This Research Paper, “Conceptualising the behaviour of 3D-printed concrete structures in fire”, was presented at the Young Concrete Researchers, Engineers and Technologists Symposium, 2023, organised by Cement & Concrete SA (CCSA). Researchers Selicia Pillay and Richard S Walls, both from the Department of Civil Engineering, Stellenbosch University, and Johann E van der Merwe from the Department of Civil Engineering, University of Pretoria, investigated the thermal performance and thermally induced stressors of three-dimensional printed concrete (3DPC) walls when exposed to elevated temperatures. 

 Abstract 

 In recent years, 3DPC has been developed into practical construction applications, including structures such as facade panels, bridges, houses and office buildings, but existing knowledge on the structural behaviour of 3DPC in fire is limited. 

 Using free-body diagrams of typical 3DPC walls to evaluate numerical results, it was shown that significant internal stresses can develop, which are likely to cause extensive cracking with vertical cracking (perpendicular to printed concrete layers) being dominant rather than horizontal cracking, as may be expected.  

 Adoption of new technologies 

 Faced with skill scarcity, an expanding number of projects and the need to limit its environmental impact, the construction industry is progressing to automation to improve productivity, cost efficiency and reduce material waste.  

 Thermal and mechanical properties 

 The Eurocode for the design of concrete in fire, EN 1992-1-2, provides the thermal properties at elevated temperatures for traditional concrete materials. Research showed an almost equivalent relationship between the proposed thermal properties of 3DPC and EN-1992-1-2 properties at elevated temperatures, which was used in this study to develop the finite element model.  

 Development of the finite element model  

 A 3D finite element numerical model was developed in Abaqus to evaluate the structural response of 3DPC wall panels with various cross-sectional patterns. A heat-transfer time step with a time of 3 600 seconds (60 minutes) was created for the evaluation period, with an ISO 834 standard time temperature used to define the exposed face gas temperature. This duration was considered adequate to illustrate the objectives of the research and is sufficient for most buildings (such as offices, residential and institutional buildings), except high-rise buildings.  

Wall panel temperatures  

 A comparison of the exposed surface temperatures of each cross-section is shown in Figure 3, with respect to time after 60 minutes of fire exposure. As expected, a maximum temperature of 919°C is observed in cross-section C (cellular), i.e., close to the standard fire curve temperature of 945°C. This cellular configuration results in the lowest resultant effective conductivity. The minimum exposed face temperature was observed in the cast solid section, with a temperature of 895°C at 60 minutes. 

Figure 4 depicts the unexposed surface panel of the various wall cross-sections. All the cross-sections investigated pass the fire-resistance insulation criterion of the unexposed face, with the highest temperature (43°C = 23°C above ambient) still being significantly below the requirement of about 140°C above ambient. Hence, from a thermal perspective a two-hour fire rating could potentially be attained by these samples. However, the very high thermal gradients encountered is of great concern, with changes in the temperature of almost 900°C across sections, which will lead to extensive internal stresses. 

Stress-strain relationship of 3DPC cavities  

 Figure 5 illustrates the fundamental structural mechanics that occur in a typical cavity cross-section, which drives cracking. From this diagram we see the first onset of cracking occurring in the internal middle layer, as depicted in Figure 5(e). Based on finite element models, the time to the onset of cracking ranged from three minutes for cross-section A and D to 15 minutes for cross-section B. This was influenced by the geometric distance from the exposed panel, resulting in the different onset of cracking times. 

It is of significant concern that simple models and a free-body diagram, based on fundamental structural mechanics, indicate that vertical cracking is likely to occur very quickly in sections. Also, it appears that cracking will start within the sample, rather than at the sample face, due to the thermal gradients and section geometries.

Conclusion  

 This research highlights a fundamental issue with the use of 3DPC and the thermal behaviour associated with it. The simplified illustrations of the internal stress and strain profiles are important for understanding why internal cracking may occur, and the predicted stress profiles can be applied to various cross-sections in the future work to identify the onset of cracking. This is fundamentally important for post-fire damage assessment and consideration in the design process of the structure.  

Read the full paper online.   

Three-dimensional printed concrete (3DPC) is becoming a growing alternative to traditional construction practices. With 3DPC walls typically having low tensile strength (i.e., no reinforcement), its thermal and associated mechanical behaviours should be investigated to consider whether 3DPC structures are as robust in fire as might be expected. 

Full acknowledgement and thanks go to:

www.cemcon.sa.org.za for the information in this editorial.