基于模式识别的桥梁故障检测

文章主要任务是研究如何从挠度数据出发,综合运用模式识别技术和数据分析技术,能有效检测桥梁中某些类型的损伤和隐患。将模式识别中的近邻算法与K-均值算法相结合,应用到桥梁检测中,提出了一种新的桥梁异常检测方法。首先,用K均值算法对数据进行聚类分析,寻找数据之间的相似性,识别出模式,将桥梁挠度数据归类;然后,在分析阶段运用改进的近邻算法进行分析检测。由于最近邻和k-近邻的交替使用,保证了检测结果的准确率。将上述方法应用于重庆菜园坝大桥的模拟检测,取得了较好的效果,并具有一定理论价值和潜在应用价值。

An Identification Method for Bridge Based on Pattern Recognition

This paper addresses the issue of detecting the abnormal behaviors of a bridge by interpreting the deflection data.By combining the near-neighborhood method and the K-mean method,a new bridge detection method is proposed,which is composed of two phases.In the first phase,the historical deflection data are classified by using the K-mean algorithm,while the second phase performs detection by employing a modified near-neighborhood method.The rate of correct detection is guaranteed by alternating applying these...