Optimization and prediction of a geofoam-filled trench in homogeneous and layered soil
Naghizadehrokni, Mehran; Ziegler, Martin (Thesis advisor); Fuentes Gutierrez, Raul (Thesis advisor); Vrettos, Christos (Thesis advisor)
Aachen : RWTH Aachen University (2022, 2023)
Dissertation / PhD Thesis
Dissertation, Rheinisch-Westfälische Technische Hochschule Aachen, 2022
In metropolitan cities, man-made ground vibrations generated by external dynamic sources including traffic, blasting, machine foundations, and other constructional activities have recently become a major concern for the nearby structures and inhabitants. Most of the transmitted vibration energy is propagated by Rayleigh waves close to the soil surface. The rest of the energy is transmitted by body waves. Therefore, implementing a wave barrier in the transmission path is a suitable solution as it allows scattering the ground-borne vibrations. This study presents the performance of geofoam-filled trenches in mitigating ground vibration transmissions by the means of a comprehensive parametric study. Fully automated 2D and 3D numerical models are applied to evaluate the screening effectiveness of the trenches in the near field and far field schemes. The validated model is used to investigate the influence of geometrical and dimensional features on the trench with three different configurations including single, double, and triangular wall obstacles. The parametric study is based on complete automation of the model through coupling finite element analysis software (Plaxis) and Python programming language to control input, change the parameters, as well as to produce output and calculate the efficiency of the barrier. The main assumption during the parametric study is treating each parameter as an independent variable and keeping other parameters constant. An optimization model is also presented to optimize the governing factors of geofoam or concrete-filled trenches as a wave barrier. A genetic algorithm code is implemented with coupling the Python software and the finite element program (Plaxis) for optimization of all parameters mutually. Furthermore, three different configurations including single, double, and triangular wall systems are evaluated with the same cross-sectional area for considering the effect of the shape of the barrier in attenuating the incoming waves. A usual assumption for the study of ground-borne vibration is considering soil as homogeneous, which is unrealistic. Therefore, it is necessary to find the effect of non-homogeneity of the soil on the efficiency of the geofoam-filled trench. A comprehensive parametric study has been performed automatically by coupling Plaxis and Python under the assumption of treating each parameter as an independent variable. The results showed that some parameters have a considerable impact on each other. Therefore, the interaction of all governing parameters on each other is also evaluated through the response surface methodology method. In addition, a genetic algorithm code is presented for optimizing all parameters mutually in homogeneous and layered soil. The results showed that layered soil requires a deeper trench for reaching the same value of the efficiency as in homogeneous soil. An artificial neural network model and a quartic polynomial equation are developed in order to estimate the efficiency of the geofoam-filled barrier. The agreement between the results of numerical modelling and the developed models demonstrated the capability of the models in predicting the efficiency of the geofoam-filled trench. Finally, an application has been developed to easily use and share all developed models and data. The user can install and use the app to access all data, predicting the efficiency of the trench and optimizing the governing parameters.
- Chair of Geotechnical Engineering and Institute of Geomechanics and Underground Technology