scieee Science in your language
[en] (orig)

Additive manufacturing capabilities for heat switch technology: Key challenges & knowledge gaps

Abstract

The paper is to provide an overview of the key challenges and knowledge gaps in additive manufacturing of metals applied to two parts of a novel heat switch technology – a Baseplate and a Flexible thermal structure. Additive design, optimization and manufacturing capabilities as well as quality of produced parts were investigated. The key challenge was to manufacture bio-inspired structure without internal supports and assure 190 surfaces to be in a contact at the same time for efficient heat transfer. Finally, modern trends in additive as multi-material design and manufacturing or effect of defects are discussed for further evolution of parts.

Read accessible full text

Additive manufacturing capabilities for heat switch technology: Key challenges & knowledge gaps

Author: Mašek, Jakub; Löffelmann, František; Popela, Robert; Kubík, Petr; Šebek, František; Koutný, Daniel; Malý, Martin; Pantělejev, Libor; Pambaguian, Laurent
Publisher: EUCASS association
Year: 2022
DOI: 10.13009/EUCASS2022-6196
Source: https://dspace.vut.cz/bitstreams/b8fffc54-6958-4398-8907-94d99eb7ff47/download
Copy igh © 2022 by Jakub Mašek e al. Published by he EUCASS associa ion wi h pe mission.
Addi i e manu ac u ing capabili ies o hea swi ch
echnology: Key challenges & knowledge gaps
MAŠEK Jakub1*
1 B no Uni e si y o Technology, Ins i u e o Ae ospace Enginee ing, Technicka 2, 61669 B no, Czech Republic
* Jakub.Masek@ u b .cz
LÖFFELMANN F an išek, POPELA Robe , KUBÍK Pe , ŠEBEK F an išek, KOUTNÝ Daniel, MALÝ Ma in,
PANTĚLEJEV Libo 2, PAMBAGUIAN Lau en 3
2 B no Uni e si y o Technology, Facul y o Mechanical Enginee ing, Technicka 2, 61669 B no, Czech Republic,
F an isek.Lo elma[email p o ec ed]; pop[email p o ec ed] .cz; [email p o ec ed] b .cz; sebek@ me. u b .cz;
Daniel.Kou[email p o ec ed]; Ma in.Maly2@ u .cz; pan eleje @ me. u b .cz
3 Eu opean Space Agency (ESA - ESTEC), Ma e ials and P ocesses Sec ion (TEC-MSP), Keple laan 1, NL-2200 AG
Noo dwijk, The Ne he lands, Lau en .Pamb[email p o ec ed]
Abs ac
The pape is o p o ide an o e iew o he key challenges and knowledge gaps in addi i e manu ac u ing
o me als applied o wo pa s o a no el hea swi ch echnology – a Basepla e and a Flexible he mal
s uc u e. Addi i e design, op imiza ion and manu ac u ing capabili ies as well as quali y o p oduced
pa s we e in es iga ed. The key challenge was o manu ac u e bio-inspi ed s uc u e wi hou in e nal
suppo s and assu e 190 su aces o be in a con ac a he same ime o e icien hea ans e . Finally,
mode n ends in addi i e as mul i-ma e ial design and manu ac u ing o e ec o de ec s a e discussed
o u he e olu ion o pa s.
1. In oduc ion
Addi i e manu ac u ing o me als, as a g owing ma ke , has a lo o o e in e ms o design-manu ac u ing capabili ies
and sus ainabili y [1]. The e o e, i a ac s key playe s and en ep eneu s in he space indus y o each be e
pe o mance o pa s a a lowe mass. One o he game-changing applica ions is he mal ha dwa e [2; 3].
The mal managemen is an in eg al pa o each spacec a . I is necessa y o ensu e he he mal egula ion o
in e nal elec onics and payload o allow hem o wo k in sui able condi ions. A phase change ma e ial-based (PCM)
hea swi ch [4; 5] is a one ype o many [6; 7] s udied in li e a u e. The swi ch echnology can be di ided in passi e
and ac i e de ices ha di e in empe a u e and conduc i i y pa ame e s as well as in he physical p inciple o unc ion.
Selec i e lase mel ing (SLM) echnology is a cu en ly leading inno a i e manu ac u ing p ocess among he
addi i e echnologies ha accele a es changes in many enginee ing a eas, pa icula ly he ae ospace [8; 9]. Se e al
s udies in es iga ed space componen s op imized and p oduced by SLM echnology om AlSi10Mg powde , as
b acke s [10; 11], CubeSa ame [12], an ennas and wa eguides [13] o hea pipes [14]. The bene i o p in ing he
componen s laye -by-laye enables a di e en app oach in design and manu ac u ing p ocess. Cos s, weigh and
ma e ial was ing can be educed while p oducing pa s wi h he same o e en be e mechanical p ope ies [15; 16].
Howe e , only ew addi i e pa s we e quali ied o space ligh [11; 17; 18]. The echnology i sel c ea es mic o and
mac o s uc u e de ec s, po osi y, he mal de o ma ions and poo su ace quali y [19; 20; 21]. E ec o de ec s [22]
and ela ed inspec ion and pos -p ocessing echnologies a e among he mos c i ical pa ame e s o la ge space
applica ion and quali ica ion o addi i e pa s. Highe amoun o po osi y, lowe geome y and su ace p ecision a e
he key disad an ages o aluminium-based ma e ials despi e o easie SLM manu ac u ing and la ge knowledge base
o ma e ial p ope ies [23; 24].
DOI: 10.13009/EUCASS2022-6196
9ᵀᴴ EUROPEAN CONFERENCE FOR AERONAUTICS AND SPACE SCIENCES (EUCASS)
Jakub Mašek e al.
2
A minia u ized hea swi ch [25; 26] (Figu e 1) is
a de ice unde he de elopmen dedica ed o e icien ,
au onomous and powe less empe a u e egula ion
based on he mo-physical p ope ies o pa a in. The
echnology epea edly couples o decouples he hea
sou ce wi h ex e nal adia o ha dissipa es he excess
o ene gy o he en i onmen . I shall ha e wo-posi ion
on/o he mal ans e up o 10 W. The swi ch shall
wo k in deep space and Ma ian condi ions o
wi hs and empe a u es om – 125 °C o + 60 °C. The
implemen a ion o addi i e echnology is expec ed o
enhance he design owa ds an ope a ional sample.
The pape is o p o ide an o e iew o he key
challenges and knowledge gaps in addi i e SLM
manu ac u ing applied o wo pa s o he no el hea
swi ch echnology – a Basepla e and a Flexible he mal
s uc u e. In addi ion, o p o ide a con ex o mode n
ends in addi i e o u he e olu ion o pa s.
1.1 Legacy issues o pa s
Basepla e (BP) pa ep esen s a hal o pa a in p essu e con aine (Figu e 1a). The componen (Ø32 mm and 17.6 mm
in heigh ) o mass oughly 30 g shall wi hs and an in e nal p essu e o up o 16 MPa, while he de o ma ions each less
han 0.02 mm. The Basepla e shall allow pa a in mel ing and on he con a y, p e en i s leakage. The la ness o he
pla e is essen ial o p o ide a con ac in e ace o an ex e nal hea sou ce o e icien hea ans e . Howe e , pa a in
p essu e d i ing he ac ua o causes in he ‘on’ mode swelling o he swi ch so ha he con ac su ace migh be
signi ican ly educed (Figu e 2b) [26]. The goal is he e o e o op imize he he mal-s i ness p ope ies by adding
in e nal ein o cemen s a a minimum weigh .
Figu e 2: Basepla e pa ; (a) o me design; (b) empe a u e- o-de o ma ion dependence based on pa a in p essu e
The second pa , Flexible he mal s uc u e (FTS), ep esen s he pa h o hea ans e . The componen (Ø56 mm and
24.1 mm in heigh ) o mass oughly 70 g has a wo-posi ion on/o mode allowing a e ical linea mo emen and a
he mal conduc i i y highe han 1.5 W⋅K-1. When he empe a u e o pa a in ises, he ac ua o pushes he lexible
s uc u e o p olong by 1.7 mm (7.6 %) and c ea es he hea conduc i e pa h.
Figu e 3: Flexible he mal s uc u e; (a) o me design; (b) design pa ame e s
The Flexible s uc u e (Figu e 3a) was ini ially made o coppe b aid solde ed o wo coppe pla es. Howe e , he
welded join and he coppe b aid p o ed o be unaccep able. A mic o-compu ed omog aphy inspec ion e ealed ha
mel ed coppe had no pene a ed in be ween he wi es and ins ead c ea ed a mechanical join . I esul ed in a low
Tempe a u e o
pa a in [°C]
Pa a in p essu e*
De lec ion
[MPa]
[mm]
20
7.8
0.122
40
12.5
0.193
60
16.4
0.251
* MHS in e nal pa a in p essu e was es ima ed based on he measu ed de lec ions; CAD
models we e analysed by ini e elemen analysis FEA a di e en loads o pa a in p essu e
o each he same de o ma ions.
Pa ame e
Cu en alue
Requi emen
The mal conduc i i y
0.365 W∙K-1
> 1.5 W∙K-1
FTS weigh
(3-pa s assembly)
55.1 g
< 55.1 g
Speci ic he mal a io
6.6 W∙K-1∙kg-1
> 27.2 W∙K-1∙kg-1
Figu e 1: Minia u ized hea swi ch schema (dimensions
in millime es)
(a)
(b)
(a)
(b)
DOI: 10.13009/EUCASS2022-6196
ADDITIVE MANUFACTURING CAPABILITIES FOR HEAT SWITCH TECHNOLOGY:
KEY CHALLENGES & KNOWLEDGE GAPS
3
measu ed he mal conduc i i y o 0.36 W∙K-1, mo e han 4- imes lowe han equi ed (Figu e 3b). Addi ionally, he
leng h o each indi idual wi e was wice he heigh o he pa and he hea was ans e ed along he wi es a he han
ac oss he poin s whe e he wi es we e in con ac [25].
2. Ma e ial and Me hods
2.1 AlSi10Mg powde
The powde AlSi10Mg was p oduced by SLM Solu ions G oup AG, Lübeck, Ge many using a gas a omiza ion in
ni ogen a mosphe e. The mo phology o simila powde was e alua ed in he s udy o V ána e al. [27] based on he
scanning elec on mic oscopy (Philips XL30 SEM, Ams e dam, The Ne he lands) and showed almos a sphe ical shape
o he pa icles. The powde pa icles had a la ge a iabili y in size wi h a mean alue o 41.4 μm and 90 % we e up o
58.0 μm as in es iga ed by he lase di ac ion analysis (Ho iba LA-960, Kyo o, Japan). Chemical composi ion de ined
by he endo is shown in Table 1.
Table 1: Chemical composi ion (w %) o SLM ma e ial AlSi10Mg [28]
Elemen
Si
Fe
Cu
Mn
Mg
Zn
O he s
Al
SLM powde
9.0-11.0
0.55
0.05
0.45
0.20-0.45
0.10
0.45
Balance
Be o e he componen ab ica ion s a ed, he me al powde was d ied o less han 5 % o esidual humidi y and
subsequen ly illed in o a ecoa e . All he samples we e buil on a pla o m (280 x 280 mm) made o aluminium alloy
Al 3.3547 ha was p ehea ed o 150 °C. A 50 μm laye was applied co esponding o he pa icle size dis ibu ion.
2.1.1 Mechanical and he mal p ope ies
The AlSi10Mg aluminium alloy me al powde was chosen o e o he ma e ials due o i s he mo-mechanical speci ic
p ope ies a e a i icial ageing. Di e en hea ea men s we e in es iga ed by Va e ka e al. [29]. Two s age hea
ea men - annealing and a i icial ageing (T6: 520 °C/ 6 hou s/ wa e quench/ 175 °C/ 4 hou s) showed highe alues
o yield s ess (YS) and on he con a y by 30 % lowe ul ima e ensile s ess (UTS) compa ed o as-build.
Table 2: Mechanical p ope ies a e hea ea men – a i icial ageing [29]
YS
UTS
Elonga ion
R.S.*
The mal conduc i i y
Rp0.2 [MPa]
Rm [MPa]
[%]
[MPa]
[W∙m-1∙K-1]
258
288
2.8
-17
163.1
* Comp essi e esidual s ess acco ding o ini e elemen (FE) analysis
2.1.2 The mal conduc i i y analysis
The mal conduc i i y was in es iga ed o SLM ab ica ed and a i icially aged samples 10x10x2 mm o AlSi10Mg
alloy. Two main su aces we e polished and ma ed wi h g aphi e powde . Two samples we e measu ed a he Ins i u e
o Plasma Physics, Czech Technology Agency (IPP-CAS) by a lase - lash me hod in acuum (Linseis LFA 1000,
Robbins ille, USA) wi h 5 measu emen s pe sample. A de ia ion o measu emen s pe one sample was ± 4 %. The
samples we e analysed by mic o-compu ed omog aphy o e eal he po osi y o 0.10 % a 10 μm esolu ion. The esul
o a e age 163.1 W∙m-1∙K-1 ma ches well he s udy o Sélo e al. [30] ha e e s o alues o 145 ÷ 173 W∙m-1∙K-1 o
samples hea ea ed by annealing.
2.2 Manu ac u ing and inspec ion
2.2.1 Selec i e lase mel ing
Addi i e p oduc ion o bo h echnology demons a o s was pe o med on an indus ial 3D p in e SLM 280HL
(SLM Solu ions G oup AG, Lübeck, Ge many) equipped wi h 400 W y e bium ib e lase YLR-400-WC-Y11
(IPG Pho onics, Ox o d, USA) wi h a ocus diame e o 82 µm and a Gaussian beam in ensi y dis ibu ion.
All pa s we e p oduced on wo pla o ms unde he p ocess pa ame e s (Figu e 4c) de eloped o achie e a
homogeneous componen , low po osi y and high p oduc i i y. Ni ogen gas low was used du ing ab ica ion and he
oxygen le el was kep unde 0.2 % h eshold.
DOI: 10.13009/EUCASS2022-6196
Jakub Mašek e al.
4
Lase powe
Scanning speed – bo de s
Scanning speed – olume
Beam compensa ion
Laye hickness
Ha ch dis ance
Beam diame e
350 W
500 mm∙s-1
930 mm∙s-1
150 μm
50 μm
150 μm
82 μm
(a)
(c) [27]
Figu e 4: SLM manu ac u ing; (a) pla o m wi h Basepla e and Flexible he mal s uc u e pa s a e SLM
manu ac u ing; (b) wo walls wi h di e en hickness - lase scanning s a egy gene a ed as de aul o hin wall
s uc u es; (c) SLM lase p ocess pa ame e s
The ‘’O se illing’’ scanning s a egy was used o he hin walls and suppo s uc u es, Figu e 4b. All pa s we e
p in ed 5 mm abo e he pla o m on suppo s consis ing o pe o a ed blocks and cones. Suppo s uc u es we e
gene a ed o suppo su aces wi h an inclina ion lowe han 35° and wi h an a ea highe han 0.1 mm2.
2.2.2 Po osi y analysis
In e nal po osi y was analysed using a mic o-compu ed omog aphy (μCT, GE phoenix | ome|x L240, GE, Wuns o ,
Ge many). By each μCT measu emen , wo Basepla es, wo Flexible s uc u es o ou he mal samples we e join ly
analysed. The linea oxel size esolu ion o 1/1000 o he la ges sample size was 60 μm o he Basepla e and 75 μm
o he Flexible s uc u e. The pos -p ocessing o econs uc ed da a was pe o med in he so wa e VGS udio
MAX 3.1, including he po osi y analysis module. In o de o inspec he in e nal in isible su aces o pa s, scanned
3D da a we e also expo ed in STL o ma o dimensional and de o ma ion analyses in GOM Inspec so wa e.
2.2.3 Dimensional and de o ma ion analysis
Visible su aces o he Basepla e and Flexible s uc u e we e digi ized by he op ical 3D scanne A os T iple Scan 8M
(GOM GmbH, B aunschweig, Ge many) wi h he op ics MV170 o unde s and he de o ma ions o pa s in as-build
and a e a i icial ageing. To success ully scan Al-alloy ma e ial ha is highly e lec i e, pa s we e ma ed by Ti02
i anium powde wi h a hickness o 2 - 3 μm. The blue-ligh scanne esolu ion was hen ± 0.005 mm. A e he op ical
measu emen , GOM Inspec so wa e was used o analyse he dimensional de ia ions om 3D CAD model.
2.2.4 Pos -p ocessing o pa s a e SLM ab ica ion
A e he p oduc ion o SLM, all pa s unde wen a sequence o pos -p ocesses: de-powde ing - inspec ion - hea
ea men T6 - inspec ion - cu -o pa s om pla o m - suppo s emo al - inspec ion and su ace ea men by sand
blas ing. Two inspec ions by he blue ligh 3D scanning we e planned be o e & a e hea ea men o e alua e he
ising inaccu acies. The suppo s uc u es we e emo ed by a combina ion o Wi e Elec ic Discha ge Machining
(Wi e cu EDM, CHMER G32S, Taiwan) and manually. Subsequen ly, he pa s we e inspec ed by he mic o-
omog aphy o unde s and he p ecision o SLM p oduc ion as well as he in e nal de ec s and in isible su aces. The
ou e su aces we e ea ed by sandblas ing (Suc ion Blas Cabine SBC 420L wi h manual blas gun) o isually uni y
he pa s. The di ec ion o beam and ime exposu e was con olled manually by using ab asi e o co undum pa icles.
2.3 Design o addi i e pa s o hea swi ch
2.3.1 Basepla e design de ini ion
The la ness o he Ho in e ace is essen ial o p o ide a con ac su ace o e icien hea ans e . S uc u al
ein o cemen encou ages couple o design solu ions, such as opology, pa ame ic o mul i-ma e ial op imiza ion. The
aim was o each he bes mass- o-s i ness a io and o demons a e he capabili ies o con en ional so wa e ools and
Con ou
Fill-con ou
(o se illing)
Possible lack
o usion
Pa h o lase scanning
(b)
Ha ch
DOI: 10.13009/EUCASS2022-6196
ADDITIVE MANUFACTURING CAPABILITIES FOR HEAT SWITCH TECHNOLOGY:
KEY CHALLENGES & KNOWLEDGE GAPS
5
manu ac u ing p ocesses. Based on he s i ness equi emen (de lec ion < 0.02 mm), he heigh o he design domain
was a i icially inc eased by 5 mm compa ed o he o me pa , see dimensions in Figu e 5.
2.3.2 Flexible he mal s uc u e design de ini ion
FTS mechanism, as pa o he Minia u ized hea swi ch echnology, ep esen s he pa h o hea ans e . I shall be a
he mally op imized s uc u e wi h a speci ic he mal conduc i i y > 2.72 W∙K-1∙kg-1. The design space (Figu e 6) was
de i ed om he p io design wi h h ee cu -ou s o pylons. Ou e diame e was enla ged compa ed o he baseline
design, aking in o accoun he new addi i e echnology applica ion and a high he mal conduc ance equi emen . The
c oss-sec ion a ea o 1400 mm2 was aken as he maximum size o he o me hea swi ch [25].
(a) (b)
Figu e 5: Basepla e non-design and design geome y (a ailable o op imiza ion); (a) BP baseline de ini ion and
compa ison o o me design (dimensions in millime es); (b) ISO iew o BP geome y
(a) (b)
Figu e 6: FTS design space a ailable; (a) design space geome y c oss-sec ion [25]; (b) ISO iew
2.4 FE nume ical models
The opology op imiza ion and ini e elemen analysis (FEA) o he Basepla e we e done using Tosca 2019 and Ansys
2019 R3. One six h symme y was used. The model con ained he hal o he pylon made om To lon and simpli ied
sc ew, all disc e ized by hexahed al elemen s C3D8R o 0.035 mm. No mal displacemen s we e cons ained o nodes
on he planes o symme y. The p essu e o 16 MPa was applied o inne su aces o he pa a in box. The nodes in
uppe su ace o he Basepla e we e ied wi h he e e ence poin (Figu e 7a) which enabled only e ical axial
displacemen . A eac ion o ce was applied o he e e ence poin in he axial di ec ion and was equi alen o he
dis ibu ed p essu e load. The pylon, sc ew and base pla e we e connec ed wi h ie cons ain . Ma e ial p ope ies a e
lis ed in Table 3.
Objec i e unc ion o he opology op imiza ion was de ined o minimize mass subjec ed o cons ain s o
de o ma ion (< 0.02 mm), maximum s ess (σ onMises ≤ 110 MPa) and manu ac u abili y (minimum wall hickness
0.3 mm o 2.0 mm p esc ibed as minimum elemen size). Res ic ion o maximum e ical ela i e displacemen
dz [mm] was p esc ibed o nodes on he Ho in e ace (Figu e 7a) acco ding o equa ion
−0.02 mm < ∆(dzmax – dzmin) < 0.02 mm. (1)
Pylon pads
Fo me design
Non-design
geome y
Pylon and sc ew
connec ion
Ho in e ace - la con ac
Design
geome y
Pa a in box cap and sealing
DOI: 10.13009/EUCASS2022-6196

Jakub Mašek e al.
6
(a)
(b)
Figu e 7: FE model se -up used o analyses and opology op imiza ion; (a) BP bounda y condi ions; (b) FTS
bounda y condi ions
Hea ans e analysis o he Flexible s uc u e was ca ied ou by Ansys 2019 R3 sol e . Rough mesh consis s o
second o de e ahed al elemen s TET10 wi h elemen size 0.5 mm. Hea lux Q [W] is p esc ibed o one o he
in e aces (10 W) and empe a u e (T1 = 15 °C) is p esc ibed o he opposi e in e ace. The a e age su ace empe a u e
(T2) is used o calcula e he mal conduc i i y q [W∙K-1] o he FTS s uc u e om empe a u e di e ence acco ding o
equa ion
𝑞 = 𝑄
𝑇2−𝑇1 [W∙K-1]. (2)
The he mal con ac wi hin he mechanism is se o bonded wi h su ace hea ans e coe icien o 2127.9 W∙m-2⋅K-1
which equals o Al-Al con ac o 0.4 MPa con ac p essu e and bo h su ace oughness o Ra 1.6 µm. Hea leakage by
adia ion and con ec ion was no conside ed.
Table 3: Ma e ial p ope ies used in opology op imiza ion and FEA analyses [28]
AlSi10Mg (addi i e)
To lon 4203®
AISI 316L
The mal Conduc i i y
[W∙m-1∙K-1]
150
0.26
16
Densi y
[g∙cm-3]
2.69
1.42
7.8
Young´s Modulus
[GPa]
70.9
4.48
210
Poisson´s Ra io
[-]
0.33
0.45
0.27
3. Resul s
3.1 Basepla e Design
Mul i-ma e ial design o Basepla e was success ully op imized, howe e was no o mee design equi emen s [26].
The e o e, ou o many ials, h ee compe i i e designs we e selec ed. Two geome y concep s a e based on opology
op imiza ion wi h di e en es ic ion o minimum wall hickness 0.3 and 2.0 mm (Figu e 8a,b) and one geome y is
based on he con en ional design o ein o cing ibs based on he designe ’s expe ience (Figu e 8c).
To imp o e he s i ness o geome y BP(2) and BP(3), ISO g id (w/h 1:3) was designed wi h heigh o 1.5 mm
and spacing 1.43 mm. The g id unde 75° angle was limi ed by 1/6 symme y and hus each line is no con inuous in
he neighbou ing sec ion. Gene a ed ISO g id can be obse ed in Figu e 8-1c (pa ially also in Figu e 8-2,3b) and was
iden ical o bo h a o emen ioned geome ies. Design pa ame e s o he ISO g id we e no op imized.
3.1.1 Fini e elemen analysis o Basepla e
Fini e elemen analyses o he design concep s we e pe o med aking in o accoun i s 1/6 symme y (Table 4). BP(1)
and BP(3) concep s mee he design equi emen o de lec ion in e ical axis, whe eas BP(2) eaches imp o emen by
oughly 12 % due o he ISO g id and has s i ness lowe by 50 % compa ed o he design a ge . Ne e heless, he
geome y could be imp o ed by he inc eased heigh o he ISO g id and by he applica ion o ein o cemen nea he
connec ion o he pylon pads o he e ical wall.
Analyses o a igue li e we e pe o med bu he s ess concen a o s we e no supp essed, p ima ily o he
unce ain y in he p e-selec ed pa a in p essu e. The e o e, low-cycle- a igue esul s a e much lowe han he equi ed
one hund ed housand. I co esponds o esul s o he s a ic equi alen s ess ( on-Mises) ha show s ong
concen a ions o s ess. Mo e speci ically, loca ed a he edges o powde holes and in a eas whe e small o none
adius ( ille adius) exis .
Re e ence poin
½ o pylon and
simpli ied sc ew
Max. displacemen
on Ho in e ace
Hea lux
Q = 10 W
T1 = 15 °C
Al-Al bonded he mal
con ac wi h 0.4 MPa
p essu e
DOI: 10.13009/EUCASS2022-6196
ADDITIVE MANUFACTURING CAPABILITIES FOR HEAT SWITCH TECHNOLOGY:
KEY CHALLENGES & KNOWLEDGE GAPS
7
BP(1)
BP(2)
BP(3)
(a)
(b)
(c)
Figu e 8: Th ee Basepla e compe i i e geome ies (a) BP(1) opology op imized wi h min. wall hickness 0.3 mm; (b)
BP(2) opology op imized wi h min. wall hickness 2.0 mm; (c) BP(3) wi h con en ional design o ein o cing ibs;
(1) suppo s s uc u es de aul se ings; (2) p e-p in modi ica ions and sup essed in e nal suppo s; (3) op ical 3D
scan esul s a e hea ea men – a i icial ageing
Table 4: Resul s o Basepla e design concep s - FEA analyses o e iew
BP(1)
BP(2)
BP(3)
The mal Conduc i i y
[W∙K-1]
2.99
2.61
2.72
Weigh
[g]
30.56
26.55
28.78
De lec ion
[mm]
0.0166
0.0299
0.0192
Speci ic de lec ion
[mm∙kg-1]
0.543
1.126
0.667
Fa igue Li e
[-]
1600
5500
25000
3.1.2 Basepla e design o be p in ed
P io o he ab ica ion, geome y was modi ied o manu ac u ing cons ain s, elimina ing closed ca i ies, oo small
ea u es < 0.3 mm and su ace angles < 35°. On he con a y, holes o powde emo al, iden i ica ion labels and
suppo s uc u es we e added ( ed a ows in Figu e 8-1,2). The holes o powde emo al and pa a in illing we e cu
in o a pa aboloid wall in geome y BP(1) and BP(2); 3 holes pe pa , adially o ien ed in a loca ion o pylon pads
(Figu e 8-2a,b). Analyses showed in his egion lowe loads and he e o e also less ein o cemen s uc u e appea s.
The cu -ou s o holes do no a ec inne ein o cemen and he e o e do no dec ease he mechanical p ope ies o BP.
Based on he o ien a ion o s u s, walls and su aces, he mos p omising o ien a ion o 0° was selec ed o
minimize he olume o suppo s needed. Gene a ed suppo s solely inside he BP design a e no emo able by de aul
(Figu e 8-1a,b). The e o e, he p ima y aim o success ul manu ac u abili y was o sup ess he necessi y o in e nal
suppo s s uc u es and a oid a powde locking inside he closed ca i ies.
3.2 Design o Flexible he mal s uc u e
Ou o many design concep s and op imiza ion ials, a wo-pa mechanism design wi h 5 con ac le els was selec ed
(Figu e 9). The design based on s ems and b anches in a “ ee shape” wi h a iable hickness is o mee he design
equi emen s. All con ac su aces shall be in con ac a he same ime, which demands s ic equi emen s o he
(1)
(2)
(3)
DOI: 10.13009/EUCASS2022-6196
Jakub Mašek e al.
8
addi i e manu ac u ing i sel and he p ecision o he su ace inishing ope a ions. The highe he su ace oughness,
he highe is he he mal con ac esis ance and he lowe he e iciency o he FTS mechanism. The build o ien a ion
has also a undamen al e ec on he su ace quali y and manu ac u abili y.
(b)
(c)
Figu e 9: FTS mechanism; (a) FTS geome y modi ica ion acco ding o su ace ea men echnology; (b)
FTS(WEDM) in 0° o ien a ion a e SLM ab ica ion; (c) FTS(EDM) in 90° o ien a ion – p epa a ion o suppo s;
(d) FTS(ECM) in 90° o ien a ion a e hea ea men – op ical scanning esul s
3.2.1 Fini e elemen analysis o FTS mechanism
Figu e 10c shows, among o he s, he speci ic he mal a io ( he mal conduc i i y di ided by weigh ) o mee he design
a ge o 27.2 W∙K-1∙kg-1. The mal conduc i i y and con ac o ce we e designed oughly 30 % highe han equi ed
howe e ha ing nega i e impac on weigh . The speci ic he mal a io ep esen s he key p ope y o each FTS s uc u e
(i.e. o he dis ibu ion o ma e ial in ho izon al and e ical di ec ion wi hin he design olume). Assuming ha he
speci ic pa ame e is cons an and ha he con ac o ce and weigh a e di ec ly p opo ional o he conduc i i y, he
FEA esul s we e ecalcula ed wi h a a ge conduc i i y o 1.5 W∙K-1. The pa ame e s dec eased by 27.9 % and a e
newly o mee he a ge alues. The heo e ical analy ical calcula ion would be hen con e ed o a eal design
modi ica ion by elimina ing some he mal s em-b anch s uc u es. The speci ic he mal a io is 10.3 % highe han
equi ed and migh mi iga e ma e ial and manu ac u ing unce ain ies.
Addi ionally, a a iable hickness o s ems inc eases he o e all pa ´s he mal conduc i i y by oughly 10.9 % by
op imiza ion o he hea lux while keeping he same weigh , see Figu e 10a,b. This e ec was no included in he
esul s o FEA p esen ed in Figu e 10c.
FTS
FTS
( heo e ical)
Weigh
[g]
69.26
49.98
The mal
conduc i i y
[W∙K-1]
2.08
1.50
Speci ic
The mal Ra io
[W∙K-1∙kg-1]
30.01
30.01
Fo ce*
[N]
1071.0
773.1
*To c ea e 0.4 MPa con ac p essu e
(a)
(b)
(c)
Figu e 10: FTS mechanism; (a) hea lux o FTS geome y cu -ou ; (b) op imized hea lux o FTS geome y cu -ou
wi h a iable hickness o s ems; (c) FEA esul s o Flexible s uc u e
S ems
B anches
Lowe and uppe
pla e (in e aces)
Machining
allowances
Suppo s – cones and pe o a ed blocks
Cu -ou s
o pylons
Technological holde s- cubes
0°
90°
WEDM modi ica ion
EDM/ECM modi ica ion
(a)
(d)
0
DOI: 10.13009/EUCASS2022-6196
ADDITIVE MANUFACTURING CAPABILITIES FOR HEAT SWITCH TECHNOLOGY:
KEY CHALLENGES & KNOWLEDGE GAPS
9
3.2.2 Flexible s uc u e design o be p in ed
To ensu e ha all 190 su aces a e a he same ime in con ac o e icien hea ans e , only h ee uncon en ional
su ace- inishing echnologies we e iden i ied as applicable – Elec ical Discha ge Machining (EDM), Wi e-EDM and
Elec o-Chemical Machining (ECM). Howe e , each inishing echnology demands i s own speci ic geome y
modi ica ion. The geome y o WEDM echnology was p in ed as one pa and subsequen ly was cu by he wi e in o
wo sec ions. On he con a y, o ECM and EDM echnologies, he mechanism was p in ed as wo pa s ‘assembled’
oge he wi h a p e-de ined clea ance (Figu e 9a). All manu ac u ing s eps a e leading o he same geome y a he end
o he p oduc ion cycle.
Fou con igu a ions o mechanism, WEDM modi ica ion in 0° and 90° (Figu e 11e) o ien a ions and EDM and
ECM in 90° o ien a ion (Figu e 11d), we e selec ed o assu e manu ac u abili y, emo al o suppo s and su ace
ea men , while conside ing he posi ion o all local su aces. Despi e he simila geome y, i was no possible o p in
he EDM and ECM con igu a ions in 0° o ien a ion due o he posi ion o he b anches in he uppe pa .
P io o he ab ica ion, each geome y was modi ied o manu ac u ing cons ain s, adding machining allowances,
suppo s uc u es, echnological holde s-cubes, cu -ou s o pylons and iden i ica ion labels o assu e ixa ion,
clamping, easie machining and acking o pa s du ing all p ocesses. Machining allowances we e added o all con ac
su aces o be subsequen ly machined (Figu e 9b,c and 11e, ).
3.3 SLM manu ac u ing
Elemen s o concu en design and manu ac u ing enginee ing we e implemen ed o sa e ime o e he de elopmen
cycle o pa s [31]. The e o e, di e en geome y-based concep s o he Basepla e and manu ac u abili y-based
modi ica ions o he Flexible s uc u e we e designed and subsequen ly manu ac u ed by SLM. P oduc ion o pa s
(Figu e 11) was oughly 8.5 hou s pe pla o m, consis ing o 480 laye s o he Basepla e and 882 o 1220 laye s o
he Flexible s uc u e acco ding o o ien a ion.
(a)
(b)
(c)
(d)
(e)
( )
Figu e 11: SLM manu ac u ed pa s; (a) BP(1) wi h 0.3 mm wall hickness cons ain ; (b) BP(2) wi h 2.0 mm wall
hickness cons ain ; (c) BP(3) wi h ibs and ISO g id; (d) FTS build in 90° o ien a ion in WEDM/EDM/ECM
modi ica ion; (e) FTS build in 0° o ien a ion in WEDM modi ica ion; ( ) FTS a e machining
3.4 Dimensional and po osi y inspec ion
An o e iew o dimensional inaccu acy p og ession a e he hea ea men is gi en in Table 5. Based on he 3D
scanning o pa s a ached o he pla o m, majo i y o de ia ions in as-build a e wi hin ± 0.10 mm, while a e he hea
ea men la ge de o ma ions we e measu ed mos ly lowe han ± 0.20 mm. Howe e , he de ia ions in speci ic hin-
wall loca ions each up o ± 0.25 mm, as in Figu e 8-3c. The pa ECM (90°) appa en ly eleased s ong esidual s esses
in o an ex ensi e de o ma ion o he op- ee end, up o + 0.19 mm in as-build and up o ± 0.70 mm a e a i icial
ageing (Figu e 9d). The e ec is alike o all pa s, bu in smalle scale. No h eshold was se -up o maximum
de o ma ion accep ance. No c ack was obse ed.
WEDM
ECM
EDM
Iden i ica ion labels
0°
90°
WEDM
DOI: 10.13009/EUCASS2022-6196