基于CNN与改进XGBoost的采煤机健康状态识别  

Health Status Identification of Shearer Based on CNN and Improved XGBoost

作　　者：张金坡 李曼[2] 金楠 郑永涛 Zhang Jinpo;Li Man;Jin Nan;Zheng Yongtao(Shaanxi Railway Institute,Weinan 714000,China;Xi’an University of Science and Technology,Xi’an 710054,China)

机构地区：[1]陕西铁路工程职业技术学院,陕西渭南714000 [2]西安科技大学,西安710054

出　　处：《煤矿机械》2025年第2期183-187,共5页Coal Mine Machinery

摘　　要：采煤机结构复杂,工作环境恶劣,其健康状态评估比较困难。提出一种基于卷积神经网络(CNN)和改进XGBoost的采煤机健康状态识别方法。结合评估数据时间、参数间相关性,将原始数据重构为矩阵形式作为CNN模型输入,降低了网络的训练难度;针对CNN特征提取优势明显但分类能力有限的问题,利用CNN实现特征提取,并使用分类性能良好的XGBoost模型完成采煤机健康状态识别;针对XGBoost算法最优参数难以获取、专家经验法和网格搜索等常用调参方法容易陷入局部最优等问题,利用乌燕鸥优化算法(STOA)进行XGBoost模型参数优化,提高模型对数据的分析学习能力。利用采煤机仿真数据对该模型进行实验验证,并与其他模型进行对比,结果显示,该模型识别准确度可达98.2%,可为采煤机的状态预知和维护提供理论支持。The structure of shearer is complex and the working environment is bad,so it is difficult to evaluate its health status.A shearer health status identification method based on convolutional neural network(CNN)and improved XGBoost was proposed.Combined the assessment data time and correlation between parameters,the raw data are reconstructed into a matrix form as the input of CNN model,which reduces the training difficulty of the network.Aiming at the problem of CNN feature extraction with obvious advantage but limited classification ability,the feature extraction is realized by CNN,and the shearer health status identification is completed by using XGBoost model with good classification performance.Aiming at the problems that the optimal parameters of the XGBoost algorithm are difficult to obtain,and that the common methods of parameter adjustment such as expert experience method and grid search are easy to fall into local optimization,sooty tern optimization algorithm(STOA)is used to optimize the parameters of the XGBoost model to improve the data analysis and learning ability of model.This model was experiment verified by the simulation data of shearer and compared with other models.The results show that the identification accuracy of this model can reach 98.2%,which can provide theoretical support for condition prediction and maintenance of shearer.

关 键 词：采煤机 健康状态识别 CNN STOA XGBoost 

分 类 号：TD421.6[矿业工程—矿山机电]