从体系结构、硬件构成、网络结构和软件设计等几个方面来看,大坝监测自动化系统采集的数据会失真,针对以往监测数据验证方法存在的缺陷,提出了基于相空间重构和贝叶斯框架最小二乘支持向量机(BLS-SVM)相结合的监测物理量数据验证方法。采用LS-SVM构建了3种预测器,应用于单监测物理量和多监测物理量输出系统。实例结果证明了所提出的方案的有效性。 From the aspects of architecture, hardware structure, network structure and software design, the data collected by dam monitoring automation system will be distorted. In view of the shortcomings of previous monitoring data verification methods, a new phase space reconstruction A framework of least square support vector machine (BLS-SVM) combination of monitoring physical quantity data validation method. Using LS-SVM, three kinds of predictor are constructed, which are applied to single monitoring physical quantity and multi-monitoring physical quantity output system. The result of the example proves the validity of the proposed scheme.