采煤机作为煤矿生产的重要机械设备，其安全可靠性关系着整个综采工作面的生产效率，因此对采煤机状态进行监测以及寿命预测显得尤为重要。但采煤机系统结构复杂，运行工况恶劣，目前没有针对采煤机整机进行寿命预测的相关方法，从采煤机的单部件对其整机进行寿命预测已经不能满足采煤机剩余寿命预测需求，并且寿命预测时未考虑部件间关联性。针对以上问题，本文在分析采煤机整机机械结构的基础上，对采煤机进行了传感器测点部署，通过特征提取，数据融合等方法对整机状态指标进行构建，提出了针对于采煤机整机的剩余寿命预测方法，建立基于Self-Attention-LSTM的采煤机剩余寿命预测模型，并通过实验进行验证。本文主要内容如下：

首先，对基于深度学习的剩余寿命预测方法框架及相关技术进行详细介绍，面向设备剩余寿命预测准确度较低、关联性分析较少等问题，对相应技术及方法做出改进。

其次，分析采煤机的整机结构，根据采煤机结构特点选取传感器测点位置，使用小波折中阈值降噪对信号进行去噪处理，运用时域、频域以及集合经验模态分解（EEMD）特征提取方法对采煤机的振动信号进行特征提取，并对其温度、电流等信号进行时序处理，提出自注意编码网络（Self-Attention-Encoder，SFAE）数据融合方法，构建采煤机状态指标。

然后，针对目前设备剩余寿命预测方法中忽略部件间关联性的问题，提出基于Self-Attention-LSTM的设备剩余寿命预测方法。通过Self-Attention对输入的多部件信息进行关联性分析，对数据进行有效调整后输入至LSTM中，模型既考虑输入数据之间的关联关系又可进行高效预测。试验通过C-MAPSS公开数据集进行论证，论证结果表明本文提出的方法具有较好的预测效果。

最后，通过搭建采煤机全寿命周期数据采集模拟平台对所提出的方法进行实验验证。将采集到的设备数据进行筛选、降噪、特征提取，构建针对采煤机的剩余寿命预测模型，并对模型进行实验论证，论证结果表明，本文所提出的基于Self-Attention-LSTM的采煤机剩余寿命预测方法在实际应用中能达到良好的预测效果。

综上所述，本文提出面向采煤机的基于深度学习的剩余寿命预测方法，有效解决目前采煤机整机寿命预测方法面临的技术性问题，对煤矿企业安全稳定生产起到了良好的预知性维护的作用。

Shearer as an important coal production machinery and equipment, and its safety and reliability of the entire relationship mechanized mining face production efficiency, monitor the condition of the shearer and lifetime prediction is very important. However, the shearer system has a complex structure and poor operating conditions. At present, there is no relevant method for predicting the life of the entire shearer. The life prediction of the whole machine from the single component of the shearer can no longer meet the demand for the remaining life prediction of the shearer  Life expectancy demand, and the correlation between components is not considered in the life expectancy. In response to the above problems, this paper analyzes the mechanical structure of the shearer machine, deploys the sensor measurement points of the shearer machine, and constructs the state index of the whole machine through methods such as feature extraction and data fusion. The method for predicting the remaining life of the shearer is based on the Self-Attention-LSTM based prediction model of the remaining life of the shearer, and is verified by experiments. The main contents of this article are as follows:

Firstly, the framework of the remaining life prediction method based on deep learning and related technologies are introduced in detail, and the corresponding technologies and methods are improved for the problems of low accuracy of equipment remaining life prediction and less correlation analysis.

Secondly, analyze the whole machine structure of the shearer, select the position of the sensor measurement point according to the structural characteristics of the shearer, use the wavelet compromise threshold denoising to denoise the signal, and use the time domain, frequency domain and ensemble empirical mode decomposition (EEMD) feature extraction method extracts the features of the vibration signal of the shearer, and processes its temperature, current and other signals in time series, and proposes a self-attention-encoder (SFAE) data fusion method to construct the shearer Status indicator.

Then, aiming at the problem of ignoring the correlation between components in the current equipment remaining life prediction methods, a Self-Attention-LSTM-based equipment remaining life prediction method is proposed. The correlation analysis of the input multi-component information is carried out through Self-Attention, and the data is effectively adjusted and then input into the LSTM. The model considers the correlation between the input data and can also make efficient predictions. The experiment is demonstrated through the C-MAPSS data set, and the result of the demonstration shows that the method proposed in this paper has a better predictive effect.

Finally, the proposed method of experimental verification by Shearer build the whole life cycle simulation data collection platform. The collected equipment data is screened, noise-reduced, and feature extracted, and a residual life prediction model for the shearer is constructed, and the model is demonstrated experimentally. The demonstration results show that the Self-Attention-LSTM-based coal mining proposed in this paper The prediction method of the remaining life of the machine can achieve a good prediction effect in practical applications.

In summary, this paper proposes a deep learning-based remaining life prediction method for shearers, which effectively solves the technical problems faced by the current shearer life prediction methods, and plays a good predictive maintenance role for the safe and stable production of coal mining enterprises.

[1]王国法,刘峰,孟祥军,等. 煤矿智能化(初级阶段)研究与实践[J].煤炭科学技术,2019,47(08):1-36.

[2]刘建林,李泉新,杨伟锋,等. 采煤工作面煤岩界面探测定向孔设计施工与数据处理方法[P]. 陕西省：CN111485825A,2020-08-04.

[3]国家能源局新闻发言人. 国家能源局2021年一季度网上新闻发布会文字实录[J].中国电业,2021(02):26-29.

[4]王国法,张德生. 煤炭智能化总裁技术创新实践与发展展望[J].中国矿业大学学报,2018,47(3):459-467.

[5]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].工况自动化,2018,44(2):5-12.

[6]康传峰. 煤矿综机设备故障统计分析对策[J].内蒙古煤炭经济,2018(12):72+76.

[7]X. Xi, M. Chen and D. Zhou. "Remaining useful life prediction for non-Markovian degrading systems with multiple working conditions[C]," 2018 IEEE International Conference on Prognostics and Health Management (ICPHM), Seattle, WA, 2018, pp. 1-8.doi: 10.1109/ICPHM.2018.8448673.

[8]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.

[9]毛君,郭浩,陈洪月. 基于深度自编码网络的采煤机截割部减速器故障诊断[J].煤炭科学技术,2019,47(11):123-128.

[10] 中国煤炭工业协会.2018煤炭行业发展年度报告[R].北京:中国煤炭工业协会.2019.

[11] 李杰其,胡良兵. 基于机器学习的设备预测性维护方法综述[J].计算机工程与应用,2020,56(21):11-19.

[12] Hafiz Waqar Ahmad,Jeong Ho Hwang,et al. Probabilistic Fatigue Life Prediction of Dissimilar Material Weld Using Accelerated Life Method and Neural Network Approach[J].Computation,2019,7(1).

[13] 李媛媛,陈捷,洪荣晶,等. 基于模糊C均值的转盘轴承剩余寿命预测[J].轴 承,2017(03):50-55.

[14] 吴祎,王友仁. 基于变分模态分解和高斯过程回归的锂离子电池剩余寿命预测方法[J].计算机与现代化,2020(02):83-88.

[15] 于宁,孙业新,陈洪月. 基于多源数据融合的采煤机截割载荷预测方法[J/OL].中国机械工程:1-9[2021-03-15].

[16] 刘文溢,刘勤明,周林森. 基于高阶隐半马尔科夫模型的设备剩余寿命预测[J/OL].计算机集成制造系统:1-22[2021-03-15].

[17] 罗俊海,杨阳. 基于数据融合的目标检测方法综述[J].控制与决策,2020,35(01):1-15.

[18] 袁媛,方红彬,殷忠敏. 基于多数据融合的电机故障诊断方法研究[J/OL].电气传动:1-6[2021-03-15].

[19] 闫书法,马彪,郑长松,等. 基于劣化数据的综合传动装置剩余寿命预测[J].北京理工大学学报,2018,38(11):1126-1133.

[20] 张红,程传祺,徐志刚,等. 基于深度学习的数据融合方法研究综述[J].计算机工程与应用,2020,56(24):1-11.

[21] Dai Yi,Liu Bin. Robust video object tracking via Bayesian model averaging-based feature fusion[J].OPTICAL ENGINEERING,2016,55(8):083102.

[22] 古莹奎,承姿辛,朱繁泷. 基于主成分分析和支持向量机的滚动轴承故障数据融合分析[J].中国机械工程,2015,26(20):2778-2783.

[23] Shaoke Wan,Xiaohu Li,YanjingYin,et al.Milling chatter detection by multi-feature fusion and Adaboost-SVM[J].Mechanical Systems and Signal Processing,2021,156(7):107671.

[24] 于明,张吉俊,郭迎春,等. 基于多层级注意力融合的图像美学重定向[J/OL].激光与光电子学进展,2021:1-15.

[25] 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.

[26] 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.

[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,2020,40(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] Wang W. A model to predict the residual life of rollingelement bearings given monitored condition information to date[J]. IMA Journal of Management Mathematics,2002,13(1):

3-16.

[30] 石慧,宋仁旺,张岩,等. 基于核密度估计和随机滤波理论的齿轮箱剩余寿命预测方法[J].计算机集成制造系统,2020,26(03):632-640.

[31] 马晓勇,翟晓鹏,张艺馨,等. 用数理统计方法预测油套管腐蚀剩余寿命[J].断块油气田,2013,20(04):513-515.

[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] 赵申坤,姜潮,龙湘云. 一种基于数据驱动和贝叶斯理论的机械系统剩余寿命预测方法[J].机械工程学报,2018,54(12):115-124.

[34] 刘红梅,李连峰等. 一种基于工况识别和相似性匹配的变工况下航空发动机整机剩余寿命预测方法[P].G06F17/50;G06K9/62.2016-10-12.

[35] Zhang, B., Wang, H., Tang, Y. et al. Residual Useful Life Prediction for Slewing Bearing Based on Similarity under Different Working Conditions. Exp Tech 42, 279–289 (2018) doi:10.1007/s40799-018-0235-4

[36] 高山. 变工况滚动轴承剩余使用寿命研究[D].郑州大学,2019.

[37] Z. Gebraeel, M. A. Lawley. A neural network degradation model for computing and updating residual life distributions[J]. IEEE Transactions on Automation Science and Engineering,2008,5(1):154-163.

[38] 孟文俊,张四聪,淡紫嫣,等. 滚动轴承寿命动态预测新方法[J].振动.测试与诊断,2019,39(03):652-658+678.

[39] 陈自强. 基于LSTM网络的设备健康状况评估与剩余寿命预测方法的研究[D].中国科学技术大学,2019.

[40] YANG Yu, ZHANG Na, CHENG Jun-sheng. Global parameters dynamic learning deep belief networks and its application in rolling bearing life prediction[J]. Journal of vibration and shock,2019,38(10):199-205+249.

[41] WANG Yu-jing,WANG Shi-da,KANG Shou-qiang, et al. Prediction Method of Remaining Useful Life of Rolling Bearings Based on Improved GcForest[J/OL]. Proceedings of the Chinese Society of Electrical Engineering,2020,40(15) : 5032-5042.

[42] Lipton ZC,et al. A critical review of recurrent neural networks for sequence learning[EB/OL].CoRR2015:1–38.

[43] Wennian Yu, II Yong Kim, Chris Mechefske. An improved similarity-based prognostic algorithm for RUL estimation using an RNN autoencoder scheme[J]. Reliability Engineering & System Safety,2020,199(7):106926.

[44] 裴洪,胡昌华,司小胜,等. 基于机器学习的设备剩余寿命预测方法综述[J].机械工程学报,2019,55(08):1-13.

[45] 彭开香,皮彦婷,焦瑞华,等. 航空发动机的健康指标构建与剩余寿命预测[J]. 控制理论与应用,2020,37(04):713-720.

[46] 张继冬,邹益胜,邓佳林,等. 基于全卷积层神经网络的轴承剩余寿命预测[J]. 中国机械工程,2019,30(18):2231-2235.

[47] K.G.,R.K,S.J.,et al. LSTM: A Search Space Odyssey[J]. IEEE Transactions on Neural Networks and Learning Systems,2017,28(10):2222-2232.

[48] 唐旭,徐卫晓,谭继文,等. 基于LSTM的滚动轴承剩余使用寿命预测[J]. 机械设计,2019,36(s1):117-119.

[49] Jianjing Zhang, Peng Wang, Ruqiang Yan, et al. Long short-term memory for machine remaining life prediction[J]. Journal of Manufacturing Systems,2018,48(7):78-86.

[50] Jinglong Chen, Hongjie Jing, Yuanhong Chang, et al. Gated recurrent unit based recurrent neural network for remaining useful life prediction of nonlinear deterioration process[J].Reliability Engineering & System Safety,2019,185(6):372-382.

[51] Ya Song, Guo Shi, Leyi Chen, et al. Remaining Useful Life Prediction of Turbofan Engine Using Hybrid Model Based on Autoencoder and Bidirectional Long Short-Term Memory[J]. Journal of Shanghai Jiaotong University (Science),2018, 23(12):85-94.

[52] 李京峰,陈云翔,项华春,等. 基于LSTM-DBN的航空发动机剩余寿命预测[J].系统工程与电子技术,2020,42(07):1637-1644.

[53] 胡城豪,胡昌华,司小胜,等. 基于MSCNN-LSTM的滚动轴承剩余寿命预测方法[J].中国测试,2020,46(09):103-110.

[54] 陆继翔,张琪培,杨志宏,等. 基于CNN-LSTM混合神经网络模型的短期负荷预测方法[J].电力系统自动化,2019,43(08):131-137.

[55] 张永峰,陆志强. 基于集成神经网络的剩余寿命预测[J].工程科学学报,2020,42(10):1372-1380.

[56] Xiaolin Wang, Narayanaswamy Balakrishnan, Bo Guo. Residual life estimation based on nonlinear-multivariate Wiener processes[J]. Journal of Statistical Computation and Simulation, 2015, 85(9):1742-1764.

[57] ZHANG Hanwen, CHEN Maoyin,ZHOU Donghua. Remaining useful life prediction for a nonlinear multi-degradation system with public noise[J]. Journal of Systems Engineering and Electronics,2018,29(02):429-435.

[58] 林伟杰. 基于粒子滤波的多部件退化建模与剩余寿命预测方法研究[D].电子科技大学,2019.

[59] 白灿,胡昌华,张建勋,等. 随机冲击影响的多部件退化设备寿命预测方法[J].中国测试,2018,44(07):124-131.

[60] Sheng Xiang, et al. Long short-term memory neural network with weight amplification and its application into gear remaining useful life prediction[J]. Engineering Applications of Artificial Intelligence.2020,91(5):103587.

[61] Yuanhang Chen, et al. A novel deep learning method based on attention mechanism for bearing remaining useful life prediction[J]. Applied Soft Computing Journal.2020,86:105919.

[62] Y.Q.,S.X,Y.C.,H.C.Macroscopic–Microscopic Attention in LSTM Networks Based on Fusion Features for Gear Remaining Life Prediction[J].IEEE Transactions on Industrial Electronics, 2020,67(12):10865-10875.

[63] 马宏伟,吴少杰,曹现刚,等. 基于AGB组合模型的采煤机运行状态预测方法研究[J].煤矿机械,2019,40(06):37-39.

[64] 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.

[65] 秦超. 采煤机镐型截齿力学有限元仿真及剩余寿命预测研究[D].山东科技大学,2018.

[66] 刘文广,杨兰柱,秦波. LSTM网络在采煤机摇臂齿轮箱寿命预测中的应用[J].机电工程技术,2020,49(11):168-170.

[67] 丁华,杨亮亮,杨兆建,等. 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.

[68] Rui Zhao, Ruqiang Yan, Zhenghua Chen,et al. Deep learning and its applications to machine health monitoring[J]. Mechanical Systems and Signal Processing, 2019,115(01):213-237.

[69] 邓文韬. Attention机制在神经机器翻译中的作用探究[J].计算机产品与流通,2020(09):108-109.

[70] 李康康,张静. 基于注意力机制的多层次编码和解码的图像描述[J/OL].计算机应用:1-6[2021-03-15].

[71] 杨威,胡燕. 混合CTC/attention架构端到端带口音普通话识别[J].计算机应用研究,2021,38(03):755-759.

[72] 廖文雄,曾碧,徐雅芸. 结合一维扩展卷积与Attention机制的NLP模型[J].计算机工程与应用,2021,57(04):114-119.

[73] 徐瑞聪,刘瑞峰,曹佐,等. 图片类别的确定方法、装置、电子设备及存储介质[P]. 北京市：CN111858983A,2020-10-30.

[74] 邵清,马慧萍. 融合Self-Attention机制的卷积神经网络文本分类模型[J].小型微型计 算机系统,2019,40(06):1137-1141.

[75] Xinjian Gao, Zhao Zhang, Tingting Mu,et al. Self-attention driven adversarial similarity learning network[J]. Pattern Recognition,2020,105(9):107331.

[76] Sepp Hochreiter, Jürgen Schmidhuber. Long short-term memory[J]. Neural computation, 1997,9(8):1735–1780.

[77] ZHENG S,et al. Long short-term memory network for remaining useful life estimation[C]. 2017 IEEE International Conference on Prognostics and Health Management (ICPHM), 2017: 88-95.

[78] 采煤机电机工况特点及提高其寿命的初步意见[J].煤炭科学技术,1979(08):42-45.

[79] 崔海滨. 采煤机工况监测及故障诊断系统的研究开发[J].机械工程师,2016(09):225-227.

[80] 洪民江. 基于小波变换的语音信号去噪算法研究[D].南京邮电大学,2018.

[81] 刘立芳,杨海霞,齐小刚. 基于线性判别分析的时频域特征提取算法[J].系统工程与电子技术,2019,41(10):2184-2190.

[82] 李大中,赵杰,刘建屏,等. 基于EMD超声缺陷信号故障特征提取方法[J].华北电力技术,2015(07):1-6.

[83] 唐静,王二化,朱俊,等. 基于EEMD的特征提取及其在齿轮裂纹故障诊断中的应用[J].机床与液压,2020,48(20):161-166.

[84] Ahmed Elsheikh, Soumaya Yacout, Mohamed-Salah Ouali. Bidirectional handshaking LSTM for remaining useful life prediction[J]. Neurocomputing,2019,323(1):148-156.

[85] Xiang Li, Qian Ding, Jian-Qiao Sun. Remaining useful life estimation in prognostics using deep convolution neural networks[J]. Reliability Engineering & System Safety,2018, 172(4):1-11.

[86]Saxena A, Goebel K, Simon D,et al. Damage propagation modeling for aircraft engine run-to-failuresimulation[C].2008 international conference on prognostics and health management. IEEE, 2008,4711414(12): 1-9.

[87] Kunyuan Deng, Xiaoyong Zhang, Yijun Cheng, et al. A remaining useful life prediction method with long-short term feature processing for aircraft engines[J]. Applied Soft Computing,2020,93(8):106344.

[88] Saxena A, Geoebel K. PHM08 Challenge Data set[R]. NASA Ames Prognostics Data Repository NASA Ames Research Center, Moffett Field,CA,2008.