基于嵌入式神经网络的电梯钢丝绳弱磁检测系统  

Weak Magnetic Detection System for Elevator Steel Wire Rope Based on Embedded Neural Network

作　　者：马铭阳 MA Mingyang(School of Electronic Information Engineering,Shenyang Aerospace University,Shenyang,Liaoning 110136,China)

机构地区：[1]沈阳航空航天大学电子信息工程学院,辽宁沈阳110136

出　　处：《自动化应用》2024年第17期168-171,181,共5页Automation Application

摘　　要：为实时监测和排查电梯钢丝绳运行中存在的安全问题,提出了剩余磁场激励的弱磁检测方法。该方法采用开环永磁体磁化方式将被测钢丝绳充分磁激励,励磁后的钢丝绳缺陷部位会产生微弱漏磁场,采用霍尔传感器可将被测钢丝绳周围空间中磁感应强度的变化情况转化为电信号。检测系统采用多传感器融合的RBF-BP组合神经网络算法结合RT-Thread实时操作系统在单片机上运行。结果显示,钢丝绳各类损伤判断准确率达到了80%以上,检测效果良好,为电梯的安全运行提供了一条解决路径。A weak magnetic detection method using residual magnetic field excitation is proposed for real-time monitoring and troubleshooting of safety issues in elevator wire rope operation.This method uses an open-loop permanent magnet magnetization method to fully excite the tested steel wire rope.After excitation,a weak leakage magnetic field will be generated at the defective part of the steel wire rope.The Hall sensor can convert the changes in magnetic induction intensity in the surrounding space of the tested steel wire rope into electrical signals.The detection system adopts a multi-sensor fusion RBF-BP combination neural network algorithm combined with RT Thread real-time operating system to run on a microcontroller.The results showed that the accuracy of various types of damage detection for steel wire ropes reached over 80%,and the detection effect was good,providing a solution path for the safe operation of elevators.

关 键 词：钢丝绳 神经网络 嵌入式 霍尔元件 剩余磁场激励 操作系统 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]