街屋耐震評估模型之研究－以人工智慧及敏感度分析理論為研究方法

To Interpret the Seismic Assessment Models of Street Houses Using Artificial Intelligence and Sensitivity Analysis Theories

10.53106/101632122022060120002

陳清山(Ching-Shan Chen)

街屋 ； 人工智慧 ； 耐震評估 ； 基因表達規劃法 ； 類神經網路 ； Street House ； Artificial Intelligence ； Seismic Assessment ； Gene Expression Programming ； Artificial Neural Network

建築學報

120期（2022 / 06 / 30）

17 - 38

繁體中文

街屋為台灣常見的住宅型態，據營建署2015年住宅調查統計，街屋約占住宅類型中49.20%之比例，可見街屋為台灣重要的住宅類型。此種低矮型建築型態在歷次地震中破壞嚴重，遭受極大之生命及財產損失。如何快速評估現有街屋耐震能力，以發揮街屋應有的功能，乃一件刻不容緩的工作。除此之外，目前研究人員以人工智慧推論建築物耐震能力時，對於如何決定適當的因子數目，以及如何判定推論模型的優劣，亦常感到困擾，這也是一個值得探討的課題。本論文以崩塌地表加速度代表街屋耐震能力，由於崩塌地表加速度之計算常須耗費大量的時間及金錢，且必須倚賴該領域專家之知識才能訂定，非一般工程人員可以勝任。為解決上述課題，並保存專家寶貴的知識，採用主成份分析法、資料探勘以及人工智慧中的灰色理論、類神經網路和基因表達規劃法，推論街屋之耐震因子及耐震能力；並應用敏感度分析，測試於不同耐震因子數目下，各耐震評估模型之推論能力，結果顯示，13個耐震因子的推論能力最佳，類神經網路和基因表達規劃法之推論結果亦頗為良好。本研究主要採用人工智慧理論為研究方法，以不同的面向探究街屋耐震領域，希望能以不同的研究角度獲得耐震評估的新思維。研究成果可提供建築專業者使用，所發展的研究方法亦可供學術界後續研究的參考。

Street house is a common type of housing in Taiwan. According to the statistics of the 2015 housing survey of Construction and Planning Agency, street houses occupied 49.20% of the housing types. It can be seen that street house is an important housing type in Taiwan. This type of low-rise building was severely damaged in previous earthquakes and suffered great loss of life and property. Therefore, how to quickly assess the seismic performance of existing street houses are critical issues that deserved to further investigate. Besides, when the researchers using artificial intelligence to infer the seismic performance of street houses, are often confused about how to determine the appropriate number of factors and judge the pros and cons of the seismic inference model. These are also main topics worthy of discussions. This paper adopted the collapse ground acceleration as the seismic performance of street houses. Owing to the calculation of the acceleration often takes lots of time and budget, and must rely on the knowledge of experts in the field to determine, it is difficult for general engineers to accomplish this job. In order to solve the above problems and preserve the valuable knowledge of experts, the principal component analysis, data mining, grey theory, artificial neural network (ANN) and gene expression programming (GEP) method were used to infer the seismic factors and assessment models of street houses. Furthermore, this paper also used the sensitivity analysis to experiment the seismic assessment model under different number of seismic factors. Results show that when using 13 seismic factors, the seismic assessment model is the best, and the inference results of the ANN and GEP are also good. This paper mainly used artificial intelligence theories as the research methods, exploring the seismic assessments of street houses with different aspects, hoping to obtain new vision about seismic evaluation from different research perspectives. The research results can be used by construction professionals, and the developed research methods also can be referenced for subsequent researches in academia.

1. 杜怡萱, Y. H.,葉桐, T.(2020)。以 2018 花蓮地震震害建物探討耐震初評方法之有效性。結構工程，35(2)，43-64。
   連結：
2. 張嘉祥, J. S.,王貞富, C. F.(2001)。九二一集集地震磚造歷史街屋震害調查研究。建築學報，37，93-113。
   連結：
3. 陳清山, C. S.(2017)。以人工智慧理論探討臺中市中小學校舍耐震因子及耐震能力。建築學報，100，95-116。
   連結：
4. 陳清山, C. S.(2020)。規劃設計階段考量中小學體育館之耐震因子及耐震能力─以多變量及人工智慧理論為研究方法。建築學報，111，55-75。
   連結：
5. 賴昱志, Y. C.,賴濤, T.,鍾立來, L. L.,黃國倫, G. L.,楊耀昇, Y. S.,曾建創, C. C.,林聖學, S. H.,張筑媛, C. Y.(2018)。老舊建物耐震能力之簡易詳細評估。結構工程，33(2)，69-87。
   連結：
6. Chen, C. S.(2020).Seismic performance assessments of school buildings in Taiwan using artificial intelligence theories.Engineering Computations,37(9),3321-3343.
7. Chiu, C. K.,Wu, C. H.,Sung, H. F.,Liao, W. I.,Lin, C. H.(2020).Application of post-embedded piezoceramic sensors for force detection on RC columns under seismic loading.Applied Sciences,10(15),5061.
8. Deng, J. L.(1982).Control problem of grey system.Systems & Control Letters,1(5),288-294.
9. Ferreira, C.(2001).Gene expression programming: A new adaptive algorithm for solving problems.Complex Systems,13(2),87-129.
10. Güllü, H.(2012).Prediction of peak ground acceleration by genetic expression programming and regression: a comparison using likelihood-based measure.Engineering Geology,141,92-113.
11. Güneyisi, E. M.,D'Aniello, M.,Landolfo, R.,Mermerdaş, K.(2013).A novel formulation of the flexural overstrength factor for steel beams.Journal of Constructional Steel Research,90,60-71.
12. Hansapinyo, C.,Latcharote, P.,Ketsap, A.,Limkatanyu, S.(2020).Seismic building damage prediction from GIS-based building data using artificial intelligence system.Frontiers in Built Environment,6,178.
13. Kaiser, H. F.(1958).The varimax criterion for analytic rotation in factor analysis.Psychometrika,23(3),187-200.
14. Khaleghi, M.,Salimi, J.,Farhangi, V.,Moradi, M. J.,Karakouzian, M.(2021).Application of artificial neural network to predict load bearing capacity and stiffness of perforated masonry walls.Civil Engineering,2(1),48-67.
15. Londhe, S. N.,Kwatra, N.(2014).Application of artificial neural networks for dynamic analysis of building frames.Computers and Concrete,13(6),765-780.
16. Mansouri, I.,Güneyisi, E. M.,Mosalam, K. M.(2021).Improved shear strength model for exterior reinforced concrete beam-column joints using gene expression programming.Engineering Structures,228,111563.
17. Momeni, M.,Hadianfard, M. A.,Bedon, C.,Baghlani, A.(2020).Damage evaluation of H-section steel columns under impulsive blast loads via gene expression programming.Engineering Structures,219,110909.
18. Nikose, T. J.,Sonparote, R. S.(2021).Application of artificial neural network for predicting dynamic along‐wind response of tall buildings.The Structural Design of Tall and Special Buildings,e1837.
19. Shahrara, N.,Çelik, T.,Gandomi, A. H.(2017).Gene expression programming approach to cost estimation formulation for utility projects.Journal of Civil Engineering and Management,23(1),85-95.
20. Shapiro, B. P.(1977).Can marketing and manufacturing coexist?.Harvard Business Review,55,104-114.
21. Vapnik, V. N.(1995).The Nature of Statistical Learning Theory.NY, USA:Springer.
22. Vazirizade, S. M.,Nozhati, S.,Zadeh, M. A.(2017).Seismic reliability assessment of structures using artificial neural network.Journal of Building Engineering,11,230-235.
23. Yamashita, T.,Kohiyama, M.,Watanabe, K.(2021).Deep neural network for detecting earthquake damage to brace members installed in a steel frame.Japan Architectural Review,4(1),56-64.
24. 邱聰智, T. C.,黃世建, S. J.,宋嘉誠, J. C.,鍾立來, L. L.(2014)。低矮型街屋耐震能力快速評估法之開發與驗證。結構工程，29(4)，65-87。

1. 陳清山(2023)。考量多目標最佳化及建築師規劃偏好之街屋規劃效率評估。建築學報，124，1-20。