论文部分内容阅读
基于最小二乘支持向量机方法,提出了小波核函数最小二乘支持向量机结构损伤模式识别方法。该方法以结构固有频率变化率为特征向量,利用有限的荷载和动力响应数据,便能得到良好的结构损伤识别效果,弥补了传统方法对结构固有频率不敏感的缺陷,克服了其他动力参数数据采集误差大的问题。数值模拟计算表明该方法具有较高的识别精度。对一个12层的混凝土框架结构模型进行了振动台试验,研究分析结果表明,理论识别结果与实验现象基本吻合,说明该方法具有良好的识别效果和实际应用前景。
Based on the least square support vector machine method, a wavelet kernel function least square support vector machine structural damage pattern recognition method is proposed. The method takes the rate of change of structure natural frequency as the eigenvector, and uses the limited load and dynamic response data to get the good identification effect of structural damage, which makes up for the shortcomings of the traditional method that is insensitive to the natural frequency of the structure and overcomes the other dynamic parameter data Collection of large errors. Numerical simulation shows that the method has higher recognition accuracy. A shaking table test was carried out on a 12-story concrete frame structure model. The research results show that the theoretical recognition results are in good agreement with the experimental ones, which shows that the proposed method has good recognition effect and practical application prospect.