Noise-Resilient Deep Learning for Post-Earthquake Strain Prediction in Steel Frames with Setbacks

Abstract

A systematic engineering framework is presented that translates established machine-learning techniques into a noise-resilient deep neural network (NR-DNN) for post-earthquake building assessment using noisy strain sensor data. The proposed model is built upon four major mechanisms: (1) noise‑aware training using a multiplicative synthetic noise model (calibration error, thermal drift, random perturbations), (2) dropout, (3) Bayesian hyperparameter tuning with K‑fold cross‑validation (CV), and (4) ensemble averaging. An internal ablation study is performed to show that simultaneous incorporation of these mechanisms yields the best results. Random Forest (RF) is used to identify the best locations for strain monitoring. The Performance of the model is investigated on two SMRF case studies using nonlinear time history analyses (NTHAs) of 58 ground motion records at immediate occupancy (IO) and life safety (LS) performance levels, plus 43 out-of-sample collapse level records. The model predicts full field strains with a limited number of strain sensors under highly nonlinear structural responses caused by unseen out-of-sample earthquake excitations. Compared with a conventional DNN, the proposed model reduces root mean squared error (RMSE) by up to 90 % under noisy conditions and maintains high damage state classification accuracy. The sensitivity analyses validate the framework's stability both under varying noise levels and under threshold variations in damage state classification. The results confirm that, unlike a DNN trained without appropriate noise-handling mechanisms, the proposed NR-DNN model remains numerically stable under noisy conditions. This validation is numerical, based on a synthetic noise model; experimental field validation is left for future work.

Keywords

• Noise resilient deep neural network
• post-earthquake strain prediction
• sensor placement optimization
• structural health monitoring

References

1. Sivasuriyan A, Vijayan DS, Górski W, Wodzyński Ł, Vaverková MD, Koda E. Practical implementation of structural health monitoring in multi-story buildings. Buildings 2021;11:263. https://doi.org/10.3390/buildings11060263.
2. Tsuchimoto K, Narazaki Y, Spencer BF. Development and validation of a post-earthquake safety assessment system for high-rise buildings using acceleration. Mathematics 2021;9:1758. https://doi.org/10.3390/math9151758.
3. Tsuchimoto K, Narazaki Y, Hoskere V, Spencer BF. Rapid postearthquake safety evaluation of buildings using sparse acceleration measurements. Struct Health Monit 2021;20:1822–40. https://doi.org/10.1177/1475921720936296.
4. Damikoukas S, Chatzieleftheriou S, Lagaros ND. First level pre- and post-earthquake building seismic assessment protocol based on dynamic characteristics extracted in situ. Infrastructures (Basel) 2022;7:115. https://doi.org/10.3390/infrastructures7090115.
5. Wang Y, Cui L, Zhang C, Chen W, Xu Y, Zhang Q. A two-stage seismic damage assessment method for small, dense, and imbalanced buildings in remote sensing images. Remote Sens (Basel) 2022;14:1012. https://doi.org/10.3390/rs14041012.
6. Kaplan O, Kaplan G. A response spectra-based Post-earthquake rapid structural damage estimation approach aided with remote sensing data: 2020 Samos earthquake. Buildings 2021;12:14. https://doi.org/10.3390/buildings12010014.
7. Mangalathu S, Sun H, Nweke CC, Yi Z, Burton H V. Classifying earthquake damage to buildings using machine learning. Earthquake Spectra 2020;36:183–208. https://doi.org/10.1177/8755293019878137.
8. You T, Wang W, Fang C, Chen Y. Rapid probabilistic loss assessment of buildings based on post-earthquake structural deformation conditions. Journal of Building Engineering 2022;45:103629. https://doi.org/10.1016/j.jobe.2021.103629.

9. Roohi M, Hernandez EM. Performance-based post-earthquake decision making for instrumented buildings. J Civ Struct Health Monit 2020;10:775–92. https://doi.org/10.1007/s13349-020-00416-1.
10. Lin J-F, Li X-Y, Wang J, Wang L-X, Hu X-X, Liu J-X. Study of building safety monitoring by using cost-effective MEMS accelerometers for rapid after-earthquake assessment with missing data. Sensors 2021;21:7327. https://doi.org/10.3390/s21217327.
11. Ibrahim A, Eltawil A, Na Y, El-Tawil S. A machine learning approach for structural health monitoring using noisy data sets. IEEE Transactions on Automation Science and Engineering 2020;17:900–8. https://doi.org/10.1109/TASE.2019.2950958.
12. Salem M, Gomaa A, Tsurusaki N. Detection of earthquake-induced building damages using remote sensing data and deep learning: a case study of Mashiki town, Japan. IGARSS 2023 - 2023 IEEE International Geoscience and Remote Sensing Symposium, IEEE; 2023, p. 2350–3. https://doi.org/10.1109/IGARSS52108.2023.10282550.
13. Foroughnia F, Macchiarulo V, Berg L, DeJong M, Milillo P, Hudnut KW, et al. Quantitative assessment of earthquake-induced building damage at regional scale using LiDAR data. International Journal of Disaster Risk Reduction 2024;106:104403. https://doi.org/10.1016/j.ijdrr.2024.104403.
14. Ghosh Mondal T, Jahanshahi MR, Wu R, Wu ZY. Deep learning‐based multi‐class damage detection for autonomous post‐disaster reconnaissance. Struct Control Health Monit 2020;27. https://doi.org/10.1002/stc.2507.
15. Xia H, Wu J, Yao J, Zhu H, Gong A, Yang J, et al. A deep learning application for building damage assessment using ultra-high-resolution remote sensing imagery in Turkey earthquake. International Journal of Disaster Risk Science 2023;14:947–62. https://doi.org/10.1007/s13753-023-00526-6.
16. Sri Preethaa KR, Munisamy SD, Rajendran A, Muthuramalingam A, Natarajan Y, Yusuf Ali AA. Novel ANOVA-statistic-reduced deep fully connected neural network for the damage grade prediction of post-earthquake buildings. Sensors 2023;23:6439. https://doi.org/10.3390/s23146439.
17. Özerol Özman G, Arslan Selçuk S, Arslan A. Image classification on post-earthquake damage assessment: a case of the 2023 Kahramanmaraş earthquake. Engineering Science and Technology, an International Journal 2024;56:101780. https://doi.org/10.1016/j.jestch.2024.101780.
18. Yang Y, Xie C, Tian B, Guo Y, Zhu Y, Yang Y, et al. Large-scale building damage assessment based on recurrent neural networks using SAR coherence time series: A case study of 2023 Turkey–Syria earthquake. Earthquake Spectra 2024;40:2285–305. https://doi.org/10.1177/87552930241262761.
19. Jia J, Ye W. Deep learning for earthquake disaster assessment: objects, data, models, stages, challenges, and opportunities. Remote Sens (Basel) 2023;15:4098. https://doi.org/10.3390/rs15164098.
20. Akhyar A, Asyraf Zulkifley M, Lee J, Song T, Han J, Cho C, et al. Deep artificial intelligence applications for natural disaster management systems: A methodological review. Ecol Indic 2024;163:112067. https://doi.org/10.1016/j.ecolind.2024.112067.
21. Choi SW, Park K, Kim Y, Park HS. A numerical study on the strain based monitoring method for lateral structural response of buildings using FBG sensors. Journal of the Computational Structural Engineering Institute of Korea 2013;26:263–9. https://doi.org/10.7734/COSEIK.2013.26.4.263.
22. Torfs T, Sterken T, Brebels S, Santana J, van den Hoven R, Spiering V, et al. Low power wireless sensor network for building monitoring. IEEE Sens J 2013;13:909–15. https://doi.org/10.1109/JSEN.2012.2218680.
23. Kurata M, Li X, Fujita K, Yamaguchi M. Piezoelectric dynamic strain monitoring for detecting local seismic damage in steel buildings. Smart Mater Struct 2013;22:115002. https://doi.org/10.1088/0964-1726/22/11/115002.
24. Li L, Wang W, Shi P. A visual data-informed fiber beam-column model for the analysis of residual hysteretic behavior of post-earthquake damaged RC columns. Journal of Earthquake Engineering 2024;28:946–71. https://doi.org/10.1080/13632469.2023.2226222.
25. Tusun ME, Metin EM, Tigrel F, Ünsalan C. Embedded machine learning system design for post-earthquake structural health assessment. 2024 32nd Signal Processing and Communications Applications Conference (SIU), IEEE; 2024, p. 1–4. https://doi.org/10.1109/SIU61531.2024.10600859.
26. Yu X-Y, Li X, Li Y-W. Estimating strain responses of beams in steel moment-resisting frames using limited acceleration data. J Civ Struct Health Monit 2025;15:1709–32. https://doi.org/10.1007/s13349-025-00907-z.
27. Zhang P, Teramoto A, Ohkubo T, Seike T, Hibino Y, Kusunoki K. Application of various sensing techniques for post-earthquake damage assessment of exterior wall tiles. Eng Struct 2025;343:120990. https://doi.org/10.1016/j.engstruct.2025.120990.
28. Chen X, Li H, Zhao C, Zhou G, Li W, Zhang X, et al. Smartphone‐based method for measuring maximum peak tensile and compressive strain. Computer-Aided Civil and Infrastructure Engineering 2024;39:2330–49. https://doi.org/10.1111/mice.13204.
29. Mokarram V, Banan MR. Strategic sensor placement in steel frames using random forests and deep neural networks for post-earthquake assessment. Structures 2025;73:108413. https://doi.org/10.1016/J.ISTRUC.2025.108413.
30. Di Michele F, Stagnini E, Pera D, Aloisio R, Marcati P. Random Forest based estimate to assess the damages of future earthquakes: preliminary results. Annals of Geophysics 2024;66. https://doi.org/10.4401/ag-8935.
31. He J, Aguilar E, Eger M, Sotoudeh Z, Zhuang L. Sensor optimization for damage detection within a digital twin framework. Volume 3: Advanced Materials: Design, Processing, Characterization and Applications; Advances in Aerospace Technology, American Society of Mechanical Engineers; 2024. https://doi.org/10.1115/IMECE2024-144479.
32. Boutahri Y, Tilioua A. Machine learning-based predictive model for thermal comfort and energy optimization in smart buildings. Results in Engineering 2024;22:102148. https://doi.org/10.1016/j.rineng.2024.102148.
33. Alammar A, Alymani A, Jabi W. Building energy efficiency estimations with random forest for single and multi-zones. Data-Driven Intelligence 2024;2:365–74. https://doi.org/10.52842/conf.ecaade.2024.2.365.
34. Xu W, Tu J, Xu N, Liu Z. Predicting daily heating energy consumption in residential buildings through integration of random forest model and meta-heuristic algorithms. Energy 2024;301:131726. https://doi.org/10.1016/j.energy.2024.131726.
35. Genuer R, Poggi J-M. Random Forests with R. 1st ed. Cham, Switzerland: Springer Nature Switzerland; 2020. https://doi.org/10.1007/978-3-030-56485-8.
36. Glisic B. Concise historic overview of strain sensors used in the monitoring of civil structures: the first one hundred years. Sensors 2022;22:2397. https://doi.org/10.3390/s22062397.
37. Omidbeike M, Routley B, Fleming AJ. Independent estimation of temperature and strain in tee-rosette piezoresistive strain sensors. 2018 IEEE/ASME International Conference on Advanced Intelligent Mechatronics (AIM), IEEE; 2018, p. 131–5. https://doi.org/10.1109/AIM.2018.8452304.
38. Xue F, Peng Q, Ding R, Li P, Zhao X, Zheng H, et al. Ultra-sensitive, highly linear, and hysteresis-free strain sensors enabled by gradient stiffness sliding strategy. Npj Flexible Electronics 2024;8:1–8. https://doi.org/https://doi.org/10.1038/s41528-024-00301-7.
39. Huang X, Liu L, Lin YH, Feng R, Shen Y, Chang Y, et al. High-stretchability and low-hysteresis strain sensors using origami-inspired 3D mesostructures. Sci Adv 2023;9. https://doi.org/10.1126/sciadv.adh9799.
40. Shen Z, Zhang Z, Zhang N, Li J, Zhou P, Hu F, et al. High-stretchability, ultralow-hysteresis conducting polymer hydrogel strain sensors for soft machines. Advanced Materials 2022;34. https://doi.org/10.1002/adma.202203650.
41. James RW. Strain-gage misalignment errors—a special case of interest. Exp Tech 1988;12:28–9. https://doi.org/10.1111/j.1747-1567.1988.tb02182.x.
42. Montero W, Farag R, Díaz V, Ramirez M, Boada BL. Uncertainties associated with strain-measuring systems using resistance strain gauges. Journal of Strain Analysis for Engineering Design 2011;46:1–13. https://doi.org/https://doi.org/10.1243/03093247JSA661.
43. Xiao X, Li H, Cui H, Bao X. Evaluation of thermal-drift effect on strain measurement of energy pile. Constr Build Mater 2023;362:129663. https://doi.org/10.1016/j.conbuildmat.2022.129663.
44. Burr T, Hamada MS. A Multiplicative Model for Gauge R & R Studies. Qual Reliab Eng Int 2015;31:801–9. https://doi.org/10.1002/QRE.1638.
45. Iliopoulos AP, Andrianopoulos NP. An approach to analyse errors introduced in the random grid strain measurement method. Strain 2010;46:258–66. https://doi.org/10.1111/j.1475-1305.2008.00444.x.
46. Bishop CM. Training with noise is equivalent to Tikhonov regularization. Neural Comput 1995;7:108–16. https://doi.org/10.1162/neco.1995.7.1.108.
47. Shorten C, Khoshgoftaar TM. A survey on image data augmentation for deep learning. J Big Data 2019;6:60. https://doi.org/10.1186/s40537-019-0197-0.
48. Karl WC. Regularization in image restoration and reconstruction. Handbook of Image and Video Processing. 2nd ed., Elsevier; 2005, p. 183–202. https://doi.org/10.1016/B978-012119792-6/50075-9.
49. Srivastava N, Hinton G, Krizhevsky A, Salakhutdinov R. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research 2014;15:1929–58.
50. Zhao D, Yu G, Xu P, Luo M. Equivalence between dropout and data augmentation: a mathematical check. Neural Networks 2019;115:82–9. https://doi.org/10.1016/j.neunet.2019.03.013.
51. Snoek J, Larochelle H, Adams RP. Practical Bayesian Optimization of Machine Learning Algorithms. Adv Neural Inf Process Syst 2012;4:2951–9.
52. Shahriari B, Swersky K, Wang Z, Adams RP, De Freitas N. Taking the human out of the loop: A review of Bayesian optimization. Proceedings of the IEEE 2016;104:148–75. https://doi.org/10.1109/JPROC.2015.2494218.
53. Rasmussen CE, Williams CKI. Gaussian processes for machine learning. In: Dietterich T, Bishop C, Heckerman D, Jordan M, Kearn M, editors. Adaptive computation and machine learning, Cambridge, Massachusetts: the MIT Press; 2006.
54. Ganaie MA, Hu M, Malik AK, Tanveer M, Suganthan PN. Ensemble deep learning: A review. Eng Appl Artif Intell 2022;115:105151. https://doi.org/10.1016/j.engappai.2022.105151.
55. Zhang X, Liu S, Wang X, Li Y. A fragmented neural network ensemble method and its application to image classification. Sci Rep 2024;14:2291. https://doi.org/10.1038/s41598-024-52945-0.
56. Kingma DP, Ba JL. Adam: a method for stochastic optimization. 3rd International Conference on Learning Representations, ICLR 2015 - Conference Track Proceedings 2015.
57. Mokarram V, Banan MR. A computationally efficient framework for inelastic multi-objective performance-based seismic design using surrogate linear dynamic models. Iranian Journal of Science and Technology, Transactions of Civil Engineering 2026;50:333-51. https://doi.org/10.1007/s40996-025-01862-x.
58. ASCE/SEI 7-22. ASCE standard ASCE/SEI 7-22 Minimum design loads for buildings and other structures. Reston, Virginia: American Society of Civil Engineers; 2022.
59. PEER. PEER Ground Motion Database-Pacific Earthquake Engineering Research Center 2025. https://ngawest2.berkeley.edu/ (accessed May 2, 2025).
60. Pacific Earthquake Engineering Research Center. (2024). OpenSees version 3.7.0 [Software]. Berkeley, CA: PEER.
61. Mokarram V, Banan MR. A new Fibonacci-based algorithm for locating peak intensity measure of IDA curves. Iranian Journal of Science and Technology, Transactions of Civil Engineering 2019;43:15–22. https://doi.org/10.1007/s40996-018-0201-5.