BP神经网络在风电塔筒裂纹AE源定位中的应用

Application of BP neural network in localization of crack acoustic emission source in wind vane tower barrel

针对常规无损检测方法难以对在役风电塔筒进行实时监测及检测效率低等问题,为实现在役风电塔筒的在线监测和塔筒裂纹源的定位。该文提出一种声发射检测技术与BP神经网络相结合的智能定位方法,研究声发射传感器分布对智能定位输出结果的影响。研究结果表明:该方法可有效解决风电塔筒裂纹声发射源的定位问题,为在役风电塔筒运行过程中的实时在线监测提供可靠依据;对于焊接结构对称且具有多条环焊缝的风电塔筒,声发射传感器的分布位置应避开塔筒构件的对称位置,以此提高智能定位输出结果的准确性,为在役风电塔筒检测时声发射传感器的合理、有效布置提供参考。

As the conventional nondestructive testing can hardly have real-time monitoring of the wind vane tower barrel in service and the testing efficiency is low, to have online monitoring of the condition of wind vane tower barrel in service and locate the crack source of the tower barrel, this paper puts forward an intelligent localization method combining with acoustic emission testing technique and BP neural network and studies the influences of the distribution of acoustic emission sensors on the intelligent localization of output results. The results show that this method can locate the crack acoustic emission （AE） source in wind vane tower barrel effectively and can be used for the real-time online monitoring of the wind vane tower in service. For wind vane tower barrels with symmetrically welded structure and several circumferential welds, the distribution of acoustic emission sensors should avoid the symmetric position of tower barrel components to improve the accuracy of the intelligent localization of output results, providing a basis for the rational and effective arrangement of acoustic emission sensors on wind vane tower barrels in service.