Application of grey wolf optimizer to develop new global GMPE for estimating peak ground acceleration

Ground motion prediction equations (GMPEs) are open challenge problems that have been developed since 1964. Parametric and nonparametric methods predict ground motion characteristics such as peak ground acceleration (PGA), velocity, displacements, and spectral accelerations. In the present study, the grey wolf optimization (GWO) algorithm was used to obtain a new and developed GMPE for predicting PGA. Data from recorded earthquakes from all over the world were collected, and after filtering of Mw and distance parameters, close to 2000 data were used for modelling. Three parameters of Mw (4–7.9), epicentral distance (0.25–115 km) and geological conditions (soft soil, stiff soil, rock) were used as input parameters for estimating PGA. Many previous studies classified geological conditions based on shear wave velocity at the top 30 m (Vs30), without taking into account the effect of Vs30 at each group. In this study, the effects of Vs30 were considered separately for each geological group too. Results showed that PGA decreased by increasing Vs30 and moving from soft soil toward rock. Finally, the relationship was compared with the other two relations suggested for the local region and global earthquakes, and despite the simplicity of the suggested relation gained by the GWO method, it estimated PGA in terms of accuracy to a good and acceptable level.

---

Opracowanie rekordu ze środków MNiSW, umowa nr SONP/SP/546092/2022 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2024).