Advancing Construction 3D Printing with Predictive Interlayer Bonding Strength: A Stacking Model Paradigm
Open Access
Abstract
To enhance the quality stability of 3D printing concrete, this study introduces a novel machine learning (ML) model based on a stacking strategy for the first time. The model aims to predict the interlayer bonding strength (IBS) of 3D printing concrete. The base models incorporate SVR, KNN, and GPR, and subsequently, these models are stacked to create a robust stacking model. Results from 10-fold cross-validation and statistical performance evaluations reveal that, compared to the base models, the stacking model exhibits superior performance in predicting the IBS of 3D printing concrete, with the R2 value increasing from 0.91 to 0.96. This underscores the efficacy of the developed stacking model in significantly improving prediction accuracy, thereby facilitating the advancement of scaled-up production in 3D printing concrete.Keywords
Funding Information
- Guangdong Basic and Applied Basic Research Foundation (2023A1515012136)
This publication has 46 references indexed in Scilit:
- 3D printing with cementitious materials: Challenges and opportunities for the construction sectorAutomation in Construction, 2022
- Meta‐learning and the new challenges of machine learningInternational Journal of Intelligent Systems, 2021
- Interlayer bonding improvement of 3D printed concrete with polymer modified mortar: Experiments and molecular dynamics studiesCement and Concrete Composites, 2020
- A critical review of 3D printing in construction: benefits, challenges, and risksArchives of Civil and Mechanical Engineering, 2020
- Design and implementation of a parallel geographically weighted k-nearest neighbor classifierComputers & Geosciences, 2019
- Hardened properties of 3D printed concrete: The influence of process parameters on interlayer adhesionCement and Concrete Research, 2019
- Effects of interlocking on interlayer adhesion and strength of structures in 3D printing of concreteAutomation in Construction, 2017
- Automatic plaque segmentation based on hybrid fuzzy clustering and k nearest neighborhood using virtual histology intravascular ultrasound imagesApplied Soft Computing, 2017
- Single imputation with multilayer perceptron and multiple imputation combining multilayer perceptron and k-nearest neighbours for monotone patternsApplied Soft Computing, 2015
- Estimation of prediction error by using K-fold cross-validationStatistics and Computing, 2009