Influence of 3D Printing Parameters on Mechanical Properties of the PLA Parts Made by FDM Additive Manufacturing Process

Paper presents the results of factorial experiments made to establish the influence of technological parameters of 3D printing using the Fused Deposition Modeling (FDM) Technology, on the mechanical properties of the material deposited with ULTRA PLA filaments. By planning the experiments and the statistical processing of the results, mathematical relations were established regarding the dependence between the objective functions, the controllable factors and their interactions. To obtain high tensile strengths of the components made by 3D printing, regimes are required to allow the deposition of layers as thin as possible (h = 0.1 mm) and temperatures as high as possible, close to 300°C. The choice of inappropriate values of these process parameters can lead to a significant decrease in tensile strength, reaching even up to 30-40% of the maximum possible value to be reached. The experiments reveal that the printing speed does not have a major influence on the mechanical properties. Practical, the printing speed is limited by the technical characteristics of the printer used. The results of the experimental research obtained on a number of 30 process variants led to the establishment of optimal 3D printing variants that correspond to the requirements imposed on the objectively analysed functions (tensile strength, dimensional accuracy, speed of execution, surface quality).


Introduction
Fused Deposition Modeling (FDM) additive manufacturing technology is constantly evolving. New printing possibilities is implemented, high-performance prints have been made and new materials have appeared. However, although it has been used for many years, the PLA (PolyLactic Acid) filament has not lost ground to the new materials, being one of the most used filaments in 3D printing with thermoplastic materials due to the fact that it behaves properly in the printing process. It combines mechanical characteristics with those of plasticity at an acceptable level, it is a biodegradable material, its decomposition being achieved at temperatures above 60 °C in certain humidity conditions and in the presence of microorganisms that accelerate the process. The technological process of 3D printing is a complex technological system, being characterized by a relatively large number of independent variables, which requires their optimization in order to obtain high productivity and quality of products. A modern variant of industrial process optimization is the use of factorial experiments that consider the analysis of the influence of certain controllable factors on the outputs, as response functions. In this way, the necessary resources and workload can be considerably reduced, by resorting to the statistical processing of experimental results and establishing relationships of dependence between objective functions, controllable factors and their interactions.
The effect of printing parameters on the mechanical properties of parts fabricated by 3D printing is a topic of interest for specialists in the field, being known a series of previous studies [1 -6]. It has been found that 3D printing by FDM is a complex phenomenon and the mechanical properties of the printed parts are strongly depending on the 3D printing process parameters and on the filament material composition. On the other hand, since FDM deposits material directionally, the result is a layered parts with anisotropic behaviour. Also, normally distributed errors were achieved between the predicted and experimental values. The experiments have shown [3] that the pre-processing technologies such as vacuum-assisted FDM and laser or infrared heating could be used to improve the quality of the 3D printed parts due to positive impact on bonding between the layers. Also, the part cooling rate, environmental conditions and post-processing need to be considered as well since they have a significant effect on characteristics of the 3D printed components. Thus, research must be done to find optimal solutions [7][8][9][10][11] for the combinations of parameters to adapt them based on the application (since various optimal parameters could be obtained based on dimensional accuracy, printing time, flexural, impact or tensile strength as well as surface roughness).
In this paper, in order to establish the influence of technological parameters on the mechanical properties of the material deposited by printing using ULTRA PLA filament, the Design of Experiments approach (DOE) has been used [12]. Thus, a factorial experiment was designed and implemented aiming the following 5 objective functions: tensile strength, roughness of the printed surface, test piece thickness a, test piece width b and printing duration. The main factors (technological parameters) that significantly influence the results obtained are: print temperature, print speed (speed of movement of the print head relative to the printed component) and the height of the layer deposited at each pass (determined by the distance between the print head and component).

Material and methods
ULTRA filaments (PLA) were used in the experimental program on the influence of technological parameters on the mechanical properties of the material deposited by 3D printing. Table 1 shows the main characteristics of this material, according to the manufacturer's specification (Roboze). Description Material designed to make high-definition parts, ideal for prototypes and small series that require both precision and mechanical rigidity.
In order to determine the limits of variation of the parameters specific to the 3D printing process and to design the factorial experiments, preliminary tests were performed in a first stage of the research. It was taken into account that these parameters influence the mechanical strength characteristics and the dimensions of the printed specimen, thus there is a need to establish optimal variants of 3D printing regimes depending on the different requirements for the practical applications.
Preliminary tests have established the ranges in which the main parameters called controllable factors of the 3D printing process can vary so that the process runs smoothly. The ranges of values obtained based on the results of the preliminary research carried out, are presented in Table 2.

8
Engineering Innovations Vol. 2 Simplify3D software was used to design and simulate the printing of test specimens for the experiment program. An example of a simulation is shown in Fig. 1. The software allows to load from a database the 3D drawing of the component to be printed and the selection of process parameters: extrusion temperature, printing speed, layer height, how is made the "shell" of the component and selecting the mode of " filling" of the component. Also, the software can estimate the printing time of the component, the length of the filament consumed when making it, the weight of the component, as well as the cost of the material. The printing of the specimens was made with Roboze one printer (Fig. 2). To evaluate the surface quality of 3D printed components, roughness measurements were performed using a Mitutoyo SJ -201P device, Fig. 3.

Experimental testing program
Within the experimental program, specimens for tensile tests according to [13] were printed.     (Table 4). Next, samples for the experimental program were printed using the parameters presented above. Before being tested, the specimens were conditioned at a temperature of 23 °C ± 1 °C and a relative humidity of 50% for 8 hours, after which they were subjected to a mechanical tensile test using Zwick equipment with maximum load of 5 kN.

Results and Discussion
The aspect of the 3D printed specimens, after the tensile tests, is shown in Fig. 5.

Figure 5:
The appearance of 3D printed specimens, after the tensile test

Engineering Innovations Vol. 2
It is found that both the failure mode and the appearance of the fracture surfaces are related to the 3D printing parameters used to make the specimens. Thus, specimens 1, 3 and 5 show a ductile fracture with a partial or complete tearing of the printed structure (Fig. 6). The other specimens show a brittle fracture, specimen 4 presenting the most brittle behaviour the since the failure occurred suddenly with the detachment of a portion from the specimen (Fig. 7).
a) Specimen 1 b) Specimen 3 c) Specimen 5   Fig. 9b presents the response surfaces regarding to the influence of printing process parameters on tensile strength (Rm). Fig. 9a shows that in order to obtain the highest values of tensile strength of the printed components, it is favourable to be selected a temperature as close as possible to 300°C, correlated with the use of speeds close to the extreme values, (printing speed of 80 mm/s). Fig. 9b shows the need to select the highest possible values of temperature (T), close to 300°C and a thickness of the layer heigh (h) as small as possible, of 0.1 mm, to obtain high tensile strengths exceeding 25 N/mm 2 . It is also observed that in the case of lower temperature values, towards the minimum limit of 200°C and if a layer heigh close to 0.3 mm are used, a significant decrease of the tensile strength, up to 10 N/mm 2 can occur.
From the graphs shown in Fig. 9c and Fig. 9d, data can be obtained that confirm the information previously presented in the surface graphs. Thus, to obtain the highest possible values of tensile strength (Rm), the highest possible values of the printing temperature, close to 300°C, the layers heigh as thin as possible and the printing speeds close to the extreme values should be selected.  To obtain the lowest possible values of surface roughness, it is recommended to choose temperatures as close as possible to 300°C, layers heigh as thin as possible, of 0.1 mm and printing speeds close to 60 mm/s or 80 mm/s. From the graphs in the form of contour plots, shown in Fig. 11c and Fig. 11d, more accurate additional information can be obtained regarding to the selection of appropriate values of the process parameters, when it is desired to obtain a surface roughness as low as possible.
Thus, e.g., if the value of Ra is to be between 10 to 12.5 µm, it is recommended that the printing temperature be higher than 270°C and the layer height be selected less than 0.15 mm. (Fig.  11d). In the similar way Fig. 12 and Fig. 13 presents the response surfaces and the contour plots regarding to the influence of the printing process parameters on the dimensions of the specimen (a and b). In this case, there is an influence on the objective function (tp) only from a smaller number of controllable factors and their interactions, in particular it can be seen that there is a decrease in printing time with increasing of printing speed and layer height.
Given that this size of the specimen needs to be as close as possible to a standardized value, in this case 10 mm, it is recommended to select as much as possible a set of parameters with lower values for temperature and layer thickness, speed printing has a relatively small influence on this objective function. In order to obtain values of the test piece width (b) in the range 10.0-10.1 mm, it is shown that the process parameters can vary between relatively wide limits.
The optimization of the process as a whole was performed based on the mathematical models obtained for each objective function. In order to optimize the process, the following criteria have been established for each objective function: • tensile strength by Rm -maximum value • surface roughness Ra -minimum value • the test piece thickness a -target value of 4 mm • the test piece width b -target value of 10 mm • printing time t p -minimum value The optimal values of the printing process parameters for the different combinations of objective functions considered are presented in Table 5. Given some requirements that may arise in industrial practice, based on the results obtained in the experimental four other optimal variants of the printing process parameters were established as following:

18
Engineering Innovations Vol. 2 a) Component quality (dimensional accuracy and superior mechanical characteristics)optimization with four objective functions (Rm, Ra, a, b); b) Execution speed -optimization with three objective functions (tp, a, b); c) Component strength -optimization with an objective function (Rm); d) Degree of surface finish -optimization with an objective function (Ra).

Conclusions
A 1 st order factorial experiment was conducted on 30 test pieces taking into consideration the main process parameters of 3D printing process using 3 controllable factors (T, v and h) with 2 levels, 3 replicas for each combination of controllable factors and 6 replicas in the central point.
Based on the experimental results obtained, regression models for correlation between the 3D printing process parameters and objective functions: tensile strength, surface roughness and component dimension (a and b) were obtained. Regression models can be used to select the appropriate process parameters to obtain required characteristics of 3D printed components, depending on the loading and operation conditions. Experimentally, optimal variants of parameters were determined that correspond to minimum printing speeds, maximum tensile strength, high surface quality or high dimensional accuracy.