<b>Fault Identification Technology of Wind Turbine Blade Damage Sound Source Localization and Multimodal Data Fusion Based on Voiceprint Monitoring</b>

nt. Finally, damage sound source localization and fault classification are achieved using the Time Difference of Arrival (TDOA) algorithm and Convolutional Neural Network (CNN). Experimental results show that the proposed method achieves an accuracy and recall rate of over 85% under various wind turbine fault conditions, with a localization error within 0.15m. Additionally, a real - time monitoring system based on an embedded hardware platform is designed. It can perform online fault identification via data stream processing and issue timely fault warnings to ensure the safe operation of wind turbines.As wind energy rapidly develops globally, wind turbine blades, as crucial components, require effective damage detection to ensure the reliability and longevity of wind turbines.nt. 最后，本研究借助到达时间差（Time Difference of Arrival, TDOA）算法与卷积神经网络（Convolutional Neural Network, CNN），实现了损伤声源定位与故障分类。实验结果表明，所提方法在各类风力涡轮机故障工况下，准确率与召回率均超过85%，定位误差控制在0.15米以内。此外，本研究设计了一款基于嵌入式硬件平台的实时监控系统，该系统可通过数据流处理实现在线故障识别，并及时发出故障预警，以保障风力涡轮机的安全运行。随着风能在全球范围内快速发展，作为核心关键部件的风力涡轮机叶片亟需开展有效的损伤检测工作，以保障风力涡轮机的可靠性与使用寿命。