基于激光多普勒测振技术的建筑房屋外围护结构缺陷预测研究  

作　　者：徐海华 陈文杰 李肖 

机构地区：[1]浙江省建设工程质量检验站有限公司,浙江杭州310000 [2]浙江省建筑科学设计研究院有限公司,浙江杭州310000 [3]中天美好集团有限公司,浙江杭州310000

出　　处：《住宅产业》2025年第3期88-90,共3页Housing Industry

摘　　要：本文基于激光多普勒测振技术的建筑房屋外围护结构缺陷预测研究,利用激光多普勒测振技术,通过发射特定频率的线偏振光并接收反射光,精确捕捉到建筑房屋外围护结构的微小振动信号。将结构振动信号转换到频域内,结合离散频谱校正及卡尔曼滤波器进行降噪,获取径向振动速度和加速度信号的频域表示作为特征。构建时间序列模型,采用自回归模型与回归树—支持向量机回归综合方法,依据特征以及设定的阈值预测结构是否存在缺陷。实验结果表明,该方法能精准预测建筑物外围护结构缺陷,与实际信号高度一致,预测精度高,显著提升缺陷预测准确性。A research on defect prediction of building envelope structures is proposed based on Laser Doppler Vibrometry(LDV)technology.By emitting linearly polarized light of a specific frequency and receiving reflected light,the tiny vibration signals of the building envelope structures can be accurately captured through the LDV technology,which can convert the structural vibration signal into the frequency domain,combine discrete spectrum correction and Kalman filter for noise reduction,and obtain the frequency domain representation of radial vibration velocity and acceleration signals as features.This paper constructs a time series model using autoregressive model and regression tree-support vector machine regression synthesis method,and predicts whether the structure has defects based on features and set thresholds.The experimental results show that this method can accurately predict the defects of the building envelope structures,which is consistent with the actual signal height and has high prediction accuracy,significantly improving the accuracy of defect prediction.

关 键 词：激光多普勒测振技术 建筑房屋外围护结构 提取信号特征 建立预测模型 缺陷预测 

分 类 号：TN9[电子电信—信息与通信工程]