基于集成神经网络和改进极限学习机的矿井移动机器人故障检测  

Fault Detection of Mine Mobile Robot Based on Integrated Neural Network and Improved Extreme Learning Machine

作　　者：郑伦川 梁新元[2] 袁乖宁 ZHENG Lunchuan;LIANG Xinyuan;YUAN Guaining(College of Big Data,Chongqing Water Resources and Electric Engineering College,Chongqing 402160,China;School of Artificial Intelligence,Chongqing Technology and Business University,Chongqing 400067,China)

机构地区：[1]重庆水利电力职业技术学院大数据学院,重庆402160 [2]重庆工商大学人工智能学院,重庆400067

出　　处：《金属矿山》2024年第6期159-164,共6页Metal Mine

基　　金：2020年度市教委科学技术研究计划项目(编号:KJQN202003806);重庆市教委教学改革研究项目(编号:203633)。

摘　　要：矿井移动机器人作为一种自主运动的智能设备,广泛应用于采矿、输送和装载等工作中。然而,由于其在恶劣环境下运行,往往长时间无法得到检修维护,导致故障频发,影响了井下安全高效生产。如何及时准确地对机器人进行故障检测,提高其可靠性和生产效率成为一个亟待解决的问题。提出了一种基于集成神经网络和改进极限学习机的矿井移动机器人故障检测方法。该方法融合了多个神经网络模型,并通过改进极限学习机算法来提高检测精度和效率。首先,基于集成学习思想将传统卷积神经网络、递归神经网络和自编码器等多个预训练模型集成为一个更强大的检测模型。其次,在极限学习机的基础上引入了自适应权重调整策略,提高了算法的自适应能力和准确性。将所提出的方法在某矿山数据集上进行了试验,结果表明:该方法在检测区分度较低或异常数据较多的情况下性能优异,有助于实现高精度和高效率的故障检测。As a kind of intelligent equipment with autonomous movement,mine mobile robot is widely used in mining,conveying and loading.However,due to its operation in harsh environments,it is often unable to be repaired and maintained for a long time,resulting in frequent failures and affecting safe and efficient production underground.How to detect the fault of the robot timely and accurately,improve its reliability and production efficiency has become an urgent problem to be solved.A fault detection method of mobile robot in mine is proposed based on integrated neural network and improved extreme learning machine.This method integrates several neural network models,and improves the detection accuracy and efficiency by improving the algorithm of extreme learning machine.Secondly,based on the extreme learning machine,the adaptive weight adjustment strategy is introduced to improve the adaptive ability and accuracy of the algorithm.The proposed method is tested on a mine data set,the results show that the method has excellent performance in the case of low detection differentiation or more abnormal data,and is helpful to achieve high precision and high efficiency fault detection.

关 键 词：矿井移动机器人 故障检测 集成神经网络 改进极限学习机 

分 类 号：TD67[矿业工程—矿山机电]