DOI QR코드

DOI QR Code

Prediction of Rheological Properties of Asphalt Binders Through Transfer Learning of EfficientNet

EfficientNet의 전이학습을 통한 아스팔트 바인더의 레올로지적 특성 예측

  • Ji, Bongjun (Department of Industrial and Management Engineering, Pohang University of Science and Technology)
  • 지봉준 (포항공과대학교 산업경영공학과)
  • Received : 2021.09.10
  • Accepted : 2021.09.24
  • Published : 2021.09.30

Abstract

Asphalt, widely used for road pavement, has different required physical properties depending on the environment to which the road is exposed. Therefore, it is essential to maximize the life of asphalt roads by evaluating the physical properties of asphalt according to additives and selecting an appropriate formulation considering road traffic and climatic environment. Dynamic shear rheometer(DSR) test is mainly used to measure resistance to rutting among various physical properties of asphalt. However, the DSR test has limitations in that the results are different depending on the experimental setting and can only be measured within a specific temperature range. Therefore, in this study, to overcome the limitations of the DSR test, the rheological characteristics were predicted by learning the images collected from atomic force microscopy. Images and rheology properties were trained through EfficientNet, one of the deep learning architectures, and transfer learning was used to overcome the limitation of the deep learning model, which require many data. The trained model predicted the rheological properties of the asphalt binder with high accuracy even though different types of additives were used. In particular, it was possible to train faster than when transfer learning was not used.

도로 포장에 널리 사용되는 아스팔트는 도로가 노출되는 환경에 따라 요구되는 물리적 특성이 상이하다. 이에 따라 첨가제의 배합에 따라 아스팔트가 어떤 물리적 특성을 나타내는지 평가하고 도로의 교통, 기후 환경에 맞추어 적절한 배합을 선택하는 것이 아스팔트 도로의 수명을 확보하기 위해 필수적이다. 아스팔트의 다양한 물리적 특성 중 소성변형에 대한 저항성을 측정하기 위해서는 Dynamic shear rheometer(DSR) 테스트를 주로 사용한다. 하지만 DSR 테스트는 실험 세팅에 따라 결과가 상이하고 특정 온도 범위 내에만 측정이 가능한 단점이 있다. 따라서 본 연구에서는 DSR 테스트의 단점을 극복하고자, Atomic force microscopy로부터 수집된 이미지를 학습하여 레올로지적 특성을 예측하고자 했다. 딥러닝 아키텍처 중 하나인 EfficientNet을 통해 이미지를 학습하였고 딥러닝 모델의 한계인 많은 데이터를 요구한다는 점을 극복하기 위해 전이학습을 이용하여 학습을 진행하였다. 학습된 모델은 이종의 첨가제를 사용하였음에도 높은 정확도로 아스팔트 바인더의 레올로지적 특성을 예측하였다. 특히, 전이학습을 사용하지 않았을 때와 비교하여 빠르게 학습이 가능했다.

Keywords

References

  1. Bang, S.H., Ak, R., Narayanan, A., Lee, Y.T., Cho, H. (2019). A survey on knowledge transfer for manufacturing data analytics, Computers in Industry, 104, 116-130. https://doi.org/10.1016/j.compind.2018.07.001
  2. Divya, P.S., Krishnan, J.M. (2019). How to consistently collect rheological data for bitumen in a Dynamic Shear Rheometer?, SN Applied Sciences, 1(1), 1-5. https://doi.org/10.1007/s42452-018-0001-3
  3. Hanz, A.J., Faheem, A., Mahmoud, E., Bahia, H.U. (2010). Measuring effects of warm-mix additives: use of newly developed asphalt binder lubricity test for the dynamic shear rheometer, Transportation Research Record, 2180(1), 85-92. https://doi.org/10.3141/2180-10
  4. He, K., Zhang, X., Ren, S., Sun, J. (2016). Deep residual learning for image recognition, In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 770-778.
  5. Huang, B., Mohammad, L.N., Graves, P.S., Abadie, C. (2002). Louisiana experience with crumb rubber-modified hot-mix asphalt pavement, Transportation Research Record, 1789(1), 1-13. https://doi.org/10.3141/1789-01
  6. Ji, B.J., Lee, S.J., Mazumder, M., Lee, M.S., Kim, H.H. (2020). Deep regression prediction of rheological properties of SIS-modified asphalt binders, Materials, 13(24), 5738. https://doi.org/10.3390/ma13245738
  7. Kim, H.H., Lee, M.S., Lee, S.J. (2019b). Performance evaluation of polymer modified asphalt(PMA) binders containing ground tire rubber(GTR), International Journal of Pavement Research & Technology, 12(2). 215-222. https://doi.org/10.1007/s42947-019-0027-y
  8. Kim, H.H., Lee, S.J. (2015). Effect of crumb rubber on viscosity of rubberized asphalt binders containing wax additives, Construction and Building Materials, 95, 65-73. https://doi.org/10.1016/j.conbuildmat.2015.07.066
  9. Kim, H.H., Mazumder, M., Lee, M.S., Lee, S.J. (2019a). Evaluation of high-performance asphalt binders modified with SBS, SIS, and GTR, Advances in Civil Engineering, 2019.
  10. Kim, H.H., Mazumder, M., Lee, S.J. (2018). Recycling of aged asphalt binders with wax warm additives, Road Materials and Pavement Design, 19(5), 1203-1215. https://doi.org/10.1080/14680629.2017.1294102
  11. Mazumder, M., Siddique, A., Ahmed, R., Lee, S.J., Lee, M.S. (2020). Rheological and morphological characterization of styreneisoprene-styrene(SIS) modified asphalt binder, Advances in Civil Engineering, 2020.
  12. McDaniel, R.S. (2001). Asphalt Additives to Control Rutting and Cracking, Ph.D Thesis, Purdue University, United States.
  13. Tan, M., Le, Q. (2019). Efficientnet: rethinking model scaling for convolutional neural networks, In International Conference on Machine Learning, 6105-6114.
  14. Tayfur, S., Ozen, H., Aksoy, A. (2007). Investigation of rutting performance of asphalt mixtures containing polymer modifiers, Construction and Building Materials, 21(2), 328-337. https://doi.org/10.1016/j.conbuildmat.2005.08.014