摘要
高压输电铁塔健康监测对保证输电线路安全运行具有重要意义。以500kV输电塔为例,建立多尺度模型,引入小波包能量曲率差作为损伤指标,提出了一种基于多尺度模型的输电铁塔裂纹损伤识别方法。研究结果表明,该方法能准确判断裂纹的损伤位置,具有良好的抗噪能力。将该指标结合支持向量机(SVM)方法对损伤程度进行识别,与最小二乘支持向量机(LSSVM)方法对比分析发现,LIBSVM方法更逼近目标值,识别效果更准确。
It is important of transmission tower health monitoring to ensure the safe operation of transmission lines.Taking 500 kV transmission tower for an example,based on the multi-scale model,a crack damage identification method of transmission tower is proposed by introducing the wavelet packet energy curvature difference as damage index.The results show that this method can accurately identify the damage location,and this index does well in noise immunity.Combining this index with SVM to identify the degree of damage,comparative studies show that the LIBSVM method is more close to the target value than that of the last squares support vector machine method,and the result is more accurate.
出处
《水电能源科学》
北大核心
2016年第2期177-181,216,共6页
Water Resources and Power
基金
国家自然科学基金项目(50978049
51278091
51308094)
关键词
输电塔
多尺度
小波包能量曲率差
支持向量机
裂纹
损伤识别
transmission tower
multi-scale
wavelet packet energy curvature difference
SVM
crack
damage identification
