Your browser does not support the PubReader view.
Go to this page to see a list of supporting browsers.
한국강구조학회 학술지영문홈페이지
[ Article ]
Journal of Korean Society of Steel Construction - Vol. 36, No. 6, pp.443-450
ISSN: 1226-363X (Print) 2287-4054 (Online)
Print publication date 27 Dec 2024
Received 22 Sep 2024 Revised 20 Nov 2024 Accepted 06 Dec 2024
DOI: https://doi.org/10.7781/kjoss.2024.36.6.443

기계학습 모델을 활용한 공장지붕형 태양광 구조물의 안전성 평가

김시원1 ; 한경민2 ; 이창환3 ; 박민재4, *
1석사과정, 국립부경대학교, 건축ㆍ소방공학부
2대표, 제일구조기술사사무소
3부교수, 국립부경대학교, 건축공학과
4조교수, 국립부경대학교, 건축공학과
Structural Safety Assessment of Factory Rooftop Solar Structures Using Machine Learning Models
Kim, Shi Won1 ; Han, Kyoung Min2 ; Lee, Chang-Hwan3 ; Park, Min Jae4, *
1Graduate Student (Master’s Course), Division of Architectural and Fire Protection Engineering, Pukyong National University, Busan, 48513, Korea
2CEO, Jaeil Structure Eng.Co, 48059, Busan
3Associate Professor, Department of Architectural Engineering, Pukyong National University, 48513, Busan
4Assistant Professor, Department of Architectural Engineering, Pukyong National University, 48513, Busan

Correspondence to: *Tel. +82-51-6739-8739 Fax. +82-51-629-7084 E-mail. mjp@pknu.ac.kr

Copyright © 2024 by Korean Society of Steel Construction

초록

태양광 발전설비 설치 시 비전문가들의 초기 설계와 태양광 구조검토 시 반복적인 설계변경으로 인해 생기는 시간 비용을 해결하기 위한 기계학습 모델을 제안한다. 기계학습 모델로는 K최근접이웃(K-Nearest Neighbors)과 로지스틱회귀(Logistic Regression)를 적용하였다. 최적의 하이퍼 파라미터를 탐색하기 위하여 그리드서치(Grid Search)를 사용했다. 표준화 변환(Standard Scaler)을 사용하여 모델의 성능을 높였다. F1스코어(F1-score)를 사용하여 모델을 평가했다. 2가지 모델은 80 % 이상의 정확도를 보여주었지만, 안전 클래스에 대한 정확도가 낮았다. 2가지 클래스의 불균형으로 인한 데이터 편향이 생긴 것으로 예상된다. 데이터 세트의 조정 및 전처리, 다른 기계학습 모델의 사용을 통해 데이터 편향에 대한 추가적인 연구가 필요하다.

Abstract

This paper proposes a machine learning model to address the time and cost inefficiencies caused by the initial designs of non-experts and repeated design changes during the structural design of photovoltaic installations. The study applies K-Nearest Neighbors (KNN) and Logistic Regression as the machine learning models. GridSearch was used to find the optimal hyperparameters, while StandardScaler was employed to improve model performance. The F1-score was utilized to evaluate the performance of the models. Both models demonstrated a low prediction accuracy for the safety class, likely due to data bias caused by class imbalance. Further research is needed to mitigate this data bias through dataset adjustment and preprocessing, and the exploration of alternative machine learning models.

Keywords:

Steel structures, Solar structures, Machine learning, KNN, Logistic Regression

키워드:

강구조, 태양광구조물, 머신러닝, K최근접이웃, 로지스틱 회귀

Acknowledgments

이 논문은 부경대학교 자율창의학술연구비(2023년)에 의하여 연구되었습니다.

References

  • Kim, C. (2021) A Review of the Deployment Programs, Impact, and Barriers of Renewable Energy Policies in Korea, Renewable and Sustainable Energy Reviews, Elsevier, Vol.144, 110870. [https://doi.org/10.1016/j.rser.2021.110870]
  • Kang, I.S., Moon, J.W., and Park, J.C. (2017) Recent Research Trends of Artificial Intelligent Machine Learning in Architectural Field -Review of Domestic and International Journal Papers-, Journal of the Architectural Institute of Korea Structure & Construction, AIK, Vol.33, No.4, pp.63-68 (in Korean). [https://doi.org/10.5659/JAIK_SC.2017.33.4.63]
  • Baduge, S.K., Thilakarathna, S., Perera, J.S., Arashpour, M., Sharafi, P., Teodosio, B., Shringi, A., and Mendis, P. (2022) Artificial Intelligence and Smart Vision for Building and Construction 4.0: Machine and Deep Learning Methods and Applications, Automation in Construction, Elsevier, Vol.141, 104440. [https://doi.org/10.1016/j.autcon.2022.104440]
  • Kim, B.I. (2012) Optimum Design of Steel Structures Using Genetic Algorithms, Journal of Korean Society of Steel Construction, KSSC, Vol.24, No.6, pp.701-710 (in Korean). [https://doi.org/10.7781/kjoss.2012.24.6.701]
  • Kim, B.I. (2015) Optimum Design for Sizing and Shape of Truss Structures Using Harmony Search and Simulated Annealing, Journal of Korean Society of Steel Construction, KSSC, Vol.27, No.2, pp.131-142 (in Korean). [https://doi.org/10.7781/kjoss.2015.27.2.131]
  • Hashemi, S.S., Sadeghi, K., Fazeli, A., and Zarei, M. (2019) Predicting the Weight of the Steel Moment-Resisting Frame Structures Using Artificial Neural Networks, International Journal of Steel Structures, KSSC, Vol.19, No.1, pp.168-180. [https://doi.org/10.1007/s13296-018-0105-z]
  • Gupta, A.K., Chakroborty, S., Ghosh, S.K., and Ganguly, S. (2023) A Machine Learning Model for Multi-Class Classification of Quenched and Partitioned Steel Microstructure Type by the K-Nearest Neighbor Algorithm, Computational Materials Science, Elsevier, Vol.228, 112321. [https://doi.org/10.1016/j.commatsci.2023.112321]
  • Lee, S., Lee, Y., Lee, K., and Lee, J. (2022) Compressive Strength Prediction of CFST Columns Using Machine Learning Methods, Journal of Korean Society of Steel Construction, KSSC, Vol.34, No.2, pp.101-110 (in Korean). [https://doi.org/10.7781/kjoss.2022.34.2.101]
  • Ha, M.-G., Kwon, T.-Y., Kim, R.-H., and Ahn, J.-H. (2024) Methods of Condition Evaluation and Assessment for paint Chalking of Steel Bridge Depending on Paint Coating Deterioration, Journal of Korean Society of Steel Construction, KSSC, Vol.36, No.1, pp.13-24 (in Korean). [https://doi.org/10.7781/kjoss.2024.36.1.013]
  • Duong, T.H., Le, T.-T., and Le, M.V. (2023) Practical Machine Learning Application for Predicting Axial Capacity of Composite Concrete-Filled Steel Tube Columns Considering Effect of Cross-Sectional Shapes, International Journal of Steel Structures, KSSC, Vol.23, No.1, pp.263-278. [https://doi.org/10.1007/s13296-022-00693-0]
  • Jang, A., Kim, J., Park, M.J., Ju, Y.K., and Kim, S.J. (2022) Analysis of Machine Learning for Detect Concrete Crack Depths Using Infrared Thermography Technique, Proceedings of IABSE Symposium Prague 2022, International Association for Bridge and Structural Engineering, Czech, pp.758-765. [https://doi.org/10.2749/prague.2022.0758]
  • Kim, J., Jang, A., Park, M.J., and Ju, Y.K. (2021) Comparison Analysis of Machine Learning for Concrete Crack Depths Prediction Using Thermal Image and Environmental Parameters, Journal of Korean Association for Spatial Structures, KASSS, Vol.21, No.2, pp.99-110 (in Korean). [https://doi.org/10.9712/KASS.2021.21.2.99]
  • Jeong, S.-G., Jang, A., Park, M.J., and Ju, Y.K. (2023) Defect Evaluation Technology for Welded Parts Using Thermal Image Based on Machine Learning, Proceedings of 2023 Autumn Annual Conference of AIK, Architectural Institute of Korea, Vol.43, No.2, pp.367-368 (in Korean).
  • Architectural Institute of Korea (2019) Building Structure Standards-Design Load, KDS 41 10 15 : 2019, Ministry of Land, Infrastructure and Transport (in Korean).
  • Architectural Institute of Korea (2019) Steel Structures Design Standard, KDS 41 30 10 : 2019, Ministry of Land, Infrastructure and Transport (in Korean).
  • Thai, H.-T. (2022) Machine Learning for Structural Engineering: A State-of-the-Art Review, Structures, Elsevier, Vol.38, pp.448-491. [https://doi.org/10.1016/j.istruc.2022.02.003]