Design optimization of cement grouting material based on adaptive boosting algorithm and simplicial homology global optimization. (15th May 2022)

• Record Type: Journal Article
• Authors: Ren, Jiaolong; Zhao, Hongbo; Zhang, Lin; Zhao, Zedong; Xu, Yinshan; Cheng, Yuanyuan; Wang, Meng; Chen, Jingchun; Wang, Jian
• Is Part Of: Journal of building engineering. Volume 49(2022)
• Publication Date: 2022-05-15
• Subjects: AdaBoost-based surrogate model -- Design optimization -- Simplicial homology global optimization -- Cement grouting material -- Machine learning; Building -- Periodicals; 690.05
• DOI: 10.1016/j.jobe.2022.104049
• Languages: English
• ISSNs: 2352-7102

Abstract

Abstract: Cement grouting materials have been widely used in maintenance and reinforcement engineering. Various design methods have been developed to determine the optimal compositions of the cement grouting material. However, it is a challenging task to determine the compositions of the materials due to the complexity of cement grouting. In this study, a novel intelligent optimal method by combing the simplicial homology global optimization method (SHGO), Adaptive Boosting algorithm (AdaBoost), and laboratory experiments were developed to determine the optimal composition of the cement grouting material. The AdaBoost presented the complex interrelations of different material compositions and their properties. The SHGO was adopted to search for a reasonable and optimal composition to improve the performance. Compared to the compositions determined by the traditional method, the optimal composition determined by the developed method provides improved strength, especially for the early strength. The flexural/compressive strengths of the grouting material determined by the proposed method are 3.90/12.12 MPa, 9.71/28.86 MPa, 12.11/38.12 MPa at the curing age of 1-day, 3-day, and 7-day, respectively, which are 1.27/1.12 times, 1.17/1.09 times, and 1.11/1.08 times higher than the grouting material determined the traditional method. The developed intelligent optimal framework provides a helpful, valuable, and promising tool for determining the optimal composition of cement-based materials. Highlights: Developed a novel compositions optimal design approach. Adopted AdaBoost to build a surrogate model. Determined the optimal compositions with an excellent performance. Provided a promising way for determining the optimal compositions.