论文部分内容阅读
为避免基于结构频率的损伤检测法因温变影响而出现误判,提出一种自联想神经网络(AANN)和奇异检测技术相结合的损伤检测方法,利用某桥梁结构健康监测(SHM)Benchmark试验结构的数值模型,分析温变和损伤对结构频率的影响,用温变影响下健康结构的前十阶竖向弯曲模态频率训练AANN来提取频率和温度的关系。为消除温变影响,构造网络输出与输入间的欧式距离作为损伤识别指标,对比结构未知状态和健康状态的指标值以判定结构是否存在损伤。通过在桥梁Benchmark结构中模拟多级损伤来验证该法的有效性,检测结果表明:该法不仅能可靠地检测温变影响下结构损伤的存在,且能定性地判别损伤程度的大小,并具有较强的抗噪声鲁棒性,可为实际桥梁结构的在线健康监测提供参考。
In order to avoid misjudgment of damage detection method based on structure frequency due to temperature change, a damage detection method based on AANN and singularity detection is proposed. By using a bridge structural health monitoring (SHM) Benchmark test Structure of the structure of the numerical model to analyze the effects of temperature changes and damage on the structural frequency of the health structure with the temperature change of the first ten vertical bending modal training AANN to extract the relationship between frequency and temperature. In order to eliminate the influence of temperature change, the Euclidean distance between the output of the network and the input is used as the damage identification index, and the unknown value of the structure and the index value of the health status are compared to determine whether the structure is damaged. The validity of this method is verified by simulating multistage damage in the Benchmark structure of the bridge. The test results show that this method not only can reliably detect the existence of structural damage under the influence of temperature change, but also can qualitatively determine the extent of damage and has Strong anti-noise robustness can provide reference for online health monitoring of actual bridge structures.