Development of Free-Form Assembly-Type Mold Production Technology Using 3D Printing Technology

Abstract

Free-form molds are used for one-time curve configuration, and because they are produced through manpower, they have issues with reduced precision and the occurrence of errors. In this study, 3D printing technologies were used to ensure precision, and polylactic acid and reusable eco-friendly materials to develop free-form assembly-type side-mold production technologies. In verifying the side mold, a free-form concrete panel was produced to check whether deformation occurs due to lateral pressure. Therefore, in this study, to verify this, a free-form concrete panel was produced and 3D-scanned to analyze the error at the side mold and the cause of the error to confirm the performance of the mold. The results showed that the error at each part was small, with a standard deviation of 1.627 mm, and there was little error at the panel joint part, around 1°. Such research is expected to be used in studies related to mold production technologies using 3D printers and on the production of free-form side molds.

1. Introduction

Free-form buildings have different structures, materials, functions, designs, etc., compared to other more general buildings. Different curved surfaces and curvatures are used for the exterior, thus giving it a unique design, and they are commonly used in public art facilities, libraries, museums, etc. Due to such characteristics, these buildings can serve as landmarks and have economic and social advantages; therefore, the demand for free-form buildings is growing [1]. But because the exteriors of free-form buildings are enormous, they are difficult to configure at once. This is the reason why they are panelized in sizes that are easy to produce for the exterior during construction, and a panel made of concrete is defined as a free-form concrete panel [2]. The production of free-form molds is required to configure the panelized molds as free form. Free-form molds have different free-form shapes depending on where they are used. Therefore, free-form molds are made in different shapes. Until now, wood or steel have been widely used because of their good formative properties. However, these materials are impossible to reuse in the case of manufacturing free-form molds with these materials. This also causes environmental pollution and increases the cost. In addition, since these molds are manufactured by humans, errors may occur [3,4]. Panels produced by free-form molds that are not precise and cannot be joined due to errors during the combining process.

Dongdaemun Design Plaza (DDP) is a free-form building located in Seoul, Korea. It is a landmark building and creates social and economic profits [5]. These buildings use a total of 45,133 aluminum panels, such as flat, curved, and double-curved types, to configure geometric shapes, such as the curves and twists of external panel surfaces. It took approximately 1 year and 11 months for the panel design, production, and cladding. But double-curved materials require highly difficult technologies for configuring curves and errors occur frequently. Because of this, the molds were produced again, thus resulting in delayed construction time and increased construction costs, and the cost rose from the initially estimated KRW 330 billion to KRW 500 billion [6,7]. Figure 1 is an example of joining defects due to the non-precision of the side curves between the free-form panels. Cracks occurred in the building due to reduced precision in the curved parts. In the event that such construction defects occur, there are risks of reduced construction quality, leaks, and production outcomes that are different from the design intent.

Table 1. Free-form mold limitation [8].

Division		Limitations of Free-Form Mold
Production aspects of free-form molds	Cost	Recycling of mold is not possible
		High dependence on manpower
		Increased construction period and construction cost
	Quality	Reproduction due to lack of shape precision
Performance aspects of free-form molds	Durability	Lack of durability due to concrete lateral pressure
	Eco-friendliness	Generates a lot of construction waste
	Shape precision	Occurrence of free-form shape errors

shows the limitations of free-form molds summarized through the above information. The following two factors summarize the issue.

(1)

Free-form molds must be custom-made for single use, and after use, this causes construction waste and environmental pollution.

(2)

There is reduced precision in the free-form panel side curves due to manual work.

Therefore, in this study, 3D printing technologies are used for the precise production of molds and polylactic acid (PLA) is used for 3D printing technologies. This is an ecofriendly plastic made using lactic acid processed from glucose extracted from corn as its raw material [9,10]. Furthermore, since it is produced using 3D printing, it is a technology with excellent performance for producing free-form molds due to its out-standing workability and precision [11]. In fact, there are studies that configured medium–large free-form construction products using a large 3D printer and verified the possibility for reducing production time and configuring the desired quality [12]. The order of this study is as shown in Figure 2. First, this study aimed at analyzing preceding studies related to the production of free-form side molds and FCPs (free-form concrete panels). FCPs produced based on analysis of preceding studies are selected. The free-form assembly-type side mold is designed using CAD based on the selected FCP shape and is printed using a 3D printer. FCP is produced to verify the performance of the produced free-form assembly-type side mold. The lower shape is configured using the lower CNC equipment according to the FCP shape to be produced, and the free-form assembly-type side mold is installed on the upper part. Concrete is placed on the installed mold to 3D-scan the completed FCP to analyze the side shape error. By doing so, this study aims at verifying the performance of free-form assembly-type mold.

2. Preceding Studies Related to Free-Form Side Molds

FCP shapes are diverse and require different curvatures and curved surfaces de-pending on the location of use. In addition, FCPs are used being fixed to a frame of a free-form building. If precision is insufficient between panels, this can lead to errors and thus result in construction defects. Studies using wood, ice, 3D printers, etc., were carried out to produce free-form molds [13,14,15]. However, there are few studies on free-form sides that are important in the mold joining process. Thus, preceding studies were analyzed to select the requirements necessary for developing free-form side mold prior to carrying out this study. Huyghe and Schoofs (2009) conducted a study for producing double-curved concrete panels. They used wood and steel wires to produce variable molds. Steel wires were placed in 3 mm grooves on the upper and lower parts of the processed wood, which were supported by rods, to configure the lower shape. The upper part of the wood was covered with silicone plates to produce the panels. However, though it rose at the part where the rod was supported, it lacked flexibility of wood resulting in limitations for configuring the curves. Accordingly, this caused defects on the sides, thus resulting in errors in the joining process of the produced panels [16].

Janssen (2011) developed the double-curved flexible mold that can configure both curved and double-curved shapes. This method places concrete on flat and even silicone mold, and raises the wood strip according to the free-form shape. But because it is fixed with a wood strip and plywood, flexures can occur. The lack of flexibility of wood during the process of configuring the curves caused high compression stress on the outer circumference of the mold surface. This resulted in buckling between the actuator, thus leading to wrinkles forming in the corners and errors in the shape. A method that does not cause buckling in the process is needed for configuring the curves of side molds [17].

Jeong (2020) developed FCP manufacturing technology based on CNC equipment. This is composed of two-sided multi-point press equipment and side mold control equipment that configure the upper and lower curves. It rises as much as the curve that requires the rod of the lower multi-point press equipment according to designs. The side mold control equipment is composed of the variable side mold and the ball bearing rod attached with a magnet that fixes and supports this [18]. Youn (2022) developed a variable side mold using steel plates and conducted experiments as shown in Figure 3. This is used for the side mold control equipment of the FCP manufacturing technology and it is joined with the magnets attached on the end of the rod and its magnetism. In this study, it was not installed with side mold control equipment to check the rigidity of the variable side mold, and FCP production experiments were carried out. By 3D-scanning the completed FCP, the error occurrence rate of the center and end was compared to verify performance. The error in the variable side mold was located within the allowable error range, and thus had sufficient rigidity. But because the variable side mold is assembled in a manner where two steel plates cross each other, errors such as impressions of steel plates on the side occurred. Therefore, variable side molds can configure various shapes and be reused, but it requires improvements to the shape errors [19].

3. Development of 3D Printing Free-Form Assembly-Type Side Molds

Based on the aforementioned preceding study, side molds used for free-form molds must be able to configure precise shapes and must enough power against side pressure that occurs during concrete placing. Furthermore, it should not have to be thrown away after one-time use, but must be possible to reuse. Therefore, this study aims at producing free-form molds with PLA using a 3D printer. PLA is an eco-friendly material and as it can be reused, it is suitable as a material for side molds. For the types of free-form panels, if the curve of the side is one-direction, it is categorized as single-curved, and if the curves have two directions, they are categorized as double-curved. This study is basic research on producing free-form molds using 3D printers, and therefore, a one-direction, single-curved panel will be produced. The specifications of the single-curved panels to be produced were set at 500 mm (horizontal) × 500 mm (vertical) × 30 mm (height). As the specifications of the lower CNC equipment used in this study were 600 mm × 600 mm, 500 mm was selected, and the height is the part that corresponds to the curvature, and so the height of 40 mm of the curvature used in the FCP production experiment previously by Yun was selected [20]. This is as shown in Figure 4.

The developed free-form assembly-type side molds were categorized as A-type and B-type depending on the joint connecting the sub-materials of the mold, as shown in Figure 5. A-type was designed using a slide method that is assembled by coming down from the top of the B-type. The corner joints of the side mold curved sub-materials are half-circle shapes and were designed for flexible joining and to make configuring curves when layering easier.

The 3D printer used in this study was MASTER-S, manufactured by NEXTOP, and its characteristics are shown on

Table 2. MASTER-S 3D printer specifications.

Hardware	Exterior	Steal, Acryl, ABS
	Bed material	Aluminum PCB
	Bed maximum temperature	110 °C
Nozzle	Size	0.4 mm
	Maximum temperature	260 °C
	Layer thickness	0.06 mm–0.3 mm
		(Recommend 0.1 mm–0.2 mm)
Software	Program used	Cura
	Supported operating system	Windows/Mac OS/LINUX
	Supported file	STL/OBJ/DAE/AMF

. It uses the FDM (fused deposition modeling) method to melt plastic filaments and solidify on the layer through a nozzle. Two types of plastic filaments, PLA and ABS (acrylonitrile butadiene styrene), are mainly used in the FDM method. In addition, comparing various physical and mechanical properties of the filaments that was used. PLA has higher strength and lower shrinkage, and is better for printing large products in the manufacturing free-form molds [21]. The bed temperature of the 3D printer was set at 55 °C and the nozzle temperature at 210 °C, and the printed mold is shown in Figure 6. Also, Raft was used with the functions of Cura when printing the mold. Raft is a function that fixes the printed object by extruding the filament layer between the printed object and the bed. Therefore, it prevented the risks of peeling and wrapping that occur during the printing process.

4. FCP Production Experiment Using Free-Form Assembly-Type Side Molds

To verify the performance of the produced free-form assembly-type side molds, free-form concrete panel production experiments were conducted. A lower CNC was used for FCP production. The lower CNC was composed of 36 rods and silicone caps were attached to each rod, and silicone plates were installed on top of it. The rod rises according to the free-form design shape and the silicone plate is pushed to configure the lower curve. The silicone cap supports the silicone plate and prevents drooping caused by the load that occurs when placing concrete for producing FCPs. Figure 7 shows the silicone plate attached to the top and the CNC equipment used for this experiment.

The lower CNC equipment used in this study manually rotated bolts to configure the curvature. As it is operated manually, there was possibility for errors. In order to minimize this, this study converted the curvature value of the free-form shape into the number of rod rotations and arranged them as shown in

Table 3. Rod rotation calculation.

Line	1	2	3	4	5	6
Distance from ground to panel (mm)	20	32.83	39.2	39.2	32.83	20
Distance/2.45° = 1 Wheel	=0.2	=13.4	=16	=16	=13.4	=8.2
Covert the number of wheels to degrees	0.2 Wheels =360 × 0.2 =72°	0.4 Wheels =360 × 0.4 =144°	-	-	0.4 Wheels =360 × 0.4 =144°	0.2 Wheels =360 × 0.2 =72°
Number of wheels and Angle (°)	8 Wheels 72°	13 Wheels 144°	16 Wheels	16 Wheels	13 Wheels 144°	8 Wheels 72°

. For the method, a distance of 100 mm was placed based on the point where the rod is installed, and a total of 36 rod rising values were arranged. The number of rod rotations and rotation angles were calculated and divided by 2.45° per rotation, and the rod was rotated according to the computed value to configure the lower shape.

The size of the FCP to be produced in this study was (500 × 500 × 30) ${\mathrm{mm}}^3$, and a PVA (polyvinyl alcohol) fiber was added to the used concrete. FCPs cannot perform compacting due to their free-form shape. Therefore, concrete placed in the mold must be filled evenly in the mold separately and not by external force. In this process, PVA fibers improve the fluidity of the concrete, thus making it possible for the PVA concrete to fill itself up inside the free-form mold. Accordingly, the mixing ratio used for this research experiment is shown in

Table 4. PVA fiber concrete mixing ratio.

Size ($\mathbf{m}{\mathbf{m}}^3$)	W/C (%)	Cement (g)	Sand (g)	Water (g)	PVA (g)
500 × 500 × 30	40	9750	9750	3900	146.25

.

The mixing ratio mentioned above was used to perform the FCP production experiment, as shown in Figure 8. The experiment raises the rod of the lower CNC equipment according to the design shape. A free-form assembly-type side mold is installed on the upper part of the configured silicone plate. After placing the concrete inside the mold, it is cured for one day and removed to produce the FCP. Finally, quality inspections were conducted through 3D scanning to analyze the shape error for this experiment.

In this study, Go! SCAN SPARK, which is a 3D scanner by Proto 3000, was used for analyzing the shape error. This is an optical scanner and its specifications are shown in

Table 5. Go! SCAN SPARK specifications.

Hardware
Size	89 × 114 × 346 ${\mathrm{mm}}^3$
Weight	1.25 kg
Scan range	390 × 390 ${\mathrm{mm}}^2$
Software
Program used	VXelements
Support file	STL/TXT/WRL/X3D/X3DZ

.

The scan analysis program, VXelements, was used to overlap the scanned files acquired using the 3D scanner and the design shape CAD file, and comparative analysis was conducted to find the error value. After removing bubbles and holes formed in the scanned panel shape, evening work was performed for accurate analysis. After this process, the error range was analyzed to calculate the error value of the upper part and sides of the panel. The error values of the panel are as shown in

Table 6. Free-form concrete panel shape error.

Error Position	Least (mm)	Max (mm)	±(mm)	SD (mm)
Side error	−4.626	4.326	8.952	1.627
Total error	−4.563	3.940	8.503	1.302

. The overall error value was 5.503 mm and side error 8.952 mm, thus having an average 8.7 mm strain rate compared with the initial design shape.

Figure 9 overlaps the scanned shape and free-form design shape, and measures the angles at the parts where the free-form assembly-type side molds are joined. If the rigidity of the free-form assembly-type side mold is sufficient, resistance to side pressure that occurs when placing concrete is possible, so it must be measured as 0°. But errors occurred at each part and it was concluded that the rigidity of the joints that fix the joining areas of the mold was insufficient.

Figure 10 verified the precision of the joint parts of the free-form assembly-type side mold. For this, the corner angles of the produced panels were measured based on the design shape with panel corner angles of 90°; (a) and (d) were measured to have an average of 91°, and (b) and (c) were measured at an average of 89°. Overall, this is an error of approximately 1°. As the free-form assembly-type side mold is installed without a separate fixing method with the silicone plate on top of the lower CNC, it is vulnerable against side pressure that occurs on the curved parts when placing concrete. Due to such side pressure, the mold is pushed and this causes a 1° error.

5. Discussion

Free-form molds are made in different shapes and used according to the required curved shape, but they can only be used one time, thus having limitations in that it causes construction wastes and increases the total construction costs. In order to resolve this, this study used 3D printers and PLA to develop free-form assembly-type side molds.

Free-form assembly-type side molds use PLA, which is an eco-friendly material, according to the required curves, and printed the mold precisely via 3D printing for assembly. In order to verify the produced mold, one-direction free-form concrete panels were produced and scanned. The scanned FCP shape and designed shape were overlapped to analyze the size and cause of the errors that occurred on the sides to verify the performance of free-form assembly-type side molds, and the results are as follows.

The max value of the side error value of the panels was 4.326 mm and minimal value 4.626, thus being approximately 8.925 mm, and the SD was measured at 1.627 mm.

For the shape error analysis of free-form concrete panels, the free-form shape allowable error standard of 3.000 mm was used. In the case of the free-form concrete production sector, because there are no regulations on the allowable error, 3% (3.000 mm), which is the average wall thickness allowable error in Korea, was used. Based on this, the error of the free-form panel produced in this study was 1.627 mm, thus satisfying the allowable error. In order to verify whether errors occurred in the joining parts, the error value of the central joining parts and peak parts of the mold were checked [20].

Based on 0°, the angle of the free-form mold’s central joining part error values of 0.196°, 0.134°, 0.146°, and 0.942° were measured. Errors occurred at each part and the cause of this was that the mold printed with PLA had an internal density of 20%, thus lacking strength to resist the side pressure of concrete. The density of the mold can be adjusted, so it is possible to secure the necessary strength. However, the higher the internal density, the longer it takes for the production time, which can cause waste of materials; therefore, it is necessary to select an appropriate internal density of free-form molds.

Based on the 90° angle of the panel peak, it was measured as 91.046°, 88.966°, 88.933°, and 91.032°. The overall error was about 1°, and it was concluded that this was caused by pushing due to side pressure that occurred during concrete placing because it was installed without a fixing method of the lower CNC and free-form assembly side mold. It was deduced that this can be supplemented by fixing the mold using free-form lower equipment and side equipment.

6. Conclusions

In this study, PLA was used with 3D printing technologies to develop free-form assembly-type side molds. Existing free-form molds were used for a single time and produced using manpower, thus having limitations in reduced precision. These limitations are shown in

Table 7. Solutions to the limitations of existing molds using free-form assembly-type molds.

Division		Limitations of Free-Form Molds	Solution by Application of This Technology
Production aspects of free-form molds	Cost	Recycling of mold is not possible	Molds can be recycled by using PLA filaments
		High dependence on manpower	Applying automation technologies such as 3D printing technology
		Increased construction period and construction cost	Securing productivity of FCPs by using PLA molds
	Quality	Reproduction due to a lack of shape precision	Realization of free-form shapes using a 3D printer
Performance aspects of free-form molds	Durability	Lack of durability due to concrete lateral pressure	Resistance to lateral pressure through the combined structure of free-form assembly-type molds
	Eco-friendliness	Generates a lot of construction waste	Reducing construction waste by recycling PLA filaments
	Shape precision	Occurrence of free-form shape errors	Verification with a 3D scanner for the completed free-form panel

and are expected to be resolved by applying the technology developed in this study. The free-form assembly-type mold printed using a 3D printer is a new mold that solves the limitations of the existing free-form molds. In this study, PLA materials were used with no mixture for the molds. After being used, the mold can be crushed and reproduced as filaments to print the mold again. Also, as 3D printing technologies make it easy to configure shapes, designed shapes can be printed precisely as free-form molds.

However, production takes a long time, and while steel wires are used to connect the parts while the mold is being assembled, it is necessary to develop a new method that has higher strength and can ensure complete fixing. Furthermore, it was produced as a one-direction panel in this study, and more research related to double-curved free-form panel production is needed. It is anticipated that this study will be used in research related to mold production technologies using 3D printers and on the production of free-form side molds within the construction sector.