SBS Content Detection for Modified Asphalt Using Deep Neural Network
This study proposes a prediction model for accurately detecting styrene-butadiene-styrene (SBS) content in modified asphalt using the deep neural network (DNN). Traditional methods used for evaluating the SBS content are inaccurate and complicated because they are prone to produce errors by manual c...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2020-01-01
|
| Series: | Advances in Materials Science and Engineering |
| Online Access: | http://dx.doi.org/10.1155/2020/2513147 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1841524472555765760 |
|---|---|
| author | Zhixiang Wang Jiange Li Zhengqi Zhang Youxiang Zuo |
| author_facet | Zhixiang Wang Jiange Li Zhengqi Zhang Youxiang Zuo |
| author_sort | Zhixiang Wang |
| collection | DOAJ |
| description | This study proposes a prediction model for accurately detecting styrene-butadiene-styrene (SBS) content in modified asphalt using the deep neural network (DNN). Traditional methods used for evaluating the SBS content are inaccurate and complicated because they are prone to produce errors by manual computation. Feature data of SBS content are derived from the spectra, which are obtained by the Fourier-transform infrared spectroscopy test. After designing DNN, preprocessed feature data are utilized as training and testing data and are fed into the DNN via a feature matrix. Furthermore, comparative studies are conducted to verify the accuracy of the proposed model. Results show that the mean square error value decreased by 68% for DNN with noise and dimension reduction. The DNN-based prediction model showed that the correlation coefficient between the target value and the mean predicted value is 0.9978 and 0.9992 for training and testing samples, respectively, indicating its remarkable accuracy and applicability after training. In comparison with the standard curve method and the random forest method, the precision of DNN is greater than 98% for the same test conditions, achieving the best predicting performance. |
| format | Article |
| id | doaj-art-904813a85f934d1997fee1f2a8434cc2 |
| institution | Kabale University |
| issn | 1687-8434 1687-8442 |
| language | English |
| publishDate | 2020-01-01 |
| publisher | Wiley |
| record_format | Article |
| series | Advances in Materials Science and Engineering |
| spelling | doaj-art-904813a85f934d1997fee1f2a8434cc22025-02-03T05:53:10ZengWileyAdvances in Materials Science and Engineering1687-84341687-84422020-01-01202010.1155/2020/25131472513147SBS Content Detection for Modified Asphalt Using Deep Neural NetworkZhixiang Wang0Jiange Li1Zhengqi Zhang2Youxiang Zuo3School of Highway, Chang’an University, Xi’an 710064, ChinaSchool of Highway, Chang’an University, Xi’an 710064, ChinaSchool of Highway, Chang’an University, Xi’an 710064, ChinaGuangdong Provincial Key Laboratory of Digital Signal and Image Processing and the Department of Electronic Engineering, Shantou University, Shantou 515063, ChinaThis study proposes a prediction model for accurately detecting styrene-butadiene-styrene (SBS) content in modified asphalt using the deep neural network (DNN). Traditional methods used for evaluating the SBS content are inaccurate and complicated because they are prone to produce errors by manual computation. Feature data of SBS content are derived from the spectra, which are obtained by the Fourier-transform infrared spectroscopy test. After designing DNN, preprocessed feature data are utilized as training and testing data and are fed into the DNN via a feature matrix. Furthermore, comparative studies are conducted to verify the accuracy of the proposed model. Results show that the mean square error value decreased by 68% for DNN with noise and dimension reduction. The DNN-based prediction model showed that the correlation coefficient between the target value and the mean predicted value is 0.9978 and 0.9992 for training and testing samples, respectively, indicating its remarkable accuracy and applicability after training. In comparison with the standard curve method and the random forest method, the precision of DNN is greater than 98% for the same test conditions, achieving the best predicting performance.http://dx.doi.org/10.1155/2020/2513147 |
| spellingShingle | Zhixiang Wang Jiange Li Zhengqi Zhang Youxiang Zuo SBS Content Detection for Modified Asphalt Using Deep Neural Network Advances in Materials Science and Engineering |
| title | SBS Content Detection for Modified Asphalt Using Deep Neural Network |
| title_full | SBS Content Detection for Modified Asphalt Using Deep Neural Network |
| title_fullStr | SBS Content Detection for Modified Asphalt Using Deep Neural Network |
| title_full_unstemmed | SBS Content Detection for Modified Asphalt Using Deep Neural Network |
| title_short | SBS Content Detection for Modified Asphalt Using Deep Neural Network |
| title_sort | sbs content detection for modified asphalt using deep neural network |
| url | http://dx.doi.org/10.1155/2020/2513147 |
| work_keys_str_mv | AT zhixiangwang sbscontentdetectionformodifiedasphaltusingdeepneuralnetwork AT jiangeli sbscontentdetectionformodifiedasphaltusingdeepneuralnetwork AT zhengqizhang sbscontentdetectionformodifiedasphaltusingdeepneuralnetwork AT youxiangzuo sbscontentdetectionformodifiedasphaltusingdeepneuralnetwork |