Machine learning-based prediction of seismic response of structures
Identyfikator grantu: PT00998
Kierownik projektu: Neda Asgarkhani
Politechnika Gdańska
Wydział Inżynierii Lądowej i Środowiska
Gdańsk
Data otwarcia: 2022-10-07
Streszczenie projektu
The vulnerability of a building can be evaluated either by the in-situ technique of data analysis with non-constructive methods, known as structural health monitoring, or numerical analysis of structural models. The main idea of using such methods is to evaluate the performance of a building in its operating condition. Although the in-situ technique can provide a wide range of data, some practical limitations such as implementing the sensors and mechanical problems during the time can prevent the performance assessment of a building. Therefore, this method can be improved by response prediction methods for buildings subjected to seismic excitations.
Nowadays, the seismic probabilistic assessment of a building needs to perform complicated analysis using a precise finite element model, which is hard for the civil engineering community. Due to the unpredictable nature of ground motions, it is necessary to estimate the nonlinear structural response during seismic loads to take precautions for reducing the probability of collapse risk. The most common ways of estimating seismic response are conducting nonlinear time history analysis and Incremental Dynamic Analysis (IDA) using prior seismic events and finite element methods. The prediction of seismic response using these approaches needed to model complex models and perform time-consuming analysis, while using simplified models (e.g. single-degree of freedom model) are computationally efficient with low performance and behaviour compared to the real structures. Therefore, there is a need to introduce a novel ML-based method to efficiently and accurately predict the seismic response of structures. Predicting the seismic limit-state capacities of structures can help engineers to find a preliminary estimation for the performance levels of the designed structure.
This research aims to implement the most well-known Machine Learning (ML) algorithms in Python software to propose a solution for predicting seismic responses. To prepare the training and testing datasets for developing data-driven decision techniques, IDAs were performed considering Reinforced Concrete Moment-Resisting Frames (RC MRFs) and steel MRFs with different story levels and structural plans assuming seismic excitations introduced by seismic provisions. Then, whole models should be models in Opensees and analysis should be done based on this software, which may take some days for each model. Then, important structural features were considered in datasets to train and test the ML-based prediction models to find the most precise algorithms for seismic probabilistic prediction of steel and RC MRFs. We submitted three papers regarding this method.
In the doctoral thesis, we have many models of structures, in which seismic analysis should be performed using cloud systems. To do this, an algorithm to use Matlab and Opensees software simultaneously was developed. Each model should run with a system and it takes between 3 to 5 days to have results, depending on the kind of system. Therefore, we need at least 8 systems which can help to perform analysis. Software needed: Windows 64-bit, Matlab 2020 or newer version, TCL editor, Notepad++, Python and Opensees 2.5.0, which are open access and can be installed in any system. The virtual machines (CI TASK cloud computing) can be 8 systems with selected information on the website (4 VCPU 10 GB RAM 40 GB SSD). We also need a system with the information of "8 VCPU 20 GB RAM 80 GB SSD".
Nowadays, the seismic probabilistic assessment of a building needs to perform complicated analysis using a precise finite element model, which is hard for the civil engineering community. Due to the unpredictable nature of ground motions, it is necessary to estimate the nonlinear structural response during seismic loads to take precautions for reducing the probability of collapse risk. The most common ways of estimating seismic response are conducting nonlinear time history analysis and Incremental Dynamic Analysis (IDA) using prior seismic events and finite element methods. The prediction of seismic response using these approaches needed to model complex models and perform time-consuming analysis, while using simplified models (e.g. single-degree of freedom model) are computationally efficient with low performance and behaviour compared to the real structures. Therefore, there is a need to introduce a novel ML-based method to efficiently and accurately predict the seismic response of structures. Predicting the seismic limit-state capacities of structures can help engineers to find a preliminary estimation for the performance levels of the designed structure.
This research aims to implement the most well-known Machine Learning (ML) algorithms in Python software to propose a solution for predicting seismic responses. To prepare the training and testing datasets for developing data-driven decision techniques, IDAs were performed considering Reinforced Concrete Moment-Resisting Frames (RC MRFs) and steel MRFs with different story levels and structural plans assuming seismic excitations introduced by seismic provisions. Then, whole models should be models in Opensees and analysis should be done based on this software, which may take some days for each model. Then, important structural features were considered in datasets to train and test the ML-based prediction models to find the most precise algorithms for seismic probabilistic prediction of steel and RC MRFs. We submitted three papers regarding this method.
In the doctoral thesis, we have many models of structures, in which seismic analysis should be performed using cloud systems. To do this, an algorithm to use Matlab and Opensees software simultaneously was developed. Each model should run with a system and it takes between 3 to 5 days to have results, depending on the kind of system. Therefore, we need at least 8 systems which can help to perform analysis. Software needed: Windows 64-bit, Matlab 2020 or newer version, TCL editor, Notepad++, Python and Opensees 2.5.0, which are open access and can be installed in any system. The virtual machines (CI TASK cloud computing) can be 8 systems with selected information on the website (4 VCPU 10 GB RAM 40 GB SSD). We also need a system with the information of "8 VCPU 20 GB RAM 80 GB SSD".
Publikacje
- Farzin Kazemi, Neda Asgarkhani, Robert Jankowski, Machine learning-based seismic fragility and seismic vulnerability assessment of reinforced concrete structures, Soil Dynamics and Earthquake Engineering 166, (2023) 1
- Farzin Kazemi, Neda Asgarkhani, Robert Jankowski, Predicting seismic response of SMRFs founded on different soil types using machine learning techniques, Engineering Structures 274, (2023) 1
- Farzin Kazemi, Neda Asgarkhani, Robert Jankowski, Machine learning-based seismic response and performance assessment of reinforced concrete buildings, Archives of Civil and Mechanical Engineering 23, (2023) 1
- Farzin Kazemi, Neda Asgarkhani, Robert Jankowski, Probabilistic assessment of SMRFs with infill masonry walls incorporating nonlinear soil-structure interaction, Bulletin of Earthquake Engineering 21, (2023) 1
- Neda Asgarkhani, Farzin Kazemi, Robert Jankowski, Machine learning-based prediction of residual drift and seismic risk assessment of steel moment-resisting frames considering soil-structure interaction, Computers & Structures 289, (2023) 1
- Farzin Kazemi, Neda Asgarkhani, Ahmed Manguri, Natalia Lasowicz, Robert Jankowski, Introducing a computational method to retrofit damaged buildings under seismic mainshock-aftershock sequence, International Conference on Computational Science 1, (2023) 180-187
- Benyamin Mohebi, Farzin Kazemi, Neda Asgarkhani, RETROFITTING DAMAGED BUILDINGS UNDER SEISMIC MAINSHOCK-AFTERSHOCK SEQUENCE, COMPDYN 2023 1, (2023) 1
- Farzin Kazemi, Benyamin Mohebi, Neda Asgarkhani, Atefeh Yousefi, Advanced Scalar-valued Intensity Measures for Residual Drift Prediction of SMRFs with Fluid Viscous Dampers, International Journal of Structural Integrity 12, (2023) 20-25