Muhammad Saeed Zafar , Farid Javadnejad , Maryam Hojati
{"title":"通过集成机器学习优化3D打印胶凝材料的流变性能","authors":"Muhammad Saeed Zafar , Farid Javadnejad , Maryam Hojati","doi":"10.1016/j.addma.2025.104889","DOIUrl":null,"url":null,"abstract":"<div><div>The complex interaction between rheology-modifying admixtures and fresh cementitious mix printability limits 3D printing applications in construction. To optimize the properties of 3D printable concrete, this study presents a machine learning (ML)-based, knowledge-guided framework that integrates data-driven modeling with expert validation. A structured workflow uses a small dataset to predict and refine optimal mix designs. A total of 77 lab samples were prepared with varying amounts of nano-clay (NC), silica fume (SF), bentonite volclay (BC), and methylcellulose (MC). Their rheological properties, including plastic viscosity (VIS), dynamic yield stress (DYS), and static yield stress (SYS), were measured using a rheometer. Ensemble ML models were developed through automated preprocessing, cross-validated hyperparameter tuning, and RMSE-based selection. The top five models per rheological responses were combined using a voting regressor, improving predictive accuracy while mitigating overfitting. Predictions were visualized using contour maps from gridded synthetic data, revealing nonlinear interactions among input features. A key innovation is applying expert ratings to contour maps to guide the selection of high-performing mixes. This step allows domain knowledge to define acceptable printability ranges and helps address ML uncertainty from limited training data. Optimized mixes were selected based on rating maps and re-evaluated through additional rheology and 3D printing tests. The results demonstrated that the mixes met satisfactory extrudability and buildability requirements, confirming the validity of the defined expert rating criteria and the practical utility of the framework in optimizing 3D printable concrete mixes containing the defined additives. The proposed approach ensures both predictive robustness and practical applicability. It enables iterative refinement of models as new data becomes available and offers a systematic approach to navigating complex mix interactions. Overall, combining ensemble modeling, contour visualization, and knowledge-driven evaluation provides a powerful tool for advancing 3D concrete printing mix design.</div></div>","PeriodicalId":7172,"journal":{"name":"Additive manufacturing","volume":"109 ","pages":"Article 104889"},"PeriodicalIF":11.1000,"publicationDate":"2025-07-05","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Optimizing rheological properties of 3D printed cementitious materials via ensemble machine learning\",\"authors\":\"Muhammad Saeed Zafar , Farid Javadnejad , Maryam Hojati\",\"doi\":\"10.1016/j.addma.2025.104889\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<div><div>The complex interaction between rheology-modifying admixtures and fresh cementitious mix printability limits 3D printing applications in construction. To optimize the properties of 3D printable concrete, this study presents a machine learning (ML)-based, knowledge-guided framework that integrates data-driven modeling with expert validation. A structured workflow uses a small dataset to predict and refine optimal mix designs. A total of 77 lab samples were prepared with varying amounts of nano-clay (NC), silica fume (SF), bentonite volclay (BC), and methylcellulose (MC). Their rheological properties, including plastic viscosity (VIS), dynamic yield stress (DYS), and static yield stress (SYS), were measured using a rheometer. Ensemble ML models were developed through automated preprocessing, cross-validated hyperparameter tuning, and RMSE-based selection. The top five models per rheological responses were combined using a voting regressor, improving predictive accuracy while mitigating overfitting. Predictions were visualized using contour maps from gridded synthetic data, revealing nonlinear interactions among input features. A key innovation is applying expert ratings to contour maps to guide the selection of high-performing mixes. This step allows domain knowledge to define acceptable printability ranges and helps address ML uncertainty from limited training data. Optimized mixes were selected based on rating maps and re-evaluated through additional rheology and 3D printing tests. The results demonstrated that the mixes met satisfactory extrudability and buildability requirements, confirming the validity of the defined expert rating criteria and the practical utility of the framework in optimizing 3D printable concrete mixes containing the defined additives. The proposed approach ensures both predictive robustness and practical applicability. It enables iterative refinement of models as new data becomes available and offers a systematic approach to navigating complex mix interactions. Overall, combining ensemble modeling, contour visualization, and knowledge-driven evaluation provides a powerful tool for advancing 3D concrete printing mix design.</div></div>\",\"PeriodicalId\":7172,\"journal\":{\"name\":\"Additive manufacturing\",\"volume\":\"109 \",\"pages\":\"Article 104889\"},\"PeriodicalIF\":11.1000,\"publicationDate\":\"2025-07-05\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"Additive manufacturing\",\"FirstCategoryId\":\"5\",\"ListUrlMain\":\"https://www.sciencedirect.com/science/article/pii/S2214860425002532\",\"RegionNum\":1,\"RegionCategory\":\"工程技术\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, MANUFACTURING\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"Additive manufacturing","FirstCategoryId":"5","ListUrlMain":"https://www.sciencedirect.com/science/article/pii/S2214860425002532","RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, MANUFACTURING","Score":null,"Total":0}
Optimizing rheological properties of 3D printed cementitious materials via ensemble machine learning
The complex interaction between rheology-modifying admixtures and fresh cementitious mix printability limits 3D printing applications in construction. To optimize the properties of 3D printable concrete, this study presents a machine learning (ML)-based, knowledge-guided framework that integrates data-driven modeling with expert validation. A structured workflow uses a small dataset to predict and refine optimal mix designs. A total of 77 lab samples were prepared with varying amounts of nano-clay (NC), silica fume (SF), bentonite volclay (BC), and methylcellulose (MC). Their rheological properties, including plastic viscosity (VIS), dynamic yield stress (DYS), and static yield stress (SYS), were measured using a rheometer. Ensemble ML models were developed through automated preprocessing, cross-validated hyperparameter tuning, and RMSE-based selection. The top five models per rheological responses were combined using a voting regressor, improving predictive accuracy while mitigating overfitting. Predictions were visualized using contour maps from gridded synthetic data, revealing nonlinear interactions among input features. A key innovation is applying expert ratings to contour maps to guide the selection of high-performing mixes. This step allows domain knowledge to define acceptable printability ranges and helps address ML uncertainty from limited training data. Optimized mixes were selected based on rating maps and re-evaluated through additional rheology and 3D printing tests. The results demonstrated that the mixes met satisfactory extrudability and buildability requirements, confirming the validity of the defined expert rating criteria and the practical utility of the framework in optimizing 3D printable concrete mixes containing the defined additives. The proposed approach ensures both predictive robustness and practical applicability. It enables iterative refinement of models as new data becomes available and offers a systematic approach to navigating complex mix interactions. Overall, combining ensemble modeling, contour visualization, and knowledge-driven evaluation provides a powerful tool for advancing 3D concrete printing mix design.
期刊介绍:
Additive Manufacturing stands as a peer-reviewed journal dedicated to delivering high-quality research papers and reviews in the field of additive manufacturing, serving both academia and industry leaders. The journal's objective is to recognize the innovative essence of additive manufacturing and its diverse applications, providing a comprehensive overview of current developments and future prospects.
The transformative potential of additive manufacturing technologies in product design and manufacturing is poised to disrupt traditional approaches. In response to this paradigm shift, a distinctive and comprehensive publication outlet was essential. Additive Manufacturing fulfills this need, offering a platform for engineers, materials scientists, and practitioners across academia and various industries to document and share innovations in these evolving technologies.