Hydrostatic seasonal state model for monitoring data analysis of concrete dams

Based on the idea that time-effect deformation can be regarded as a state vector indicating the characteristics of concrete dams, a hydrostatic seasonal state (HSS) model is proposed for improving the accuracy of the data analysis of concrete dams. According to the theory of non-stationary time series, the state equation of the HSS model is represented for objectively extracting the time-effect deformation from monitoring data of total deformation and to interpret its evolution accurately. In this study, the HSS model is applied to an arch dam case, and it is estimated by using the expectation-maximisation algorithm. The results illustrate that the HSS model can efficiently estimate the time-effect deformation from monitoring data of total deformation. Compared with the hydrostatic seasonal time model, the proposed HSS model shows an improvement in fitting precision and prediction ability.