摘要
提出了一种利用振动响应加速度信号结合主成分分析和支持向量机算法进行结构损伤识别的方法。首先,提取结构加速度振动响应信号;然后,采用主成分分析方法对上述提取到的结构加速度振动响应信号进行分析,提取前n阶主成分并将其按照某种规则组合成一维向量,作为结构损伤特征向量;最后,利用支持向量机的分类和回归算法分别进行损伤位置和损伤程度识别研究。该方法可以在样本较少的条件下取得较好的损伤识别结果,且具有一定的抗噪声能力。白噪声激励下的五层框架结构的数值模拟结果证明了该方法的有效性。
A method for structural damage identification using vibration acceleration signal based on principal component analysis and support vector machine algorithm was presented Firstly, the acceleration response signal of the structure was extracted; then, vibration response signal of the structure was analyzed using principal component analysis method, the principal components of the first n - rank were extracted to combine into one-dimensional vector according to a certain rule, and wer6 served as the structure damage characteristic vector; In the end, identification research of damage location and severity were conducted respectively using support vector machine classification and regression algorithms. Through this method better identification results can be obtained under less sample conditions, and it has certain anti-noise ability. The three-dimensional numerical simulation results of five layer frame structure under white noise proved the validity of the method.
作者
门雪
刘晓春
李国华
武梦龙
Men Xue1, Liu Xiaochun2 , Li Guohua1 , Wu Menglong1(1.School of Civil Engineering, Qingdao University of Technology, Qingdao 266033, Shandong ; 2. Shandong Lanke Engineering Consulting Co. , Ltd. , Yantai 264000, Shandong)
出处
《工程建设》
2018年第3期35-40,共6页
Engineering Construction
关键词
损伤识别
主成分分析
支持向量机
损伤位置识别
损伤程度识别
damage identification
principal component analysis
support vector machines
damage location identification
damage degree identification
