Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms

Vicky-Lee Taljaard; Dewald Z. Gravett; Christos Mourlas; George Markou; Nikolaos Bakas; Manolis Papadrakakis

doi:10.7712/120121.8748.18534

Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms

Vicky-Lee Taljaard
, Dewald Z. Gravett
, Christos Mourlas
, George Markou
, Nikolaos Bakas
, Manolis Papadrakakis

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The fundamental period of a structure is one of the key parameters utilized in the design phase to compute the seismic-resistant forces. Although the importance of seismic-resistant buildings is well understood it has been found that the current design code formulae, which are used to predict the fundamental period of reinforced concrete (RC) buildings are quite simplistic, failing to accurately predict the natural frequency, raising many concerns with regards to their reliability. The primary objective of this research project was to develop a formula that has the ability to compute the fundamental period of an RC structure, while taking into account the soil-structure interaction phenomenon. This was achieved by using a computationally efficient and robust 3D detailed modelling approach for modal analysis obtaining the numerically predicted fundamental period of 475 models, producing a dataset with numerical results. This dataset was then used to train a machine learning algorithm to formulate three fundamental period formulae using a higher-order, nonlinear regression modelling framework. The three newly proposed formulae were evaluated during the validation phase to investigate their performance using 60 new out-of-sample modal results, where, in this work, additional validation models are created and used to test the predictive abilities of the proposed fundamental period formulae. The findings of this research report suggest that the proposed fundamental period formulae exhibit exceptional predictive capabilities for the under-study RC multi-storey buildings, where they outperform all existing de-sign code fundamental period formulae currently in effect.

Original language	English
Title of host publication	COMPDYN 2021 Proceedings
Editors	M. Papadrakakis, M. Fragiadakis
Publisher	ECCOMAS Proceedia
Pages	3801-3809
Number of pages	9
Volume	2
ISBN (Electronic)	9786188507258
DOIs	https://doi.org/10.7712/120121.8748.18534
Publication status	Published - Jun 2021
Event	COMPDYN 2021: 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering - Athens, Greece Duration: 28 Jun 2021 → 30 Jun 2021 https://2021.compdyn.org/

Publication series

Name
ISSN (Electronic)	2623-3347

Conference

Conference	COMPDYN 2021
Country/Territory	Greece
City	Athens
Period	28/06/21 → 30/06/21
Internet address	https://2021.compdyn.org/

Keywords

Fundamental Period Formula
Soil-Structure Interaction
Machine Learning Algorithms
Modal Analysis
Finite Element Method
Reinforced Concrete

Access to Document

10.7712/120121.8748.18534

Cite this

Taljaard, V.-L., Gravett, D. Z., Mourlas, C., Markou, G., Bakas, N., & Papadrakakis, M. (2021). Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms. In M. Papadrakakis, & M. Fragiadakis (Eds.), COMPDYN 2021 Proceedings (Vol. 2, pp. 3801-3809). Article 8748 ECCOMAS Proceedia. https://doi.org/10.7712/120121.8748.18534

@inproceedings{4f18aad7183d49a192e8056bd9cbcfff,

title = "Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms",

abstract = "The fundamental period of a structure is one of the key parameters utilized in the design phase to compute the seismic-resistant forces. Although the importance of seismic-resistant buildings is well understood it has been found that the current design code formulae, which are used to predict the fundamental period of reinforced concrete (RC) buildings are quite simplistic, failing to accurately predict the natural frequency, raising many concerns with regards to their reliability. The primary objective of this research project was to develop a formula that has the ability to compute the fundamental period of an RC structure, while taking into account the soil-structure interaction phenomenon. This was achieved by using a computationally efficient and robust 3D detailed modelling approach for modal analysis obtaining the numerically predicted fundamental period of 475 models, producing a dataset with numerical results. This dataset was then used to train a machine learning algorithm to formulate three fundamental period formulae using a higher-order, nonlinear regression modelling framework. The three newly proposed formulae were evaluated during the validation phase to investigate their performance using 60 new out-of-sample modal results, where, in this work, additional validation models are created and used to test the predictive abilities of the proposed fundamental period formulae. The findings of this research report suggest that the proposed fundamental period formulae exhibit exceptional predictive capabilities for the under-study RC multi-storey buildings, where they outperform all existing de-sign code fundamental period formulae currently in effect.",

keywords = "Fundamental Period Formula, Soil-Structure Interaction, Machine Learning Algorithms, Modal Analysis, Finite Element Method, Reinforced Concrete",

author = "Vicky-Lee Taljaard and Gravett, \{Dewald Z.\} and Christos Mourlas and George Markou and Nikolaos Bakas and Manolis Papadrakakis",

note = "Copyright: {\textcopyright} 2021 The Authors. Published by Eccomas Proceedia.; COMPDYN 2021 : 8th International Conference on Computational Methods in Structural Dynamics and Earthquake Engineering ; Conference date: 28-06-2021 Through 30-06-2021",

year = "2021",

month = jun,

doi = "10.7712/120121.8748.18534",

language = "English",

volume = "2",

publisher = "ECCOMAS Proceedia",

pages = "3801--3809",

editor = "M. Papadrakakis and M. Fragiadakis",

booktitle = "COMPDYN 2021 Proceedings",

url = "https://2021.compdyn.org/",

}

Taljaard, V-L, Gravett, DZ, Mourlas, C, Markou, G, Bakas, N & Papadrakakis, M 2021, Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms. in M Papadrakakis & M Fragiadakis (eds), COMPDYN 2021 Proceedings. vol. 2, 8748, ECCOMAS Proceedia, pp. 3801-3809, COMPDYN 2021, Athens, Greece, 28/06/21. https://doi.org/10.7712/120121.8748.18534

Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms. / Taljaard, Vicky-Lee; Gravett, Dewald Z.; Mourlas, Christos et al.
COMPDYN 2021 Proceedings. ed. / M. Papadrakakis; M. Fragiadakis. Vol. 2 ECCOMAS Proceedia, 2021. p. 3801-3809 8748.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms

AU - Taljaard, Vicky-Lee

AU - Gravett, Dewald Z.

AU - Mourlas, Christos

AU - Markou, George

AU - Bakas, Nikolaos

AU - Papadrakakis, Manolis

PY - 2021/6

Y1 - 2021/6

N2 - The fundamental period of a structure is one of the key parameters utilized in the design phase to compute the seismic-resistant forces. Although the importance of seismic-resistant buildings is well understood it has been found that the current design code formulae, which are used to predict the fundamental period of reinforced concrete (RC) buildings are quite simplistic, failing to accurately predict the natural frequency, raising many concerns with regards to their reliability. The primary objective of this research project was to develop a formula that has the ability to compute the fundamental period of an RC structure, while taking into account the soil-structure interaction phenomenon. This was achieved by using a computationally efficient and robust 3D detailed modelling approach for modal analysis obtaining the numerically predicted fundamental period of 475 models, producing a dataset with numerical results. This dataset was then used to train a machine learning algorithm to formulate three fundamental period formulae using a higher-order, nonlinear regression modelling framework. The three newly proposed formulae were evaluated during the validation phase to investigate their performance using 60 new out-of-sample modal results, where, in this work, additional validation models are created and used to test the predictive abilities of the proposed fundamental period formulae. The findings of this research report suggest that the proposed fundamental period formulae exhibit exceptional predictive capabilities for the under-study RC multi-storey buildings, where they outperform all existing de-sign code fundamental period formulae currently in effect.

AB - The fundamental period of a structure is one of the key parameters utilized in the design phase to compute the seismic-resistant forces. Although the importance of seismic-resistant buildings is well understood it has been found that the current design code formulae, which are used to predict the fundamental period of reinforced concrete (RC) buildings are quite simplistic, failing to accurately predict the natural frequency, raising many concerns with regards to their reliability. The primary objective of this research project was to develop a formula that has the ability to compute the fundamental period of an RC structure, while taking into account the soil-structure interaction phenomenon. This was achieved by using a computationally efficient and robust 3D detailed modelling approach for modal analysis obtaining the numerically predicted fundamental period of 475 models, producing a dataset with numerical results. This dataset was then used to train a machine learning algorithm to formulate three fundamental period formulae using a higher-order, nonlinear regression modelling framework. The three newly proposed formulae were evaluated during the validation phase to investigate their performance using 60 new out-of-sample modal results, where, in this work, additional validation models are created and used to test the predictive abilities of the proposed fundamental period formulae. The findings of this research report suggest that the proposed fundamental period formulae exhibit exceptional predictive capabilities for the under-study RC multi-storey buildings, where they outperform all existing de-sign code fundamental period formulae currently in effect.

KW - Fundamental Period Formula

KW - Soil-Structure Interaction

KW - Machine Learning Algorithms

KW - Modal Analysis

KW - Finite Element Method

KW - Reinforced Concrete

UR - https://www.scopus.com/pages/publications/85120807043

U2 - 10.7712/120121.8748.18534

DO - 10.7712/120121.8748.18534

M3 - Conference contribution

VL - 2

SP - 3801

EP - 3809

BT - COMPDYN 2021 Proceedings

A2 - Papadrakakis, M.

A2 - Fragiadakis, M.

PB - ECCOMAS Proceedia

T2 - COMPDYN 2021

Y2 - 28 June 2021 through 30 June 2021

ER -

Development Of A New Fundamental Period Formula By Considering Soil-structure Interaction With The Use Of Machine Learning Algorithms

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this