DEGREE PROJECT TECHNOLOGY, FIRST CYCLE, 15 CREDITS
, STOCKHOLM SWEDEN 2017
Repeatability of Additive
KTH ROYAL INSTITUTE OF TECHNOLOGY
Saab Surveillance in Järfälla constructs complex products, such as radars and electronic sup-port measures. Saab sees an advantage in manufacturing details with additive manufacturing as it enables a high level of complexity. Additive manufacturing is relatively new in the in-dustry and consequently there are uncertainties regarding the process. The purpose of this bachelor thesis was to improve the knowledge of the repeatability of additive manufactured parts as well as compare additive manufactured test rods in two different directions, horizon-tally and vertically, to subtractive manufactured test rods with a vibration test. The vibration test was conducted to simulate the operative environment where the additive manufactured parts might be implemented in the future. Before the vibration test could be performed, the test rods were designed in a 3D-modeling program and analysed with a finite element method to achieve the required natural frequency range of 100 - 200 Hz and a maximal bending stress of 60 - 80 MPa in the notched area of the test rod.
It was concluded that the subtractive manufactured test rods had the highest repeatability. The horizontally additive manufactured test rods had a higher repeatability than the vertically additive manufactured test rods, but the vertically additive manufactured test rods had the highest overall strength. It was also concluded that more studies are needed to ensure that additive manufactured parts can be produced with high repeatability while maintaining the structural integrity.
Keywords: Additive Manufacturing, 3D-printing, Repeatability, Vibration Test, Aluminium
Saab Surveillance i Järfälla konstruerar komplexa försvarsprodukter som till exempel radarsys-tem. Additiv tillverkning i metall möjliggör tillverkning av produkter med hög komplexitet, men då tillverkningsprocessen är relativt ny i industrin finns det en stor osäkerhet kring pro-cessen. Syftet med detta kandidatexamensarbete var att få en bättre förståelse för repeter-barheten hos additivt tillverkade delar samt att jämföra additivt tillverkade provstavar kon-struerade i två olika riktningar, horisontellt och vertikalt, med svarvade provstavar med hjälp av ett vibrationstest. Vibrationstestet genomfördes för att simulera den operativa miljön där de additivt tillverkade detaljerna skulle kunna implementeras i framtiden. Innan vibrationstestet kunde utföras simulerades provstavarnas design i en mjukvara för 3D-modellering. En finit element-analys utfördes även för att få en egenfrekvens inom intervallet 100 - 200 Hz och en maximal böjspänning mellan 60 - 80 MPa i anvisningen på provstaven.
Slutsatsen drogs att de traditionellt bearbetade stavarna hade den högsta repeterbarheten. De horisontellt additivt tillverkade stavarna hade högre repeterbarhet än de vertikalt addi-tivt tillverkade stavarna, men att de vertikalt addiaddi-tivt tillverkade stavarna hade ett längre ut-mattningsliv. Det kunde även konstateras att fler studier inom ämnet behövs för att kunna säkerställa repeterbarheten hos additivt tillverkade delar utan att behöva kompromissa med hållfastheten.
Nyckelord: Additiv tillverkning, 3D-printing, Repeterbarhet, Vibrationstest, Aluminiumlegering
Contents1 Introduction 1 1.1 Background . . . 1 1.2 Aim . . . 1 2 Manufacturing Processes 3 2.1 Additive Manufacturing . . . 3 2.1.1 Design . . . 4
2.1.2 Manufacturing with Selective Laser Melting . . . 5
2.1.3 Material . . . 5
2.1.4 After Treatments . . . 6
2.2 Subtractive Manufacturing . . . 7
3 Mechanics 9 3.1 Fatigue . . . 9
3.2 Analysing Random Vibration Fatigue . . . 9
3.3 Natural Frequency . . . 10
3.4 Finite Element Method . . . 10
4 Method 11 4.1 Software . . . 11
4.1.1 ANSYS . . . 11
4.1.2 Siemens NX . . . 11
4.2 Design of Test Rod and Vibration Plate . . . 11
4.3 Manufacturing . . . 13
4.4 Scanning Electron Microscope . . . 16
4.5 Vibration Test . . . 16
5 Results 19 5.1 Placement During Manufacturing . . . 19
5.2 Composition of Metal Powder . . . 20
5.3 Vibration Test . . . 22
6 Discussion 25 6.1 Sustainability . . . 26
7 Conclusions 29
A Material Data - Aalco Aluminium Alloy 6082 - T6 B Material Data - EOS AlSi10Mg
C Material Data - Test Report of EOS AlSi10Mg D Blueprints of Test Rod and Vibration Plate E Table of Test Rod Data
Saab Surveillance, a business area of Saab, is one of the world’s premier suppliers of airborne, land based and naval radar, electronic support measures and self-protection systems. Their products are performing in harsh operative environments and should be compact and easy to operate, which results in an increased complexity within their portfolio. This report is car-ried out for one of the mechanical departments at Saab Surveillance in Järfälla, where they design enclosures for electronics to protect them from ambient environments, such as rain, temperature and vibration. Their products need to be designed in a way to endure thermal-, environmental- and structural loads as well as being easy to assemble and maintain.
Saab’s products normally have complex designs and therefore Saab sees an advantage in inves-tigating the possibilities of using additive manufacturing. This manufacturing process enables even more complex designs compared to conventional manufacturing, a type being subtrac-tive manufacturing. Improvement of mechanical properties in their enclosures, for example increased flow of cooling fluid, is one potential benefit. Additive manufacturing, commonly known as 3D-printing, with metal is also a relatively new manufacturing process in the in-dustry. Consequently, there are a lot of questions regarding the process, especially with the repeatability of the products. Repeatability is the variation in measurements acquired under the same conditions and is important as it ensures a high predictability of the products’ per-formance. The repeatability of additive manufactured parts needs to be investigated before it can be applied in products with a high safety and security demand, as is the case with Saab’s products.
Chapter 1. Introduction
To achieve this, a sketch of a test rod will be simulated in ANSYS, a structural analysis software, and then modelled in Siemens NX, a 3D-modeling software. A set number of test rods will then be manufactured. The test rods will be subjected to a vibration test in one direction, where the time of fracture will be measured. The purpose of the vibration test is to simulate the operative conditions where additive manufactured parts might be implemented in the future. The repeatability will be analysed using a Weibull distribution graph with the results from the vibration test.
This report will focus on the following:
• The parameters are determined to achieve test rods as consistent as possible for valid comparisons.
• One type of aluminium alloy is considered for each process: Aalco Aluminium alloy 6082 - T6 for the subtractive manufactured test rods and EOS AlSi10Mg for the additive manufactured test rods.
Additive manufacturing is a term which includes all methods that repeatedly adds material together to form the finished product, more commonly known as 3D printing. Additive manu-facturing offers a new way of designing and manumanu-facturing complex details in a more efficient manner compared to more conventional methods such as subtractive manufacturing.  In the industry, additive manufacturing was first used for rapid prototyping as it gave a quicker view of how real products could look before having them produced with more conventional methods. Plastics are mainly used in rapid prototyping as they are time and cost efficient. Met-als, on the other hand, are becoming more frequently applied in the industrial sector due to their benefits with complex designs and a reduced lead time. The automotive, aerospace, tool-ing and medicine industries are some examples of where the additive manufacturtool-ing method with metals is becoming more common. 
There are many possibilities with additive manufacturing, mainly because the process enables very complex designs. For example, a detail with a complex internal structure can be more easily manufactured with additive manufacturing but is nearly impossible to produce with subtractive manufacturing. This is possible with the help of computational optimization, where details can be designed to withstand heavy loads but with less material. 
Chapter 2. Manufacturing Processes
Certain aspects need to be considered when designing a product for the additive manufactur-ing process. The minimum angle that can be self-supportmanufactur-ing is between 30 - 45◦, see Figure 2.1, and varies for different materials. This is because the metal powder needs support during the building process. Aluminium, which is the material taken into consideration in this report, has a minimum angle of 45◦. A support structure is needed if the angle is less than that. 
Complex details are often manufactured together with support structures in order to increase the rigidity of the geometry and to reduce the possibility of collapse due to heat building up during manufacturing. Support structures help disperse the heat that is generated as well as help avoid deformation of the newly built construction. However, an ideal design has no support structures as material waste and energy consumption can be highly reduced during the manufacturing process. If support structures cannot be avoided, they should be minimized. 
FIGURE2.1: The angle that can be self-supporting varies for different materials. 
Another aspect to consider is the ratio between the height and the width, which should be no more than 8:1. The roller moves quickly over the part and therefore the taller the product, the higher the chance for the geometry to collapse due to collision. 
It is possible to create threads with additive manufacturing, but it depends on the thread’s size and direction. In order to verify that the threaded area has the right tolerance, enough space around the thread is also needed.  With today’s additive manufacturing technology, it is recommended that threads, especially smaller threads, are post-processed by machining. During the additive manufacturing process the surface can vary between the upside and the downside, also known as up-skin and down-skin. It is of high importance to distribute the loads between the screw and detail correctly in threads, since differences in surface structure will affect the load distribution. 
2.1. Additive Manufacturing
2.1.2 Manufacturing with Selective Laser Melting
The principles of selective laser melting can be illustrated in Figure 2.2. A layer of metal powder is flattened out by a roller and a laser solidifies the powder, which creates one layer of the product. The building platform is then lowered and the roller spreads a new even layer of metal powder. This process is repeated until the product is finished. 
FIGURE2.2: A principal overview of the selective laser melting method. 
Excessive metal powder is removed from the completed part, which leaves the finished product adhered to the building plate. Because of the high temperatures needed in additive manufac-turing, stress can build up in the produced parts. The outcomes of these stresses could, for example, be delamination of the layers, cracks or distortion in the material. However, this can be prevented with a stress relieving cycle before separating the product from the building plate, to relieve the product from any mechanical stresses that have been built up during the process (see Appendix B). Support structures are removed thereafter, and the finished product can then undergo various treatments to smoothen out the surfaces. 
An important factor to consider with additive manufacturing machines is the placement of the products during manufacturing. The roller might touch the solid layer with force, causing the geometry to collapse. This could happen if, for example, a thin section is placed parallel to the roller, since the roller will very likely hit the section and make the construction fail, see Figure 2.3. The surface needs to be shifted with at least 5◦to withstand any force from the roller. 
Chapter 2. Manufacturing Processes
FIGURE2.3: How a thin section should be placed and not be placed during addi-tive manufacturing. 
for example steel, aluminium and titanium. There is also a large spectrum of alloys which allows for an even greater variation of applications. 
The metal powder used in this report is AlSi10Mg, an aluminium alloy with 10 wt% Si, which has been provided to Lasertech LSH AB by EOS. According to the data sheet (see Appendix C) 85.9% of the particles have a diameter between 25 - 45 µm. The metal particles of AlSi10Mg are melted and re-solidifies during the additive manufacturing process at a very short time which makes the mechanical properties similar to the correspondent material 6082 which are T6 heat-treated (see Appendix B).
2.1.4 After Treatments
The stress relieving cycle, as mentioned above, is a type of heat treatment where details in AlSi10Mg are annealed in a furnace at 300◦C for two hours. As the built-in stresses are re-duced with this treatment, a certain anisotropy from the layer-wise building process is also reduced. This can be seen under "Mechanical properties of the parts" in Appendix B where the yield strength in both the horizontal and vertical direction have the same value after the heat treatment.
2.2. Subtractive Manufacturing
Cutting processing, also known as subtractive manufacturing, is the most conventional man-ufacturing method and includes all types of machining where raw material is removed dur-ing controlled forms to achieve a desired shape of an object. Metal products are very often produced with machining, but plastic, wood, ceramics and composites products can also be manufactured with this method. 
The most common cutting processes are turning, milling and drilling, see Figure 2.4. Turning operations are normally performed with a lathe, where the workpiece is rotated around its axis and machines tools are applied to the workpiece, creating a symmetric cut around the object’s axis. Milling and drilling operations are performed by milling machines and drill presses re-spectively, where the machine tools are rotating instead of the workpiece. Drilling operations could also be achieved in lathes or mills in order to be more time efficient by not having to move the workpiece to a drill machine. 
FIGURE 2.4: An overview of the most common cutting processes. From left to right: turning, drilling and milling. 
The surface roughness of machined objects can vary from Ra > 10µmfor rough machining to Ra<0.2 µm for polished surfaces. For machined products, an average surface finish of Ra0.4 - 1.6 µm is normally achieved. 
One of the most common causes for breakdowns and fracture in constructions is fatigue failure. Even way below its tensile strength, repeatedly cycling with change of loads over time can generate a fracture. Due to high concentrations in stress, microcracks are initiated and will over time slowly propagate into larger cracks. The construction will fail when these cracks reach a critical point, causing either a ductile or brittle final fracture. 
It is important to test the fatigue life of materials and is normally done with cyclic stress. The results can be plotted with a logarithmic Stress-Number diagram, commonly known as a Wöh-ler curve, which shows the stress levels and the number of cycles to failure on the material. This curve gives an indication that the fatigue life is shorter for a metal with a higher level of amplitude cycles compared to a lower level of amplitude cycles. 
The accumulated damage of a part from various cyclic loads can be calculated using Palmgren-Miners’ rule. The rule states that a summation of all the partial damages from different stress cycles can be done to estimate the total damage on the part which is useful for parts exposed to, for example, random vibration. 
Analysing Random Vibration Fatigue
It is relatively simple to decide a part’s fatigue under periodic vibration, due to the vibration being greatly predictable at any point in time. On the other hand, determining the damage of a part’s fatigue under random vibration is altogether much more difficult; the vibration is vastly unpredictable and there is no evident pattern for the stress amplitudes. However, analysing random vibration fatigue is crucial in many applications, especially in the aeronautic industry. 
Chapter 3. Mechanics
to this being a simplification of the random vibration, data and information are lost. However, PSD gives a very good indication of the vibration and an estimate of the fatigue life. 
All linear elastic mechanical systems have a natural frequency when vibrated freely, but the vibrations must first be initiated by an external disturbance. There is an infinite number of nat-ural frequencies if the system is continuous, but in most technical applications only the lowest frequencies are normally of interest. A system can, through appropriate models and simplifi-cations, be described with a determined number of natural frequencies. Forced vibration, in comparison to natural vibration, occurs when an outer mechanical load is applied.  If the natural and forced frequencies coincide then the system’s oscillating amplitude could greatly increase, which is also known as resonance. 
An important aspect to consider during examination of vibration and natural frequency is damping, which is always present in oscillating systems and affects the amplitude of the oscil-lation. For low levels of damping, the effects of the natural frequencies will also be low.  A damping ratio at nearly 0% means that the system would continue oscillating to nearly infinity while a high level of damping would decrease the amplitude of the oscillation greatly.
Finite Element Method
The following subsections will describe the software that have been used in the method of this report.
ANSYS is an engineering software company which has many different simulation programs depending on the application. The simulation software used in this report is ANSYS Mechan-ical, a structural analysis software. This software has many finite element analysis tools with which the simulations can becomes more accurate and effective. The simulations include static, dynamic and thermal problems.  ANSYS DesignModeler and ANSYS Workbench, two ap-plications in ANSYS Mechanical, are used during the analysis of the test rods in this report.
4.1.2 Siemens NX
Computer Aided Design, also known as CAD, is a software system for constructional design. CAD enables and simplifies almost all desired properties of a product for documentation and manufacturing purposes, including exact measurements, automated calculations and generat-ing tool paths for NC machines. The most commonly known modern 3D CAD-systems are NX, Solid Edge, Creo, Inventor, Catia and SolidWorks.  In this report, Siemens NX is used.
Design of Test Rod and Vibration Plate
Chapter 4. Method
The PSD was simulated and can be seen in Table 4.1, where the damping ratio was set to 2.5%.
TABLE 4.1: Simulated power spectral density (PSD). The damping ratio was set to 2.5%. Frequency [Hz] PSD [G2/Hz] 10 0.035 20 0.07 1000 0.07 2000 0.035
The analysis was then made with ANSYS Workbench where multiple lengths and notch ra-diuses were evaluated. The analysis was completed when the natural frequency was in the range of 100 - 200 Hz and the maximal bending stress in the notch, 60 - 80 MPa, had been achieved.
A 3D-model was then created in NX, with the results from the ANSYS analysis, and exported to ANSYS in order to give an estimate on how well the design achieved the required values of natural frequency and maximal bending stress in the notch.
Some minor adjustments were done with the 3D-model to fulfil necessary manufacturing pa-rameters, for example, the 45◦ minimum self-supporting building angle and the minimum notch radius of 0.7 mm.
A vibration plate was designed to match the vibrator used in the test as well as the rod’s attach-ment to the plate. The threaded holes of the test rods were placed to ensure that the test rods would have enough vibration space. The total weight limit of the vibrator was also considered in this process, which was 40 kg.
FIGURE4.2: Final NX design of the vibration plate, in mm.
A final design of the test rod was completed in NX and the mechanical requirements were confirmed in ANSYS. All important parameters, measurements and design requirements were examined, including the total weight of all 30 rods and the test plate. The final NX design of the test rod and the vibration plate can be seen in Figure 4.1 and Figure 4.2 respectively.
After the required mechanical properties and the additive manufacturing design parameters of the 3D-model were achieved, blueprints with measurements, tolerances and other instruc-tions were created and sent to manufacturing (see Appendix D). Twenty test rods were created with additive manufacturing by Lasertech LSH AB in Karlskoga, where 10 of those were man-ufactured horizontally and 10 vertically. The additive manman-ufactured test rods were processed with EOS M 290, a standard selective laser melting machine, and was heat treated afterwards at 300◦C for two hours.
Ten test rods were also manufactured from aluminium bars by Alumbra AB in Järfälla. These test rods were lathed using a NC machine. All the test rods, including the additive manu-factured rods, were threaded and notched by Alumbra. This was done in order to achieve as consistent notches on all the test rods, due to the notch being an important parameter in this report.
Chapter 4. Method
FIGURE4.3: Left to right: horizontally additive manufactured, vertically additive
manufactured and subtractive manufactured. The leftover support structures can be seen on the two additive manufactured test rods.
FIGURE 4.5: The vertically additive manufactured test rods, a total of 9 as one collapsed during manufacturing.
Chapter 4. Method
Scanning Electron Microscope
A Scanning Electron Microscope (SEM), which is a type of electron microscope, was used to analyse the metal powder’s particle size and composition in order to understand the mate-rial and compare the accuracy of the given matemate-rial data sheet of the metal powder used by Lasertech (see Appendix C). The sample was prepared by covering the sample holder’s carbon tape with a small amount of the powder and then using compressed air to press the powder to the holder and to remove any loose material. The sample was then placed in the SEM for anal-ysis. A few particles were measured to give an estimate of the overall amount of each particle size. Mapping was done afterwards on larger areas to analyse the composition of the powder.
Each individual test rod was weighed and data was gathered in a table (see Appendix E). The vibration test plate was mounted to the vibrator, see Figure 4.7. All test rods were then attached to the test plate by first applying grease to the threads and then using an adjustable spanner and a torque wrench to achieve the same torque, 16 N m, on all the test rods, see Figure 4.8.
FIGURE4.7: The vibration plate mounted on the vibrator.
4.5. Vibration Test
FIGURE4.8: Arbitrarily mounting the test rods to the vibration plate.
FIGURE4.9: Everything mounted before the vibration test.
Chapter 4. Method
Placement During Manufacturing
The placement of the additive manufactured test rods can be seen in Figure 5.1. During the manufacturing of the vertically printed test rods, the construction failed twice. All of the test rods collapsed during the first round and one test rod collapsed during the second round, leaving a total of 9 vertically printed test rods instead of 10. During the first round the test rods were placed with the threaded side down and the point mass up. One of the test rods collapsed and subsequently caused all of the other ones to collapse. Hence, the placement was changed in between the two construction rounds, showing the final placement in Figure 5.1b.
(A) The horizontally additive manufactured rods. (B) The vertically additive manufactured rods.
FIGURE5.1: The placement of the additive manufactured test rods on the build-ing plates before manufacturbuild-ing, illustrated from above. The crossed-out parts
Chapter 5. Results
Composition of Metal Powder
The overall particle size in the metal powder compared to the given metal powder data sheet (see Appendix C) that was used in the production of the additive manufactured test rods seemed to be very similar, where 85.9% of all particles’ size were 25-45 µm according to the data sheet. An overview of the powder and the measurement of a particle’s size can be seen in Figures 5.2 and 5.3 respectively.
The mapping of the metal powder can be seen in Figure 5.4 and a comparison with the given material data sheet can be seen in Table 5.1.
FIGURE5.2: Overview of the metal powder.
5.2. Composition of Metal Powder
FIGURE5.4: Mapping of the larger area above using an energy dispersive spec-troscopy (EDS).
TABLE5.1: The result from mapping of a larger area (see Figure 5.4) compared to the given material data sheet (see Appendix C).
Element Result from EDS [wt%] Material Data Sheet [wt%]
Chapter 5. Results
The results from the vibration test can be seen in Figure 5.5, showing the number of failed test rods over time (see Appendix F for tabulated data). Test rod No 9, horizontally printed, was excluded from the diagram due to its abnormal result. The total vibrational time was one hour and 54 minutes at a level of -9 dB. The data was further analysed by adapting a Weibull distribution graph, which shows the probability of a test rod to fracture at a certain time, see Figure 5.6.
The accelerometer on test rod Nos 19 and 28 fell off in the beginning of the vibration test, but the accelerometer on test rod No 9 was attached for one hour and 13 minutes. Test rod No 18, vertically printed, was deformed during the mounting of the test rods to the vibration plate, see Figure 5.7.
FIGURE5.5: Result from the vibration test, excluding test rod No 9. AM (additive manufactured) and SM (subtractive manufacture). See Appendix F for tabulated
5.3. Vibration Test
FIGURE5.6: A Weibull distribution graph of the fractured test rods. AM (additive manufactured) and SM (subtractive manufacture)
The analysis of the Weibull distribution graph, see Figure 5.6, indicates that the subtractive manufactured test rods have the highest repeatability. This is due to the probability for a test rod to fracture at a certain time being higher with a smaller time interval. The vertically addi-tive manufactured test rods shows the highest overall strength to the vibrations, but the lowest repeatability. Even though the strength is high, high repeatability is normally preferred in most applications due to a higher predictability.
The horizontally additive manufactured test rods were all roughly faced the same direction. It is possible that these test rods had anisotropic properties as a result of being manufactured layer-wise even though they had been heat treated afterwards, but no correlation between the results of the vibration test and the vibrational direction can be concluded. This would, how-ever, be an interesting aspect to consider in future studies. The vertically additive manufac-tured test rods had the highest overall strength, as they lasted the longest.
There were two test rods that stood out among the samples: test rods Nos 9 and 18. Test rod No 9 lasted the longest before fracture at nearly double the time compared to the majority of the other test rods. This test rod also had the accelerometer attached to itself during most of the testing time, compared to test rods Nos 19 and 28 that also had accelerometers attached in the beginning of the vibration test. It is very likely that this affected the already low damping and it can therefore be concluded that the damping affected all the test rods. The other abnormal sample was test rod No 18 that was bent during mounting. This could potentially have affected the strength positively due to deformation hardening in the material, but this was not further investigated.
Chapter 6. Discussion
It can be determined from the SEM analysis that the metal powder mostly matched the mate-rial data from EOS regarding particle size and composition, save a small amount of Ag in the batch. The material data comes from the original powder batch, but during the additive man-ufacturing process most of the powder is reused multiple times. It is possible that the small amount of Ag comes from another batch or that there has been an energy overlap in the EDS spectra. However, this should not have affected the results from the vibration test.
It was difficult to determine the same Ra-factor for the subtractive manufactured test rods and the additive manufactured test rods, due to manufacturing differences. An average surface roughness for additive manufactured products is approximately Ra6 - 10 µm, but this was not a confirmed parameter in this study due to lack of accessible equipment. The surface rough-ness can be minimized with after treatments, for example sandblasting, polishing, peening or subtractive manufacturing. However, this was not a focus in this report and thus no after treat-ments were done to the surface of the additive manufactured test rods except to the notch and the thread of each rod.
It was important to determine the parameters, which is why all of the test rods’ threading and notching were post-processed by Alumbra. If these parameters are conducted by the same manufacturer, the variations in tools and machines will be mostly eliminated. This will, for example, give a higher possibility for all of the test rods to have the same placement and Ra value in the notch. The coarse surface from the additive manufacturing process, with up-skin and down-skin surfaces, was also avoided with post-processing. It is crucial for threads to have an even surface on both sides in order to distribute to loads evenly.
There are a lot of benefits using additive manufacturing through a sustainable perspective, the main one being that additive manufacturing enables complex internal designs of products to be more efficiently constructed with less material. Thus, the energy cost for both raw and support structure material can be reduced. A more complex geometry costs neither more time nor more money with additive manufacturing, compared to subtractive manufacturing. Nevertheless, additive manufacturing also requires high levels of energy and is relatively expensive.
Sources of Error
The main source of error is that too few test rods of each kind were manufactured, which made it difficult to draw valuable conclusions. As there were only 9 vertically additive manufactured test rods, 9 horizontally additive manufactured test rods and 10 subtractive manufactured test rods tested in total, not enough data was gather in order to be conclusive.
6.2. Sources of Error
to give a good implication on the results. Thus, this should not have greatly affected the final results.
During the preparation of the vibration test, the accelerometers’ results determined the vibra-tion level of -9 dB of the test. The placement of the accelerometers on the test rods greatly affects the achieved responses, as the values were used in the calculation of the material damping and gave an estimation of the starting vibration level. The results would have been remarkably different if the accelerometers were to be placed at slightly different positions.
There is a possibility that the point mass of the test rods contributed to a circular motion during the vibration test instead of the test rods vibrating in one direction. This could have lowered the maximal bending stress in the notch of the test rods, but was not further studied.
The purpose of this report was to improve the knowledge of the repeatability of additive man-ufactured parts for Saab as well as compare additive manman-ufactured test rods printed in two different directions, horizontally and vertically, to subtractive manufactured test rods after a vibration test. Through the results of the vibration test, the following conclusions were made:
• The subtractive manufactured test rods had the highest repeatability.
• The horizontally additive manufactured test rods had a higher repeatability than the ver-tically additive manufactured test rods.
• The vertically additive manufactured test rods had the highest overall strength.
We would like to give a warm thank you to our supervisors at Saab Surveillance in Järfälla, Molly Ericson and Anders Molin, for their wonderful support, knowledge and for inspiring us on our future career paths. We have learned a lot during this semester and truly appreciate all of your help.
We would also like to thank Peter Hedström, our supervisor at KTH, for his great support and feedback; Jan Gilldorf and Conny Svensson, material experts at Saab, for helping us with our material analysis; and Mauricio Saldes, Mechanical Engineering Manager at EW Systems, for providing us with this opportunity.
 K. Maxey, “Metal Sintering Meets Industrial Needs With the EOS
M 290”, Engineering.com. Internet: http://www.engineering.
Jun. 2014 [2017-04-25].
 S. S. Muthu and M. M. Savalani, Handbook of Sustainability in Additive Manufacturing, Springer, ISBN: 978-981-10-0547-3.
 CrucibleDesign, “Design Guidelines For Direct Metal Laser Sintering (DMLS)”, 2014, Inter-net: https://www.crucibledesign.co.uk/guides/ [2017-05-03].
 Interview with Karolina Johansson, AM-engineer at Lasertech LSH AB, May 2017.  EOS, “EOS M 290”, Internet: https://www.eos.%info/eos-m290 [2017-05-05].
 EOS, “Additive Manufacturing, Laser-Sintering and Industrial 3D Printing - Bene-fits and Functional Principle”, https://www.eos.info/additive_manufacturing/
 EOS, “Materials For Metal Additive Manufacturing”, Internet: https://www.eos.
 Höganäs AB, “Additive Manufacturing of Small and Complex Metal Parts”, Mar.
2016. Internet: https://www.hoganas.com/globalassets/media/documents/
 A. E. W. Jarfors, T. Carlsson, A. Eliasson, H. Keife, C.-M. Nicolescu, B. Rundqvist, M. Be-jhem and B. Sandberg, Tillverkningsteknologi. Studentlitteratur AB, 2010, vol. 4, pp. 77–79, 297–299, 300, 351, 364, 370, ISBN: 978-91-44-07039-1.
 H. Lundh, Grundläggande hållfasthetslära. Stockholm: KTH, Hålfasthetslära, 2000, pp. 243–244, ISBN: 978-91-972860-2-2.
 A. Ekberg, “Damage Accumulation”, Am.chalmers.se, Dec. 1998. Internet: http://www.
 B. Sundström, Elementär strukturanalys - svängningar i diskreta system, 2nd ed. KTH, Hållfasthetslära, 1988, pp. 133–136.
 J. S. Nilsson, T. Eriksson, and K.-O. Olsson, “Nationalencyklopedin, resonans.”, Internet:
 G.R. Liu and S. S Quek, The Finite Element Method A Practical Course, 2nd. Burlington : Elsevier Science, 2014.
 ANSYS, “Simulation Software Products”, Internet: http://www.ansys.com/
Aluminium alloy 6082 is a medium strength alloy with excellent corrosion resistance. It has the highest strength of the 6000 series alloys. Alloy 6082 is known as a structural alloy. In plate form, 6082 is the alloy most commonly used for machining. As a relatively new alloy, the higher strength of 6082 has seen it replace 6061 in many applications. The addition of a large amount of manganese controls the grain structure which in turn results in a stronger alloy. It is difficult to produce thin walled, complicated extrusion shapes in alloy 6082. The extruded surface finish is not as smooth as other similar strength alloys in the 6000 series.
In the T6 and T651 temper, alloy 6082 machines well and produces tight coils of swarf when chip breakers are used.
6082 is typically used in: ~ Highly stressed applications ~ Trusses ~ Bridges ~ Cranes ~ Transport applications ~ Ore skips ~ Beer barrels ~ Milk churns CHEMICAL COMPOSITION BS EN 573-3:2009 Alloy 6082 Element % Present Silicon (Si) 0.70 - 1.30 Magnesium (Mg) 0.60 - 1.20 Manganese (Mn) 0.40 - 1.00 Iron (Fe) 0.0 - 0.50 Chromium (Cr) 0.0 - 0.25 Zinc (Zn) 0.0 - 0.20 Others (Total) 0.0 - 0.15 Titanium (Ti) 0.0 - 0.10 Copper (Cu) 0.0 - 0.10 Other (Each) 0.0 - 0.05 Aluminium (Al) Balance
Aluminium alloy 6082 also corresponds to the following standard designations and specifications but may not be a direct equivalent: AA6082 HE30 DIN 3.2315 EN AW-6082 ISO: Al Si1MgMn A96082 TEMPER TYPES
The most common tempers for 6082 aluminium are: T6 - Solution heat treated and artificially aged •
O - Soft •
T4 - Solution heat treated and naturaly aged to a substantially stable condition
T651 - Solution heat treated, stress relieved by stretching then artificially aged
Alloy 6082 T6 & T651 is typically supplied as Plate and Shate.
GENERIC PHYSICAL PROPERTIES
Density 2.70 g/cm³ Melting Point 555 °C Thermal Expansion 24 x10-6 /K Modulus of Elasticity 70 GPa Thermal Conductivity 180 W/m.K
Electrical Resistivity 0.038 x10-6 Ω .m
BS EN 485-2:2008 Plate
6.00m to 12.5mm
Proof Stress 255 Min MPa Tensile Strength 300 Min MPa Elongation A50 mm 9 Min %
Hardness Brinell 91 HB
Properties above are for material in the T6 and T651 condition
BS EN 485-2:2008 Plate
12.5mm to 100.00mm
Proof Stress 240 Min MPa Tensile Strength 295 Min MPa Hardness Brinell 89 HB
Properties above are for material in the T6 and T651 condition
BS EN 485-2:2008 Plate
100.00mm to 150.00mm
Proof Stress 240 Min MPa Tensile Strength 275 Min MPa Hardness Brinell 84 HB
Elongation A 6 Min %
Properties above are for material in the T6 and T651 condition
6082 has very good weldability but strength is lowered in the weld zone. When welded to itself, alloy 4043 wire is recommended. If welding 6082 to 7005, then the wire used should be alloy 5356.
Weldability – Gas: Good Weldability – Arc: Good Weldability – Resistance: Good Brazability: Good
Workability - Cold: Good Machinability: Good
Address: Please make contact directly with your local service centre, which can be found via the Locations page of our web site
Datasheet Updated 11 January 2016
This Data is indicative only and as such is not to be relied upon in place of the full specification. In particular, mechanical property requirements vary widely with temper, product and product dimensions. All information is based on our present knowledge and is given in good faith. No liability will be accepted by the Company in respect of any action taken by any third party in reliance thereon.
Please note that the 'Datasheet Update' date shown above is no guarantee of accuracy or whether the datasheet is up to date.
The information provided in this datasheet has been drawn from various recognised sources, including EN Standards, recognised industry references (printed & online) and manufacturers’ data. No guarantee is given that the information is from the latest issue of those sources or about the accuracy of those sources.
Material supplied by the Company may vary significantly from this data, but will conform to all relevant and applicable standards.
As the products detailed may be used for a wide variety of purposes and as the Company has no control over their use; the Company specifically excludes all conditions or warranties expressed or implied by statute or otherwise as to dimensions, properties and/or fitness for any particular purpose, whether expressed or implied.
Advice given by the Company to any third party is given for that party’s assistance only and without liability on the part of the Company. All transactions are subject to the Company’s current Conditions of Sale. The extent of the Company’s liabilities to any customer is clearly set out in those Conditions; a copy of which is available on request.
6082 - T6~T651 Plate
Material data sheet
EOS GmbH - Electro Optical Systems
EOS Aluminium AlSi10Mg
EOS Aluminium AlSi10Mg is an aluminium alloy in fine powder form which has been specially
optimised for processing on EOSINT M systems
This document provides information and data for parts built using EOS Aluminium AlSi10Mg
powder (EOS art.-no. 9011-0024) on the following system specifications:
- EOSINT M 280
with PSW 3.6 and Original EOS Parameterset AlSi10Mg_Speed 1.0
- EOS M 290 400Watt
with EOSPRINT 1.0 and Original EOS Parameterset AlSi10Mg_Speed 1.0
AlSi10Mg is a typical casting alloy with good casting properties and is typically used for cast
parts with thin walls and complex geometry. It offers good strength, hardness and dynamic
properties and is therefore also used for parts subject to high loads. Parts in EOS Aluminium
AlSi10Mg are ideal for applications which require a combination of good thermal properties and
low weight. They can be machined, spark-eroded, welded, micro shot-peened, polished and
coated if required.
Material data sheet
EOS GmbH - Electro Optical Systems
General process and geometrical data
Typical achievable part accuracy   100 µm
Smallest wall thickness   approx. 0.3 – 0.4 mm
approx. 0.012 – 0.016 inch Surface roughness, as built, cleaned   Ra 6 - 10 µm, Rz 30 - 40 µm
Ra 0.24 - 0.39 x 10-³ inch
Rz 1.18 - 1.57 x 10-³ inch
- after micro shot-peening Ra 7 - 10 µm, Rz 50 - 60 µm
Ra 0.28 - 0.39 x 10-³ inch
Rz 1.97 - 2.36 x 10-³ inch
Volume rate  7.4 mm³/s (26.6 cm³/h)
 These properties were determined on an EOSINT M 270.
 Based on users' experience of dimensional accuracy for typical geometries. Part accuracy is subject to appro-priate data preparation and post-processing, in accordance with EOS training.
 Mechanical stability dependent on the geometry (wall height etc.) and application
 Due to the layerwise building, the surface structure depends strongly on the orientation of the surface, for example sloping and curved surfaces exhibit a stair-step effect. The values also depend on the measurement method used. The values quoted here given an indication of what can be expected for horizontal (up-facing) or vertical surfaces.
Material data sheet
Physical and chemical properties of the parts
Material composition Al (balance)
Si (9.0 - 11.0 wt-%) Fe ( 0.55 wt-%) Cu ( 0.05 wt-%) Mn ( 0.45 wt-%) Mg (0.2 - 0.45 wt-%) Ni ( 0.05 wt-%) Zn ( 0.10 wt-%) Pb ( 0.05 wt-%) Sn (. 0.05 wt-%) Ti ( 0.15 wt-%)
Relative density approx. 99.85 %
Density 2.67 g/cm³
Material data sheet
EOS GmbH - Electro Optical Systems
Mechanical properties of the parts
As built Heat treated 
Tensile strength 
- in horizontal direction (XY) 460 20 MPa
66.7 2.9 ksi 345 10 MPA 50.0 1.5 ksi
- in vertical direction (Z) 460 20 MPa
66.7 2.9 ksi
350 10 MPa 50.8 1.5 ksi Yield strength (Rp 0.2 %) 
- in horizontal direction (XY) 270 10 MPa
39.2 1.5 ksi 230 15 MPa 33.4 2.2 ksi
- in vertical direction (Z) 240 10 MPa
34.8 1.5 ksi
230 15 MPa 33.4 2.2 ksi Modulus of elasticity
- in horizontal direction (XY) 75 10 GPa
10.9 0.7 Msi 10.2 0.7 Msi 70 10 GPa
- in vertical direction (Z) 70 10 GPa
10.2 0.7 Msi 8.7 0.7 Msi 60 10 GPa Elongation at break 
- in horizontal direction (XY) (9 2) % 12 2%
- in vertical direction (Z) (6 2) % 11 2%
Hardness  approx.119 5 HBW
Fatigue strength  
- in vertical direction (Z) approx. 97 7 MPa
approx. 14.1 1.0 ksi
 Mechanical strength tested as per ISO 6892-1:2009 (B) annex D, proportional specimens, specimen diameter 5 mm, original gauge length 25 mm (1 inch).
 Hardness test in accordance with Brinell (HBW 2.5/62.5) as per DIN EN ISO 6506-1. Note that measured hard-ness can vary significantly depending on how the specimen has been prepared.
 Fatigue test with test frequency of 50 Hz, R = -1, measurement stopped on reaching 5 million cycles without fracture.
 Stress relieve: anneal for 2 h at 300 °C (572 °F).
Material data sheet
Thermal properties of parts
As built  Heat treated  
Thermal conductivity (at 20 °C)
- in horizontal direction (XY) approx. 103 5 W/m°C approx. 173 10 W/m°C
- in vertical direction (Z) approx. 119 5 W/m°C approx. 173 10 W/m°C
Specific heat capacity
- in horizontal direction (XY) approx. 920 50 J/kg°C approx. 890 50 J/kg°C - in vertical direction (Z) approx. 910 50 J/kg°C approx. 890 50 J/kg°C
The data are valid for the combinations of powder material, machine and parameter sets referred to on page 1, when used in accordance with the relevant Operating Instructions (including Installation Requirements and Maintenance) and Parameter Sheet. Part properties are measured using defined test procedures. Further details of the test procedures used by EOS are available on request.
The data correspond to our knowledge and experience at the time of publication. They do not on their own provide a sufficient basis for designing parts. Neither do they provide any agreement or guarantee about the specific properties of a part or the suitability of a part for a specific application. The producer or the purchaser of a part is responsible for checking the properties and the suitability of a part for a particular application. This also applies regarding any rights of protection as well as laws and regulations. The data are subject to change without notice as part of EOS' continuous development and improvement processes.
EOS, EOSINT and DMLS are registered trademarks of EOS GmbH.
Appendix E. Table of Test Rod Data
TABLE E.1: Data of the test rods. Test rod Nos 1-10 are horizontally additive
manufactured, Nos 11-19 are vertically additive manufactured and Nos 21-30 are subtractive manufactured. No 20 does not exist.
Test Rod Weight [g] Diameter [mm]
VIBRATION TEST Time of
Horizontally Additive Manufactured Test Rods Vertically Additive Manufactured Test Rods Subtractive Manufactured Test Rods