Research On Stress Monitoring Model Of Concrete Arch Dam Of Xiluodu Hydropower Station

The spatial stress field is constructed based on the basic assumption of mechanics materials.Take the 16# section of Xiluodu arch dam as an example,analysing the stress behavior from storage until now.Then using the stepwise regression analysis and BP neural network to construct effective stress monitoring model of Xiluodu Hydropower Station,including single-point model and multi-points model.The conclusions are as follows:(1)During the storage process,due to the increased thrust of the reservoir water level,the radial stress is compressive and increasing.Arch action increases with rising of water level,leading to arch stress is compressive and increasing.And the bending moment action caused by the upstream water level makes the vertical stress of the upstream side is compressive stress decreased and the downstream side is compressive stress increased.(2)For the period of operation,the hydration heat has been basically dissipated.And the concrete temperature and periodic terms can be used as temperature component factors to simplify the model.(3)One-point stress statistical model,one-dimensional multi-points stress statistical model and two-dimensional multi-points stress statistical model of Xiluodu constructed in this paper is effective.It can be used in pretreatment of stress data of Xiluodu Arch Dam.It is also helpful for the establishment of stress statistics model of other arch dams.(4)The accuracy of BP neural network model is higher than that of statistical model.But the BP neural network model can not analyze the influence of each component on the stress.(5)The single point model has high fitting precision and the multi-dimensional multi-point measurement model has a clear physical concept and spatial characteristics,strong anti-interference ability,can reflect the spatial stress distribution,and the obtained model is more realistic than the actual physical meaning.