ICCM Conferences, The 14th International Conference of Computational Methods (ICCM2023)

Font Size: 
Construction and verification of composite intensity measures for multi-storey subway station structures based on partial least squares regression
wei yu, zhiyi chen

Last modified: 2023-07-27

Abstract


In order to quantitatively evaluate the seismic damage potential of ground motions (GMs) on multi-storey subway station structures and solve the problem of that single intensity measures (IMs) are inadequate in accurately representing the seismic damage potential of GMs, this paper takes the input GMs in the seismic design of multi-storey subway station structures as the research object and constructs composite intensity measures (IMs) that can effectively characterize the damage potential of input GMs on a typical four-storey and three-span subway station structures. Considering the various seismic response of underground structures in different characteristic period type sites, the soil-subway station structure system is divided into four period bands according to the characteristic period of the system. Then, four representative period soil-structure systems in four period bands are selected. Subsequently, 40 GMs were selected for elastoplastic numerical analysis, and the results of numerical analysis were used as sample data for partial least squares regression (PLSR) to construct composite IMs corresponding to each period bands soil-structure systems. Finally, 40 addition GMs are used to verify the correlation between the composite IMs and the seismic damage of multi-storey subway station structures. The results show that the correlation between the composite IMs and the seismic damage of multi-storey subway station structures is better than that of commonly used single IMs.

 


Keywords


subway station structures, damage potential of ground motions, composite intensity measures, Characteristic periods, partial least squares regression

An account with this site is required in order to view papers. Click here to create an account.