基于多尺度特征融合的风机叶片裂缝自动检测技术研究

Research on Automatic Detection Technology of Fan Blade Crack Based on Multi-scale Feature Fusion

风力发电是一种利用自然风能的基础,以较低风速为驱动,将风轮所经过的动能转化成机械动力来带动发电机旋转。因此在电力系统中占据着非常重要地位。从智能技术多尺度特征融合中研究风机叶片裂缝自动检测技术,目的是实时监测风机的故障及防范风力发电的停止。主要应用了实验法和案例分析的方式对风机叶片裂缝检测自动检测技术及多尺度特征融合进行研究。实验表明:多特征提取方式在风机叶片横向裂缝检测的准确率达到95%,基本符合要求。

Wind power generation is based on the use of natural wind energy and driven by low wind speed,which converts the kinetic energy passed by the wind wheel into mechanical power to drive the generator to rotate.Therefore,it occupies a very important position in the power system.This paper intends to study the automatic detection technology of fan blade crack from the multi-scale feature fusion of intelligent technology,in order to monitor the fault of fan in real time and prevent the stop of wind power generation.This paper mainly studies the automatic detection technology of fan blade cracks and multi-scale feature fusion by means of experiment and case analysis.The experimental results show that the accuracy of multi feature extraction method in fan blade transverse crack is 95%,which basically meets the requirements.