Fused Deposition Modelling (FDM) is a 3D-printing technique that is used for prototyping. Because of FDM application in various industries such as medical instruments and aerospace, the produced part surface quality, dimensional accuracy and mechanical properties are brought to attention in recent years. Optimizing the process parameters helps to enhance the mentioned products properties. In this paper, influential process parameters including print speed, infill percentage and nozzle temperature were selected to examine. That was concluded nozzle temperature and infill percentage increasing and print speed decreasing enhance dimensional accuracy. Density of the FDM part was affected mostly by infill percentage. Desirability method was used in order to conduct optimization. That was resulted nozzle temperature of 240°C, infill percentage of 50% and print speed of 54 mm/min are the optimum parameters. Besides, verification tests were carried out and maximum error of 7.5% were reported which shows the optimization confidence. In the following, part geometry influence on the final part surface roughness and dimensional accuracy was examined. The experiments demonstrated cube geometry has the least surface roughness and dimensional error that was due to independent motion of the axes. Also cone geometry has the highest dimensional error and surface roughness which was due to simultaneous axes motion and lateral face with lower than 90°.
Research Article
Practical examining performance of the FDM 3D printed parts
https://doi.org/10.21203/rs.3.rs-1598555/v1
This work is licensed under a CC BY 4.0 License
Version 1
posted
You are reading this latest preprint version
Fused Deposition Modelling
Dimensional accuracy
Surface roughness
Density
Optimization
Part geometry
Additive Manufacturing (AM) processes reference technologies that produce an object by constructing one layer at a time. AM is very cost effective and time consuming. Using AM which is also called 3D printing is rising because of the process flexibility and ability to produce complex structures in less time[1, 2]. One of the common methods in AM is Fused Deposition Modeling (FDM). FDM was introduced in 1990s by Stratasys Inc.[3]. FDM is a 3D printing process that continuous filament or thermoplastic material is fed by a moving and heated nozzle (print extruder head) and deposited on a growing part. The extruder head moves in two directions in order to create one layer horizontally. Afterwards, head or work table moves vertically to create other layers. The movements and speeds are controlled by computer programming. Figure 1 demonstrates schematic of FDM process.
Various complex products that is produced by FDM requires suitable process condition settings. Inappropriate parameter selection causes poor dimensional accuracy and mechanical properties, waste of energy, unacceptable surface quality and increased cost and time[4]. Hence, optimization of the input parameters in FDM is compulsory in order to obtain the requested process output and simultaneously adjust the process parameters. In FDM products, surface quality, part strength and dimensional accuracy are the most important parameters that determine the process performance. These main subjects have attracted numerous researchers. Different investigations have been conducted in this regard. There were different goals in the researches such as surface roughness[5], dimensional error[6], elastic performance[7], residual stress[8], tensile strength[9] and compress strength[10].
Surface roughness parameter has been further studied compared to other FDM products properties. Corral et al.[11] studied on effect of print orientation on surface roughness in FDM. Peng et al. [12] conducted a multi-objective analysis aimed at surface roughness and energy optimization in FDM process. The FDM parameters include printing speed, layer thickness and infill percentage. Vahabli et al.[13] carried out an optimization using neural network method and predicted surface roughness in varied surface building angles. Kim et al.[14] studied on effects of FDM parameters including nozzle – substrate distance, filament inflow speed and nozzle speed on surface roughness and extracted a polynomial correlation in order to estimate surface roughness. Khan et al.[15] improved surface quality for FDM built parts from ABS material. They statistically investigated on Raster angle, raster width, air gap, time, temperature and smoothening cycle effects on surface roughness.
Another prerequisite for an AM product to be useful is dimensional accuracy. In recent years, many medical applications are carried out via desktop FDM devices which is due to cost reduction. Therefore higher dimensional accuracy is required. It is needless to say that part's accuracy depends on the customer requirements. In order to achieve better dimensional accuracy in FDM, that is necessary to examine the process parameters effect and adjust them. Generally, the process parameters are selected from handbooks but those information do not fulfill the required desires. Sahu et al.[16] optimized FDM process in order to achieve higher dimensional accuracy by using fuzzy logic and found part orientation has the most effect on dimensional accuracy. Azhikannickal et al.[17] carried out a research on the effects of layer height and infill ratio on dimensional stability of the FDM part. The part material was ABS. Melenka et al.[18] declared analyzed layer thickness, infill ratio and print orientation influence on dimensional accuracy. Up to 6% dimensional error was observed. More detailed review related to researches about FDM products surface quality and dimensional accuracy can be found in [19, 20].
The next parameter that influences on the FDM part strength is density value. The part with different densities have the same exterior apparent. Porosity of the final FDM product can be measured via final part density. The higher the density, the less porosity the part has. Porosity parameter has not been investigated deeply by the researchers and needs to be studied more.
Different materials have different printing conditions in order to be accurate and have suitable properties. Most of the researches have been carried out on ABS (Acrylonitrile Butadiene Styrene)[21, 22]. Also, nylon have been used in investigations[23]. Zharylkassyn et al. [20] stated nylon parts tend to be accurate by higher thickness of the layers which is also correct for ASA (Acrylonitrile Styrene Acrylate) material[24] whereas for ABS parts, preferable thickness of the layer is between 0.1 and 0.2 mm. Nowadays, PLA (Poly-lactic Acid) material was also introduced which is appropriate for biodegradable medical devices. Unlike other polymers, PLA is a material primarily meant for AM.
Finding optimum process condition has a main role in order to enhance surface quality, dimensional accuracy and part strength in FDM process. Aim of this investigation is to define effects of important FDM process parameters on dimensional accuracy and 3D-printed part density. In the following, an optimization using desirability method is conducted to give a suggestion for 3D-printing the PLA material in the best condition. On the other hand, geometrical shape of the FDM product effect on surface roughness and dimensional accuracy of 3D-printed PLA material is discussed for the first time.
This investigation is in two sections: 1- Examining FDM process parameters influence on dimensional accuracy and part density. 2- Examining part geometry influence on final part quality.
In 1st section, initially 3D model of the cylindrical part is modelled by Solidworks software. (Fig. 1).
Afterwards the files converted into STL file and imported in Cura software. Cura is an open source and user-friendly slicing application that is used for 3D printers. This application is owned by Ultimaker Company. By this software, the operator can control the printing parameters such as infill density of the layers, nozzle head temperature and print speed of the 3D printing machine. Also, via Cura software, G-code files are produced. One G-code file involves commands in a language that describes how a 3D FDM should print a job. The nozzle movement in different axis, nozzle speed and temperature are received from G-code file. In the following, the files are transmitted to the machine through SD card or USB cable. All the parts are printed by 3D FDM (Quantum 2035). Figure 2 demonstrates the FDM setup. General specifications of the used printing machine is presented in Table 1. Poly Lactic Acid (PLA) is used in this investigation. That is one of materials produced by AM process. This material is also called plant-based thermoplastic. PLA is a strong and light material. PLA material specifications are also presented in Table. 2.
| Specification | Quantum 2035 |
|---|---|
| Print size (cm) | 20×20 |
| Print accuracy (µm) | 100 |
| Layer thickness (µm) | 60–400 |
| Number of nozzles | 1 |
| Nozzle diameter (mm) | 0.2 |
| Materials | ABS, PLA, WOOD, METAL, PA,PC, PP, POM, HIPS, PVA |
| Material diameter (mm) | 1.75 |
| Print speed (mm/s) | up to 120 mm/s |
| Nozzle temperature (°C) | up to 280 |
| Nozzle speed (mm/s) | up to 180 |
| Print table temperature (°C) | up to 120 |
| Specification | PLA |
|---|---|
| tensile strength (MPa) | 0.88 |
| Elastic modulus (GPa) | 8.6 |
| Elongation at break (%) | 3–30 |
| Flexural strength (MPa) | 70 |
| Melt Flow Index (g/min) | 0.85 |
Next, in order to design the experiments, full factorial method is used. This method covers all the affecting factors and presents a robust system. Thereupon, the experiments are conducted based on the provided Design Of Experiments (DOE). Process variables with levels are listed in Table. 3.
| No. | Variable | Symbol | Levels |
|---|---|---|---|
| 1 | Print speed (mm/min) | PS | 50-100-150 |
| 2 | Infill Percentage (%) | IP | 10-30-50 |
| 3 | Nozzle Temperature (°C) | NT | 195–240 |
In order to examine the process performance, two important results consisting dimensional accuracy and part density are measured. By CMM (Coordinate Measuring Machine) device (manufactured by Tesa 3D), interior and outer ovality of the samples (representative of the dimensional accuracy) are calculated. Minitab software is used in order to correlate a relationship between input and output parameters. Moreover, statistical analysis and regression analysis are carried out. The regression model accuracy is also examined through R-squared (R2) parameter and Analysis Of Variance (ANOVA) that determines input parameters significance. R2 helps to check the obtained regression model fitting with the experimental results.
In 2nd section, 3D models of four part geometries including cube, cylinder, cone and pyramid are modelled by Solidworks software. In order to conduct the experiments, part geometry is changed (cube, cone, cylinder, pyramid) and examined in three PS (50, 100 and 150 mm/min). In order to examine the process performance, total dimensional accuracy and surface roughness are measured.
Table. 4 gives experimental results with control and response parameters.
| Table 4. Experimental results | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No. | Control parameters | Response parameters | ||||||||
| PS (mm/min) | Infill (%) | NT (°C) | Outer Ovality (mm) | Interior Ovality (mm) | Density (g/cm3) | |||||
| 1 | 50 | 10 | 195 | 0.89 | 0.03 | 0.07 | ||||
| 2 | 100 | 10 | 195 | 0.9 | 0.06 | 0.1 | ||||
| 3 | 150 | 10 | 195 | 0.92 | 0.09 | 0.16 | ||||
| 4 | 50 | 30 | 195 | 0.96 | 0.05 | 0.1 | ||||
| 5 | 100 | 30 | 195 | 0.95 | 0.06 | 0.1 | ||||
| 6 | 150 | 30 | 195 | 0.96 | 0.08 | 0.17 | ||||
| 7 | 50 | 50 | 195 | 1 | 0.02 | 0.07 | ||||
| 8 | 100 | 50 | 195 | 1.02 | 0.03 | 0.09 | ||||
| 9 | 150 | 50 | 195 | 1 | 0.05 | 0.12 | ||||
| 10 | 50 | 10 | 240 | 0.9 | 0.04 | 0.09 | ||||
| 11 | 100 | 10 | 240 | 0.92 | 0.03 | 0.06 | ||||
| 12 | 150 | 10 | 240 | 0.89 | 0.03 | 0.07 | ||||
| 13 | 50 | 30 | 240 | 0.93 | 0.06 | 0.11 | ||||
| 14 | 100 | 30 | 240 | 0.97 | 0.01 | 0.03 | ||||
| 15 | 150 | 30 | 240 | 0.97 | 0.04 | 0.09 | ||||
| 16 | 50 | 50 | 240 | 1.02 | 0.06 | 0.13 | ||||
| 17 | 100 | 50 | 240 | 1.03 | 0.03 | 0.07 | ||||
| 18 | 150 | 50 | 240 | 1.02 | 0.04 | 0.06 | ||||
In the following, each parameter is discussed separately.
Ovality
Ovality parameter is measured in two positions: inner diameter and outer diameter. ANOVA results for inner and outer ovality parameter are listed in Tables. 5 and 6, which shows effects of input variables consisting PS, IP and NT. According to the data in Tables 5 and 6, PS has the most influence on inner and outer ovality individually by contribution of 43.92% and 66.6%, respectively. On the other hand, the simultaneous influence of multiple factors on each response value should be examined in order to understand effects of 3D printing variables together. In this regard, ANOVA results help to study interaction effects of the input variables on the resultant parameter. Interaction of NT and PS has significant influence on inner ovality parameter with p-values < 0.003. This is also the same for outer ovality. Based on Tables 5 and 6, IP and NT has the same influence on outer ovality (14.6%) but in case of inner ovality, IP has more influence (36.3%) compared to NT (11.93%). That is dedicated IP is more influential on inner section of the 3D-printed parts. Hence, for the parts with inner holes, IP parameter will be more significant.
| Source | Degree of Freedom | Contribution | Sum of squares | Mean | F-Value | P-Value |
|---|---|---|---|---|---|---|
| Regression | 8 | 99.44% | 0.006913 | 0.000864 | 201.28 | 0.000 |
| NT | 1 | 11.93% | 0.000048 | 0.000048 | 11.23 | 0.009 |
| IP | 1 | 36.30% | 0.000018 | 0.000018 | 4.11 | 0.073 |
| PS | 1 | 43.92% | 0.000014 | 0.000014 | 3.37 | 0.100 |
| IP*IP | 1 | 1.94% | 0.000135 | 0.000135 | 31.34 | 0.000 |
| PS*PS | 1 | 3.02% | 0.000210 | 0.000210 | 48.98 | 0.000 |
| NT*IP | 1 | 0.81% | 0.000056 | 0.000056 | 13.12 | 0.006 |
| NT*PS | 1 | 0.98% | 0.000068 | 0.000068 | 15.88 | 0.003 |
| IP*PS | 1 | 0.54% | 0.000038 | 0.000038 | 8.82 | 0.016 |
| Error | 9 | 0.56% | 0.000039 | 0.000004 | ||
| Total | 17 | 100.00% |
| Source | Degree of Freedom | Contribution | Sum of squares | Mean | F-Value | P-Value |
|---|---|---|---|---|---|---|
| Regression | 8 | 98.60% | 0.023720 | 0.002965 | 79.31 | 0.000 |
| NT | 1 | 14.63% | 0.000083 | 0.000083 | 2.21 | 0.171 |
| IP | 1 | 14.60% | 0.000048 | 0.000048 | 1.30 | 0.284 |
| PS | 1 | 66.60% | 0.000694 | 0.000694 | 18.57 | 0.002 |
| IP*IP | 1 | 0.17% | 0.000041 | 0.000041 | 1.09 | 0.324 |
| PS*PS | 1 | 0.02% | 0.000005 | 0.000005 | 0.14 | 0.713 |
| NT*IP | 1 | 0.33% | 0.000078 | 0.000078 | 2.10 | 0.181 |
| NT*PS | 1 | 1.49% | 0.000358 | 0.000358 | 9.57 | 0.013 |
| IP*PS | 1 | 0.76% | 0.000183 | 0.000183 | 4.90 | 0.054 |
| Error | 9 | 1.40% | 0.000336 | 0.000037 | ||
| Total | 17 | 100.00% |
Figure 3 demonstrates main effects values for each factor that influences on inner and outer ovality, respectively. This plot shows influence of NT, infill and PS. Increasing NT causes lower ovality value that means improved dimensional accuracy. That is due to further adhesion of 3D printed layers in case of larger nozzle heat which creates stronger and more stable structure. Moreover, higher NT is required in order to release the material from the extruder without delay.
Based on Fig. 3, IP increment produces 3D printed structure with higher dimensional accuracy. This can be interpreted as result of higher 3D printed segment structure at higher IP. In case of low IP, the produced segment will be more porous. Hence, anisotropic shrinkage is happened and dimensional accuracy is reduced.
Figure 3 show that higher PS causes more ovality error amount. That could be due to low cooling time and residual heat from the bottom layer that induces unstable structure in 3D printed segment. On the other hand, Anitha et al.[25] proved that lower PS results in better surface quality. Therefore, lower PS improves both dimensional accuracy and surface quality.
Besides, based on the information obtained from ANOVA, interaction of NT and PS influences on ovality parameter. That is concluded in order to acquire better dimensional accuracy, higher NT and lower PS are more influential to be set together. In higher NT, the material is extruded easily and fills the porosities in the part which induces stronger structure with higher dimensional accuracy. In addition to this, lower PS helps to fill the part further.
As a result, the best dimensional accuracy is achieved in NT of 240°C, IP of 50% and PS of 150 mm/min.
In the following, regression correlations were established in order to estimate inner and outer ovality error (Eq. 1–2). In order to calculate the goodness of fit for Eq. 1–2, R-squared tool can be utilized. Higher R-squared indicates the established regression model strength. The calculated R-squared for Eqs. 1 and 2 is 97.79% and 95.37%, respectively which imply the best fitness of the second order regression correlation.
Inner Ovality = 0.0947–0.000234 NT − 0.000685 IP + 0.000265 PS − 0.000014 IP*IP + 0.000003 PS*PS + 0.000005 NT*IP − 0.000002 NT*PS − 0.000002 IP*PS R2 = 97.79% (1)
Outer Ovality = 0.0973–0.000307 NT − 0.001134 IP + 0.001837 PS − 0.000008 IP*IP + 0.000000 PS*PS + 0.000006 NT*IP − 0.000005 NT*PS − 0.000005 IP*PS R2 = 95.37% (2)
Previously, the regression model should be checked to be adequate. In this regard, normal probability plot of the residuals is used. This plot is a graphical tool in order to compare the response data (dimensional accuracy) with normal distribution. Fig. shows normal probability of residuals for the regression model in Eqs. 1 and 2. That is obvious most of the points are set along a straight line which implies normal distribution of errors of the established model.
Density
ANOVA results for density parameter is presented in Table 7 which shows effects of input parameters consisting PS, IP and NT on final density of the 3D printed part. According to Table 7, IP has the most influence on density individually by contribution of 92.94%. PS and NT are in the next orders but with very low contribution (0.75% and 0.35%, respectively).
On the other hand, ANOVA results help to study interaction effects of the input variables on the resultant parameter. In this regard, interaction of NT and IP have the most significant influence on density (0.52%).
| Source | Degree of Freedom | Contribution | Sum of squares | Mean | F-Value | P-Value |
|---|---|---|---|---|---|---|
| Regression | 8 | 95.73% | 0.038531 | 0.004816 | 25.21 | 0.000 |
| NT | 1 | 0.35% | 0.000021 | 0.000021 | 0.11 | 0.748 |
| IP | 1 | 92.94% | 0.000016 | 0.000016 | 0.08 | 0.779 |
| PS | 1 | 0.75% | 0.000196 | 0.000196 | 1.02 | 0.338 |
| IP*IP | 1 | 0.06% | 0.000025 | 0.000025 | 0.13 | 0.726 |
| PS*PS | 1 | 0.99% | 0.000400 | 0.000400 | 2.09 | 0.182 |
| NT*IP | 1 | 0.52% | 0.000208 | 0.000208 | 1.09 | 0.324 |
| NT*PS | 1 | 0.00% | 0.000000 | 0.000000 | 0.00 | 1.000 |
| IP*PS | 1 | 0.12% | 0.000050 | 0.000050 | 0.26 | 0.621 |
| Error | 9 | 4.27% | 0.001719 | 0.000191 | ||
| Total | 17 | 100.00% |
As demonstrated in Fig. 5, main effects plot for density is presented. NT and PS are not very influential on final segment density and have a little bit effect which can be neglected. On the other hand, IP increase leads to density increase because of lower part porosity in higher IP.
On the other hand, as obtained from ANOVA, NT and IP influence together on density. That was understood in order to obtain higher density, both higher temperature and IP are required. That seems in higher temperatures, the material releases quickly from the extruder and fills more porosities in the produced part.
In the following, regression correlation is established in order to estimate density of the part (Eq. 3). As described previously, in order to calculate the Eq. 2 goodness of fit, R-squared tool can be utilized. The calculated R-squared for Eq. 1 is 95.73% which implies the best fitness of the second order regression correlation.
Density = 0.861–0.000154 NT + 0.00065 IP + 0.000975 PS + 0.000006 IP*IP − 0.000004 PS*PS
+ 0.000009 NT*IP + 0.000000 NT*PS − 0.000003 IP*PS R2 = 81.95% (3)
In the following, adequacy of the established regression model (Eq. 3) is examined by normal probability plot of the residuals. This graphical tool helps to compare the density data with normal distribution. Figure 6 shows normal probability of residuals for the regression model in Eq. 3. That is observed most of the density points are set along a straight line which implies normal distribution of errors of the established model.
Optimization
Multi objective optimization is required to assess the best process condition aimed at the least dimensional error and the largest density. In this regard, desirability method will be applied. The relationships to obtain desirability value is as the following:
While di is the desirability value for each process response; Y is response value; Low and High are the minimum and maximum amount of the resultant response; r is number of the experimental tests; and w is the weight factor.
Optimized process condition is illustrated in Fig. 7. Red lines in Fig. 7 show optimum values for each effective parameter on Density, inner and outer ovality and finally in total process (named by composite desirability). Since composite desirability (D) value is 0.9606 and near to 1, that means the acquired condition is precise. That is shown composite desirability will be occurred in IP of 50%, PS of 50 mm/min and NT of 240° C.
On the other hand, confirmation evaluation should be conducted to verify the optimization procedure. According to Table 8, the verification test carried out and the error values between experimental and predicted results are 2.9%, 6.6% and 7.5% for density, inner and outer ovality, respectively. The error values are in acceptable range which indicate the optimization method is suitable for FDM process control.
Confidential interval demonstrates the probability that a certain parameter value will be in the defined range. The interval is calculated at the designated confidential level that is 95% in this investigation. Wider confidence range shows less confidential level of predicting the process responses. Both the predicted and experimental values are in confidential interval. Prediction interval is an estimation of interval that future results will occur. Prediction interval is always larger than confirmation interval. As shown in Table 8, the experimental results are in the range of prediction interval that shows the model prediction strength.
| Parameter | Predicted result | Experimental result | Error amount (%) | Confidence interval (%) | Prediction interval (%) |
|---|---|---|---|---|---|
| Density | 1.01 | 0.98 | 2.97 | (0.994,1.041) | ( 0.9788,1.0569) |
| Inner ovality | 0.015 | 0.014 | 6.6 | (0.0116,0.0187) | (0.0093,0.0215) |
| Outer ovality | 0.040 | 0.037 | 7.5 | (0.0296,0.0503) | (0.0227,0.0572) |
Part geometry effect
Because of axes moving and their interference in FDM process, the product shape influences on the process performance. In this regard, four geometrical shapes are selected for 3D printing: cube, cone, cylinder and pyramid. Influential parameters in determining 3D printing performance are surface quality and dimensional accuracy[26] which selected to examine the part geometrical effect. The produced parts are shown in Fig. 8.
In order to calculate dimensional and geometrical accuracy of the part, all dimensional errors are measured by CMM and the maximum error is reported. As demonstrated in Fig. 9, the cubic part has the best dimensional accuracy comparing to others. In cubic part, the X and Y axes move independently from each other. The error in the movements can be due to friction between the components. That is shown cylindrical part produces more dimensional error compared to cubic one. That could be due to axes movement simultaneously which produces further movement error. Also the parts that the lateral faces are perpendicular to XY plane (cube and cylinder), the dimensional error is lower comparing to conical and pyramidal shapes (lateral faces have lower than 90° angle with XY plane).
Moreover, Fig. 10 shows part geometry effect on surface roughness. The cone and pyramid shaped parts have the worst surface quality. That could be due to slicing procedure. Based on Fig. 11, because of larger dimensional error in nozzle movement, the gap will be produced in different heights which destroys the surface quality. This phenomenon is lower in cube and cylinder parts. Also the parts with circular cross-section (cone and cylinder) have higher surface roughness compared to rectangular cross-section ones (pyramid and cube) which is due to nozzle circular motion that induces further dimensional error and lower slices arrangement.
This experimental investigation presents the effects of FDM parameters including NT, IP and PS on dimensional accuracy and part density. In this regard, ovality of the cylindrical part was selected as dimensional accuracy criteria. Furthermore, geometrical shape of the produced part influence on the FDM process performance was examined. The main inferences of the investigation are as follows:
-
According to ANOVA analysis, PS has the most effect on dimensional accuracy of the 3D-printed part by 43.92% contribution. Moreover, interaction of NT and PS is significant for determining dimensional accuracy.
-
In order to improve dimensional accuracy in FDM process, NT and IP should be increased and PS should be decreased.
-
Based on ANOVA analysis, the most influential parameter on the produced part density was IP with 92.94% contribution.
-
The regression models were formulated for estimating dimensional accuracy and density whose R-squared were 97.79% and 95.73% which point high fitness of the second order regression correlations.
-
FDM process can be well optimized by adjusting the process parameters. The desirability method was utilized for optimization which resulted the highest density and dimensional accuracy can be acquired by NT of 240°C, IP of 50% and PS of 54 mm/min.
-
Confirmation test was conducted and resulted 2.97% and 6.6% error value for density and dimensional error, respectively. That means the desirability method is suitable for FDM process optimization.
-
In FDM, movement of X and Y axes independently produces the parts with higher dimensional accuracy which was shown in cubic parts compared to other geometrical shaped ones.
-
The parts which lateral faces are not perpendicular to XY plane, have lower surface quality and dimensional accuracy due to gap in different heights.
Ethical approval: It is approved that the paper is original and has been written based on the authors' own findings. All the figures and tables are original and every expression from other published works were acknowledged and referenced.
Consent to participate: It is confirmed that all the authors are aware and satisfied with the authorship order and correspondence of the paper
Consent to Publish: All the authors are satisfied that the last revised version of the paper is published without any change
Authors Contributions: Mohammad Baraheni had 35% contribution in conducting the research and analyzing the results; Mohammad Reza Shabgard had 25% contribution in supervising the research; Ali Mahdavi Tabatabaee had 25% contribution in providing facilities and Amir Hossein Adhami had 15% contribution in conducting the research experiments.
Funding: This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Competing Interests: It is declared that there is no conflict of interest in authorship and data provided.
Availability of data and materials: It is confirmed that the data and materials can be available after publication on the basis of springer nature rights and access.
- Mohamed OA, Masood SH, Bhowmik JL (2015) Optimization of fused deposition modeling process parameters: a review of current research and future prospects. Adv Manuf 3(1):42–53
- Baraheni M, Shabgard MR, Amini S (2021) Evaluating the hole quality produced by vibratory drilling: additive manufactured PLA+. Int J Adv Manuf Technol 117(3):785–794
- Gebhardt A (2011) Understanding additive manufacturing.
- Rahmati S, Vahabli E (2015) Evaluation of analytical modeling for improvement of surface roughness of FDM test part using measurement results. Int J Adv Manuf Technol 79(5):823–829
- Thrimurthulu K, Pandey PM, Reddy NV (2004) Optimum part deposition orientation in fused deposition modeling. Int J Mach Tools Manuf 44(6):585–594
- Zhang JW, Peng AH (2012) Process-parameter optimization for fused deposition modeling based on Taguchi method. in Advanced Materials Research. Trans Tech Publ
- Laeng J, Khan ZA, Khu S (2006) Optimizing flexible behaviour of bow prototype using Taguchi approach. J Appl Sci 6(3):622–630
- Zhang Y, Chou K (2008) A parametric study of part distortions in fused deposition modelling using three-dimensional finite element analysis. Proceedings of the Institution of Mechanical Engineers, Part B: Journal of Engineering Manufacture. 222(8): p. 959–968
- Rayegani F, Onwubolu GC (2014) Fused deposition modelling (FDM) process parameter prediction and optimization using group method for data handling (GMDH) and differential evolution (DE). Int J Adv Manuf Technol 73(1):509–519
- Ang KC, Leong KF, Chua CK, Chandrasekaran M (2006) Investigation of the mechanical properties and porosity relationships in fused deposition modelling-fabricated porous structures. Rapid Prototyping Journal
- Buj-Corral I, Domínguez-Fernández A, Durán-Llucià R (2019) Influence of print orientation on surface roughness in fused deposition modeling (FDM) processes. Materials 12(23):3834
- Peng T, Yan F (2018) Dual-objective analysis for desktop FDM printers: energy consumption and surface roughness. Procedia CIRP 69:106–111
- Vahabli E, Rahmati S (2016) Application of an RBF neural network for FDM parts’ surface roughness prediction for enhancing surface quality. Int J precision Eng Manuf 17(12):1589–1603
- Kim MK, Lee IH, Kim H-C (2018) Effect of fabrication parameters on surface roughness of FDM parts. Int J Precision Eng Manuf 19(1):137–142
- Khan M, Mishra S (2020) Minimizing surface roughness of ABS-FDM build parts: An experimental approach. Materials Today: Proceedings. 26: p. 1557–1566
- Sahu RK, Mahapatra S, Sood AK (2013) A study on dimensional accuracy of fused deposition modeling (FDM) processed parts using fuzzy logic. J Manuf Sci Prod 13(3):183–197
- Azhikannickal E, Uhrin A (2019) Dimensional Stability of 3D Printed Parts: Effects of Process Parameters.Ohio Journal of Science.119(2)
- Melenka GW, Schofield JS, Dawson MR, Carey JP (2015) Evaluation of dimensional accuracy and material properties of the MakerBot 3D desktop printer. Rapid Prototyping Journal
- Turner BN, Gold SA (2015) A review of melt extrusion additive manufacturing processes: II. Materials, dimensional accuracy, and surface roughness. Rapid Prototyping Journal
- Zharylkassyn B, Perveen A, Talamona D (2021) Effect of process parameters and materials on the dimensional accuracy of FDM parts. Materials Today: Proceedings. 44: p. 1307–1311
- Nidagundi VB, Keshavamurthy R, Prakash C (2015) Studies on parametric optimization for fused deposition modelling process. Materials Today: Proceedings. 2(4–5): p. 1691–1699
- Vishwas M, Basavaraj C, Vinyas M (2018) Experimental investigation using taguchi method to optimize process parameters of fused deposition Modeling for ABS and nylon materials. Materials Today: Proceedings. 5(2): p. 7106–7114
- Basavaraj C, Vishwas M (2016) Studies on effect of fused deposition modelling process parameters on ultimate tensile strength and dimensional accuracy of nylon. in IOP conference series: materials science and engineering. IOP Publishing
- Camposeco-Negrete C (2020) Optimization of printing parameters in fused deposition modeling for improving part quality and process sustainability. Int J Adv Manuf Technol 108(7):2131–2147
- Anitha R, Arunachalam S, Radhakrishnan P (2001) Critical parameters influencing the quality of prototypes in fused deposition modelling. J Mater Process Technol 118(1–3):385–388
- Saad MS, Nor AM, Baharudin ME, Zakaria MZ, Aiman A (2019) Optimization of surface roughness in FDM 3D printer using response surface methodology, particle swarm optimization, and symbiotic organism search algorithms. Int J Adv Manuf Technol 105(12):5121–5137