3D Printing as an Alternative Manufacturing Method for the Microgas Turbine Heat Exchanger

(1)

3D Printing as an Alternative Manufacturing

Method for the Microgas Turbine Heat Exchanger

Wolfgang Seiya and Sherry Zhang

July 2015

Department of Energy Technology Royal Institute of Technology Stockholm, Sweden Pratt School of Engineering, Smarthome Program Duke University Durham, North Carolina, USA

(2)

Acknowledgements

We would like to thank InnoEnergy, Compower, and the ‘‘STandUP for Energy’’ project for providing resourceful background information for this study. We owe our deepest gratitude to our advisor, Anders Malmquist for his continuous support for this study throughout the summer. His guidance, motivation, and expertise were invaluable assets in all areas of the study. Our sincere thanks goes to Joachim Claesson at KTH for his time and knowledge on the subject of heat exchangers. We take this opportunity to thank all the company correspondents that took their time and interest to help us with the vast information needed for this study. Lastly, we are immensely grateful to Duke Smart Home Program and its director Jim Gaston for providing us with the opportunity and necessary funds to live in Sweden while conducting this study.

(3)

I.

List of Figures ……….… 6 List of Tables ……….. 7 Abbreviations and Equation Nomenclature ……… 8 Abstract ………...… 9 Introduction ………. 9 Methodology ……….. 10 Materials ……….. 10 Manufacturing ……….. 13 Materials ……….... 14 Material Overview ………... 15 Silicon Carbide ……….... 18 i. Creep and Oxidation ………... 18 ii. Fatigue Toughness ………. 18 iii. Crack Healing ………... 18 iv. Compatibility with 3D Printing ……… 19 v. Sustainability ………. 19 Inconel alloys ……….. 19 i. Impact Strength ……….. 19 ii. Fatigue Strength ……… 20 iii. Creep and Rupture Properties ……….. 20 iv. Corrosion resistance ………. 20 v. Compatibility with 3D printing ………. 21 vi. Sustainability ……… 21 Haynes 214 Alloy ……… 21 i. Oxidation and Creep ……….…….. 21 ii. Cracking ……….……… 21 iii. Compatibility with 3D Printing ………....… 22 iv. Sustainability ……….... 22 Stainless Steel 304 ………... 22 i. Creep and CreepFatigue ……… 22 ii. Oxidation ………... 23 iii. Compatibility with 3D Printing ……… 23 iv. Sustainability ……… 23 Evaluating hightemperature thermal and tensile properties ………... 24 Ranking results ……….... 27 3D Printing / Additive Manufacturing ……….. 28 Manufacturing Techniques Overview ……….… 29

(4)

PowderBased Fusion Processes (PBF) ……….. 30 i. SolidState Sintering ……….………. 31 ii. ChemicallyInduced Sintering ……….. 32 iii. LiquidPhase Sintering (LPS) ……….. 32 iv. Full Melting ……….. 32 v. Electron Beam / Electron Beam Melting (EB / EBM) ……….. 32 vi. Sustainability ……… 33 Binder Jetting Additive Manufacturing (BJAM) ……….... 33 Laminated Object Manufacturing (LOM) ………... 33 Electron Beam Additive Manufacturing (EBAM) ………..… 34 Ultrasonic Consolidation (UC) ……….... 35 Ranking Techniques ……… 36 Companies / Researchers ………. 37 Arcam AB ………. 38 Aurora Labs 3D ………. 38 Ceralink ………. 39 EOS ….……….. 39 ExOne ………....……… 39 Fabrisonic ……….. 40 Renishaw ………...…… 40 Sciaky ……….... 40 Other Companies ………..…. 40 Growth ……….… 42 Heat Exchanger Design ……….… 43 Classification of Heat Exchangers ……….. 43 i. From Construction ………. 43 ii. Based on the Heat Transfer Process ………. 45 iii. Based on Surface Compactness ………...… 45 iv. Based on Flow Arrangements ……….….… 45 Thermal Hydraulic Performance of Heat Exchangers ……… 46 i. Heat Transfer Mechanisms in Heat Exchangers ……… 46 ii. Conduction ……….... 46 iii. Convection ………...… 46 iv. Heat Transfer and Temperature Difference ……….… 47 Pressure Drops ……… 49 Effectiveness and Rating ……….... 50 Design Recommendations ……….. 52 Economics ……….... 54 Barriers ………. 57

(5)

Recycling ……….……….… 57 Discussion and Recommendations ………...…… 58 Sensitivity Analysis ……….. 59 Conclusion ……… 61 References ……… 62 Appendix ………..… 69

(6)

II. List of Figures

Figure 1 AspenPlus model of Compower’s ET10 ………...….. Figure 2 Ashby’s Material and Selection Chart ……….... Figure 3 Internal oxidation attack in Inconel 601 ………. Figure 4 Thermal conductivities of the top five materials with temperature ……….…... Figure 5 Specific heat capacities measured from a reference of 20°C ……….. Figure 6 Coefficient of linear expansion of the top five materials with temperature ….... Figure 7 Yield strengths of top 5 materials with temperature ………... Figure 8 Elastic modulus of top five materials with temperature ………. Figure 9 Schematic of selectivelaser sintering ………. Figure 10 Schematic of solidstate sintering ………. Figure 11 The laminated object manufacturing process ………... Figure 12 Schematic of the ultrasonic consolidation process ………... Figure 13 Different types of tubular heat exchangers ………... Figure 14a Different types of plate heat exchangers ………. Figure 14b Different types of plate heat exchangers ………. Figure 15 Parallel, counter and cross flow configurations ……….... Figure 16 LMTD Method for parallel and counterflow heat exchangers ………. Figure 17 Temperature profile along a crossflow heat exchanger ……….. Figure 18 Brayton cycle ………... Figure 19 Plot of overall performance coefficient P vs. relevance factor α ……….. Figure 20 Cost in US dollars of commercial computers from 1950 to 1985 ………. Figure 21 Price index of personal computer and peripheral equipment ……….... pg. 10 pg. 15 pg. 21 pg. 24 pg. 24 pg. 25 pg. 26 pg. 26 pg. 31 pg. 32 pg. 34 pg. 35 pg. 44 pg. 44 pg. 45 pg. 45 pg. 48 pg. 48 pg. 49 pg. 52 pg. 54 pg. 55

(7)

III. List of Tables

Table 1 Weights for each criteria used to rank the materials ………... Table 2 Ranking criteria for top five materials and their respective weights ………….. Table 3 Weights for each criteria used to rank manufacturing techniques ……….. Table 4 All the materials initially considered for the heat exchanger design ………….. Table 5 The top 22 materials and their material properties ………. Table 6 Ranking results for the top five materials ……….. Table 7 Overview of additive manufacturing techniques compatible with plastics, metals, and ceramics ……….. Table 8 Ranking and totals for each manufacturing technique ………... Table 9 Machine data from various companies ……….. pg. 11 pg. 12 pg. 14 pg. 15 pg. 1617 pg. 28 pg. 30 pg. 37 pg. 41

(8)

IV. Abbreviations and Equation Nomenclature

Abbreviations ABCSiC = AluminumBoronCarbonSilicon Carbide AM = Additive Manufacturing BJAM = Binder Jetting Additive Manufacturing BJT = Binder Jetting Technology CAD = ComputerAided Design CAGR = Compound Annual Growth Rate CNC = Computer Numerical Control DMLM = Direct Metal Laser Melting DMLS = Direct Metal Laser Sintering EBAM = Electron Beam Additive Manufacturing EB / EBM = Electron Beam / Electron Beam Melting HX= Heat Exchanger LMTD = Log Mean Temperature Difference LOM = Laminated Object Manufacturing LPS = LiquidPhase Sintering PBF = Powder Based Fusion Processes R&D = research and development SiC = Silicon Carbide SLA = Stereolithography SLM = Selective Laser Melting SLS = Selective Laser Sintering SS = Stainless Steel UAM = Ultrasonic Additive Manufacturing UC = Ultrasonic Consolidation XaaS = everythingasaservice Equation Nomenclature A = Area (m2₎ C = Heat capacity (J/kg.K) dT/dx = thermal gradient (K/m) D = Hydraulic diameter (m) f = Friction factor h = Specific enthalpy (J/kg) h_c = Convective heat transfer coefficient (W/m2_K) k = Material’s thermal conductivity constant (W/mK) L = Length (m) m = Mass (kg) m’ = Mass flow (kg/s) N_s= Number of entropy production units Nu = Nusselt number Q’ = Rate of heat transfer (J/s) ΔP = Pressure drop (%) Pr = Prandtl number Re = Reynold’s number St = Stanton number T = Temperature (K) V = Flow velocity (m/s) = Density (kg/m3₎ ρ = Viscosity (N.s/m2₎ μ

(9)

V. Abstract

A variety of materials for high temperature applications were studied. Best materials for constructing heat exchangers were selected using models based on preferential weights. Current additive manufacturing techniques and industries were also studied and rated to determine the best materialprintingtechnique combination. Although the rating models do not include every important criterion, the results were expected to be the same if the state of the 3D manufacturing industries and user preferences do not change. Design recommendations for a compact airtoair heat exchanger were made without considering manufacturing limitations. An economical assessment of 3D manufacturing techniques was made to determine whether 3D manufacturing could be a better alternative for heat exchangers. Although very promising, the choice to print heat exchangers with 3D techniques would not be economical at the moment. Future predictions of the additive manufacturing industry were made having studied related industries.

VI. Introduction

3D printing technologies are new, growing, and disruptive to the manufacturing industries. Their inherent design freedom gives them an immediate competitive edge over traditional manufacturing techniques. Their use in industries has, however, been a limited one due to other factors such as cost. Polymerbased printing processes have been more popular than in the case of printing with metals and ceramics. The study aims to evaluate the feasibility of 3D printing as an alternative method for manufacturing heat exchangers that are used in Compower’s ET10 microturbine. More information about the microturbine can be found from a Master of Science thesis by Loshan Palalayangoda (Palalayangoda 2010). As seen from the model in Figure 1, the heat exchanger under investigation will be operational in Compower’s externally fired microgasturbine for current cogeneration purposes (electricity and domestic heating). The company is aiming to integrate other energy sources to the microturbine system in the future (Compower 2015). The heat exchanger in consideration passes hot exhaust gases from the burner on the hot side while having ambient air from the compressor to the turbine in its cold side. As the heat exchanger replaces the combustion process in an idealized Brayton cycle for the turbine system, its performance will greatly affect the overall performance of the system. The study will investigate the best materials to be used for the heat exchangers and available 3D printing technologies for hightemperature materials. An indepth look into other industrial aspects like growth over time, economics, and barriers will assist further in evaluating the feasibility question.

(10)

Figure 1: AspenPlus model of Compower’s ET10 (Palalayangoda 2010) The study will also approach heat exchanger design in the light of the design freedom that comes with 3D printing. The main focus will be on the design that improves the thermal performance of the heat exchanger while prioritizing the aim of achieving minimum pressure losses in operation.

VII. Methodology

In order to assess the feasibility of using 3D printing as an alternative method of manufacturing for the heat exchanger, the problem was broken down into two smaller problems: material and 3D manufacturing technique. After analyzing the two subproblems, a final conclusion was made about whether or not 3D printing is a viable option while simultaneously keeping design and costs in mind. A. Materials The material selection process began with extensive research on past and present materials used for hightemperature heat exchangers (HX). The preliminary list of potential materials (Table 4) contains all materials capable of being used regardless of price. The preliminary list of 38 materials was narrowed down to 22 materials based on eliminating materials with exorbitant prices and limited research. Using a decision matrix (Table 5), the 22 materials were ranked and five materials were analyzed in detail. The criteria used to judge these materials were: maximum service temperature, density, yield strength, yield strength at 600°C, thermal conductivity at 1000°C, and price per pound.

(11)

Each criterion was given a weight between one and ten one being the least important and ten being the most important. Each criterion was also given a target value in order to assess how the materials’ natural properties measured up to the necessary requirements. Table 1 summarizes the weights and targets for each criterion. Maximum service temperature is critical to the efficiency of the complete microturbine and therefore was given a weight of nine. In order for the material to be considered a good viable option, it needs to have a maximum service temperature above 850°C. Light parts are preferred to heavier ones for mechanical design and logistical reasons. Lighter parts require less support system and are easier to carry and transport. For any desired HX volume, materials with lower densities are therefore better choices than those with higher densities. Density was given a weight of three (negative) because lowdensity material is preferred, but not required for the success of the heat exchanger. Yield strength is very important in determining the life cycle of the heat exchanger. Higher yield strength can tolerate higher system stresses, which will correlate to less fractures and failures. Ideally, material properties for yield strength would be found at 600°C. However, due to the lack of information, another method was used to determine the material’s qualification at 600°C. The yield strength at room temperature was taken as one aspect of the overall yield strength. Knowing that material properties change at high temperatures, a second aspect was taken into consideration. Each material was given a score of one or two depending on if it passed 200 MPa at 600°C. If the material passed, then it was given a two; if the material did not pass, then it was given a one. Ones and twos were used to fit the format of the sum equation used for the final score shown in equation 1. Similarly, thermal conductivity and cost were evaluated for all the materials with thermal conductivity having a weight of seven and cost having a weight of negative ten. Max service temp Density Yield Strength @ Room Does it pass 200 MPa @ 600°C Thermal Conductivity Cost/lb Weight 9 3 5 10 7 10 Target 850°C 8 g/cm3 _300 MPa ₁ _27 W/m°C _$2 Table 1: Weights for each criteria used to rank the materials

core1 SUM[W eight ]

S = *(V alue−Target)_Target

core1 9 0 0

S _{= *}(V alue−850)₈₅₀ _{− 3 *}(V alue−8)₈ _{+ 5 *}(V alue−300)₃₀₀ _{+ 1 *}(V alue−1)₁ _{+ 7 *}(V alue−27)₂₇ _{− 1 *}(V alue−2)₂

(12)

A second equation (Eqn. 2), based on ratios between material values to target values, was used to verify the results from the first equation. Only score 1 is shown in Table 5.

core2 SUM[W eight ]

S = _*(V alue)_Target (Eqn. 2)

After determining the top five materials, further analysis on their high temperature mechanical and thermal properties was done to determine the best material to use. The analysis focused on high temperature failure mechanisms and corrosion. Sustainability and current printability with 3D printing techniques were also added as additional factors in ranking the top five materials. The ranking system was similar to the one discussed above, the main difference is that the weights were multiplied by relative scores, scaled from 1 to 5. Table 2 shows the ranking criteria used for the top five materials and their respective weights.

core SUM[W eight Relative score in each criterion]

S = _* (Eqn. 3) Criterion for high temperature applications Weight (110) Printability (with current machines & possible methods) 7 Density 3 Thermal conductivity 8 Expansivity 7 Yield Strength 6 Ductility 4 Oxidation resistance 9 Corrosion from fuel combustion products 7 Creep & rupture strengths 5 Fatigue 4 Sustainability 10 Cost 10 Table 2: Ranking criteria for top five materials and their respective weights

(13)

High temperature mechanical failure criteria (creep, rupture, yield, and fatigue) had lower weights because the top five materials displayed good overall mechanical properties at high temperatures. Most of the materials could withstand exposures to significant stresses for a long time. This is seen in the material property discussion. B. Manufacturing In order to select the best and most appropriate manufacturing technique, extensive research was conducted for all current and progressing techniques. There are currently 11 additive manufacturing techniques, but only six methods were found to be relevant to this case. The six methods are laser melting, laser sintering, electron beam melting, binder jetting technology, laminated object manufacturing, and ultrasonic consolidation. Electron beam additive manufacturing was also briefly discussed in the Manufacturing section, but was ultimately excluded from the analysis due to the limited information on the technology. The six techniques were ranked from one to six (one being the worst and six being the best) in six categories: energy input, time required, bond quality / density, design freedom, accuracy, and maintenance cost. Each category was also given a weight for importance from one to ten one being not very important to ten being very important. Being mindful of cost and the environment, lower energy consumption would be ideal and was given a weight of eight. This criterion was assessed qualitatively by energy input needed to complete each process. For example, melting takes more energy than sintering. Following a similar logic, the lowest energy input system would receive a ranking of six. Time is an essential criterion because it also feeds into cost and was given a weight of six. The weight of six is justified by the rationale that 3D printing will not be a large production volume process and therefore is only somewhat important. However, the less time it takes to manufacture a product, the more product can be produced. Time was also assessed qualitatively with the fastest production time given a ranking of six. The major advantage of 3D printing is the design freedom it gives to engineers so three criteria that affect the product design are density, design freedom, and accuracy. The sturdiness and safety of the structure is highly important and therefore was given a weight of nine. The technique with the best bonding between the materials was given a ranking of six. Since the major advantage of 3D printing is design freedom, the category was given a weight of seven where techniques with virtually no limitation were given a ranking of six. The accuracy to which each technique can produce the finished product affects the mechanical properties of the heat

(14)

exchanger. Although efficiency is important, the main focus of the study is on feasibility and therefore accuracy was given a weight of five. The most precise technique with low surface roughness was given a ranking of six. Lastly, maintenance cost is important in terms of determining economic viability. Techniques were ranked by the perception of how much maintenance would be needed. For example, lasers are highly sensitive and require special maintenance and therefore would cost more than other techniques. Binder jetting involves fluids that may get clogged or contaminated, which would require replacement. A summary of the weights is shown in Table 3. Equation 4 shows how the total score was calculated.

_{Energy Input Time Bond Quality / Density Design Freedom Accuracy Maintenance}

Weight 8 6 9 7 5 5 Table 3: Weights for each criteria used to rank manufacturing techniques Total = SUM(Weight*Rank) (Eqn. 4) After all the ranking and weights were appropriated, the technique with the highest score was judged to be the best theoretical method of manufacturing. Further analysis was conducted qualitatively to discuss machine cost in order to analyze everything holistically.

VIII. Materials

Material selection is one of the most crucial aspects of this project. The material used for the heat exchanger must withstand a high inlet temperature and must operate above 600°C, although ideally well above 850°C in order to achieve a good overall efficiency. As the temperature is raised, the material may creep, limiting its ability to carry loads. It may degrade or decompose, changing its chemical structure in ways that make it unusable (Granta 2009). The two current models of heat exchangers are built using Stainless Steel 347 and Inconel 718. Using Ashby’s material and selection chart for Strength vs. Maximum Service Temperature shown in Figure 2, the materials were narrowed down to nickel alloys, stainless steels, tungsten alloys, and certain technical ceramics. Certain titanium metals were also considered, but would not likely be the final choice due to the lower maximum service temperature.

(15)

Figure 2: Ashby’s Material and Selection Chart (Granta 2009) A. Material Overview All the materials listed in Table 4 were initially considered for the heat exchanger, however, only the bolded materials were analyzed further. The items in Table 4 represent materials commonly mentioned in research literature and additive manufacturing. The bolded materials were considered to be superior based on scientific experiments conducted by other researchers as well as common properties such as maximum service temperature, density, and cost.

AL2025+Nb alloy Incoloy alloy 800HT Nickel 200 / 201 Stainless Steel 316L

Alloy 230 Incoloy alloy 825 Nickel Alloy 333 Stainless Steel SS347

Alloy 242 Inconel alloy HX Nicrofer Alloy 45TM Tantalum

Alloy 602CA Inconel 600 Niobium Ti64+TiC

Alloy modified 803 Inconel 617 Rhenium Ti6Al4V PM

Cobalt Chrome Inconel 625 Silicon Carbide Tungsten

Haynes Alloy 120 Inconel 718 Silicon Nitride TZM Molybdenum

Haynes Alloy 214 Inconel alloy 601 Stainless Steel 174PH Yttrium Oxide

Incoloy alloy 800 Iron Chrome Aluminum Stainless Steel 304

Incoloy Alloy 800H Molybdenum Stainless Steel 316

Table 4: All the materials initially considered for the heat exchanger design

(16)

Using a decision matrix explained in the Methodology section above, the 22 materials were narrowed down to 5 top materials in Table 5. Material Max service temp (°C) Density (g/cm3) Yield Strength (MPa) @ Room Yield strength @ 600°C Does it pass 200 Mpa @ 600°C Thermal Conductivity @ 1000(W/mC) Cost/lb Score Silicon Carbide1 ₁₆₀₀ _4.60 _1600.0 _250.0 _2.0 _41.0 _8.00 _14.51 Inconel 718 2 ₈₇₅ _8.19 _1034.0 _980.0 _2.0 _26.7 _5.46 _5.06 Haynes Alloy 214 3 ₈₀₀ _8.05 _605.0 _2.0 _32.7 _4.30 _4.51 Inconel 601 4 ₈₇₅ _8.11 _350.0 _330.0 _2.0 _27.8 _3.52 _3.66 Stainless Steel 304 5 ₇₅₀ _7.85 _215.0 _92.0 _1.0 _27.8 _0.88 _3.39 Stainless Steel 316 6 ₈₀₀ _7.99 _205.0 _153.8 _1.0 _27.8 _1.14 _2.38 Stainless Steel 316L6 ₈₀₀ _7.99 _170.0 _153.8 _1.0 _27.8 _1.23 _1.36 Stainless Steel SS347 7 ₇₅₀ _7.96 _205.0 _176.0 _1.0 _27.8 _1.40 _0.58 Alloy modified 803 8 ₈₇₅ _7.86 _290.0 _210.0 _2.0 _27.4 _4.22 _0.87 Haynes Alloy 120 9 ₈₀₀ _8.07 _375.0 _2.0 _26.2 _4.31 _1.07 1_{MEMSnet, “Material: Silicon Carbide (SiC), bulk,”} MEMSnet, accessed July 26, 2015, https://www.memsnet.org/material/siliconcarbidesicbulk/. 2_{Elgin Fastener Group, “Inconel 718,”} Elgin Fasteners, accessed July 26, 2015, http://elginfasteners.com/resources/materialproperties/inconel718/. 3_{Haynes International, “Haynes 214 Alloy.”} Haynes International, accessed July 26, 2015. https://www.haynesintl.com/pdf/h3008.pdf. 4_{Special Metals, “Inconel alloy 601.”} Special Metals, accessed July 26, 2015. http://www.specialmetals.com/documents/inconel_alloy_601.pdf. 5_{AK Steel, “304/304L Stainless Steel.”} AK Steel, accessed July 26, 2015. http://www.aksteel.com/pdf/markets_products/stainless/austenitic/304_304l_data_bulletin.pdf. 6_AK Steel 7_{Sandmeyer Steel Company, “321 and 347.”} Sand Meyer Steel, accessed July 26, 2015. http://www.sandmeyersteel.com/images/321347SpecSheet.pdf. 8_{Special Metals}

(17)

Material Max service temp (°C) Density (g/cm3) Yield Strength (MPa) @ Room Yield strength @ 600°C Does it pass 200 Mpa @ 600°C Thermal Conductivity @ 1000(W/mC) Cost/lb Score Silicon Nitride 10 ₁₅₀₀ _2.81 _525.0 _500.0 _2.0 _43.0 _8.00 _3.27 Inconel 6178 ₈₇₅ _8.36 _320.0 _255.0 _2.0 _28.7 _5.54 _6.82 Incoloy alloy 8008 ₈₇₅ _7.94 _347.5 _205.0 _2.0 _31.9 _6.00 _7.65 Inconel 6258 ₈₇₅ _8.44 _586.0 _207.0 _2.0 _25.2 _6.42 _7.72 Nickel Alloy 333 11 ₉₀₀ _8.20 _324.0 _200.8 _2.0 _28.9 _6.00 _8.65 Nicrofer 6025 HT / Alloy 602CA 12 ₉₀₀ _7.90 _270.0 _385.0 _2.0 _27.7 _6.00 _9.75 Inconel 6008 ₈₇₅ _8.47 _255.0 _210.0 _2.0 _27.5 _5.91 _10.11 Inconel alloy HX8 ₈₇₅ _8.20 _345.0 _210.0 _2.0 _27.9 _6.80 _12.83 Incoloy Alloy 800H8 ₈₇₅ _7.94 _150.0 _205.0 _2.0 _31.9 _6.80 _14.94 Incoloy alloy 800HT8 ₈₇₅ _7.94 _150.0 _205.0 _2.0 _31.9 _7.00 _15.94 Nicrofer Alloy 45TM 13 ₉₀₀ _8.00 _240.0 _135.0 _1.0 _27.0 _6.00 _20.47 Haynes Alloy 230 14 ₈₀₀ _8.97 _395.0 _274.4 _2.0 _28.4 _9.15 _24.71 Table 5: The top 22 materials and their material properties It can be seen from Table 5 that SiC has outperformed the rest of top five materials by at least a factor of two. 10_{Azo Materials, “Silicon Nitride Properties and Applications.”} AZOM, accessed July 26, 2015. http://www.azom.com/properties.aspx?ArticleID=53 11_{Rolled Alloys, “Data Sheet RA333.”} Rolled Alloys, accessed July 26, 2015. http://content.rolledalloys.com/technicalresources/databooks/RA333_DB_US_EN.pdf. 12_{ThyssenKrupp Stainless, “Nicrofer 6025 HT alloy 602 CA.”} VDM, accessed July 26, 2015. http://www.vdmmetals.com/fileadmin/user_upload/Downloads/Datenblaetter__Data_Sheets/Data_Sheet_VDM_Alloy_602_CA.pdf. 13_{ThyssenKrupp Stainless} 14_{High Temp Metals, “Haynes 230 Technical Data.”} High Temp Metals, accessed July 26, 2015. http://www.hightempmetals.com/techdata/hitempHaynes230data.php.

(18)

B. Silicon Carbide Silicon carbide (SiC) was accidentally discovered in 1890 by Edward G. Acheson while he was running an experiment on the synthesis of diamonds. Silicon carbide occurs naturally in meteorites, though very rarely and in very small amounts. Today, SiC is produced via a solid state reaction between sand (silicon dioxide) and petroleum coke (carbon) at very high temperatures in an electric arc furnace (Poco 2002). SiC is lowweight and has high strength, hardness, and strong covalency. Combining those properties with low thermal expansion coefficient and high thermal conductivity, SiC is a promising alternative to conventional metals, alloys, and ionicbonded ceramic oxides (Poco 2002). The most common forms of SiC include powders, fibers, whiskers, coatings, and single crystals. Since these materials will be paired with 3D printing, powder production will mainly be considered (Poco 2002). i. Creep and Oxidation Creep rates of ceramics possessing a glassy grainboundary phase are degraded compared with the inherent creep resistance. Silicon carbide ceramics are generally kinetically stable in air to temperatures around 1000°C. Rapid surface oxide layers start to form in the 1000°C1150°C temperature range after which oxidation becomes passive (Poco 2002). Oxidation rates start to become significant (when oxide layer starts to form) at 1650°C. The presence of impurities, introduced by the sintering additives, often reduces the oxidation resistance of SiC as well. For temperatures of around 1500°C, alpha sintered SiC samples show creep strength values of at least ten times those exhibited by Inconel 601 samples at ~1000°C (Munro 1997). ii. Fatigue Toughness The low inherent fracture toughness of conventional SiC ceramics can be improved by producing a composite, typically by incorporating continuous fibers, whiskers, or secondphase particles. Recent research has focused on monolithic SiC hot pressed with aluminum metal as well as boron and carbon. This is often referred to as ABCSiC and it has been shown to have ambient temperature fracture toughness as high as 9 Mpa m1/2_{with strengths of 650 MPa (Chen 2000).} iii. Crack Healing An interesting behavior that occurs in sintered ceramics is the crackhealing behavior. One experiment showed that SiC sintered with scandium oxide and aluminum nitride with a surface crack of 100 micrometer fully recovered its strength at room temperature after a heat treatment at 1300°C for one hour in air under no stress and at 1200°C for five hours under an applied stress of 200 MPa (Lee 2005).

(19)

iv. Compatibility with 3D Printing SiC have been tested with powder based fusion processes and new research is being conducted with laminated object manufacturing technology. The major hindrance to SiC’s compatibility with 3D printing is purity. Sintering ceramic powder must be less than 1 micron for certain machines and this would require SiC powdered to be further milled and acidtreated to remove metallic impurities (Poco 2002). Another factor to take into consideration with 3D printing ceramics is the shrinkage that occurs during sintering. The density of SiC sintered at 1750°C is close to the theoretical values (9798%) and the mass loss and shrinkage ratio are small. SiC ceramics sintered at 2000°C have more shrinkage after sintering and achieve less density (92%) (Liu 2005). This lower density at 2000°C can be explained by the evaporation of SiC and oxide additives. Therefore, the maximum temperature capability of SiC formulation is about 1900°C. Higher temperatures may cause SiC to vaporize in the presence of nitrogen and may dissociate to pure silicon and carbon (Liu 2005). v. Sustainability Unlike stainless steels, ceramics are not easily recyclable. The best way to reuse ceramics is to smash the product and mill them back into powder (Zero 2015). Ceramics production is very energy intensive even with the EU industry halving its energy consumption over the last 25 years. Dust and gaseous emissions arise during the firing and spray drying of ceramics (European 2015). C. Inconel alloys The name Inconel is a trademark of Special Metals Corporation and it is for a group of nickelchromium based alloys that show excellent performance under high temperature and corrosive environments (Special 2015). This is largely due to their compositions and also heat treatment. Agehardened Inconel 718 and solutiontreated Inconel 601 are commonly used in high temperature applications due to their higher strengths (Special 2015). While Inconel 601 has a larger nickel base and more chromium content, Inconel 718 contains several other elements in its matrix. Additional molybdenum in Inconel 718 significantly improves its strength and the higher aluminum content in Inconel 601 achieves a high oxidation resistance (Special 2015). i. Impact Strength Alloy 601 retains high impact strengths even after long exposure to high temperatures. Solution treated samples of Alloy 601 had an average impact strength values of ~170J at room temperature. These varied from 120160J after being exposed to temperatures from 540870°C for over 1000 hours of operation (Special 2015). This shows that the alloy maintained its good

(20)

ductility properties. Aged samples of Inconel 718 had maximum impact strengths of ~60J at room temperature, a significantly lower value compared to that of Inconel 601 (Special 2015). ii. Fatigue Strength The solutiontreated Inconel 601 rod has an endurance limit of ~260 MPa for infinite stress cycles at room temperature. A similar annealed sample had an endurance limit of ~330 MPa for infinite stress cycles (Special 2015). Room temperature endurance limit of an Inconel 718 rod of similar size was ~620 MPa for infinite stress cycles (Special 2015). Lowcycle endurance limits are higher for both samples. iii. Creep and Rupture Properties For creep rates of 0.01% in an hour at 870°C, the creep limit for samples of solution treated Inconel 601 was ~30 MPa. The limit was 100 MPa at 705°C while that of agehardened Inconel 718 was ~450 MPa at the same temperature. Rupture strengths at 705°C for 1000 hours were ~100 MPa and ~ 500 MPa for alloys 601 and 718, respectively. Alloy 718 showed better creeprupture properties (Special 2015). iv. Corrosion resistance The alloys show great corrosion resistance properties especially due to their nickelchromium base composition. Nickel contributes to corrosion resistance in many organic and inorganic media (Special 2015). Chromium increases the resistance to oxidation and reactions to sulfur while additional aluminum in alloy 601 helps to combat oxidation. Molybdenum in alloy 718 is known to reduce pitting corrosion (Special 2015). A good way to compare oxidation resistances of the superalloys is by looking at their oxidation rate constants ‘Kp’ at high temperatures. The rate constants are obtained by the slopes from the graphs of the squares of the mass gained per unit surface area versus time. These tend to be linear within a given range of temperatures because oxidation is a chemical reaction that follows Arrhenius behavior (Clark 2013). Openair oxidation rate constants of Inconel 601 are generally lower than those of Inconel 718 for temperature of around 1000°C (Yang 2002). The rates in Inconel 718 increase by a factor of two per 100°C rise in temperature (Greene 2001). Inconel 601 shows good corrosion resistance around both oxidizing and reducing environments while both the alloys are performing well under hydrogen attacks (Haynes 2015). However, the oxidation rates for both the Inconel alloys are much greater in comparison to Haynes 214 that has excellent corrosion resistance in a wide range of chemical media (Yang 2002). Studies have also shown internal oxidation attacks in Inconel 601 (Special 2015), (Haynes 2015). This is when oxide regions are formed inside the surface as a result of oxygen diffusion. Other alloys have tendencies to form a protective oxide layer at the surface that reduces oxidation rates and prevents further attack inside the alloy base. Figure 3 shows oxidation results in Inconel 601 versus those in Haynes 214 (Haynes 2015). Internal attack is seen in Inconel 601.

(21)

Figure 3: Internal oxidation attack in Inconel 601 (Haynes 2015) v. Compatibility with 3D printing Standard manufacturing forms for both alloys are pipes, tubes, sheets, strips, plates, round bars, flat bars, forging stocks, hexagons, and wires. Many metal alloy powder companies do produce superalloys in powder forms of different sizes for highpurity purposes (American 2015). Thus both powderbased and laminatedbased 3D printing processes can implement the Inconel alloys. vi. Sustainability Inconel 601 is known for its excellent machinability and formability, relative to both Haynes 214 and Inconel 718 (Yang 2002). Its manufacturing costs, those associated with preparing powders and sheets, are thus expected to be relatively lower. Recycling of Inconel alloys is done by a few companies due to the growing demand of the scrap Inconel alloys in different applications (Monico 2015). D. Haynes 214 Alloy Haynes 214 alloy is a nickelchromiumaluminumiron alloy with excellent hightemperature oxidation resistance. Its intended use is at temperatures above 955°C. Under 955°C, Haynes 214 still provides oxidation resistance equal to the best nickelbased alloys (Haynes 2015). i. Oxidation and Creep When exposed to air flowing at 213.4 cm/min for 1008 hours at 1150°C, Haynes 214 showed only 8 micrometers of metal affected (Haynes 2015). The most important part of the result was the virtual absence of internal attack for the Haynes 214 (Figure 3). At 870°C, Haynes 214 can withstand 54 MPa after 1000 hours of operation until rupture. The creeprupture properties of Haynes 214 and Inconel 601 samples are generally similar for up to 10000 hours of high temperature exposures (Haynes 2015). ii. Cracking Similar to nickelbase alloys, Haynes 214 will exhibit agehardening as a result of the formation of a second phase, gamma prime (Ni₃Al) (Haynes 2015). This causes a significant loss of

(22)

intermediate and low temperature tensile ductility, which results in alloy 214 being susceptible to strainage cracking when highlystressed welded components are slowly heated. iii. Compatibility with 3D Printing Haynes 214 alloy is currently available in the form of plate, sheet, strip, billet, bar, and wire, which makes it compatible with Electron Beam Additive Manufacturing and Ultrasonic Consolidation, but not with the powderbased techniques. There are currently no companies printing with Haynes 214. iv. Sustainability Haynes alloy 214 scraps are collected by private companies such as Greystone Alloys and Monico Alloys that handle the maintenance and reselling of the material E. Stainless Steel 304 Stainless Steel 304 (SS304) is one of the most common stainless steels. It is an austenitic steel that is not particularly electrically or thermally conductive, but has a higher corrosion resistance than regular steel. SS304 and SS316 are often compared and going back to Table 5, the low cost of SS304 won out in the end. The main difference between 304 and 316 stainless steel is that SS316 contains 2%3% molybdenum and 304 has no molybdenum. The molybdenum added improves corrosion resistance to chlorides (Slipnot 2015). i. Creep and CreepFatigue The dependence of minimum creep rate on stress during hightemperature deformation of SS304 follows the power law,_{ε * α}= (σ) β, where α and β are dependent on temperature (Arcam 2009). The stress exponent β determined by one experiment was 5.6, 5.9, and 6.5 for temperatures of 700, 650, and 600°C, respectively. During hightemperature deformation, creep deformations result from not only the activation of normal slip systems, but also the translation of grains relative to one another along their boundaries. At 700°C and 76 MPa of applied stress, fracture occurs (Zhang 2014). Over 482°C, deformation under stress is plastic rather than elastic, so the yield point as determined by the shorttime tensile test is higher than the creep or creeprupture strength (Nickel 2015). A creeprate curve of annealed SS304 show ~10 MPa of stress at creep rates of 0.0001% / hr at 800°C. The number of cycles to failure decreases linearly with increasing the hold time in double logarithmic coordinate less than 650°C. At high temperature of operation above 700°C, SS304 would fail after about 60 cycles of 30 minute hold time (Zhang 2014).

(23)

ii. Oxidation The oxidation resistance of stainless steel occurs due to its ability to form a protective coating layer on the surface. The coating is a passive film, which resists further oxidation. The noncorrosive property in stainless steel derives from the existence of chromium in the alloy. SS304 is 18% chromium and SS316 is about 1618% chromium (AJMFG 2015). The kinetics of oxidation of SS304 was found to be 2.3 μg/cm2_{of oxide formed at 500°C in six} hours. The rate of increase was 0.1 μg/cm2_{/hr after the first six hours. At 800°C, a transition is} observed and the rate of oxidation followed a linear growth rate rather than the previous parabolic rate law. This transition occurred for a weight gain of 9 μg/cm2_{. At 900°C and for} weight gains above 90 μg/cm2_{, the parabolic rate law was found to hold again. The second} transition is found in the kinetics of oxidation at 1150°C (Gulbransen 1962). The maximum temperature to which SS304 can be exposed continuously without appreciable scaling is about 900°C. For intermittent cyclic exposure, the maximum exposure temperature is about 815°C (AK 2015). iii. Compatibility with 3D Printing Since SS304 comes in powder, plate, wire, and sheet form it is highly compatible with 3D printing. Many companies such as Optomec already have SS304 listed as one of their materials. More companies have explored 3D printing with SS316 & 316L powders. iv. Sustainability Stainless steel is an infinitely recyclable commodity which lends itself to be highly sustainable. The typical amount of recycled stainless steel scrap is about 65 to 80%. Since stainless steels are widely used in a variety of markets, many green steps are being taken to reduce the carbon footprint of manufacturing. Environmental legislation is also forcing petrochemical and refinery industries to recycle secondary cooling water in closed systems (Advameg 2015).

(24)

F. Evaluating hightemperature thermal and tensile properties

Figure 4: Thermal conductivities of the top five materials with temperature

(25)

Thermal performance was evaluated from looking at thermal conductivities and average specific heat capacities of the top five at a range of elevated temperatures. Figure 4 shows the superiority of the conductivity of SiC in applications from cryogenic ranges to 1100°C. While the ceramic shows declining conductivities with rising temperatures, metallic alloys showed the opposite trend owing to the increasing vibrational kinetic energies with temperature. SS304 shows a slightly better performance relative to the two Inconel alloys. The second best material was Haynes 214, which showed a significantly disproportional rise in its conductivities at temperatures above 600°C. The average specific heat capacities of SiC (Figure 5) were significantly higher than those of the metallic alloys and they increased steeply with temperatures. Despite a better heat transfer that comes with high heat capacity, SiC could be of a slight disadvantage since it would take more thermal energy inputs (than in other materials) for a given rise of temperature. The ceramic is also expected to take more time to warm up or cool down, contributing to a problem of residual heats in operations. The high heat capacity advantage, however, outweighs other penalties and could be of much advantage as discussed in the design section. Figure 6: Coefficient of linear expansion of the top five materials with temperature SiC showed the lowest values of linear expansion coefficients while SS304 had the highest. Inconel 718 had the smallest coefficient of the three remaining alloys. Material expansion at elevated temperatures could be detrimental to designs (unless accounted for) and also they add additional and unexpected stress concentration points, especially if the expansion is anisotropic. Volume expansion coefficients could also be estimated from the Figure 6 as γ~3ᵯ for cubes.

(26)

Figure 7: Yield strengths of top 5 materials with temperature Figure 8: Elastic modulus of top five materials with temperature

(27)

High temperature tensile properties of the materials were compared by using yield strength (Figure 7) and elastic moduli (Figure 8) data obtained from different specimens. These values could vary significantly depending on the method of sintering (for SiC) and temperature working. Solutiontreated Inconel 601 (on the plot) is usually used for high temperature applications because of its higher strength values than in annealed samples (Special 2015). Values of tensile strength for sintered SiC could go as high as 300 MPa as well (Poco 2002). It can be seen that Inconel 718 and Haynes 214 show excellent strengths (>250 MPa) up to temperatures around 900°C. SiC shows a constant strength at even higher temperatures. SS304 has the lowest strength values and it is essentially ineffective past 750°C. Inconel 601’s best range of application is seen to be between 600°C and 800°C. Its strength values are significantly lower than those of Inconel 718, Haynes 214, and sintered SiC until around 1000°C. SiC is, however, very brittle as indicated by its high elastic moduli relative to the metallic alloys. Brittleness could pose a challenge to the formability of ceramics despite its high strength. Ceramics are still more likely to have brittle fractures especially during sudden stress loads. SS304 has the best ductility properties at high temperatures. G. Ranking results Table 6 shows the average scores of each of the top five materials after considering a wider array of properties discussed above. The ranking method is described in the Methodology section. From the ranking method, it can be seen that SiC is the best material to use for heat exchangers despite the high cost, poor sustainability, and limited printability concerns. The excellent mechanical and thermal properties, together with its unparalleled corrosion resistance, gives SiC an edge over the others. Its density that is almost half of other top five materials also adds more benefits that come with the choice of SiC. Inconel 718 comes as a close second best material because of its high strength, extensive printability, and sustainability advantages. Haynes 214’s excellent corrosion resistance puts it in between the top two and bottom two materials. Inconel 601 and SS304 had the lowest scores because of the penalties from poor corrosion resistance and strengths at high temperatures. SS304, despite its excellent sustainability and cost advantage still had the lowest score because it performed poorly in other criteria.

(28)

Property Silicon Carbide Inconel 718 Haynes 214 Inconel 601 SS304

Printability 2 5 1 4 4 Density 5 2 2 2 3 Thermal conductivity 5 2 4 2 3 Expansivity ₅ ₃ ₂ ₂ ₁ Yield Strength 4 5 3 2 1 Ductility 1 4 3 4 5 Oxidation resistance 5 3 4 2 1 Corrosion resistance 5 3 4 2 1 Creep & rupture strengths 5 4 2 2 1 Fatigue 4 3 5 3 2 Sustainability 1 4 3 4 5 Cost 1 2 3 4 5 Total 273 264 243 226 223 Table 6: Ranking results for the top five materials

IX. 3D Printing / Additive Manufacturing

3D printing is the process of making three dimensional objects from a digital CAD file. Interest in 3D printing has boomed within the last five years raising nearly four billion USD in public offering since 2011 (Wheeler 2015). Typically used as a prototyping method to speed up the design process, 3D printing is now being considered as a manufacturing process for the final product. Unlike injection molding and other traditional methods of manufacturing, 3D printing is not limited by production restraints, which opens up for design creativity.

(29)

In February 2015, Australia’s Monash University created the first 3Dprinted jet engine (Coxworth 2015). Three months later, in May 2015, GE fired up a simple jet engine made entirely of 3Dprinted parts and revved it up to 33,000 RPM (Keller 2015). The aerospace industry has jumpstarted the race to 3Dprint parts because of the reduced lead time, lighter weight of parts, and lower production costs. These same motivations led to research of manufacturing affordable heat exchangers for the microgas turbines. The key to how additive manufacturing (AM) works is that parts are made by adding material in layers; each layer is a thin crosssection of the part derived from the original CAD data. The thinner each layer, the closer the final product will be to the original design (Gibson 2015). All commercialized AM machines to date use a layerbased approach and the major difference lies in the materials used, how the layers are created, and how the layers are bonded together. These differences determine the speed of the process, the amount of postprocessing required, the size of the AM machine used, and the overall cost (Gibson 2015). A. Manufacturing Techniques Overview Several methods of 3D printing and additive manufacturing are being researched and are listed in Table 7. The methods highlighted are viable alternatives that will be considered in this report. Many of the methods use the powder bed fusion process, which follows the same basic loading procedure (Noe 2014). Basic Procedure: 1. A designer / engineer designs a part using a CAD software 2. The part is cut into virtual slices on the horizontal plane 3. A chamber is filled with powder 4. A laser / electron beam scans the powder, solidifying a thin layer 5. Another layer of powder is added as the platform moves down 6. Layer by layer, the product is built up until it is finished 7. The leftover powder is reused

(30)

Plastics Metals Ceramics Selective Laser Sintering (SLS): does not fully melt the powder, but heats it to a point that the powder can fuse together on a molecular level. Used with alloys and porosity can be controlled Selective Laser Melting (SLM): the material is heated to a full melt and forms one homogenous form Binder Jetting Technology: spreads a binder over the powder (infuses steel with bronze) Laminated Object Manufacturing (LOM): uses a continuous sheet (paper, plastic, less common metal) that is cut by a laser and built layer by layer using a heat rollers Stereolithography (SLA): converts liquid plastic to solid layer by layer using solvent and ultraviolet oven Direct Metal Laser Sintering (DMLS) is the same as SLS, but specifically for metals Direct Metal Laser Melting (DMLM) is the same as SLM, but specifically for metals Electron Beam Melting (EBM) is the same as SLM except it uses an electron beam instead of a laser Fused Deposition Modeling (FDM): heating and extruding thermoplastic filaments Electron Beam Additive Manufacturing (EBAM): an electron beam gun deposits metals layer by layer via wire feedstock Ultrasonic Consolidation (UC): welds metal foils using CNC contour milling Table 7: Overview of additive manufacturing techniques compatible with plastics, metals, and ceramics B. PowderBased Fusion Processes (PBF) Metallic powderbased fusion in AM mimics the polymeric powderbased processes, which were amongst the first commercialized AM processes (Copra 1977). Building layers are formed by fusing thin layers of metallic powder that are deposited on the build platform. The fusion occurs either by solidstate sintering, melting, or a mix of both. All PBF processes have one or more

(31)

thermal sources to induce fusion between powder particles, a method to control that fusion in specific regions, and mechanisms for adding powder layers. Most common thermal energy sources for metal PBF processes are lasers and electron beams. These are usually assisted by preheating to maintain elevated temperatures across the powder layers. Figure 9 shows the schematic of a selective lasersintering process, a setup that is very similar to many other PBF processes. Figure 9: Schematic of selectivelaser sintering (Gibson 2015) The roller spreads thin layers of powder across the build platform. Once the powder deposition is complete the laser is directed into the powder bed and moved to thermally fuse the powder to form a slice of the cross section. The surrounding powder is left loose and can be used to support subsequent layers. After completing a slice, the build platform is lowered by one layer thickness and a new powder layer begins. The process repeats until the part is complete. Due to the reactive nature of powders and the elevated temperatures on the bed, sintering and melting techniques require inert conditions to avoid rapid oxidation or reaction with other substances. There are principally four different fusion mechanisms that are associated with PBF processes: solidstate sintering, chemically induced binding, liquidphase sintering, and full melting (Gibson 2015). i. SolidState Sintering This implies fusion of the powder particles without melting. The process usually occurs at temperatures between half of the melting point and the melting point (0.5T_m< T_s< T_m). Solid state sintering occurs to minimize the surface free energy (proportional to the total surface area)

(32)

at elevated temperatures. The particles diffuse to minimize the total surface area causing a decreased porosity and solid part formation shown in Figure 10. The process takes an abundance of time and is usually not the primary fusion mechanism in PBF processes. Figure 10: Schematic of solidstate sintering (Gibson 2015) ii. ChemicallyInduced Sintering This involves a thermally activated chemical reaction between two types of powders or a powder and a gas to form a byproduct that acts as a binder. Chemically induced sintering is mostly applied for ceramic materials. For example, the formation of silicon dioxide that binds silicon carbide in laser processing (Gibson 2015). iii. LiquidPhase Sintering (LPS) This process involves the fusion of powder particles when a portion of constituents within the powder is molten while the other remains solid. The molten part acts as a binder that holds the high temperature particles together. The binder material is usually included in the main powder as a coated substance, a composite, or a separate powder. LPS processes are the most rapid and do not require full sintering or melting (Gibson 2015). iv. Full Melting This is when the powder is fully melted beyond the depth of the layer to create a wellbonded and highly dense structure with a very low porosity (Gibson 2015). This method is commonly used for metal alloys due to strength requirements of the final parts. v. Electron Beam / Electron Beam Melting (EB / EBM) This technology uses highenergy electron beam to fuse metal powder particles instead of the typical high intensity lasers. Heating is caused by the transfer of kinetic energy from incoming electrons instead of absorption of photons from lasers. EB processes require vacuums (to preserve the kinetic energy) and have achieved higher scan speeds. EB can even be applied to moderate particle sizes (Gibson 2015). They are, however, limited to metals and metal alloys due to the high conductivity requirements. EB processes are also associated with poor surface finishes (Gibson 2015).

(33)

vi. Sustainability Metal PBF are the most energy intensive processes due to their high temperature requirements. Full melting involves remelting part of the already formed solid layer underneath, contributing to more energy costs. Additional energy costs come from the need for more sophisticated finishing (postprint) processes due to low accuracies that are associated with PBF processes (Gibson 2015). Careless powder and laser handling could easily result in serious hazards as many powders are toxic and reactive, especially under energyintensive lasers. Powder recycling in sintering processes is still a challenge as the heated powders tend to have different properties from normal ones (Gibson 2015). C. Binder Jetting Additive Manufacturing (BJAM) Binder jetting technology was developed in the early 1990’s by MIT and it has been licensed for more than five companies for commercialization (Gibson 2015). The method is very similar to that the other PBF processes. The difference is that layers are formed at the powder bed with a liquid binder. The final product is built entirely without heating. After the product is built up, postprint processes such as curing, heating in the furnace to vaporize the binder, and sintering take place. This is usually associated with some degree of loss accuracy. Some companies have also applied infiltrants to compensate for the part’s density and strength losses in the postprint heating processes (Gibson 2015). Binder jetting technologies do not require highenergy inputs or dangerous lasers and are relatively fast, depending on the flow rate of the binder (Gibson 2015). There is also a promising research being done on printing ceramics parts by using binder jetting. D. Laminated Object Manufacturing (LOM) Developed by Californiabased Helisys Inc., now Cubic Technologies, laminated object manufacturing is a 3D printing technique where layers of a material are fused together using heat and pressure and then cut into the desired shape with a controlled laser or blade. The LOM apparatus uses a continuous sheet of material (usually paper, but is now being experimented with metal and ceramics), which is drawn across a build platform by a system of feed rollers. Figure 11 shows that the material is unwound from feed roll (A) onto the stack and bonded to the previous layer using a heat roller (B). The roller melts a coating on the bottom side of the material to create the bond. The 2D profiles of the desired product are traced by a laser mounted on an XY stage (C). The excess material is cut away and fed to the takeup roll (D). The process generates a lot of smoke and localized flame; therefore, either a chimney or filtration system (E)

(34)

is required (Palermo 2013). After one layer is complete, the build platform lowers about onesixteenth of an inch, or the thickness of one layer and the process repeats. Figure 11: The laminated object manufacturing process (Palermo 2013) LOM is not ideal for complex geometries and cannot create hollow objects. Therefore, this manufacturing technique creates limitations on the design similar to traditional techniques. LOM is not highly accurate and as a result has been limited primarily to conceptual prototyping. Furthermore, the main material used in this type of manufacturing has been paper. LOM is very fast, low cost, and has ease of material handling. LOM is not very popular due to the limited materials compatible with this technology and was considered because Ceralinks is currently using this method to 3D print ceramic objects. E. Electron Beam Additive Manufacturing (EBAM) Sciaky’s Electron Beam Additive Manufacturing should not be confused with electron beam melting. Electron beam melting is a powderbased method that uses an electron beam to fuse the metal together. EBAM, on the other hand, deposits metal via wire feedstock layer by layer until the product is complete. It has a standard deposition rate ranging from seven to twenty pounds per hour (Sciaky 2015).

3D Printing as an Alternative Manufacturing Method for the Micro­gas Turbine Heat Exchanger