Search

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.

Magnesium phosphate cement (MPC), as a new type of cementitious material, is difficult to be popularized in practical applications due to its short setting time, high cost, and poor water resistance. Dredged silt (DS) is a kind of hazardous waste, which may cause serious damage to the ecological environment if it is not disposed of properly. The treatment of DS and its reuse in building materials is an economical and environmentally friendly treatment method, which not only realizes the recycling of DS but also avoids secondary pollution. Using the treated DS as a mineral admixture for modified MPC not only recycles the DS but also improves the properties of MPC. In this paper, the effect of DS modification on the properties of MPC was investigated by setting time test, mechanical property test, water resistance test, and RCM test, and also compared with the modification effect of common mineral admixtures, such as fly ash (FA) and metakaolin (Mk), and finally analyzed by combining with SEM pictures. With the increase of the dosage of FA, Mk, and DS, the compressive strength of the modified MPC all showed the trend of increasing and then decreasing, and the optimal dosage was 5%, 10%, and 15%, respectively. At the optimum dosage, Mk and FA could improve the resistance of MPC to chloride erosion, but DS could not. The three kinds of admixtures could prolong the setting time of MPC, with DS having the best effect, followed by FA, and Mk having the smallest effect. All three admixtures could improve the water resistance of MPC, and the effect of improvement was Mk > DS > FA. The microstructure of the specimen was denser when the dosage of DS was low, and when the dosage was too much, the hydration reaction of MPC was affected, resulting in incomplete hydration and reduced hydration products. This research has significant guiding significance for the treatment of DS and the modification of MPC.

The dredged silt exposed to the air is highly expected to be used to replace cement in large quantities due to its richness in aluminosilicates, which is in line with the dual-carbon background. Herein, a systematic exploration on the basic properties and hydration products of sintered silt ash-blended (SSA-blended) cement is conducted, using experiments and thermodynamic simulations. The results show that in the process of adding SSA to 50%, the compressive property and chloride ion permeation resistance are enhanced first and then weakened, whereas the flexural property is weakened gradually. Among them, with the addition of 10% SSA, the compressive strength can reach 65.2 MPa and the electric flux is 37.3% lower than that of Portland cement (PC), and the chloride ion permeability is rated as the lowest. As SSA increases, the FTIR spectra show that the reduction rates of bound water, ettringite and Ca(OH)2 are accelerated gradually; SEM observation shows that the cement-based gelling skeleton has undergone a transition from dense to loose; the thermodynamic simulation results show that C-S-H increases first and then decreases, implying that SSA has a certain pozzolanic activity. In addition, the evolution laws of other hydration products are highly consistent with the analysis results of FTIR and SEM.

3D printing brings new opportunities for the development of civil engineering. Machine learning (ML) is also widely used in various fields of 3D printing. To describe the main research issues and potential future applications of ML for construction 3D printing, it is essential to comprehensively overview the current research topics. This paper aims to review, summarize, analyze, and introduce the research progress of ML applications in construction 3D printing, followed by a discussion of the current challenges and future research scope. Firstly, the classification and working principle of 3D printing and ML technology were introduced. Secondly, in the aspects of the design of construction materials, control of printing process, and quality inspection of construction components, the application status of ML in construction 3D printing was discussed. Based on this, in terms of interlayer bond performance enhancement, real-time status monitoring, and anisotropic behavior control, the future potential and challenges of ML in construction 3D printing were summarized, to provide a reference for promoting the realization of high-efficiency, intelligence, and sustainability in the field of civil engineering.

The effective utilization rate of river-dredged silt was extremely low, and common disposal methods such as dumping it into the ocean have already threatened the ecological environment. To demonstrate that dredged silt can be used as a mineral admixture to modify magnesium potassium phosphate cement (MKPC), the mechanical properties and hydration degree of sintered silt ash (SSA)-blended MKPC in the early stage of hydration were studied systematically in this paper, with MKPC as the reference group. The mechanical experiment results showed that in the process of increasing the SSA content to 25%, the compressive strength first increased and then decreased. Among the samples, the compressive strength of cement aged by 1d and 3d with 15% content was the highest, which increased by 11.5% and 17.2%, respectively, compared with the reference group. The setting time experiment found that with the increase in SSA content, the hydration reaction rate of MKPC slowed down significantly. Its effect of delaying hydration was most obvious when the SSA content was 10–15%. The X-ray diffraction pattern showed that there was no large amount of new crystalline substances formed in the hydration product. The results obtained by scanning electron microscopy show that the microstructure tended to be denser and the hydration products tended to be plump when the SSA content was in the range of 0–15%. The non-contact electrical resistivity experiment showed that the addition of SSA delayed the early hydration of MKPC. Combined with the above experiment results, it was found that when the content of SSA was less than 15%, it not only delayed the early hydration of MKPC, but also deepened its hydration degree.