Extrusion Experiment
Austin Kaiser October 4, 2011 TA: Josh Schmale Section 304
Abstract The obecti!e o" the extrusion experiment #as to examine ho# $i""erent "actors e""ect the extrusion process% The parameters teste$ #ere $ie &eometr' (roun$ or s)uare*, lubrication, an$ extrusion $irection ("or#ar$ or re!erse* to see ho# the' e""ect the brea+throu&h "orce, $ea$ metal one sie, an$ pipe $e"ect sie o" the samples teste$% To con$uct the experiment, lea$ billets #ere place$ in an extrusion assembl' accor$in& to #hat parameters #ere bein& teste$% The assembl' #as then place$ in a tensile test machine, #here "orce #as applie$ until the specimen # as "ull' extru$e$% Throu&hout the testin& process, a $ata ac)uisition s'stem collecte$ $ata re&ar$in& "orce an$ $isplacement% The $ata #as then anal'e$ usin& -icroso"t Excel% The lab consiste$ o" "i!e main steps% The "irst step #as $'ein& an$ scorin& the lea$ billets at %100 inch inter!als, #hich #ere later use$ to !ie# the $e"ormation o" the lea$ sample% The next step #as appropriatel' placin& the billets in the extrusion assembl'% The thir$ step #as placin& the assembl' in the tensile test machine an$ settin& the proper strain rate% The "ourth step #as startin& tensile test machine an$ $ata ac)uisition s'stems% The "inal step #as remo!in& the extrusion assembl' an$ extractin& the extru$e$ billets% The lar&est "actor in $eterminin& brea+throu&h "orce #as the presence o" lubrication% .ubrication #as "oun$ to lo#er the brea+throu&h "orce b' on a!era&e 1/2 poun$s% ie &eometr' an$ the $irection o" extrusion #ere also si&ni"icant "actors in the brea+throu&h "orce, thou&h less so than lubrication% t #as "oun$ that the &eometr' o" the $ie #as the most si&ni"icant "actor in $eterminin& the sie o" the $ea$ metal one an$ presence o" pipe $e"ects% S)uare $ies &enerall' pro$uce$ lar&er $ea$ metal ones an$ resulte$ in the presence o" pipe $e"ects%
Table o" ontents
Title a&e55555555555555555%%55% a&e 1 Abstract555555555555555555%%55 a&e 2 Table o" ontents555555555555555%%%5 a&e 3 Experimental Setup 5555555555555555a&e 4 E)uipment 6se$55555555555555555%%a&e / Anal'sis55555555555555%5%%5%%55a&e 7810 onclusions5%%555555555%55555555a&e 11 Appen$ix555555555555555555%%a&e 12813
Experimental Setup & Procedure
The three parameters examine$ in this experiment #ere $ie &eometr', t'pe o" extrusion, an$ lubrication% T#o le!els "or each "actor, $irect extrusion an$ re!erse extrusion, meant ei&ht tests ha$ to be $one in total% 9irst, the billet8hal" sur"aces #ere coate$ #ith re$ la'out $'e an$ hei&ht &a&es #ere use$ to scribe cross8hatch lines 0%10 0 inches apart across the billets% .ea$ resi$ue #as then cleane$ "rom the ram, $ie, an$ chamber be"ore the billets #ere place$ in their appropriate $ie assembl'% The t#o assemblies are illustrate$ in 9i&ure 1 an$ 9i&ure 2%
Figure 1 irect Extrusion assembl' schematic
Figure 2 ;e!erse Extrusion assembl' schematic
The assembl' #as then place$ in the tensile test machine, #hich #as run a t a 3%/ rate settin&% The test machine #as stoppe$ a"ter the ram mo!e$ throu&h 0% inches or a"ter the "our sta&es o" extrusion #ere complete$% A"ter the proce$ure #as complete$ "or all ei&ht tests, the extrusions #ere examine$ in terms o" $ea$ metal one sie, amount o" pipin&, an$ brea+ throu&h "orce%
Equipment Used
The ;iehle Tensile Test -achine #as use$ to pro$uce the compressi!e "orces necessar' to extru$e the billet% The Kistler $'namometer, mo$el 0<18A, #as hoo+e$ up to the test machine an$ con!erte$ the "orce (lb"* to an electric char&e (p*% The Kistler char&e ampli"ier, mo$el /004, con!erte$ the electrical char&e to !olta&e (=*% This ampli"ie$ !olta&e #as sent throu&h an analo& to $i&ital con!erter, >ational nstruments mo$el 201, to the computer #here the .ab=ie# so"t#are, !ersion 10, e!aluate$ the $ata% See 9i&ure 3 on the "ollo#in& pa&e "or illustration%
Other E)uipment 6se$:
.a'out $'e ?ei&ht &a&e .ea$ @illets, split Extrusion ies ;ubber -allet .ubrication Shop ra&s ata ac)uisition s'stem Sa"et' &lasses isposable >itrile &lo!es
Figure 3 Schematic o" ata Ac)uisition S'stem "or Extrusion Experiment Analysis
9i&ure 4 illustrates the "our sta&es in the extrusion process in terms o" "orce !% $isplacement% Sta&e one consists o" the billet ta+in& the shape o" the chamber an$ excess air bein& "orce$ out o" the assembl'% The "orces increase $urin& sta&e t#o until stic+in& "riction is o!ercome% @rea+throu&h is #hen the billet o!ercomes "riction an$ be&ins to "lo# throu&h the $ie this brea+throu&h "orce is t'picall' the hi&hest "orce seen $urin& the extrusion process an$ it ta+es places bet#een sta&es t#o an$ three% As the billet "lo#s throu&h the $ie in sta&e three, the "orce re)uire$ b' the test machine $rops because the "riction "orces $ecrease as a result o" the re$uction in !olume that contacts the container liner% 9urther extrusion is no lon&er possible once the process reaches sta&e "our% The #or+ piece has reache$ the $ea$ metal one in this "inal sta&e in #hich the remainin& material is simpl' bein& presse$ a&ainst the $ie% Table 1 summaries the $ata an$ anal'sis "rom the ei&ht test runs%
Breakthrough Force
I
II
III
IV
Displacement Figure Braph o" the "our sta&es o" extrusion
Table 1 Summar' an$ anal'sis o" the ei&ht tests% S.O. 1
S.O. 2
S.O. 3
S.O. 4
S.O. 5
S.O. 6
S.O. 7
S.O. 8
"Random Order"
1
2
3
4
5
6
7
8
Direction
Forward
Reverse
Forward
Reverse
Forward
Reverse
Forward
Reverse
Die Geometry
S!are
S!are
Ro!nd
Ro!nd
S!are
S!are
Ro!nd
Ro!nd
!#rication
$o
$o
$o
$o
%es
%es
%es
%es
&rea' ()ro!*) Force +,#s-
13468
14/
1426
128
/4//
/57/
121
14/7
Dead 0eta, one Sie
0
S
S
0
0
S
S
0
$
0
$
0
0
mo!nt o iin*
> C >one
S C Small
1
2
- C -o$erate
3
. C .ar&e
4
9i&ure 7 Braph "rom test runs S%O%3 ("or#ar$, roun$, no lube89;>* an$ S%O%4 (re!erse, roun$, no lube8;;>*%
9i&ure / illustrates the extrusion process in terms o" "orce !% $isplacement un$er the con$itions s)uare, lube% This &raph loo+s !er' similar to the &eneral extrusion 9i&ure "or#ar$, / Braph "rom testan$ runno S%O%1
process case "rom 9i&ure 4% The "our sta&es #ere easil' i$enti"ie$, as sho#n on the &raph% The brea+throu&h "orce "or this particular sample #as 13,47 (lb"*%
n comparin& the "or#ar$ an$ re!erse extrusion processes "or the samples o" roun$ an$ no lube con$ition, it can be seen "rom 9i&ure 7 that sta&e three $i""ers "or these t#o processes% The re!erse extru$e$ sample ha$ a brea+throu&h "orce that #as 2,10 (lb"* less than the "or#ar$ extru$e$ sample an$ it also re)uire$ less "orce throu&h most o" sta&e three than the "or#ar$ sample% This is to be expecte$ because the re!erse extrusion metho$ pro$uces si&ni"icantl' less "rictional "orce than the "or#ar$ extrusion metho$% The tool stem an$ $ie push $o#n on the billet in the re!erse metho$ instea$ o" pushin& on the stationar' $ie, tool stem an$ container liner li+e in the "or#ar$ metho$% @ecause the billet $oes not nee$ to be pushe$ alon& the container liner, there is less "riction pro$uce$ an$ less o!erall "orce re)uire$ "or this process% This also explains #h' the "orce "or the re!erse metho$ $oesnDt "luctuate in sta&es three an$ "our as much as the "or#ar$ metho$ $oes the sample $oesnDt experience as much "riction alon& the container #all so the extrusion "orce re)uire$ #onDt chan&e in relation to the area o" the billet in contact #ith the container #all chan&in&% Examinin& the Statistical ata
The hal" normal probabilit' plot "or the experiment can be seen in 9i&ure o" the appen$ix% This plot helps to isolate outliers in the list o" parameters e""ectin& the extrusion% O!erall, the presence o" lubrication #as the bi&&est "actor e""ectin& the brea+throu&h "orce% O!ercomin& "riction bet#een the container #all an$ billet re)uires a lar&e amount o" "orce, the presence o" lubrication re$uces this re)uire$ "orce si&ni"icantl'% n a$$ition to the billet #all "riction, there is some "riction aroun$ the $ie% .ubrication ai$s in re$ucin& these "rictional "orces as #ell, especiall' in the roun$ $ie because it has a lar&er sur"ace area in contact #ith the billet% ;esults o" the lubricate$ samples sho#e$ an a!era&e brea+throu&h "orce o" ,47 (lb"* #here non8lubricate$ samples re)uire$ an a!era&e o" 12,474 (lb"* o" brea+throu&h "orce% The next most si&ni"icant e""ect #as the interaction e""ect bet#een the $irection o" the extrusion an$ #hether or not there #as lubrication% On a!era&e, the "or#ar$ extrusion process ha$ a brea+throu&h "orce that #as 1,30 (lb"* more than the re!erse extrusion process% This is because the "or#ar$ extrusion process pro$uces a lar&e amount o" "rictional "orce bet#een the billet an$ container #all #hich ma+es the re)uire$ brea+throu&h "orce hi&her, sho#n in 9i&ure 7% The a$$ition o" lubrication re$uces these "rictional "orces an$ as a result, the brea+throu&h "orces are $ecrease$ an$ the e""ect o" the $irection is more pre!alent% Beometr' o" the sample ha$ an e""ect on the brea+throu&h "orce as #ell% Samples usin& the roun$ $ie ha$ brea+throu&h "orces that #ere an a!era&e o" 1,113 (lb"* more than that re)uire$ "or the s)uare &eometr'% This is a result o" the sur"ace area o" the $ie the cur!e$ sur"ace o" the roun$ $ie ha$ a lar&er sur"ace area than the simple openin& use$ in the s)uare $ie% hen the billet #as extru$e$ throu&h the roun$ $ie, there #as more sur"ace area in contact #ith the billet as it passe$ throu&h this pro$uce$ hi&her "rictional "orces #hich p ushe$ a&ainst the billet an$ resulte$ in a hi&her brea+throu&h "orce%
A$$itional Anal'sis 9rictional "orces pla'e$ a maor role in the !elocit' o" the material as it passe$ throu&h the $ie% At the e$&es o" each billet, there is ero !eloc it' this is a result o" the "riction actin& bet#een the billet an$ container% The hi&hest !elocit' #as in the center o" the billet, a#a' "rom
the e""ects o" "riction% Also, the en$ o" each billet #as roun$e$ this is also because the center o" the billet ha$ a hi&her !elocit' than the e$&es% O!erall, the !elocit' pro"iles resemble$ parabolas% ictures o" the !elocit' pro"iles "or each sample can be seen in the appen$ix, 9i&ure % A $ea$ metal one is the re&ion in an extrusion #here the metal stops "lo#in& throu&h the $ie an$ the &rain $oes not appear to ha!e been stretche$ or chan&e$% At this point a$$itional "orce applie$ #ill not result in more material passin& throu&h the $ie% The s)uare $ie pro$uce$ si&ni"icantl' lar&er $ea$ metal ones than an' other parameter% The presence o" lubrication an$ the $irection o" extrusion $i$nDt seem to pla' a role in $ea$ metal sie% O!erall the s)uare, no8 lube, "or#ar$ extrusion pro$uce$ the lar&est $ea$ metal one % This is because the s)uare $ie has no transition "rom the openin& o" the $ie to the sur"ace in contact #ith the billet it is simpl' a ninet' $e&ree an&le% As the extrusion process nears completion there is less material #hich can "lo# throu&h the $ie, the remainin& material is ust bein& compresse$ bet#een the si$e o" the $ie an$ the $umm' bloc+ an$ cannot be extru$e$, "ormin& the $ea$ metal one% The roun$e$ $ie creates a cur!e "rom the openin& o" the $ie to the si$e o" the $ie in contact #ith the billet% This creates a #a' "or the material to "lo# "rom the billet throu&h the $ie as the pressure increases, resultin& in a smaller $ea$ metal one% The a $$ition o" lubrication creates less "riction bet#een the $ie an$ the billet #hich allo#s the material to "lo# more easil' an$ results in a smaller $ea$ metal one% A pipe $e"ect is a sin+ or crac+ in the center o" the billet, usuall' startin& at the $ea$ metal one an$ continuin& until the center o" the billet% ipe $e"ects are a result o" material in the center o" the billet mo!in& si&ni"icantl' "aster than the si$es o" the billet, #hich is pro!en in the !elocit' pro"iles% 9or#ar$ extru$e$ samples usin& the s)uare $ie appeare$ to pro$uce the most pipe $e"ects% This is li+el' the result o" the "act that the s)uare $ie has less sur"ace area than the roun$ $ie, #hich results in less "riction an$ less "orce #or+in& a& ainst the billet as it passes throu&h the $ie% This allo#s the billet to achie!e a "aster !elocit' pro"ile an$ $e!elop a pipe $e"ect% Conclusions
This lab "ocuse$ on the e""ects o" lubrication, $ie & eometr', an$ $irection o" extrusion on the extrusion process% t #as ma$e clear throu&h probabilit' plots that lubrication #as the most si&ni"icant "actor, #ith extrusion $irection comin& secon$ an$ $ie &eometr' last% On a!era&e,
samples that use$ lubrication ha$ brea+throu&h "orces that #ere 2/1 poun$s less than samples that $i$ not use lubrication% 9or#ar$ extrusion resulte$ in a brea+throu&h "orce that #as 130 poun$s more than the re!erse process, #hile roun$ $ie &eometr' re)uire$ 1113 more poun$s o" brea+throu&h "orce than the s)uare &eometr'% The re!erse extru$e$ sample usin& s)uare $ie an$ no lubrication pro$uce$ the lar&est $ea$ metal one% This can be conclu$e$ b' obser!ation o" the sample an$ bac+e$ up b' +no#in& #hat causes the $ea$ metal one as mentione$ earlier% Obser!ation also pro!es that &eometr' o" the $ie pla'e$ the lar&est "actor in the sie o" the $ea$ metal one% t #as also "oun$ that "or#ar$ extru$e$ samples usin& the s)uare $ie pro$uce$ the most pipe $e"ects an$ that the &eometr' o" the $ie #as the most si&ni"icant "actor in $e!elopment o" pipe $e"ects% As #ith all laborator' experiments, there #ere se!eral possible sources o" error in this test% 9irst o" all, sample #as not run throu&h the complete test, as can be !ie#e$ in the appen$ix pictures% Also, the billets mi&ht not ha!e all been uni"orm in terms o" sie an$ #ei&ht #hich coul$ lea$ to !ariation in the amount o" "orce re)uire$ "or brea+throu&h% Another source o" error coul$ be resi$ual lubrication le"t on the $ie an$ container #alls a"ter testin& a lubricate$ sample i" this #ere not cleane$ properl' it coul$ a""ect the results o" a non8lubricate$ sample% Temperature !ariation coul$ a""ect this experiment but it #as all $one at the same room temperature so the samples ha$ the same con$itions% O!erall, the experiment #as run #ell an$ pro!e$ to be a &oo$ learnin& experience% One impro!ement that coul$ be ma$e is to measure the $ea$ metal one an$ pipe $e"ects )ualitati!el'% This #oul$ lea!e less room "or bias an$ human error because as o" ri&ht no# it is ust a na+e$ e'e obser!ation%
Appendix Table 2 Extrusion Summar'
S.O. 9
D
0ain ects G
DG
nteraction ects D G
DG
&(F rom
1 2 3 4 5 6 7 8 ects
: ; : ; : ; : ; :138
SQR= +
: : ; ; : : ; ; 1113 FWD= -
: : : : ; ; ; ; :2518
; : : ; ; : : ; 361.5
N/L= -
; : ; : : ; : ; 14/1.5
; ; : : : : ; ; :2/8.5
RND= +
: ; ; : ; : : ; :258 REV= +
DS 13468 14/ 1426 128 /4// /57/ 121 14/7
LUB=
Sample alculations:
ata -anipulation8 to "in$ the brea+throu&h "orces, the e)uation F C-AG(14:123* H #as use$ "or each o" the &i!en ro#s o" "orce $ata% SE alculations8 to "in$ the e""ects, the positi!e an$ ne&ati!e si&ns #ere assi&ne$ to the correspon$in& brea+ throu&h "orce in the @T9 column, these !alues #ere then summe$ up an$ $i!i$e$ b' 4% E""ect "or : (81347I1004814270I120084I/<810210I104<*4 C 8130
9i&ure < ube plot usin& brea+ throu&h "orces as response
Half Normal Probability Plot 3000 2518
2500 s t c e f E e t u l o s b A
2000 1500 1113
1000 500 258 0
0
20
30
40
50
60
Half Normal Quanes
Figure ! ?al" normal probabilit' plot%
Figure 9 ma&es o" the ei&ht extru$e$ samples
1491.5
361.5
298.5 10
1308
70
80
90
100
