基于CNN-RVM的输电杆塔外破振动识别方案

针对现有杆塔状态检测方案缺乏外破振动识别的现状,提出了一种基于深度学习模型的杆塔外破振动识别技术。首先获取外破条件下的输电杆塔外破振动信号和不同风激励条件下的输电杆塔振动信号,应用延时嵌陷技术对振动信号预处理,将原始信号转为二维形式后送入卷积神经网络(Convolutional Neural Network, CNN)进行特征提取,并采用相关向量机(Relevance Vector Machine, RVM)分类器实现振动模式识别;通过多次实验,确定CNN的最佳参数,再利用softmax分类器和梯度下降法对CNN的权值和阈值进行调整,最终得到高精度识别结构。仿真结果表明,提出的CNN-RVM识别模型在三种振动条件实验中准确率都高于99%,相比于国内外现有其它振动识别方案,具有高精度和高效率的优点。

External Vibration Identification Scheme of Transmission Line Tower Based on CNN-RVM

According to the problem of existing tower state detection scheme lacking of the identification of external vibration, this paper proposes an external vibration identification scheme of tower based on deep learning. The proposed scheme obtains the external breaking vibration signal of transmission pole and the vibration signals of transmission towers under different wind excitation conditions firstly. The vibration signals are preprocessed by adopting delay inlay technology to turn the original signal into a two-dimensional form and sent that into convolutional neural network(CNN)to feature extraction, and achieve vibration pattern recognition by employ the relevance vector machine(RVM).Through many experiments, the best parameters of CNN can be determined, and then the weights and thresholds of CNN are adjusted by softmax classifier and gradient descent method. Finally, the structure of high precision recognition can be obtained. The simulation results show that the accuracy of the proposed method is higher than 99% in the three vibration conditions. Compared with the existing vibration recognition technology of domestic and foreign steelwork towers, this model has the advantages of high precision and high efficiency.