Experimental Research on Behaviour of 3D Printed Gripper Soft Jaws

In this paper, the problem of the behaviour of soft jaws that can be used to replace the steel jaws of grippers is studied. One of the advantages of additive manufacturing is the printing of fully functional parts. Choice of material is often related to the part strength. The mechanical properties of 3D printed parts should meet the service loading and, also, must be comparable with parts produced by traditional manufacturing techniques machined parts or injection moulding. From the specialized literature information regarding the test results for effect of various printing parameters on part strength are available made in laboratory conditions and for standard test sample. For ABS materials various values for Young module are presented varying from 1.5 GPa to 2.15 GPa, for 100% infill rate and various modified parameters such as raster orientation. In order to study the behaviour of soft gripper jaws several part were printing and the resistance to bending was tested, by simulating the way a gripper works. An experimental stand was built using a force transducer and a displacement transducer to measure the deformation of the jaw, obtained by 3D printing, under load. The mechanical elastic hysteresis loop during an experimental loading/unloading was plotted and the amount of mechanical energy lost during a cycle, dissipated because the internal friction, was determined. Finite element analysis method was applied to make a comparison with the experimental results. In the finite element analysis, several simulations were considered, varying Young's modulus for the tested material


Introduction
Grippers are used for orientation-positioning and tightening of the object / part during handling, processing, control and assembly. To ensure the fulfilment of the functions presented, the materials used in the construction of devices and grippers are usually alloy steel. The requirements of grippers are similar to those of clamping fixtures, namely: precise orientation and positioning according to machining requirements, rigidity depending on the size of the workpiece and machining forces, reliability, low maintenance, standardized components, high productivity, costs low [1]. To ensure the performance of the functions presented, the materials used in the construction of devices and grippers are usually alloy steel with an HRC hardness of up to 55-60 HRC for the clamping components.
There are also cases where "flexible" jaws / soft jaws are needed to reduce the deformation of workpieces, especially for thin-walled thin parts when clamping. Soft jaws are one of the simplest methods of customized workholding and can be machined to the negative shape of the workpiece.
Soft jaws can be manufactured by Additive Manufacturing (AM) methods by benefiting from the maximization of the flexibility offered by this method of manufacturing parts [2]. Additive Manufacturing (AM) is defined by the American Society for Testing and Materials as "the process of joining materials to make objects from 3D model data, usually layer by layer" [3,4].
One of the advantages of AM is the printing of fully functional parts. Choice of material is often related to the part strength. The mechanical properties of 3D printed parts should meet the service loading and, also, must be comparable with parts produced by traditional manufacturing techniquesmachined parts, for metal 3D printing or injection moulding thermoplastic [5,6]. Most metal AM processes create parts with poor surface finish, usually no more than 15 μm Rz and very often considerably reduced. Machine finishing is necessary as a post-process [5]. Thermoplastic 3D printing can achieve good quality surfaces [7].
From the specialized literature information regarding the test results for effect of various printing parameters such as: layer thickness, raster angle, build orientations, fill pattern, printing directions, infill rates and infill patterns air gap and model build temperature, made in laboratory conditions, for standard test sample on part strength are available [5,6,[8][9][10][11].
Dudescu tested the influence of the raster orientation and of the infill pattern and the results indicate a Young's modulus corresponding to the parts with infill percentage from 982 MPa at 20% to 1503 MPa at 100%, for ABS. The influence of the raster orientation for test parts with 100% infill rate have a constant value around 1500 MPa [6].
From experimental research Ahmed determined the flexural strength for ABS test sample and obtained 31.50 MPa Ultimate Flexural Stress at a loading of 58 daN and a Deflection at Ultimate Load of 5.96 mm [11].
Wu determine experimentally that the value for the elastic limit for ABS obtained was 22.9 MPa and the tensile strength for ABS was 27.1 MPa. The tensile properties of 3D-printed ABS test samples were lower than test samples obtained by injection-moulding by 26.2% for the elastic limit and by 26.8% for the tensile strength. The bending strength of ABS determined was 48.6 MPa. The 3D-printed ABS test samples had bending strength and bending modulus reduced by up to 8.2% and 20.8%, respectively, compared with those obtained by injection-moulding [8].

Materials and methods
In order to analyze the behaviour of the 3D printed gripper jaws, based on the presented previous research, jaw from a De-sta-co gripper model De-sta-co 84A3-3300AAAA was 3D printed, Figure 1.
Based on the versatility of the printing process the jaw was designed to clamp a can of soft drink. In the experiments ABS plus P-430-Acrylonitrile-Butadiene-Styrene production-grade thermoplastic was used to 3D print the gripper soft jaws. The mechanical properties of the ABS plus P-430 thermoplastic are presented in Table 1, [12]. The pre-processing of the STL model was realized in the 3D printer Insight command software for selecting the appropriate printing parameters. A view of the tool paths used for 3D printing is presented in Figure 2.
The printer used is Stratasys Fortus 250 mc, that uses Fused Deposition Modelling (FDM) technology to build parts from the bottom up with precisely deposited layers of modelling and support material, with the following features: ABS plus P-430production-grade thermoplastic and soluble support material, build envelope -254 x 254 x 305 mm, three layer thicknesses 0.178, 0.254 and 0.330 mm. After printing the jaws in accordance with the printing parameters mentioned above, the jaws were mounted in a test stand that simulated the loading of one in the gripper.
The acting force is measured with the force transducer and simulates the effect of the clamped part. The elastic deformation was measured by using and displacement transducer.
The measuring equipment presented in Figure 3 consists of: force transducer S9 5kN produced by Hottinger Baldwin Messtechnik; inductive displacement transducer type Wl/10mm -10 mm stroke, produced by Hottinger Baldwin Messtechnik; Spider 8 multi-channel electronic PC measurement unit produced by Hottinger Baldwin Messtechnik; computer.   Figure 4 shows the hysteresis behaviour (force F versus deformation d evolution) during three successive cycles of gripper jaw loading/unloading, with three consecutive elastic mechanical hysteresis loops characterized by relative high deformations. It is obvious that there are big similarities between these loops (shape and size). The analysis of each hysteresis loop produces some interesting results. Figure 5 shows the first mechanical elastic hysteresis loop during an experimental loading/unloading cycle with a force range: 0÷52 N and a deformation range: 0÷7.12 mm.

3.Results and discussions
First important experimental information available here is the area of hysteresis loop (0.02285N·m). This area is exactly the amount of mechanical energy -El lost during a cycle, dissipated because the internal friction (El=0.02285N·m). The input mechanical energy -Ei during the loading process used for deformation (calculated as the area between the red curve and horizontal line -hl on Figure 5, Ei=0.197795 N·m) is bigger than the output mechanical energy -Eo delivered by the gripper jaw during unloading process (calculated as the area between the blue curve and horizontal line hl on Figure 5, Eo=0.174941N·m). The difference between these two energies is evidently the lost mechanical energy El as El=Ei-Eo. Both areas (energies) were calculated here by numerical integration. The energy El can be a cumulative characterization of the behaviour of material and the gripper jaw shape during a loading/unloading cycle. We propose two more appropriate characterizations: -The ratio between the energy El and the variation of force (the amount of mechanical energy lost per unit of force, Elf=4.455·10 -4 Nm/N), with meter as measurement unit.
The ratio between the energy El and the variation of deformation (the amount of mechanical energy lost per unit of deformation, Eld=3.2207Nm/m), with Newton as measurement unit.
Thus Elf does not depend by applied force and Eld does not depend by deformation.
Second important experimental information on Figure 5 is the average bending stiffness of the gripper jaw (depending by material type, material structure, shape and loading) as the slope of the line which produces the best approximation (by computer assisted linear curve fitting, or by least square regression as well) of the evolution of force versus deformation, during loading process.
It is interesting that the loading cycle can be easily approximated with two lines by curve fitting (as Figures 5 and 6 indicate). First line, from linear fit 1 (for a loading range 0÷12N) is defined (with the slope-intercept formula F=k1d+F01 of fitting line) by: Second line, from linear fit 2 (for loading bigger than 12N) is defined (with the slope-intercept formula F=k2d+F02) by: It is also interesting to underline the value of the ratio Eld/Eld=7229.56N/m: it has the same measurement unit as the stiffness and is relative close by k2value. An analysis of the other two hysteresis loops from Figure 4 produces similar results. Figure 8 describes the second hysteresis loop and fitting lines, Figure 9 describes the residual plot of Figure 8.  Figure 11 describes the residual plot of Figure 10.
The main experimental results for each mechanical hysteresis loop are described in Table 2. There are two important observations in Table 2: -There are relative small variations of the stiffness k2. The average value of bending stiffness k2(6767.4 N/m) should be considered as the best option in gripper jaw behaviour characterization.
-There are also relative small variations of the ratio Eld/Elf. As it proved in the last column from Table  2, this ratio seems to be strictly related by the bending stiffness k2 (here with average value of the dimensionless ratio r=(Eld/Elf)/k2=1.0722). The value of the ratio Eld/Elf (having the same unit of measurement as bending stiffness k2) seems to be a type of stiffness involved in gripper jaw characterization(or any other mechanical part tested in the same way). A future approach will be focused on this topic. Also the signification of the ratio r will be privileged in a future research.
In order to characterize the behaviour of the 3D printed gripper jaws a finite element analysis was performed, considering the experimental input force values of 52N and the experimental used measuring scheme.
The definition of ABSplus-P430 material, used by the Stratsys 250 mc printer, in the ANSYS software, a generic acrylonitrile butadiene styrene (ABS) material, from the information database, was selected and the values for ABSplus-P430 were introduced, the main difference was the values of Young module 2200 MPa [12], and flexural modulus is 1.6 GPa.
The loading scheme is presented in Figure 12. The areas represented in blue represent Fixed Support (or recess), and the force of 52 N. was applied at the top of the part in the direction given by the blue arrow (according to the experiment). Figure 13 presents the Equivalent Stress (Von Mises) distribution for a modulus of 1.5 GPa. From the current research the printing parameters have a large impact on the mechanical properties of the 3D printed parts. From the current research the mechanical properties of the 3D printed parts are usually lower by 15-26% from the manufacturer datasheet specification of the material [6,9,10,11]. According to Stratasys the 3D printed prototype has up to 80% of the strength of injection moulded ABS material [12].
Considering the current research on the mechanical properties of the materials FDM, several finite element analysis were done by varying Young modulus in accordance with the experimental results presented in the literature -1.5 GPa ,1.57 GPa, 1.6 GPa, 1.7 GPa, 1.8 GPa, 2 GPa, 2.15 GPa. Figure 14 presents FEA displacement results in the measuring area. The FEA results for 52N load and variation of Young module are presented in Figure 15. The FEA value that is the closest to the experimental values is corresponding to a Young modulus of 1.8 GPa, resulting a deformation of 7.11 mm.

Conclusions
The paper presents the measured values for displacement of 3D printed gripper jawssoft jaws, considering normal working conditions of type of parts. In the performed experiment a modified jaw, in order to clamp a soda can, was 3D printed with 100% infill, from a De-sta-co gripper model De-sta-co 84A3-3300AAAA.
Several loadingunloading cycles were made and the stiffness value was determined. An analysis of the hysteresis loops indicate that loading cycle can be easily approximated with two lines by curve fitting and two values of the stiffness can be calculated, thus proving a nonlinear behaviour of the 3D printed jaws. The area of the hysteresis cycle between the loading curve and the unloading curve represents the mechanical work not recovered from the system.
A finite element analysis was performed considering the experimental loading and support conditions and the Young modulus was varies considering the values proposed in the literature between 1.5 GPa to 2.15 GPa. The values for displacement that are closer to the experimental values are corresponding to a value of 1.8 GPa for Young modulus, value that is 18 % lower that Young modulus for a part obtained by injection moulding.
Printing parameters can be selected in order to maximize the mechanical properties of the parts, but when working with complex parts the distribution of loads and stress varies during the work process. For complex parts FEA is a fast and reliable method for determining the strain and stress but in order to safely design and simulate the behaviour of 3D printed parts, lower values for Young modulus must be considered. From the specialized literature and perform experiments, Young modulus for 3D printed parts when designing a complex parts with complex working conditions, should be 20 -25% lower that the injection moulding.