. 煤炭技术， 2017, 36(05): 219-221.

[54]程亮，张步勤，张金营. 基于IGWO优化LSSVM的采煤机截割部齿轮箱故障诊断[J]. 煤矿机械，2021, 42(05): 168-171.

[55]王海花，林邓伟，霍晓丽. 基于粗糙集-RBF神经网络的采煤机截割部传动系统故障诊断[J]. 煤矿机械，2021, 42(05): 175-177.

[56]田震，荆双喜，赵丽娟，等. 采煤机噪声与振动特性研究[J]. 工矿自动化，2019, (03): 23-28.

[57]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.

[58]戴豪民，许爱强，孙伟超. 基于改进奇异谱分析的信号去噪方法[J]. 北京理工大学学报，2016, 36(07): 727-732+759.

[59]张天赐，庞新宇，杨兆建. 自适应小波阈值融合去噪法对采煤机振动信号的处理[J]. 太原理工大学学报，2016, 47(02): 170-173+177.

[60]王晓蓉，王海超，王春光.基于自适应阈值函数的铡草机振动信号去噪[J]. 农机化研究，2021, 43(11): 14-18+23.

[61]雷卓. 基于LSTM的采煤机剩余寿命预测方法研究[D]. 西安科技大学，2021.

[62]戴邵武，陈强强，丁宇. 基于时域特征的滚动轴承寿命预测[J].计算机测量与控制，2019, 27(10): 60-63.

[63]邹旺，吉畅，陈伟兴，郑凯. 基于数据增强和卷积神经网络的多轴承剩余寿命预测[J].机械设计，2021, 38(08): 84-90.

[64]郭凯维，郭传超，史耀凡，等. 基于主成分分析的GA-BP神经网络地表下沉系数预测[J]. 北京测绘，2021, 35(11): 1374-1379.

[65]陈静，谭爱国，钟建伟. 一种基于GRU神经网络的电力负荷预测[J]. 电子质量，2022 (02): 122-126.

[66]徐海兵，郭久明. 基于双向GRU模型的网络流量预测的研究[J]. 电子技术应用，2022, 48(02): 19-22+27.

[67]王璞，姬联涛，朱家浩，等. 基于VMD与GRU的抽水蓄能机组振动趋势预测[J]. 水电能源科学，2022, 40(01): 192-195+205.

[68]李京泰，王晓丹. 基于代价敏感激活函数XGBoost的不平衡数据分类方法[J/OL]. 计算机科学: 1-17[2022-03-26].

[69]罗宇，李颖，郝昕宇，等. 基于相关性的传感数据分析与处理[J]. 智能计算机与应用，2022, 12(02): 49-53.

[70]Abdel-Hamia O,Mohamed A, Hui J,et al.Applying convolutional neural networks concepts to hybrid NN-HMM model for speech recognition[C]//Proceedings of IEEE international conference on acoustics,speech and signal processing. 2012:4277-4280.

[71]ZHAO J, MAO X, CHEN L.Speech emotion recognition using deep 1D &2D CNN LSTM networks[J]. Biomedical Signal Processing and Control, 2019, 47: 312-323.

[72]曹正志，叶春明. 基于并联CNN-SE-Bi-LSTM的轴承剩余使用寿命预测[J]. 计算机应用研究，2021, 38(07): 2103-2107.