中小學校舍耐震評估模式之優化－以敏感度分析及人工智慧理論為研究方法

To Optimize the Seismic Assessment Model of School Building using Sensitivity Analysis and Artificial Intelligence Theory

10.3966/101632122021030115001

陳清山(Ching-Shan Chen)

校舍 ； 人工智慧 ； 敏感度分析 ； 基因表達規劃法 ； 支持向量機 ； School Building ； Artificial Intelligence ； Sensitivity Analysis ； Gene Expression Programming ； Support Vector Machine

建築學報

115期（2021 / 03 / 31）

1 - 20

繁體中文

中小學校舍具有教育場所及災民臨時收容所之雙重機能，地位極為重要。如何正確且快速評估現有校舍耐震能力，以發揮校舍應有的功能，乃一件刻不容緩的工作。此外目前研究人員以人工智慧推論校舍耐震能力時，對於如何選擇適當的案例數量和耐震因子數目，亦常感到困擾，因為兩者之數目過多，建立資料庫及推論時將極耗費時間和成本，數目太少則推論結果欠佳，這也是一個值得探討的課題。為解決上述課題，本研究採用數種優化耐震評估模式的研究方法:首先利用敏感度分析理論，試驗於何種校舍案例數量和耐震因子數目下，可得到較佳的耐震評估模式;再採用灰色理論探討校舍耐震因子與耐震能力之關聯度;最後應用支持向量機和基因表達規劃法推論最佳校舍耐震評估模式，並以交叉驗證法（10-Fold Cross-Validation）驗證支持向量機。從研究結果中可知，研究人員於應用人工智慧理論時，因子數目至少應取5個，若欲使人工智慧呈現較佳之推論結果，則因子數目可取10個以上。測試案例數量則至少應取因子數目之2倍，3倍以上尤佳。以支持向量機和基因表達規劃法所推論的耐震評估模式，均得到頗佳的結果。此研究成果可提供建築師及一般工程人員使用，所發展之研究方法，亦可供學術界後續研究之參考。

School buildings are designed to serve both as places of education and as temporary shelters in the aftermath of major earthquakes, such the status of school buildings are very important. Therefore, how to correctly and quickly assess the seismic performance of an existing school building is an urgent issue that deserved to further investigate. Moreover, when the researchers using artificial intelligence to infer the seismic performance of the school building, it is also a disturbing topic worthy of discussion on how to determine the appropriate number of school cases and seismic factors. Because when the number is excessive, it will take lots of time and cost to build the seismic database and infer the seismic assessment model. If the number is insufficient, the inference result will be not satisfactory. In order to solve the above subjects, thispaper used several research methods to optimize the seismic assessment model. Firstly, the sensitivity analysis was applied to test an optimal model under the consideration of number of school cases and seismic factors.Then used the grey theory to explore the relationships between the seismic factors and seismic performance of school buildings.Finally, adopted support vector machine (SVM) and gene expression programming (GEP) to deduce the optimal models,SVM were also validated by 10-Fold Cross-Validation.Results show that when researchers apply artificial intelligence theory, the number of factors should be at least five. If researchers want to get a better inference result, the number of factors can be more than ten. As for the number of testing cases, two times the number of factors should be taken, and more than three times is preferred. The seismic assessment models inferred by SVM and GEP, possess good performance. These results can be used by architects andgeneral engineers, and the developed research methodscan also be referenced for subsequent researches in academia.

1. 王文清, W.C.,廖佳俊, C.C.(2016)。學校建築耐震能力資料探勘模型評估與系統建置。災害防救科技與管理，5(1)，1-25。
   連結：
2. 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：
3. 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：
4. 葉勇凱, Y.K.,周德光, T.K.(2018)。典型校舍耐震性能與易損性曲線探討。中國土木水利工程學刊，30(2)，141-149。
   連結：
5. 鄭明淵, M.Y.,張于漢, Y.H.,Prayogo, Doddy,吳建燁, J.Y.(2018)。演化式人工智慧建立專案實獲完工工期推論模式之研究。建築學報，104，73-87。
   連結：
6. Adhikari, R.,Gautam, D.(2019).Component level seismic fragility functions and damage probability matrices for Nepali school buildings.Soil Dynamics and Earthquake Engineering,120,316-319.
7. Anelli, A.,Santa-Cruz, S.,Vona, M.,Tarque, N.,Laterza, M.(2019).A proactive and resilient seismic risk mitigation strategy for existing school buildings.Structure and Infrastructure Engineering,15(2),137-151.
8. Armaghani, D. J.,Safari, V.,Fahimifar, A.,Monjezi, M.,Mohammadi, M. A.(2018).Uniaxial compressive strength prediction through a new technique based on gene expression programming.Neural Computing and Applications,30(11),3523-3532.
9. Chao, Y. F.,Chen, H. M.(2016).Automatic modeling method for regional seismic damage assessment of school buildings in Taiwan.Journal of the Chinese Institute of Engineers,39(5),606-614.
10. Chen, C. S.,Cheng, M. Y.,Wu, Y. W.(2012).Seismic assessment of school buildings in Taiwan using the evolutionary support vector machine inference system.Expert Systems with Applications,39(4),4102-4110.
11. Chen, H. M.,Kao, W. K.,Tsai, H. C.(2012).Genetic programming for predicting aseismic abilities of school buildings.Engineering Application of Artificial Intelligence,25(6),1103-1113.
12. Cheng, M. Y.,Chen, C. S.(2011).Optimal planning model for school buildings considering the tradeoff of seismic resistance and cost effectiveness: a Taiwan case study.Structural and Multidisciplinary Optimization,43(6),863-879.
13. Cheng, M. Y.,Chen, C. S.(2014).Preliminary planning efficiency evaluation for school buildings considering the tradeoffs of MOOP and planning preferences.Journal of Civil Engineering and Management,20(2),211-222.
14. Cheng, M. Y.,Wei, H. H.,Wu, Y. W.,Chen, H. M.,Wu, C. W.(2018).Optimization of life-cycle cost of retrofitting school buildings under seismic risk using evolutionary support vector machine.Technological and Economic Development of Economy,24(2),812-824.
15. Chi, N. W.,Wang, J. P.,Liao, J. H.,Cheng, W. C.,Chen, C. S.(2020).Machine learning-based seismic capability evaluation for school buildings.Automation in Construction,118,103274.
16. Chung, L. L.,Chen, Y. T.,Sun, C. H.,Lien, K. H.,Wu, L. Y.(2012).Applicability investigation of code-defined procedures on seismic performance assessment of typical school buildings in Taiwan.Engineering Structures,36,147-159.
17. Deng, J. L.(1982).Control problem of grey system.Systems & Control Letters,1(5),288-294.
18. Drucker, H.,Burges, C.,Kaufman, L.,Smola, A.,Vapnik, V. N.(1996).Support vector regression machines.Advances in Neural Information Processing Systems,9,155-161.
19. Farzampour, A.,Mansouri, I.,Mortazavi, S. J.,Hu, J. W.(2019).Force-displacement relationship of a butterfly-shaped beams based on gene expression programming.10th International Symposium on Steel Structures,Jeju, Korea:
20. Ferreira, C.(2001).Gene Expression Programming: A new adaptive algorithm for solving problems.Complex Systems,13(2),87-129.
21. Hajihassani, M.,Abdullah, S. S.,Asteris, P. G.,Armaghani, D. J.(2019).A gene expression programming model for predicting tunnel convergence.Applied Sciences,9(21),4650.
22. Hu, B.,Wei, X.,Lv, H.,Kundu, T.,Li, N.(2019).Experimental and analytical study on seismic behavior of strengthened existing single frame structures with exterior cantilevers.Advances in Materials Science and Engineering
23. Hu, J. L.,Tang, X. W.,Qiu, J. N.(2016).Assessment of seismic liquefaction potential based on Bayesian network constructed from domain knowledge and history data.Soil Dynamics and Earthquake Engineering,89,49-60.
24. Jafari, S.,Mahini, S. S.(2017).Lightweight concrete design using gene expression programing.Construction and Building Materials,139,93-100.
25. Kohavi, R.(1995).A study of cross-validation and bootstrap for accuracy estimation and model selection.Proceedings of International Joint Conference on AI
26. Kose, M. M.,Kayadelen, C.(2013).Effects of infill walls on RC buildings under time history loading using genetic programming and neuro-fuzzy.Structural Engineering and Mechanics,47(3),401-419.
27. Luo, H.,Paal, S. G.(2019).A locally weighted machine learning model for generalized prediction of drift capacity in seismic vulnerability assessments.Computer‐ Aided Civil and Infrastructure Engineering,34(11),935-950.
28. Miano, A.,Chiumiento, G.(2020).Case study on seismic retrofit and cost assessment for a school building.Structural Engineering and Mechanics,73(1),53-64.
29. Motamedi, M.,Ventura, C. E.(2020).System identification of a full scale wood frame building specimen subjected to shake table tests.Topics in Modal Analysis & Testing,Cham:
30. Motlagh, Z. S.,Dehkordi, M. R.,Eghbali, M.,Samadian, D.(2020).Evaluation of seismic resilience index for typical RC school buildings considering carbonate corrosion effects.International Journal of Disaster Risk Reduction,46,101511.
31. Segovia-Verjel, M. L.,Requena-García-Cruz, M. V.,de-Justo-Moscardó, E.,Morales-Esteban, A.(2019).Optimal seismic retrofitting techniques for URM school buildings located in the southwestern Iberian peninsula.PloS one,14(10),e0223491.
32. Seymour, G.(1994).Predictive Inference.New York:Chapman and Hall.
33. Shiwaku, K.,Ueda, Y.,Oikawa, Y.,Shaw, R.(2016).School disaster resilience assessment in the affected areas of 2011 East Japan earthquake and tsunami.Natural Hazards,82(1),333-365.
34. Tarawneh, B.(2018).Gene expression programming model to predict driven pipe piles set-up.International Journal of Geotechnical Engineering,14(5),538-544.
35. Vapnik, V. N.(1995).The nature of statistical learning theory.New York:Springer.
36. Zhang, Z.,Hsu, T. Y.,Wei, H. H.,Chen, J. H.(2019).Development of a data-mining technique for regional-scale evaluation of building seismic vulnerability.Applied Sciences,9(7),1502.
37. 鍾立來, L.L.,黃鈞鼎, C.T.,楊耀昇, Y.S.,林聖學, S.X.,賴勇安, Y.A.,吳賴雲, L.Y.(2015)。以現地試驗驗證口湖國小含磚牆校舍之簡易側推分析。中國土木水利工程學刊，27(1)，59-68。