Deformation similarity characteristics-considered hybrid panel model for multi-point deformation monitoring of super-high arch dams in operating conditions

Abstract

This research develops a novel hybrid model for multi-point deformation monitoring of super-high arch dams in operating conditions. The weighted distances are established to characterize deformation similarity degree, and then observation point groups with similar deformation regularities are produced using the bottom-up hierarchical clustering. The hybrid hydrostatic seasonal time (HHST) panel model is proposed, and the artificial fish swarm (AFS) algorithm is improved to optimize the undetermined parameters. The confidence ellipsoid criteria are established by applying multivariate statistic and principle of small probability event. According to the example analysis, the HHST panel model achieves a better fitting performance than the HST panel model; the applicability of HHST panel model is wider than that of HHST model; the optimization performance of the improved AFS is superior to that of the conventional AFS; confidence ellipsoid compared with confidence interval possesses a stricter identification for abnormal deformations and a clearer physical significance.