基于改进PSO-SVM的噪声稳健建筑物变形监测方法

对高层、超高层建筑物进行实时,高精度的变形监测对提前预防安全隐患,保证人民生命财产安全具有重要意义。建筑物变形作为一种典型的随机性和微弱性过程,噪声等误差的存在会影响从中提取有用的变形信息。针对该问题,提出一种改进粒子群(Particle Swarm Optimization,PSO)算法优化支持向量机(Support Vector Machine,SVM)的噪声稳健建筑物变形监测方法,利用改进PSO算法的全局搜索能力对SVM的核参数进行优化,提升预测精度的同时增强算法的噪声稳健性。基于实测数据的试验结果表明,相对于传统交叉验证SVM和小波方法,所提方法可以获得更高的变形预测精度,并且在低信噪比条件下优势更加明显。

Noise Robust Method of Deformation Monitoring Based on PSO-SVM

The real-time and high-precision deformation monitoring of high-rise and super high-rise buildings is of great significance to prevent potential safety hazards in advance and ensures the safety of people's lives and properties.As a typical process of randomness and faintness,the existence of random errors such as noise affects the extraction of effective deformation information.In order to solve this problem,this paper proposed a noise robust deformation prediction method of support vector machine(SVM)optimized by particle swarm optimization(PSO)algorithm,which optimized the kernel parameters of SVM by using the global search ability of PSO,improved the prediction accuracy and enhances the noise robustness of the algorithm.The experimental results based on the measured data showed that compared with the traditional SVM and wavelet methods,the proposed method could obtain higher deformation prediction accuracy,and had more obvious advantages under the condition of low SNR.