综采智能化是煤矿智能化开采技术发展的重要方向，采煤机与刮板输送机的协同控制是实现综采装备智能化和综采工作面无人化的前提。本文以蕴含着丰富采煤机与刮板输送机协同控制策略和运行逻辑的综采装备历史运行数据为基础，研究采煤机-刮板输送机协同调速方法，通过对粗糙集理论和深度学习的研究，设计了基于粗糙神经网络的采煤机-刮板输送机协同调速方法，主要工作及研究成果如下：

（1）采煤机-刮板输送机协同调速关键技术的提出。分析采煤机-刮板输送机协同工作过程，结合大数据智能矿山背景，确定采煤机-刮板输送机协同调速基本思路，提出基于粗糙神经网络的采煤机-刮板输送机协同调速方法，建立协同调速系统框架，为后续而研究奠定基础。

（2）采煤机-刮板输送机协同调速知识获取研究。基于粗糙集理论知识获取对采煤机与刮板输送机协同过程监测进行参数选择，从繁多的特征信息中去除冗余特征，提取关键特征，获得采煤机与刮板输送机协同调速的最小知识表达，建立特征选择数据集，为后续神经网络学习做准备。

（3）采煤机-刮板输送机协同调速神经网络学习方法研究。基于深度LSTM神经网络建立粗糙神经网络，建立采煤机-刮板输送机协同调速过程模型，对不同超参数对协同调速预测精度的影响进行仿真和分析，选取优化模型参数，并与其他学习算法比较分析优化模型进行连续预测的准确性。

（4）采煤机-刮板输送机协同调速方法现场数据验证。为进一步验证本文提出的基于粗糙神经网络的采煤机-刮板输送机协同调速方法的有效性，采用榆家梁煤矿43101采煤机和刮板输送机的运行数据对算法进行验证，结果表明，本文提出的粗糙神经网络算法比未经过粗糙集理论知识获取的神经网络算法准确率更高，可以更好的拟合采煤机-刮板输送机协同工作过程，实现采煤机-刮板输送机协同稳定运行。

本文所提出的基于粗糙神经网络的采煤机-刮板输送机调速方法，为工作面协同控制提供了一条新的解决途径，为综采工作面智能化的发展做出了有益探索，为实现煤矿低碳化、智能化和无人化开采奠定了基础。

The intelligentization of fully mechanized mining is an important direction for the development of intelligent mining technology in coal mines. The coordinated control of the shearer and the scraper conveyor is the premise to realize the intelligentization of fully mechanized mining equipment and the unmanned working face of fully mechanized mining. Based on the historical operation data of fully mechanized mining equipment, which contains rich cooperative control strategies and operation logic of shearer and scraper conveyor, this paper studies the cooperative speed regulation method of shearer and scraper conveyor. In the study of learning, a shearer-scraper conveyor cooperative speed regulation method based on rough neural network was designed. The main work and research results are as follows:

(1) The key technology of shearer-scraper conveyor coordinated speed regulation is proposed. The cooperative working process of shearer-scraper conveyor is analyzed, combined with the background of big data intelligent mine, the basic idea of shearer-scraper conveyor cooperative speed regulation is determined, and the shearer-scraper conveyor collaboration based on rough neural network is proposed. Speed regulation method, establish a coordinated speed regulation system framework, and lay the foundation for subsequent research.

(2) Research on knowledge acquisition of cooperative speed regulation of shearer-scraper conveyor. Based on the knowledge acquisition of rough set theory, the parameters of the cooperative process monitoring of the shearer and the scraper conveyor are selected, the redundant features are removed from the various feature information, the key features are extracted, and the cooperative speed regulation of the shearer and the scraper conveyor is obtained. Minimal knowledge representation, build feature selection dataset, prepare for subsequent neural network learning.

(3) Study on the learning method of shearer-scraper conveyor cooperative speed regulation neural network. Based on the deep LSTM neural network, a rough neural network was established to model the cooperative speed regulation process of the shearer and the scraper conveyor. The influence of different hyperparameters on the prediction accuracy of the cooperative speed regulation model was simulated and analyzed. Select and optimize, and compare and analyze the predictive ability of the optimized collaborative speed regulation model with other shallow learning algorithms.

(4) On-site data verification of the shearer-scraper conveyor cooperative speed regulation method. In order to further verify the effectiveness of the shearer-scraper conveyor cooperative speed regulation method based on rough neural network proposed in this paper, the algorithm was verified by using the operation data of 43101 shearer and scraper conveyor in Yujialiang Coal Mine. It shows that the rough neural network algorithm proposed in this paper has higher accuracy than the neural network algorithm without rough set theory knowledge, and can better fit the cooperative working process of the shearer-scraper conveyor and realize the shearer-scraper The plate conveyors work together and stably.

The shearer-scraper conveyor speed regulation method based on rough neural network proposed in this paper provides a new solution for collaborative control of working face, and makes a useful exploration for the development of intelligentization of fully mechanized working face. The realization of low-carbon, intelligent and unmanned mining of coal mines has laid the foundation.

[1] 孙超,姜琳,袁广玉.“十四五”期间我国煤炭供需趋势分析[J].煤炭工程,2021, 53(05):193-196.

[2] 胡彦勇,张瑞,郄晓彤,刘虹.全生命周期下中国煤炭资源能源碳排放效率评价[J/OL].中国环境科学:1-18[2022-03-26].

[3] 刘闯,蓝晓梅.世界煤炭供需形势分析[J].中国煤炭,2020,46(04):99-104.

[4] 刘峰,曹文君,张建明,曹光明,郭林峰.我国煤炭工业科技创新进展及“十四五”发展方向[J].煤炭学报, 2021,46(01):1-15.

[5] 范京道.煤矿智能化开采技术创新与发展[J].煤炭科学技术,2017,45(09):65-71.

[6] 王国法.综采自动化智能化无人化成套技术与装备发展方向[J].煤炭科学技术,2014,42(09):30-34+39.

[7] 王国法,刘峰,庞义辉,任怀伟,马英.煤矿智能化——煤炭工业高质量发展的核心技术支撑[J].煤炭学报,2019,44(02):349-357.

[8] Guimei, Wang, Shanlin, Jiao, Guangxing, Cheng. Fully mechanized coal mining technology for thin coal seam under complicated geol ogical conditions[J] Energy Expl oration and Expl oitation, 2011, 29 (2): 169-178.

[9] 黄曾华,王峰,张守祥.智能化采煤系统架构及关键技术研究[J].煤炭学报,2020, 45(06):1959-1972.

[10] 付国军. 自动化综采工作面概念探讨[J].工矿自动化,2014, 06: 26-30.

[11] Wanghong. The 40 years developmental review of the fully mechanized mine roadway headingtechnology in China[J]. Journal of China Coal Society, 2010, 35(11): 1815-1820.

[12] 黄曾华.综采工作面自动化控制技术的应用现状与发展趋势[J].工矿自动化,2013,10:17-21.

[13] 徐刚.综采工作面配套技术研究[J].煤炭学报,2010,11: 1921-1924.

[14] 王金华,黄曾华.中国煤矿智能开采科技创新与发展[J].煤炭科学技术,2014,42(09):1-6+21.

[15] Zhou YH, Ma HW, Wu HY, et al. Constant C utting-Power Control of Shearer Based on Neural Network Model Predictive C ontrol[J]. Advanced M aterials Research, 2013, 823:340-344.

[16] Ming-Xing H E, Liu CY, Wang W. Study on Software Upgrading of Coal Shearer with Electromagnetic Speed Control Based on Cortex-M3 Core Coal. Science and Technology,2013,4:023.

[17] Jia J, Hong-Ying Li. Inverter Technology Shearer Speed Control System. App lications[J]. Me ik uang J ixie(C oal Mine M achinery), 2013, 34(4): 223-224.

[18] Craven A. L, Muirhead I. R. Horizon control technolo gy for selective minin g in und er ground coal mines [J]. Cim bu letin, 2000, 93(1040): 63-71.

[19] Huang Huimin, John Horst, Richard Quintero. A motion control algorithm for a continuous mining machine based on a hierarchical Real-time Control System des ign meth odology [J]. Journalof Intelligent and Robotic Systems, 1992, 5(1): 79-99.

[20] Hadi H. S, Ataie M, Khalookakaei R , Kumar U. Relabilty modeling of hydraulic systemof drum shearer mach ine [J]. Journal of Coal Science and En gineering (China), 2011, 17(4):450-456.

[21] Sahoo R., Mazid A.M. Application of opto-tactile sensor in shearer machine design torecognize rock surfaces in underground coal mining [C]. Proceedings of the 2009 IEEE International C on ference on Indu strial Technology, Churchill, AUSTRALIA, FEB 10-13,2009.

[22] Ralston J. C., Strange A. D. Thermal infrared -based seam tracking for inte ligent longwall shearer horizon control [C]. Proceedings of the 12th Coal Operators Conference, Univ Wollongong, Wollon gong, AUSTRALIA, FEB 16-17, 2012.

[23] 樊启高,李威,王禹桥,范孟豹,杨雪锋.一种采用灰色马尔科夫组合模型的采煤机记忆截割算法[J].中南大学学报(自然科学版),2011,10: 3054-3058.

[24] 欧璐璐.基于直接转矩控制的采煤机变频调速系统研究[D].淮南:安徽理工大学,2008.

[25] 刘温暖,赵永德.电牵引采煤机调速系统[J].煤矿机械,2009,05: 125-127.

[26] 贾剑,纪红英.变频技术在采煤机调速系统的应用[J].煤矿机械,2013,04: 223-224.

[27] 王玉萍,宋莹莹.采煤机调高系统的模糊神经网络自适应控制[J].煤矿机械, 2009,30(8):188-189.

[28] 陈昂.基于瓦斯参数的综采工作面采煤机自动调速技术研究[D].徐州:中国矿业大学, 2014.

[29] 张寅锋.滚简式采煤机运动轨迹跟踪及控制策略研究[D].杭州:浙江大学, 2014.

[30] 刘建功,王汝琳.采煤机电磁调速技术研究[J].中国矿业大学学报, 2005, 01: 110-114.

[31] 刘侠,谢子殿.基于模糊控制的电磁调速采煤机控制系统的设计[J].工业控制计算机, 2009,05: 55-56.

[32] Cheng Yunlong, Zhao Fan, Zhang Qinghua, et al. A survey on granular computingand its uncertainty measure from the perspective of rough set theory[J]. Granular Computing, 2021, 6: 3-17.

[33] Gheyas I.A., Smith L.S. Feature subset selection in large dimensionality domains[J] Pattern Recognition,2010,43(1): 5-13.

[34] 边肇祺,张学工. 模式识别[M].第2版.北京:清华大学出版社, 2000: 176-177.

[35] Kira K., Rendell L. A.The feature selection problem: Traditional methods and a new algorithm[C]/Proceedings of the 10th National Conference on Artificial Intelligence.USA: AAAI, 1992: 129-134.

[36] Li Zhang. A new feature selection using dynamic interaction[J].Pattern Analysis and Applications, 2021, 24(1): 203-215.

[37] Abd Elminaam D.S, Nabil A., Ibraheem S. A, et al. An efficient marine predators algorithm for feature selection[J] IEEE Access, 2021, 9: 60136-60153.

[38] Too J W., Minjalili S. General leaming equilibrium optimizer: a new feature selectionmethod for biological data classi fication[J]. Applied Arti ficial Intelligence, 2021,35(3): 247-263.

[39] Hancer E., Xue Bing, Zhang Mengjie. A suvey on feature selection approaches forclustering[J]. Artificial Intelligence Review, 2020, 53(6): 4519-4545.

[40] Solorio-Fernandez S.,Carrasco-Ochoa J.A., Martinez-Trinidad J.FA review ofunsupervised feature selection methods[J].Artificial Intelligence Review, 2020,53(2):907-948.

[41] Ziarko W. Variable precision rough set model[J]. Journal of Computer and System Sciences, 1993, 46(1): 39-59.

[42] Nanda S., Majumdar S. Fuzzy rough sets[J]. Fuzzy Sets and Systems, 1992, 45(2):157-160.

[43] Azam N,Yao Jingtao. Game-theoretic rough sets for recommender systems[J]. Knowledge- Based Systems, 2014, 72: 96-107.

[44] 胡清华,于达仁,谢宗霞.基于邻域粒化和粗糙逼近的数值属性约简[J].软件学报, 2008, 19(3): 640-649.

[45] 胡清华,于达人.应用粗糙计算[M].北京:科学出版社, 2012: 45-53.

[46] Qian Yuhua, Liang Xinyan, Wang Qi, et al. Local rough set: a solution to rough data analysis in big data[J]. International Joumnal of Approximate Reasoning, 2018, 97:38-63.

[47] Yao Yiyu. An outline of a theory of three-way decisi ons[C].Proceedings of the 8thInternational Conference on Rough Sets and Current Trends in Computing. Chengdu:Springer, 2012: 1-17.

[48] Yao Yiyu.Three-way decision: An interpretation of rules in rough set theory[C].Proceeding of the Intemational Conference on Rough Setsand Knowledge Technology. Gold Coast: Springer,2009: 642-649.

[49] Wang Guoyin, Zhao Jun, An Jiujiang, et al. Theoretical study on attribute reduction of rough set theory: comparison of algebra and information views[C]. Proceedings of the 3rd IEEE International Conference on Cognitive Informatics. Victoria: IEEE, 2004: 148-155.

[50] Guan J.W., Bell D. A. Rough computational methods for information systems[J].Artificial Intelligence, 1998, 105(1-2): 77-103.

[51] Skowron A., Rauszer C. The discernibility matrices and functions in infommation systems[M]. Intelligent Decision Support. Berlin: Springer, 1992: 331-362.

[52] 苗夺谦,王珏.粗糙集理论中概念与运算的信息表示[J].软件学报,1999,10(2):113-116.

[53] 王国胤,于洪,杨大春.基于条件信息熵的决策表约简[J]计算机学报, 2002,25(7): 759-766.

[54] 苗夺谦,胡桂荣.知识约简的一种启发式算法[J].计算机研究与发展, 1999,36(6): 681-684.

[55] 张文修,魏玲,祁建军.概念格的属性约简理论与方法[J].中国科学E辑:信息科学, 2005, 35(6): 628-639.

[56] 米据生,陈锦坤.基于图的粗糙集属性约简方法[J] 西北大学学报(自然科学版), 2019, 49(4): 508-516.

[57] Jiang Zehua, Yang Xibei, Yu Hualong, et al. Accelerator for multi-granularity atribute reduction[J]. Knowl edge Based Systems,2019, 177:145-158.

[58] Jia Xiuyi, Rao Ya, Shang Lin, et al. Similarity-based attribute reduction in rough set theory: a clustering perspective[J].International Joumal of Machine Learning and Cybemetics, 2020,11(5):1047-1060.

[59] 汪凌.一种基于知识的城市交通流状态快速识别方法及应用[C].中国智能交通年会.2013.

[60] 崔庆安.基于粗糙集的焦炉加热工序控制规则提取[J].化工自动化及仪表, 2005,32(6):20-23.

[61] 王铮. 基于知识发现的航空发动机智能监测研究[D].南京:南京航空航天大学,2011.

[62] 王国法.综采自动化智能化无人化成套技术与装备发展方向[J].煤炭科学技术, 2014,09: 30-34.

[63] 赵宇星.浅谈综采自动化及发展趋势[J]. 中国新技术新产品,2014,16: 5-7.

[64] 黄曾华.综采工作面综合自动化控制技术综述[J].煤炭科学技术,2013,S2: 282-284.

[65] 吕培.采煤机闭溜控制的设计及应用[J].科技创业家,2012(16): 121.

[66] 于玲,张晨.液压支架护帮千斤顶节流优化控制研究[J].煤矿机械,2012(33): 42 _45.

[67] 周信.综采装备协同控制关键技术研究[D].徐州:中国矿业大学,2014.

[68] 刘安平,王亚晶,杨庆江.顶板压力无线监控系统设计[J].自动化技术与应用,2016(30):93 -96.

[69] 刘艳芳.综采工作面刮板输机远程监控技术研究[D].徐州:中国矿业大学,2014.

[70] 马万祥,刘小明.6.2m采高工作面成套国产装备稳定性研究与应用[J].煤炭工程,2 017(5):