基于机器视觉的火力发电厂锅炉风机叶片故障诊断方法
Machine Vision Based Fault Diagnosis Method For Boiler Fan Blades in Thermal Power Plants
李洋1
作者信息
- 1. 内蒙古大唐国际托克托电厂,内蒙古呼和浩特 010000
- 折叠
摘要
当前火力发电厂锅炉风机叶片故障诊断结果与实际故障数目差异较大,为此研究基于机器视觉的火力发电厂锅炉风机叶片故障诊断方法.首先,对叶片振动信号进行分析,确定故障特征的信息,以此为基础来提取叶片故障特征.然后消除图像噪声,保留图像细节特征并构建基于机器视觉的火力发电厂锅炉风机叶片故障诊断模型,利用损失函数对生成的诊断结果展开优化.实验结果显示,与基于红外热成像技术的火力发电厂锅炉风机叶片故障诊断方法和基于声发射技术的火力发电厂锅炉风机叶片故障诊断方法相比,基于机器视觉的火力发电厂锅炉风机叶片故障诊断损失值最低,维持在0.2左右.这表明基于机器视觉的火力发电厂锅炉风机叶片故障诊断方法的效果更好,在实际应用中具有更大的优势.
Abstract
The current fault diagnosis results of boiler fan blades in thermal power plants differ significantly from the actual number of faults.To study a machine vision based fault diagnosis method for boiler fan blades in thermal power plants.Firstly,analyze the vibration signals of the blades to determine the information of fault characteristics,and based on this,extract the fault characteristics of the blades.Then,image noise is eliminated,image detail features are retained,and a machine vision based fault diagnosis model for boiler fan blades in thermal power plants is constructed.The generated diagnostic results are optimized using a loss function.The experimental results show that compared with the infrared thermal imaging technology and the acoustic emission technology for diagnosing faults in boiler fan blades in thermal power plants,the machine vision based method has the lowest loss value for diagnosing faults in boiler fan blades in thermal power plants,with a loss value maintained at around 0.2.This indicates that the machine vision based fault diagnosis method for boiler fan blades in thermal power plants is more effective and has greater advantages in pracical applications.
关键词
机器视角/火力发电厂/锅炉风机叶片/叶片故障诊断/风机叶片故障诊断Key words
machine perspective/thermal power plants/boiler fan blades/blade fault diagnosis/fault diagnosis of fan blades引用本文复制引用
出版年
2024
NETL
NSTL
万方数据