1. Introduction
Openings are essential in the design of concrete structures for practical applications, such as windows and door access to buildings. In 3D concrete printing (3DCP) technology, fresh concrete is deposited through a nozzle at designated locations to build a concrete structure in a layered manner. Similar to regular cast concrete structures, openings are also essential in the printed concrete structure for the post-processing of windows and doors. However, the printing of openings in concrete structures using this technology is limited due to the presence of overhanging sections [
1]. Overhangs are the sections above the opening without any support or surface to rest upon. Without support, the fresh concrete filaments will collapse due to gravitational force.
The printing of overhang structures in various types of 3D printing or additive manufacturing technologies has already been well established with the assistance of temporary supports [
2]. The overhang sections in the 3D printed parts do not only include the bridging between two vertical structures but also the area underneath an inclined wall as well as the anchoring of a floating section (
Figure 1).
In general, the maximum inclination angle from the horizontal axis at which the part can be printed without support should be at least 45 degrees (
Figure 2). If the inclination angle is more than 45 degrees, the support structure can be omitted. However, the requirement for support can vary with the different types of material and 3D printing technologies. Different solutions to overcome overhang printing in various 3D printing techniques, as well as 3DCP, are discussed in this section.
In the current state-of-the-art 3D printing technologies, various types of support structures are already widely used in different forms. In some extrusion-based polymer printers, the support material can be the same as the primary material but is printed at different printing parameters, which allows weaker bonds between the printed part and the support structure to facilitate ease of removal after the printing process is completed [
3]. Such supports are typically removed through mechanical means by breaking away the material or with the aid of machining tools [
4]. However, these means also impose a risk of damaging the printed part.
Figure 1.
Support for anchoring of floating object [
4].
Figure 1.
Support for anchoring of floating object [
4].
Figure 2.
Support structure requirement for inclined angle [
4].
Figure 2.
Support structure requirement for inclined angle [
4].
On the other hand, soluble materials for support structures have also been developed for 3D printing [
4]. Such soluble material can be easily removed with the usage of a cleaning solvent. These soluble supports also allow more complex geometries to be printed, as the supports that are inaccessible through mechanical removal can be easily removed. However, due to the multi-material configuration, a multi-nozzle setup is required for such operation, which increases the complexity of the printer setup, print path design, as well as overall printing time of the part. Furthermore, the soluble supports are non-reusable due to the materials being dissolved in the solvent.
In the powder bed 3D printing techniques, such as selective laser sintering (SLS) or selective laser melting (SLM), the excess powder that is not bonded to form the part acts as a natural support and can be removed after the printing process is completed [
4]. The removed excess powder can also be reused for the printing of another part, reducing the wastage of the raw materials.
In 3DCP technology, the printing of overhangs has been a major limitation for the further development of this application. The extrusion-based concrete printing technique prints a concrete structure by depositing fresh concrete in a layered manner. Due to the fluid state of the concrete, the technology is still incapable of printing overhangs as the filament will collapse without a physical surface to rest upon.
Previous research has proposed methods to overcome the printing of overhangs in extrusion-based 3DCP. The simplest and most common method is to design corbelling in the structure to gradually create overhangs (
Figure 3a) [
5]. This method omits the need for support by using an alternative structural design for the 3D printing application. Though this method is simple and easily adopted by print path designing, it is still limited by the slope angle, which is highly material-dependent. Ref. [
6] studied the different material formulations of concrete for 3D printing applications and demonstrated that the maximum slope angle for corbelling is 17.5 degrees from the vertical axis. Any angle greater than this will result in the failure of the specimens.
Another method of printing overhangs in 3DCP is the use of temporary partial formworks to support the material in the overhanging sections. For instance, it was reported that the house printed by Arup used scaffolding and wooden planks to temporarily support the overhang sections during the printing process, which were removed after the concrete had cured (
Figure 3b) [
7]. A similar method was also used by [
3], where a temporary external support for the overhang sections was placed during the printing process (
Figure 3c). However, the placement of the support is disruptive to the printing process which increases the time gap between the concrete layers. This increase in the time gap causes weaker bonding of the layers and results in a vulnerable point in the structure due to poor interlayer strength [
8].
The printing of supports using the same material as the primary material in 3DCP was first proposed by [
3]. This method is accomplished by increasing the nozzle speed at the areas where the support structure is located while maintaining the same flow rate. This creates an underflow of material in that area, inducing discontinuities in the filaments which act as support for the overhang structures that are to be printed above it (
Figure 4a). The discontinuities allow the support to be removed, although additional mechanical post-processing may also be required to achieve a better finish (
Figure 4b). The mechanical removal process may also pose a risk of damaging the printed part.
D-shape is a powder bed fusion additive manufacturing process that was developed for architectural applications [
9]. Unlike extrusion-based 3DCP, the granular materials and fibers are first laid out before an “ink binder” is selectively deposited on the powder bed to form the structure. This process is repeated for each layer until the printing is completed. Similar to the SLS and SLM processes, the excess powder that is not bound also acts as the support for the overhangs and is removed after the printing process is completed. With the excess powder support, complex architectural designs, as seen in
Figure 5, that are unachievable by extrusion-based concrete printing, become achievable in this process.
The printing of overhang structures has been extensively studied in various types of 3D printing technologies [
10,
11]. Nonetheless, the printing of overhangs in extrusion-based 3DCP can only be accomplished with the use of external supports, such as wooden planks. Though the use of corbelling has been proven to be feasible, the solution is limited by the slope angle and the structure design. The advantage of 3DCP technologies is the ability to fabricate free-form structures. The method, however, is restricted by the limitation of creating a free-form opening in the printed structure. Although a solution was proposed to print the support with the primary material and reduce the filament quality at the support area, by varying the travel speed of the nozzle to allow ease of removal, it has been shown that a mechanical removal process still may pose a risk of damaging the printed part.
In addition to the different techniques of overcoming the printing of overhangs, 3DCP has to tackle another problem, which is the pumpability and buildability of the material. The fresh concrete is deposited through a nozzle tracing the cross section of the design in a layered manner. The material used must possess low viscosity to achieve high pumpability, and high yield strength to achieve buildability to support subsequent layers. This requirement commands a rapid yield strength evolution of the concrete mix, which has encouraged buildability studies in 3DCP. [
12] utilized a modified green compression test model to predict the buildability of a concrete mix. The model was able to predict failure heights which can enhance control in optimizing the material design and printing process. [
13] demonstrated the use of physical measurement of shape and stability to determine the buildability criterion. The study gave an insight into the definition of printability, as well as a suite of tests to determine whether a material is printable. These approaches, collectively, contribute to a more comprehensive understanding of buildability and facilitate improved optimization techniques for concrete printing.
Much existing literature typically focuses on implementing printability optimizations through two main methods. They either introduce modifications to the concrete in small batches or apply set-on-demand interventions near the print head [
14,
15]. Ref. [
14] show that set-on-demand methods are more effective for improving buildability than mixture modifications using admixtures during the initial mixing. Ref. [
16] used an active rheology control using vibration to achieve the contradicting performance for 3DCP. The vibration was observed to improve the extrudability of the material. Regardless of the method employed, the residence time of concrete within the print head is a critical parameter that must be precisely monitored to ensure optimal mixing efficiency.
In view of this, the objective of this study is to develop a removable support that both fulfills the printability and sustainability aspects of the material characteristics. This study aims to develop a printable support material that temporarily supports the overhang sections until the primary material gains self-supporting capabilities with sustainable raw materials, such as fly ash and recycled glass powder. The support material should also allow ease of removal after the primary material has hardened.
3. Results and Discussion
3.1. Printability
The printability of the material, including pumpability and buildability parameters, was characterized. Pumpability is crucial for delivering material in a 3D printing system to ensure a continuous filament product. It measures the ease of material delivery through the system. Therefore, the volumetric flow rate of the material at a constant pump input speed was used to evaluate pumpability. As shown in
Figure 10, an increase in recycled glass content leads to a decrease in the material’s volumetric flow rate. This reduction in pumpability is attributed to the angular shape and sharp edges of the recycled glass powder particles. The irregular shapes of the recycled glass powder hinder particle movement by interlocking, which obstructs the flow of the suspension material. In contrast, the spherical shapes of fly ash powder particles provide better flow properties for the suspension material. The spherical shape of the fly ash particles reduces friction among the angular recycled glass particles, creating what is termed the “ball bearing effect” [
18].
A significant decrease in volumetric flow rate was observed when the recycled glass content exceeded 50%. The dominance of recycled glass powder in the suspension significantly impairs material pumpability. Mixtures with recycled glass powder content beyond 60%, such as FA35RG65 and FA30RG70, were found to be non-pumpable due to clogging in the pump caused by excessive recycled glass powder. This clogging results from the pump’s insufficient torque to break the interlocking of the recycled glass particles. Similar to the work presented by [
19], the incorporation of waste glass material results in a decrease in slump value as the amount of sand replacement increases. The lubricating effect of fly ash particles diminishes as fly ash content decreases and recycled glass content increases.
The 3D printing process involves a layering operation to construct the final structure. Buildability is crucial for vertical stacking, as the material in the bottom layer must resist deformation from the weight of the layers above. Buildability was evaluated based on the material’s ability to withstand the weight of accumulating layers during the printing process. Cylindrical structures with a diameter of 200 mm were printed using each mixture until the filaments exhibited high levels of deformation. A laser distance sensor was attached to the nozzle to monitor and measure the change in distance between the top surface of the filament and the nozzle opening. Deformation of the filaments increased the distance between the laser sensor and the filament surface, leading to an increase in the overall height of the structure. The printing test continued until the structure deformed beyond the laser sensor’s range (approximately 35 mm).
In contrast to pumpability, buildability improves with an increase in recycled glass powder content. Similar to existing literature, an increase in the glass-to-binder ratio can lead to a reduction in filament deformation up to the extrudable limits [
20].
Figure 11 shows the laser distance data from the printing tests. A spike in distance is observed when the nozzle moves up a layer, which can be used to identify layer increments in the raw data. As shown in
Figure 11, the FA40RG60 mixture was able to build a larger number of layers with lower levels of deformation compared to the FA60RG40 mixture. The FA60RG40 mixture could not maintain its shape and deformed rapidly as layers were added, leading to structural distortion. Increasing the recycled glass powder content helps prevent deformation due to the angular shape of the glass particles, thus improving buildability. The FA55RG45 mixture could sustain its shape upon extrusion but deformed excessively after reaching a height of 80 mm. The FA50RG50, FA45RG55, and FA40RG60 mixtures demonstrated greater buldability, with printed heights of 110 mm, 140 mm, and 170 mm, respectively, before distortion due to excessive deformation.
The pumpability and buildability of the mixtures are contrasting requirements that need to be satisfied to achieve a printable material. Though the increase in fly ash powder content allows the suspension material to become more pumpable, the buildability is decreased as the mixture deforms more easily. At 60% of fly ash content, the mixture was not able to perform sufficiently as a printable material, as it does not allow the sustaining of its shape. On the other hand, the increase in recycled glass powder content increases the buildability of the mixtures but decreases the pumpability of the mixtures. Mixtures with recycled glass powder content beyond 65% were found to be non-pumpable since the material causes clogging in the delivery system. Although the further increase of the recycled glass powder results in a potential increase in the buildability, the mixtures will not have sufficient flowability to flow through the material delivery in the 3D printing system. In a study by [
21], the mixture with higher flowability and lower stiffness was found to have a lower buildability. Other literature suggests that the buildability of the printed cement-based material is dependent on the ability of the material to gain early-age strength, with time to resist deformation from the stress of the stacking layers [
22,
23]. However, the absence of cement and reactive components in this printable mixture results in the lack of the material’s ability to gain strength beyond the initial yield stress. The non-curing nature allows the material to be removed from the structure after the primary material gains self-support abilities to form the overhang section.
3.2. Uniaxial Compression Test
The compression stress of the subsequent layers on the bottom layers causes the material to deform, resulting in the distortion of the structure shape in the 3D printing process. The compression stress of the vertical stacking layers increases along with the building up of layers in the construction process. The uniaxial compression test simulates this compression process on the bottom layer material to evaluate the buildability of the material. Moreover, for the temporary support material, the bottom layers have the requirement of sustaining gravitational stress caused by the overhanging material intended in the structure.
The rate of compressive loading of 0.27 N/s was applied to the specimens in this study. The selection of compressive load rate is assumed for the cylindrical printing process in
Section 3.1. The rate of compression loading due to the vertical stacking of layers is dependent on the rate of increase in the height of the structure [
24]. The height incremental in the printing process is further derived from various printing parameters defined in the printing process, such as the contour distance scalar per layer, thickness of the layer, and the nozzle travel velocity.
Figure 12 shows the deformation behavior of the support material under the uniaxial compression loading. The 10% and 20% deformation from the initial vertical height of the specimens were observed and the corresponding compression loads were recorded. The compression load that is required to deform the material specimens was observed to increase with the increase in recycled glass powder content but peaked at 65% of recycled glass content. The increase in compressive load with the recycled glass content is a result of the angular shape of the recycled glass particles resisting the deformation of the material. As a result, the compression load that is required to cause deformation in the material increases, and a higher buildability can be expected in the material that has the ability to resist a higher compression load.
The FA30RG70 specimen showed a lower compression load, as compared to the FA35RG65 specimen, attributed to the different failure modes of the specimens. The FA30RG70 was observed to fail by the shear plane (
Figure 13b), which was likely to cause a lower compression failure load of the specimen. Ref. [
25] also observed similar shear cracking in the later stage of the material, when the material gains early age strength and higher stiffness. Due to the high stiffness of the material of the FA30RG70 specimen, similar to the behavior of a solid, the material was not able to deform and failed in the form of the shear plane. The failure by the shear plane was described as related to the brittleness of the material by [
26]. On the other hand, the other specimens fail via the barrelling mode due to the low stiffness and ductility of the material. As depicted with white dotted line on
Figure 13a, the diagram shows the FA35RG65 specimen without any shear cracks, as compared to the FA30RG70 specimen in
Figure 13b. The other specimens with lower stiffness deform upon compression loads, which led to the barreling effect (
Figure 13a), as described in previous literature [
25,
26].
3.3. Flow Table Test
The flowability of the material is essential for the smooth delivery through the material transportation system in the 3D printer. Low flowability material can lead to discontinuous extrusion and decrease the strength of the filament to support the subsequent layers and the weight of the primary concrete material in the overhang section. Thus, a material is required to have sufficient flowability to be pumpable in the 3D printing system.
The flow table test shows the flowability measurement of the mixture via the spread diameter of the material after the table is dropped for 25 cycles. The agitated material is spread across the flow table and the spread diameter is used to evaluate the flowability of the material, as shown in
Figure 14. The larger spread diameter indicates a higher flowability of the material and the lower spread diameter indicates a lower flowability of the material. The measured spread diameter of the mixtures is shown in
Figure 15. Generally, the spread diameter decreases with the increase in recycled glass content in the mixture. The flowability of the mixtures was found to decrease at a higher rate beyond the 50% recycled glass content. This observation was also seen in the pumpability study in
Section 3.1.
The pumpable region determines the allowable flow properties of the material for the pumping action to transport material during the 3D printing process. Section describes the pumping test to determine the ease of material transportation in the 3D printing delivery system. It was observed that the mixtures beyond 60% of recycled glass powder content were non-pumpable as the material feeding resulted in clogging. The pumpable region was determined in terms of the spread diameter with regards to the results in
Section 3.2. The spread diameter of the FA35RG65 and FA30RG70 mixtures was shown to be significantly lower than the other mixtures, with almost no change in diameter from the initial mold.
The flow table test can be correlated with the pumpability of the mixture in 3DCP, as it measures the behavior of the fresh mixture under agitation by the pump. The dropping of the table in this test simulates the agitation of the material by the pumping action in the 3D printing process. The flowability of 3DCP mixtures has been characterized using the flow table method in previous literature [
21,
27,
28,
29]. Ref. [
28] used the flow table test as part of a study to explore the extrudability of different 3DCP mixtures. In this study, the mixtures between 10% and 25% increase in spread diameter achieve good extrudability and buildability. Another study by [
21] also used a flow table to determine the flowability of the mixtures and subsequently determined the printable range of mixtures consisting of copper tailings. These studies show that the flow table test complements the printing test by evaluating the flowability of the mixtures, particularly their ability to be transported through the 3DCP material delivery system.
3.4. Optimization of Mix Ratio
In order to use the response optimizer, the responses were fitted with a quadratic equation. RRG is the variable given to the different mix design ratios, which is the percentage of the recycled glass within the mixture. When RRG is equal to 40, it means that the mixture is 60% fly ash and 40% recycled glass.
The coefficient of determination (R
2) is the statistical measure of fit that represents the proportion of the variance in the dependent variable that is predictable from the independent variable in the regression model. In simpler terms, it’s a value that determines how closely the model fits to actual results. From the results obtained, the R
2 of the values for compression load and the flowability were 0.7 and 0.98, respectively, which means it has a high likelihood that future events will fall within the predicted outcome. The quadratic equations for its compression load and flowability are shown in Equations (2) and (3), respectively. The equation was also plotted in
Figure 16.
From
Figure 16a, the response of the compressive strength shows that as the ratio of the RG increases, the strength increases. On the other hand, as the ratio of the RG decreases, the flowability reduces (
Figure 16b). Since the main aim is to maximize both responses, and the compressive strength and flowability have contradicting responses with reference to the content of RG, it is expected that the optimized value will be in the middle range. With the use of these quadratic equations, the software was able to determine the best mix design, which gives the highest compressive strength and the highest flowability. The optimized R
RG value is 60, which indicates that the optimized mixture contains 60% of recycled glass content. With this value, a validation test was carried out by printing with the structure shown in
Figure 16a.
3.5. Reusability of the Material
The support material can be recovered from the post-processing of the structure and reused for the subsequent printing project due to its non-curing nature. Because of the non-reactive raw materials in the mixture, the support material does not cause any chemical reaction. Thus, the material can be 100% recovered and reused for another project for the same purpose. While sustainability is a major concern in the raw materials for 3DCP mixtures, a sustainable approach was used in the formulation of the support materials. The raw materials employed for the support were sourced from waste materials to achieve a more sustainable mixture for supporting overhang printing with 3DCP technology. Fly ash is derived from pulverized coal burning in power generation plants as a by-product, where the fine ash powder from the coal combustion is transported through the flue gas and collected in electrostatic precipitators [
30]. Recycled glass powder is processed from glass waste that would otherwise end up in landfills. Glass wastes are processed through crushing to achieve finer particle size for the convenience of recycling. The fine particles are generated from the random particle size distribution through the crushing of the glass wastes and separated through sieving from the larger glass particles.
3.6. Free-Form Overhang Structure Printing
The optimized mix design from
Section 3.4 was successfully applied to print a free-form overhang section, demonstrating the effectiveness of the temporary support material. Shown in
Figure 17, a slot opening was printed using the support material, which was then removed after the primary concrete hardened. This process was facilitated by a dual nozzle system, enabling multi-material printing.
Despite these advancements, there are limitations to the temporary support material technology in 3DCP. The support material must be carefully formulated to ensure compatibility with the primary concrete mixture, as variations in concrete composition or environmental conditions can impact its performance. Additionally, although the material is designed for easy removal, extracting it without damaging the primary structure may still present challenges, especially in intricate or densely printed sections.
Furthermore, the overhang section can cause the support material to slump, leading to a hanging bridge effect in the printed layers. Due to limitations in material buildability, only small openings can be printed using this support material solution. Despite these constraints, the technique remains useful for creating openings for pipes, fittings, and reinforcements in structures. Additionally, the support material provides temporary stability during the printing of inclined walls, offering greater design freedom in 3DCP.
4. Conclusions
This study developed a temporary support material for the 3DCP process, designed to support overhang sections during printing. Once the primary concrete hardens enough to be self-supporting, the temporary support can be easily removed. The material, made with non-reactive components like fly ash, recycled glass powder, and natural sand, remains in a paste form, ensuring easy removal even after the concrete has fully hardened. Below are the key conclusions drawn from this study:
Increased recycled glass content leads to higher dynamic and static yield stress.
Higher recycled glass content improves buildability but reduces pumpability.
Mixtures with up to 60% recycled glass were pumpable, but those with 65% and 70% content were not.
Decreased flowability is caused by a higher amount of recycled glass due to particle interlocking.
Increased fly ash content improves flowability due to the “ball bearing effect” of spherical particles.
Higher recycled glass content enhances buildability by supporting more layers.
The flow table test showed reduced flowability with higher recycled glass content.
Uniaxial compression tests indicated increased compressive strength up to 65% recycled glass content, but a decrease at 70% due to shear cracking.
Balanced optimization was achieved between flowability and buildability, resulting from a mixture with 60% recycled glass content.
The optimized support material was successfully validated through the printing of a structure with a free-form opening and overhang section. While this study primarily focused on the immediate behavior of the support material under incremental stresses, it is acknowledged that incorporating time-dependent measurements of both creep and compressive strength would offer a more comprehensive understanding of material performance. Future research should address these aspects to evaluate the long-term stability and deformation tendencies of the material under sustained loads, thereby enhancing the reliability and applicability of 3D concrete printing technologies. Nevertheless, this innovation enables 3DCP to create complex free-form designs without the need for costly, labor-intensive external supports. Also, the material provides crucial stability during the printing of inclined walls, offering a higher degree of design freedom and enhancing the overall potential of 3D concrete printing in advanced architectural applications.