Article
Construction & Building Technology
Jinrui Zhang, Wenjun Niu, Youzhi Yang, Dongshuai Hou, Biqin Dong
Summary: This study investigates the effects of various factors on the compressive strength of calcined sludge-cement composites and establishes prediction models using machine learning. The results show that curing age has the greatest impact on compressive strength, while the influence of ball milling time and CaSO4 is small.
CONSTRUCTION AND BUILDING MATERIALS
(2022)
Article
Construction & Building Technology
Jinrui Zhang, Wenjun Niu, Youzhi Yang, Dongshuai Hou, Biqin Dong
Summary: This study investigates the effects of various factors on the compressive strength of calcined sludge-cement composites through experiments, and uses machine learning to establish six different regression prediction models to predict compressive strength. The results show that CNN and Ensemble Regression models provide excellent prediction accuracy, with curing age having the greatest impact on compressive strength.
CONSTRUCTION AND BUILDING MATERIALS
(2022)
Article
Engineering, Chemical
Rajakumaran Gayathri, Shola Usha Rani, Lenka Cepova, Murugesan Rajesh, Kanak Kalita
Summary: This paper compares nine machine learning algorithms for predicting the mechanical properties of cement-based mortars. XGBoost regression is found to be the best ML metamodel, while simpler models like linear regression are insufficient for handling non-linearity. The mapping of mortar compressive strength using ML techniques will be useful for practitioners and researchers in identifying suitable mortar mixes.
Article
Chemistry, Physical
Zhengyu Fei, Shixue Liang, Yiqing Cai, Yuanxie Shen
Summary: In this study, machine learning models, including XGBoost, RF, LightGBM, and AdaBoost, were used to predict the compressive strength of RP mortar. The XGBoost model showed the highest prediction accuracy, and SHAP was used to interpret the influencing factors. The study demonstrates that machine learning technologies can provide a rapid and cost-effective evaluation of RP material properties.
Article
Construction & Building Technology
Jiandong Huang, Mengmeng Zhou, Hongwei Yuan, Mohanad Muayad Sabri Sabri, Xiang Li
Summary: This paper studies the prediction models for the compressive strength of cement-based materials containing metakaolin and compares the prediction effects of four models. The results showed that the random forest model had the best prediction effect, considering the importance of cement grade and water-to-binder ratio in predicting the compressive strength.
Article
Computer Science, Interdisciplinary Applications
Wei Liang, Wei Yin, Yu Zhong, Qian Tao, Kunpeng Li, Zhanyuan Zhu, Zuyin Zou, Yusheng Zeng, Shucheng Yuan, Han Chen
Summary: This paper predicts the compressive strength (CS) of fly ash concrete using five integrated models and aims to develop an accurate prediction method. It also investigates the impact of parameter variations on the CS and concludes that the water-to-binder ratio has the greatest influence on the CS of fly ash concrete.
ADVANCES IN ENGINEERING SOFTWARE
(2023)
Article
Construction & Building Technology
Chendi Min, Shuai Xiong, Ying Shi, Zhixiang Liu, Xinyue Lu
Summary: The unconfined compressive strength (UCS) of cemented phosphogypsum (PG) backfill is a crucial mechanical index for ensuring stope safety, but traditional mechanical tests are costly and time-consuming. This study conducted UCStests to establish a dataset for model construction, consisting of 81 UCS results with various influencing variables. Six ensemble learning models were constructed to predict the UCS of cemented PG backfill, and the results showed that the models had smaller errors and better prediction performance than standalone machine learning models. Among these models, XGBoost performed the best with the lowest errors and highest accuracy. This study recommends using XGBoost as a reliable and efficient tool for predicting the UCS of cemented PG backfill.
CASE STUDIES IN CONSTRUCTION MATERIALS
(2023)
Article
Construction & Building Technology
Hai-Van Thi Mai, May Huu Nguyen, Son Hoang Trinh, Hai-Bang Ly
Summary: This study proposes an effective approach to determine the compressive strength of recycled brick aggregate concrete using ensemble machine learning models. The findings can assist material engineers in designing the composition of recycled brick aggregate concrete.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Engineering, Environmental
Vazida Mehtab, Shadab Alam, Sangeetha Povari, Lingaiah Nakka, Yarasi Soujanya, Sumana Chenna
Summary: This study presents a machine learning based ultrafast alternative for predicting CO2 solubility in physical solvents. A database is established and several models are trained, with kernel ridge regression (KRR) being the optimum model. Key descriptors are evaluated and optimized to maximize the prediction accuracy of the reduced order KRR (r-KRR) model. The resulting r-KRR model with nine key descriptors exhibits high prediction accuracy, with minimum error values and maximum R-2.
ENVIRONMENTAL SCIENCE & TECHNOLOGY
(2023)
Article
Engineering, Civil
Panagiotis G. Asteris, Paulo B. Lourenco, Mohsen Hajihassani, Chrissy-Elpida N. Adami, Minas E. Lemonis, Athanasia D. Skentou, Rui Marques, Hoang Nguyen, Hugo Rodrigues, Humberto Varum
Summary: Masonry, a building material with a long history, remains competitive in the construction industry. The compressive strength of masonry is crucial in modern design, but estimating it accurately is still a challenge. Soft computing techniques can help identify key parameters affecting masonry compressive strength and offer better estimates than existing formulas.
ENGINEERING STRUCTURES
(2021)
Article
Construction & Building Technology
Gaoyang Liu, Bochao Sun
Summary: The mixing ratio of raw materials significantly affects the compressive strength of concrete. An Explainable Boosting Machine (EBM) is used to predict the compressive strength and explain the contribution of mix ratio factors. Bayesian optimization is employed for algorithm selection and hyperparameter optimization. The EBM algorithm shows excellent prediction performance with R2 = 0.93, RMSE = 4.33, and MAE = 3.10, allowing for the interpretation of individual feature contributions and determination of concrete compressive strength.
CASE STUDIES IN CONSTRUCTION MATERIALS
(2023)
Article
Construction & Building Technology
Junbo Sun, Shukui Liu, Zhanguo Ma, Haimin Qian, Yufei Wang, Hisham Al-azzani, Xiangyu Wang
Summary: This study aims to promote the utilization of coal gangue in shotcrete production and enhance its reusability. It investigated the impact of coal gangue aggregate-to-sand ratio, PVA fiber-to-cement ratio, and coal gangue aggregate particle size on the properties of lightweight coal gangue shotcrete. By using machine learning models and sensitivity analysis, accurate and reliable predictions were obtained.
JOURNAL OF BUILDING ENGINEERING
(2023)
Article
Construction & Building Technology
Maedeh Hosseinzadeh, Mehdi Dehestani, Alireza Hosseinzadeh
Summary: This paper investigates the use of recycled aggregates and fly ash in concrete to reduce the demand for natural resources and lower greenhouse gas emissions. Machine learning techniques, specifically regression and classification tasks, are used to accurately predict the mechanical properties of Fly Ash Recycled Aggregate Concrete (FARAC). The findings show that the algorithms used are effective in predicting compressive and tensile strength as well as slump, with the Random Forest and XGBoost algorithms achieving the highest accuracy.
JOURNAL OF BUILDING ENGINEERING
(2023)
Article
Construction & Building Technology
Xueqing Zhang, Muhammad Zeshan Akber, Wei Zheng
Summary: This study explored nine machine learning methods and found that nonlinear models generally perform better than linear models, with the random forest model of ensemble learning performing the best. The study also confirmed the usefulness of data visualization in learning, summarizing data, understanding variable relationships, and making premodeling assumptions, as well as identified the top three most significant concrete constituents affecting the seven-day compressive strength.
CONSTRUCTION AND BUILDING MATERIALS
(2021)
Article
Computer Science, Artificial Intelligence
Yunlong Gao, Tingting Lin, Jinyan Pan, Feiping Nie, Youwei Xie
Summary: This paper proposes a new technique called Fuzzy Sparse Deviation Regularized Robust Principal Component Analysis (FSD-PCA) to improve the robustness of Principal Component Analysis (PCA) to noise samples. By introducing sparse deviation and fuzzy weighting, FSD-PCA is able to process noise and principal components separately, thus enhancing its ability to retain principal component information.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2022)
Article
Construction & Building Technology
Ze Chang, Minfei Liang, Yading Xu, Zhi Wan, Erik Schlangen, Branko Savija
Summary: In this study, an experimental setup was proposed to characterize the early-age creep of 3D printable mortar. The testing protocol included quasi-static compressive loading-unloading cycles with holding periods in between. An analytical model based on a double power law was used to predict creep compliance, and it was validated by comparison to uniaxial compression tests. The results showed good quantitative agreement, although minor differences were observed, particularly at the beginning of the test, due to the load level dependence.
CEMENT & CONCRETE COMPOSITES
(2023)
Article
Construction & Building Technology
Zhi Wan, Ze Chang, Yading Xu, Branko Savija
Summary: In this paper, the vascular structure of self-healing concrete is optimized using a deep neural network. An effective input representation method is proposed to enhance the optimization process and improve peak load and toughness. The results demonstrate that the DNN model has great potential for optimizing the design of the vascular system for self-healing concrete, leading to improved mechanical properties and repairability.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Construction & Building Technology
Yu Chen, Yu Zhang, Branko Savija, Oguzhan Copuroglu
Summary: This paper investigates the effect of LC2-to-slag ratio and gypsum content on the fresh properties, hydration, and compressive strength of quaternary blended cement pastes. The increase in LC2 proportion reduced flowability and increased water retention capacity, yield stress, and plastic viscosity. The addition of gypsum had little effect on most of the fresh properties. Adjusting LC2-to-slag ratio is a feasible way to meet requirements for fresh properties and compressive strength.
CEMENT & CONCRETE COMPOSITES
(2023)
Article
Construction & Building Technology
Qing-feng Liu, Xiao-han Shen, Branko Savija, Zhaozheng Meng, Daniel C. W. Tsang, Samad Sepasgozar, Erik Schlangen
Summary: This study investigates the behavior of multi-ions, including calcium and chloride, in concrete under wastewater and seawater conditions. Multiple mechanisms, such as calcium leaching, chloride transport, and multi-ion coupling, were analyzed. The interactive ingress of multi ions was realized through simulation methods, considering the individual mechanisms and mutual influences. The study also examined the distributions of diverse ions and the evolution mechanisms of porosity, considering the interaction with calcium ions in both pore solution and solid phase. The results suggest that calcium leaching accelerates chloride transport due to coarsened pore structure, and the electrochemical coupling effect of multi ions facilitates calcium leaching in the early stage but slightly delays it in the later stage.
CEMENT AND CONCRETE RESEARCH
(2023)
Article
Construction & Building Technology
Minfei Liang, Ze Chang, Yu Zhang, Hao Cheng, Shan He, Erik Schlangen, Branko Savija
Summary: This study aims to investigate the autogenous deformation and stress evolution in high-volume ground granulated blast furnace slag (GGBFS) concrete. The macro-scale autogenous deformation and stress evolution were studied using the Temperature Stress Testing Machine and Autogenous Deformation Testing Machine. The micro-scale origin of the autogenous deformation was explored using scanning electron microscopy, X-ray diffraction, and mercury intrusion porosimetry.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Construction & Building Technology
Yu Chen, Minfei Liang, Yu Zhang, Zhenming Li, Branko Savija, Erik Schlangen, Oguzhan Copuroglu
Summary: This study examines the effects of adding superabsorbent polymer (SAP) as an internal curing agent on the flow behavior, structural build-up, hydration kinetics, compressive strength, and autogenous shrinkage of limestone-calcined clay-cement (LC3) pastes with a fixed water to binder ratio (W/B) of 0.3. Results show that SAP increases yield stress and viscosity, as well as structural build-up and hydration, but decreases compressive strength. Additionally, SAP promotes early-age expansion and effectively mitigates autogenous shrinkage of LC3 pastes for up to 7 days.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Article
Materials Science, Multidisciplinary
Minfei Liang, Shan He, Yidong Gan, Hongzhi Zhang, Ze Chang, Erik Schlangen, Branko Savija
Summary: This paper uses computer vision techniques to predict the micromechanical properties of cement paste based on Backscattered Electron (BSE) images. A dataset of 40,000 nanoindentation tests and BSE micrographs is used to train and compare a Residual Convolutional Neural Network (Res-Net) model and a table model. The Res-Net model outperforms the table model in predicting the elastic modulus and hardness of the cement paste.
MATERIALS & DESIGN
(2023)
Article
Construction & Building Technology
Jinbao Xie, Yading Xu, Zhi Wan, Ali Ghaderiaram, Erik Schlangen, Branko Savija
Summary: A novel piezoelectric energy harvester with auxetic cementitious cellular composites (ACCCs) and surface-mounted PVDF film has been designed and tested in this study. A numerical model has been established and validated to simulate the energy harvesting performance of ACCC-PVDF system under different loading conditions.
ENERGY AND BUILDINGS
(2023)
Article
Engineering, Civil
Zhi Wan, Yu Zhang, Yading Xu, Branko Savija
Summary: This study investigates the use of dissolvable Polyvinyl Alcohol (PVA) filament to fabricate vascular networks for autonomous self-healing of cementitious composites. The vertically printed PVA tubes with wax coating show good dissolution behavior, and epoxy resin is found to be an effective healing agent for achieving mechanical and water tightness recovery. Specimens embedded with 3D vascular networks have higher healing potential than those utilizing 2D vascular networks.
ENGINEERING STRUCTURES
(2023)
Article
Engineering, Civil
Yading Xu, Branko Savija
Summary: This work proposes a 3D auxetic cementitious-polymeric composite structure (3D-ACPC) that combines 3D printed polymeric shell with cementitious mortar. Experimental results show that the 3D-ACPC has the ability to overcome the brittleness of conventional cementitious material and the low compressive strength of 3D printed polymeric cellular shell, exhibiting compressive strain-hardening behavior and high energy absorption ability. The 3D-ACPC also shows significantly enhanced specific energy absorption compared to conventional cementitious materials and polymeric cellular materials.
ENGINEERING STRUCTURES
(2023)
Article
Materials Science, Multidisciplinary
Ze Chang, Minfei Liang, Shan He, Erik Schlangen, Branko Savija
Summary: This paper proposes a new numerical method to analyze the early-age creep of 3D printed segments with the consideration of stress history. The feasibility of the lattice model in early-age creep analysis under incremental compressive loading is demonstrated, and the importance of considering creep during the printing process is emphasized.
MATERIALS & DESIGN
(2023)
Review
Polymer Science
Rowin J. M. Bol, Branko Savija
Summary: Additive manufacturing (AM), particularly the fused filament fabrication (FFF) technique, has gained popularity due to its advantages over traditional subtractive manufacturing techniques. AM involves layer-by-layer deposition, which introduces anisotropy into the produced parts. It is crucial to investigate how printing process parameters affect the mechanical properties of FDM 3D-printed parts.
Review
Construction & Building Technology
Ze Chang, Yu Chen, Erik Schlangen, Branko Savija
Summary: This paper reviews different methods for quantifying the buildability of 3D concrete printing, including experimental approaches, analytical modelling, and numerical simulations. It provides a brief introduction to the printing process and discusses material properties at different stages. The paper also reviews experimental and analytical models for buildability quantification and provides an overview of numerical tools for 3D concrete printing. It concludes with a discussion on the limitations of numerical tools for buildability quantification and recommendations for improvement.
DEVELOPMENTS IN THE BUILT ENVIRONMENT
(2023)
Article
Automation & Control Systems
Zhi Wan, Ze Chang, Yading Xu, Yitao Huang, Branko Savija
Summary: Generative networks are effective tools for digital materials inverse design, but their optimization performance is limited by the discrepancy between the optimized input and the prescribed input domain. This study introduces a correction technique into generative deep neural networks and generative deep convolutional neural networks, which significantly improves the optimization effectiveness by pulling the optimized inputs back to the prescribed domain during the optimization process.
ADVANCED INTELLIGENT SYSTEMS
(2023)