Energy Consumption vs. Tensile Strength of Poly[methyl methacrylate] in Material Extrusion 3D Printing: The Impact of Six Control Settings

The energy efficiency of material extrusion additive manufacturing has a significant impact on the economics and environmental footprint of the process. Control parameters that ensure 3D-printed functional products of premium quality and mechanical strength are an established market-driven requirement. To accomplish multiple objectives is challenging, especially for multi-purpose industrial polymers, such as the Poly[methyl methacrylate]. The current paper explores the contribution of six generic control factors (infill density, raster deposition angle, nozzle temperature, print speed, layer thickness, and bed temperature) to the energy performance of Poly[methyl methacrylate] over its mechanical performance. A five-level L25 Taguchi orthogonal array was composed, with five replicas, involving 135 experiments. The 3D printing time and the electrical consumption were documented with the stopwatch approach. The tensile strength, modulus, and toughness were experimentally obtained. The raster deposition angle and the printing speed were the first and second most influential control parameters on tensile strength. Layer thickness and printing speed were the corresponding ones for the energy consumption. Quadratic regression model equations for each response metric over the six control parameters were compiled and validated. Thus, the best compromise between energy efficiency and mechanical strength is achievable, and a tool creates significant value for engineering applications.

Poly[methyl methacrylate] (PMMA) is a thermoplastic widely used in energy storage [33][34][35][36][37] and medical applications, focused on cranioplasty [38][39][40][41][42] and dentistry [43,44]. Its antimicrobial properties have also been investigated [44][45][46]. In these areas composites using PMMA as the matrix material have also been developed, aiming to improve the • Quantify the energy demands when manufacturing parts with the PMMA polymer, with the MEX 3D printing process. • Identify the effect of six generic, machine-independent 3D printing settings on the energy consumption and the mechanical properties of PMMA parts in MEX 3D printing. • Analyze and optimize the energy consumption and the mechanical properties of these six 3D printing settings studied, aiming to provide a road map for future use.

•
Contribute to the Increase of the sustainability of the MEX 3D printing process and the increase of the performance of the parts build with the process.

Materials
PMMA was procured in pellet form (brand name JULIER, Fujian, China, density 1.15-1.19 g/cm 3 , type suitable for extrusion, Figure 1a) and it was dried in a laboratory oven at 40 • C for 4 h before any further use (Figure 1b). It was then extruded into a

Materials
PMMA was procured in pellet form (brand name JULIER, Fujian, China, density 1.15-1.19 g/cm 3 , type suitable for extrusion, Figure 1a) and it was dried in a laboratory oven at 40 °C for 4 h before any further use (Figure 1b). It was then extruded into a 1.75 mm filament, compatible with the MEX 3D-printing process (3devo precision, Utrecht, The Netherlands) (Figure 1c).

Filament Examination and Thermal Properties
Before the MEX 3D-printing process, the produced filament underwent quality control for consistency in its diameter and surface quality (Figure 1e,f). Its thermal properties were investigated with thermogravimetric analysis (TGA) (Perkin Elmer Diamond TG/TDA, Waltham, MA, USA, 40-550 °C, 10 °C/min, nitrogen purge gas, Figure 1i) and differential scanning calorimetry (DSC) (Perkin Elmer Diamond DSC, Waltham, MA, USA, 50-300 °C, 10 °C/min, air atmosphere), to verify that the extrusion temperatures (filament and MEX 3D printing) did not compromise the thermal stability of the PMMA polymer.

Specimens Fabrication
Prior to the MEX 3D-printing process, filament was further dried (Figure 1d). Tensile test (dogbone) specimens were manufactured with the MEX 3D-printing process with a different set of 3D-printing parameters for each run. Five specimens were manufactured and tested for each run (135 in total). Specimens conformed with the corresponding standard (American Society for Testing and Materials-ASTM D638-02a, type V), and their

Filament Examination and Thermal Properties
Before the MEX 3D-printing process, the produced filament underwent quality control for consistency in its diameter and surface quality (Figure 1e,f). Its thermal properties were investigated with thermogravimetric analysis (TGA) (Perkin Elmer Diamond TG/TDA, Waltham, MA, USA, 40-550 • C, 10 • C/min, nitrogen purge gas, Figure 1i) and differential scanning calorimetry (DSC) (Perkin Elmer Diamond DSC, Waltham, MA, USA, 50-300 • C, 10 • C/min, air atmosphere), to verify that the extrusion temperatures (filament and MEX 3D printing) did not compromise the thermal stability of the PMMA polymer.

Specimens Fabrication
Prior to the MEX 3D-printing process, filament was further dried (Figure 1d). Tensile test (dogbone) specimens were manufactured with the MEX 3D-printing process with a different set of 3D-printing parameters for each run. Five specimens were manufactured and tested for each run (135 in total). Specimens conformed with the corresponding standard (American Society for Testing and Materials-ASTM D638-02a, type V), and their geometries are shown in Figure 2. An Intamsys, Funmat HT (Intamsys, Shanghai, China) MEX 3D printer was used for the fabrication of the specimens (Figure 1g). The different 3Dprinting parameters studied herein are schematically shown in Figure 2, along with their geometries are shown in Figure 2. An Intamsys, Funmat HT (Intamsys, Shanghai, China) MEX 3D printer was used for the fabrication of the specimens (Figure 1g). The different 3D-printing parameters studied herein are schematically shown in Figure 2, along with their values, which were selected according to the material's and the 3D printer's specifications, and the literature [72].

Energy Indicators
The energy consumption can be distinguished into three main phases: (i) machine startup, (ii) 3D-printing process, and (iii) machine shutdown, and can be calculated utilizing the equations below [32]: where: is the consumed energy by the motors of the 3D printer and of the electronics and the remaining parts of the 3D printer. The specific printing energy (SPE) index was calculated: Figure 2. 3D-printing process: graphical representation of the 3D-printing settings considered in the work, their values (right side), and the dimensions and the specimens manufactured with the process, following the corresponding standard (ASTM D638-02a).

Energy Indicators
The energy consumption can be distinguished into three main phases: (i) machine startup, (ii) 3D-printing process, and (iii) machine shutdown, and can be calculated utilizing the equations below [32]: where: E motion is the consumed energy by the motors of the 3D printer and of the electronics and the remaining parts of the 3D printer. The specific printing energy (SPE) index was calculated: The specific printing power (SPP) index was calculated: where energy printing consumption (EPC) represents the energy used by the 3D printer (E total ), w the actual weight of each specimen, and PT the actual printing time for each experimental run.

DOE and Regression Analysis
The Taguchi design of experiment was followed in the work [88] and an L25 orthogonal array was formed. Six 3D printing settings were studied for their effect on the energy consumption and the tensile properties of the MEX 3D printed parts, made with the PMMA polymer. The control parameters were selected according to the literature review, while their levels, as mentioned above, were selected according to the material's and the 3D printer's specifications and the literature [72]. The control parameters were, Raster Deposition Angle (RDA, deg), Layer Thickness (LT, mm), Infill Density (ID, %), Nozzle Temperature (NT, • C), Bed Temperature (BT, • C), and Printing Speed (PS, mm/min). Each control parameter had five levels and five replicas were tested per run.
The effect of these control parameters on eight response metrics related to the energy consumption during the 3D printing process (MEX 3D printing time-s, manufactured specimens' weight-g, EPC-MJ, SPE-MJ/g, and SPP-KW/g), and the tensile properties of the PMMA MEX 3D printed specimens (tensile strength-MPa, the tensile modulus of elasticity-MPa and tensile toughness-MJ/m 3 ), was investigated. The experimental results were statistically analyzed and then a regression analysis followed for the formation of predictive models for each one of the response indicators, as functions of the six control parameters studied herein. Two additional confirmation runs with five replicas each were performed for the evaluation of the prediction models produced.

Thermal Properties and Morphological Analysis
One out of the five samples of each run studied herein is presented in Figure 3. The top surface of each specimen is shown and images were acquired using an optical stereoscope. The differences in the 3D-printing structure between the specimens, due to the different 3D-printing parameters used in each run, are evident in the images. In all samples, a defect-free 3D-printing structure is shown, verifying that the 3D-printing parameters used and studied in the work are appropriate for the PMMA material and the 3D printer used. Figure 4 shows the corresponding images from the fracture surface of the samples (again images were taken with an optical stereoscope). The fracture surfaces significantly differ between the runs, showing the high effect of the 3D printing parameters on the behavior of the 3D printed parts. In some cases (for example run no 3, 17, and others), the fracture surface indicates a brittle failure with minimum deformation shown. In other samples (for example run no 5 and 10) a more ductile fracture behavior is observed, with deformation shown in the 3D printing structure. There are samples with mixed fracture surfaces, such as the samples of run 9 and 12. In several cases, the strands have failed in a brittle manner, but they have also been detached from each other (failure of the fusion between them) and parts of the sample have been abscinded (for example runs 1, 7, 11, and 21). The expected voids and pores in the 3D printing structure [23] are present in the samples, and filament tearing [89] is not visible in the images. In the specimens in which the fusion between the strands seems to have collapsed or partly collapsed (for example in the sample of run 12) the size of the voids is increased and can be attributed to the failure of the specimen during the tensile test. Such abnormalities in the 3D structure affect the mechanical properties of the parts [90].
The thermal properties of the PMMA polymer, as they were determined during the measurements taken, are shown in Figure 5 (TGA graph Figure 5a, DSC graph Figure 5b).
From the derived curves, it is safe to assume that the temperatures applied to the PMMA material during the filament extrusion and the MEX 3D-printing process did not affect its thermal stability or cause any degradation in the material that would affect the produced parts. Higher magnification SEM images from the side surface of the samples (Figure 5c) show a defect-free 3D-printing structure and verified the images taken with the optical stereoscope. SEM images from the fracture surface (Figure 5d), which are of higher magnification than the optical stereoscope, similarly present the fracture surface. The strands failed in a brittle way with no visible deformation. Strands that are piled one above the other maintained their fusion after the failure of the specimen in the tensile test, while the adjacent piles of strands detached, forming internal voids in the specimens. In the specific specimen shown in Figure 5d (run 13), the bottom part is more solid with minimum pores and no clearly visible 3D-printing structure up to about 1/3 of its height. Then, the structure described above is visible.  Figure 4 shows the corresponding images from the fracture surface of the samples (again images were taken with an optical stereoscope). The fracture surfaces significantly differ between the runs, showing the high effect of the 3D printing parameters on the behavior of the 3D printed parts. In some cases (for example run no 3, 17, and others), the fracture surface indicates a brittle failure with minimum deformation shown. In other samples (for example run no 5 and 10) a more ductile fracture behavior is observed, with  The thermal properties of the PMMA polymer, as they were determined during the measurements taken, are shown in Figure 5 (TGA graph Figure 5a, DSC graph Figure 5b). From the derived curves, it is safe to assume that the temperatures applied to the PMMA material during the filament extrusion and the MEX 3D-printing process did not affect its thermal stability or cause any degradation in the material that would affect the produced parts. Higher magnification SEM images from the side surface of the samples (Figure 5c) show a defect-free 3D-printing structure and verified the images taken with the optical stereoscope. SEM images from the fracture surface (Figure 5d), which are of higher magnification than the optical stereoscope, similarly present the fracture surface. The strands failed in a brittle way with no visible deformation. Strands that are piled one above the other maintained their fusion after the failure of the specimen in the tensile test, while the adjacent piles of strands detached, forming internal voids in the specimens. In the specific specimen shown in Figure 5d (run 13), the bottom part is more solid with minimum pores and no clearly visible 3D-printing structure up to about 1/3 of its height. Then, the structure described above is visible.

Taguchi Design and Experimental Results
The L25 orthogonal array compiled in the work with the six control parameters with five levels each is presented in Table 1. For each one of the twenty-five experimental runs conducted, the average value and the deviation for each response indicator studied herein is presented in Tables 2 and 3. Weight, tensile strength (sB), tensile modulus of elasticity (E), and tensile toughness are depicted in Table 2, while Table 3 presents the corresponding values for the printing time, EPC, SPE, and SPP. Figure 6 presents, in a bar chart, the average value and the deviation for the printing time, the part weight, and the EPC for the twenty-five runs. Such a presentation shows in a more vivid way the differences between the cases studied for these indicators, contributing to the comprehensiveness of the variations between the experimental results and the importance of the 3D-printing settings in the performance of the fabricated parts. The experimental results for each specimen are analytically presented in the supplementary material of the work.

Taguchi Design and Experimental Results
The L25 orthogonal array compiled in the work with the six control parameters with five levels each is presented in Table 1. For each one of the twenty-five experimental runs conducted, the average value and the deviation for each response indicator studied herein is presented in Tables 2 and 3. Weight, tensile strength (sB), tensile modulus of elasticity (E), and tensile toughness are depicted in Table 2, while Table 3 presents the corresponding values for the printing time, EPC, SPE, and SPP. Figure 6 presents, in a bar chart, the average value and the deviation for the printing time, the part weight, and the EPC for the twenty-five runs. Such a presentation shows in a more vivid way the differences between the cases studied for these indicators, contributing to the comprehensiveness of the variations between the experimental results and the importance of the 3D-printing settings in the performance of the fabricated parts. The experimental results for each specimen are analytically presented in the supplementary material of the work.     Figure 7 shows boxplot graphs for four of the response metrics studied, i.e., printing time (s), part weight (g), tensile strength (MPa), and EPC (MJ). For each response metric, the boxplots were formed against the two higher ranked control parameters for the specific metric. Scattered experimental-result values for a specific response metric for a given control parameter value indicate that this value has a strong influence on the specific response metric:   Figure 7 shows boxplot graphs for four of the response metrics studied, i.e., printing time (s), part weight (g), tensile strength (MPa), and EPC (MJ). For each response metric, the boxplots were formed against the two higher ranked control parameters for the specific metric. Scattered experimental-result values for a specific response metric for a given control parameter value indicate that this value has a strong influence on the specific response metric: The information provided in the boxplots indicates that further analysis is required to have a clearer view of the effect of the 3D-printing settings on the mechanical properties and the energy-consumption metrics.   (4) and (5), these metrics affect energy efficiency. Printing time (s) is highly affected by the PS (mm/s) and the LT (mm) control parameter only. Higher values in these control parameters lead to reduced printing-time (s) values. LT (mm) is the dominant parameter and PS (mm/s) is ranked no 2. NT (°C) is the least important parameter for the printing time (s). For the part weight (g), PS (mm/s) is the dominant parameter followed by NT (°C) (rank no 2), while BT (°C) is the least significant parameter. Higher PS (mm/s) values decrease the part-weight (g) metric up to the 50 mm/s value. Then, at the 60 mm/s value (the highest studied value for this control parameter), part weight (g) is slightly increased. The increase in NT (°C) increases the part weight (g) metric. None of the other control parameters have a linear response to the part-weight (g) metric, with values initially decreasing and then increasing, as the values increase, or the other way around, showing a significant effect on the specific metric. Figure 9 presents the MEP for the tensile-strength (MPa) and the EPC (MJ) indicators. This is a crucial plot, since it depicts the set of 3D-printing settings producing the optimum results for strength and energy consumption in the parts built with PMMA, with the MEX 3D-printing process. RDA (deg) was found to be the dominant parameter for the tensile The information provided in the boxplots indicates that further analysis is required to have a clearer view of the effect of the 3D-printing settings on the mechanical properties and the energy-consumption metrics. Figure 8 shows main effect plots (MEP) for the printing-time (s) and the part-weight (g) metrics. According to Equations (4) and (5), these metrics affect energy efficiency. Printing time (s) is highly affected by the PS (mm/s) and the LT (mm) control parameter only. Higher values in these control parameters lead to reduced printing-time (s) values. LT (mm) is the dominant parameter and PS (mm/s) is ranked no 2. NT ( • C) is the least important parameter for the printing time (s). For the part weight (g), PS (mm/s) is the dominant parameter followed by NT ( • C) (rank no 2), while BT ( • C) is the least significant parameter. Higher PS (mm/s) values decrease the part-weight (g) metric up to the 50 mm/s value. Then, at the 60 mm/s value (the highest studied value for this control parameter), part weight (g) is slightly increased. The increase in NT ( • C) increases the part weight (g) metric. None of the other control parameters have a linear response to the part-weight (g) metric, with values initially decreasing and then increasing, as the values increase, or the other way around, showing a significant effect on the specific metric. parameters in each response indicator. For this purpose, interaction plots were created and are presented in Figure 10 and the supplementary material of the work. Figure 10 shows the interaction plots for the tensile strength (MPa) and the EPC (MJ) metric. In all cases, antagonistic relations between the control parameters are observed. Only in the case of EPC (MJ), is a synergistic relation shown between the PS (mm/s) and the LT (mm) control parameters.   MEP provides information about the optimum response values for the metrics evaluated; still, they do not provide any information about the relations between the control parameters in each response indicator. For this purpose, interaction plots were created and are presented in Figure 10 and the supplementary material of the work. Figure 10 shows the interaction plots for the tensile strength (MPa) and the EPC (MJ) metric. In all cases, antagonistic relations between the control parameters are observed. Only in the case of EPC (MJ), is a synergistic relation shown between the PS (mm/s) and the LT (mm) control parameters.

Regression Analysis
The reduced quadratic regression model (RQRM) for each response was calculated: where k represents the quality output (e.g., weight, printing time, tensile strength, tensile modulus of elasticity, tensile toughness, EPC, SPE, SPP); a is the constant value; b is the coefficient of the linear terms; c is the coefficient of the quadratic terms; e is the error; and x i the seven (n = 6) control parameters, i.e., the infill density, raster deposition angle, nozzle temperature, printing speed, layer thickness and bed temperature. For each response metric, a regression table was formed along with the corresponding prediction model as a function of the control parameters. The regression tables for the printing time (s), the part weight (g), the tensile strength (MPa), and EPC (MJ) are presented in Tables 4-7, respectively. After each table, the corresponding prediction model is presented (Equations (7)-(10)). For each metric studied, the corresponding prediction model was established using regression analysis. This is an equation for the calculation of the specific metric as a function of the six control parameters studied herein. Such equations, when proven to be reliable, have industrial merit, as they can be applied to real application environments and estimate the expected value for the metric under investigation by giving as input the required values for the control parameters. For the remaining response metrics, the corresponding regression tables, and prediction models, are available in the supplementary material of the work. For all response metrics, the F value was significantly higher than four (4) and the p value was almost zero (0). Regression values were higher than 85%, reaching values of about 95% in some cases, showing that the prediction models are sufficient for the response indicators in all cases evaluated herein. Figures 11 and 12 show Pareto charts created for the response metrics to identify the statistically important parameters for each metric. Next to each Pareto chart, a graph of the predicted vs. the actual metric values is shown. For each such graph, two metrics were calculated, the mean absolute percentage error (MAPE) (the prediction accuracy, <5% is acceptable) [91] and the Durbin-Watson factor (<2 means positive autocorrelation, equal to 2 means no autocorrelation, and >2 means negative autocorrelation) [92]. Pareto charts and predicted vs actual values graphs for additional response metrics can be found in the supplementary material of the work.    For the printing time (s), all parameters are statistically important (cross the 1.98 line), except RDA and RDA 2 (Figure 11a). For the part weight (g), all parameters are statistically important (cross the 1.98 line) (Figure 11b). For both metrics, the MAPE was found to be higher than five (5), indicating reduced prediction accuracy; still, this is a quite good accuracy, and the result can be considered acceptable (less than 10, [93]). The corresponding Durbin-Watson factor for both metrics shows a positive autocorrelation of the residuals. For the tensile strength (MPa), all parameters are statistically important (cross the 1.98 line), except RDA 2 , LT, and LT 2 (Figure 12a). For the EPC (MJ) metric, statistically important parameters (crossing the 1.98 line) were found to be PS, PS 2 , LT, and LT 2 . For both metrics, the MAPE was found to be higher than five (5), indicating reduced prediction accuracy. The corresponding Durbin-Watson factor for both metrics shows the positive autocorrelation of the residuals. Figure 13 presents printing time (s), part weight (g), tensile strength (MPa), and EPC (MJ) surface graphs against two of the higher ranked control parameters for each metric.

Confirmation Run
The control parameter values for the two confirmation runs performed (numbered run 26 and 27, respectively) are presented in Table 8. The average value and the deviation for each response metric studied herein in each one of the two confirmation runs are presented in Tables 9 and 10. The analytical experimental results for each one of the five replicas in the confirmation runs can be found in the corresponding tables in the supplementary material of the work. The accuracy of the prediction models was verified by comparing the experimental values of the confirmation runs with the values calculated. The deviation (error) between those values for the tensile strength (MPa) and the EPC (MJ) metric for both confirmation runs is shown in Table 11. From the calculated absolute error, it is evident that the accuracy of the prediction models is more than sufficient.

Discussion
As shown in the literature review, 3D-printing settings significantly affect the performance of the parts built with the process. Herein, an attempt was made to produce parts with minimized energy requirements and optimized mechanical properties, with a polymer widely used in the specific type of applications, PMMA polymer, with the MEX 3D-printing process. The effect of six critical and machine-independent 3D-printing parameters was experimentally investigated and analyzed with modeling tools. Additionally, for the first time, insight is provided into the energy consumption when 3D printing PMMA parts with the MEX process. Such information affects the eco-friendliness of the process and increases its sustainability. It was found that energy and mechanical properties are not particularly controversial in this case, since the 3D-printing parameters that produced parts with the highest mechanical strength in the tensile tests required average amounts of energy, among the cases studied. Only in the case of the PS (mm/min) parameter are the results contradictory, as low PS values increase the tensile strength and the energy consumption. In this case, a value should be selected according to the requirements of each application and all the other 3D-printing settings should be adjusted accordingly.
Regarding the mechanical properties of the MEX 3D printing PMMA parts, the literature is still limited. Studies are focusing on the flexural, impact, and hardness properties of the parts made with PMMA in resin form [66,71]. Studies that have investigated the tensile properties of the 3D printed PMMA parts in the MEX [68] and the VPP process [65] are in good agreement with the results presented herein. The effect of the 3D printing parameters on the tensile properties of the build parts becomes evident by looking at the results of the work since, for example, the tensile strength of the parts built with the various parameters studied, varies from about 6 MPa to more than 40 MPa, which is more than 600% difference. Such variations in the results justify the need for the analysis conducted in the study, to comprehend the effect of the 3D printing settings on the build parts performance. The differences in the 3D printing settings affected the fracture mechanism on the samples, as it is shown in Figure 4. From the fracture surfaces it can be observed that in parts with less solid structure (more and larger voids shown in the fracture surface), such as the ones of run 4 and 5 for example, the mechanical performance was inferior to the parts with more solid structure, for example, the samples shown for runs 12 and 17. This is an expected response, as larger and more in number voids affect the mechanical performance of the parts, due to the reduced samples' area and the inferior fusion between the strands [90].
Regarding energy consumption, there is no study available in the literature for the PMMA polymer in 3D printing. So, the results presented herein cannot be directly correlated to the literature. A study on the energy consumption of MEX 3D printed ABS parts [32] concluded that high PS (mm/min) and LT (mm) values significantly decrease the energy consumption for the manufacturing of the parts. These results agree with the findings presented herein. In [32] the mechanical properties of the ABS MEX 3D printed parts were also evaluated and the ID (%) was the dominant parameter for the tensile strength which is not the case here. The mechanical test findings agree regarding the effect of RDA and do not agree on the other 3D printing parameters. This difference can be attributed to the different materials studied mainly.
In the study presented herein, the 3D printing settings used during the process highly affected the energy requirements to build the parts, as the difference between the lowest and the highest EPC values recorded in the study is higher than 200 %. Such differences justify the need for the analysis conducted herein and also for the energy indicators. Such differences also indicate that with the proper selection of the 3D printing settings, the eco-friendliness and sustainability of the MEX 3D printing process can be significantly increased. Reduced energy demands also reduce the cost of the parts built with the process. The EPC values follow a similar pattern with the printing time and the part weight metrics, showing a strong correlation between them. For example, PS and LT were the two most important parameters for the printing time and the part weight. Their increase constantly decreased these two metrics up to the highest value studied. The same effect is shown for the EPC metric, with LT being the dominant parameter of the metric, as well. It should be noted though, that these findings refer to the specific 3D printer used and the specific control parameters and levels.

Conclusions
In this study, an attempt was made to evaluate the effect of six generic, machineindependent 3D-printing settings on the energy consumption and the tensile strength of parts built with the Poly[methyl methacrylate] polymer with material extrusion 3D printing. Such an approach has merit, especially in the case of the Poly[methyl methacrylate] polymer, for which the literature in material extrusion 3D printing is still rather limited. The experimental data were processed with statistical modeling tools and regression analysis. Raster deposition angle was the dominant parameter for the tensile strength, with low values achieving increased results. The layer thickness was the dominant parameter for energy consumption, with high values reducing the metric.
The values that minimized energy consumption (print speed of 60 mm/s and layer thickness of 0.25 mm) did not produce parts with increased tensile strength. On the other hand, a zero-degree raster deposition angle and nozzle temperature of 240 • C, which produced Poly[methyl methacrylate] parts with high tensile strength, required moderate energy amounts for the process. This indicates that a compromise should be made between the highest mechanical properties and the required energy for the production of the part. Still, these two criteria are not totally independent of each other.
To derive such results, the produced prediction models were evaluated with two confirmation runs with five replicas each, making the models acceptable for use in reallife applications. In future work, the applicability of the results can be expanded, by broadening the range of the control parameter values and by investigating additional 3D-printing settings.

Data Availability Statement:
The data presented in this study are available upon request from the corresponding author.