摘要
针对选煤厂带式输送机现有防纵撕保护装置存在的"单点"判别局限性和误动作率比较高的问题,基于纵撕发生时的情况,研发了一种综合带式输送机上带面的带宽检测、下带面的异物检测以及侧带面的裂痕检测的纵撕保护系统。该系统主要利用深度学习模型YOLOv3和基于历史高斯统计模型分别实现带宽检测、异物检测和裂痕检测。将3者检测结果进行综合逻辑判断获取判断结果,当判断结果为纵撕裂故障时,直接进行实时报警和联动控制。在神东公司上湾选煤厂218带式输送机应用结果表明:该系统简单易用,模型准确率和实时性满足要求,解决了现有防纵撕保护装置存在的"单点"判别局限性和误动作率比较高的问题,减少了误动作引起的停机时间,延长了设备运行时间,提高了工作效率。
For belt conveyor of coal preparation plant existing longitudinal tear proof protection device"single point"of the existence of discriminant limitations and misoperation rate is quite high,and case according to the longitudinal tearing occurs,belt conveyor tearing protection system was developed with the bandwidth detection,the foreign body detection and side with the scratch of longitudinal tear protection system.The system mainly uses the deep learning model YOLOv3 and the historical Gaussian statistical model to realize bandwidth detection,foreign body detection and scratch detection respectively.The results are obtained by comprehensive logic judgment.When the judgment result is longitudinal tearing fault,real-time alarm and linkage control are directly carried out.Results show that the system is easy to implement,model accuracy and real-time performance meet the requirements,solving the existing protection against longitudinal tearing device limitations"single point"of the existence of discrimination and high rates of misoperation in Shendong Shangwan Coal Preparation 218 belt conreyor.Meanwhile,it reduces downtime caused the misoperation,prolongs the equipment running time,and improves the working efficiency.
作者
薛红伟
刘显望
杨娟利
高攀
王赟
XUE Hongwei;LIU Xianwang;YANG Juanli;GAO Pan;WANG Yun(CHN Energy Shendong Coal Preparation Center,Ordos 017209,China;Xi'an Huaguang Information Technology Co.,Ltd.,Xi'an 710075,China)
出处
《洁净煤技术》
CAS
北大核心
2021年第S02期68-72,共5页
Clean Coal Technology
关键词
纵撕保护
带宽检测
异物检测
裂痕检测
YOLOv3模型
误动作率
工作效率
tearing protection
bandwidth detection
foreign body detection
scratch detection
YOLOv3 model
misoperation rate
working efficiency
