Transformer test bench -implementation and usability

Master Thesis

Tranformer Test Bench

Implementation and usability

Jonas Ayddan

Emin Özbek

-implementation and usability

Emin Özbek Jonas Ayddan

eminokth.se ayddankth.se

Supervisor: Examiner:

Claes Carrander Göran Engdahl

The task of this master's thesis was to design a transformer test ben h while fo using on safety and usability. There are several safety risks when workingwithhighvoltages,thereforethehighdemandforsafetypre autions.

A ben h was worked upon, and atta hments, su h as module frames and banana onne tors were integrated into the ben h. Further, a list of omponents were modied and installed on the new ben h. A new design, using a magneti non- onta t safety interlo k swit h and an easy-to-apply ir uitry,wasproposedandimplemented.Inadditiontotheimplementation, a ode in LabVIEW was written in order to handle the measured data from the tests that an be performed on a transformer. Several tests were performed on a single-phase transformer and a three-phase transformer in order to verify the fun tion of the transformer test ben h, with respe t to thedesign.Compli ationsemergedduetorandomadditivenoise.Aftersome work, the noise was suppressed and the results showed that theequipment worked well withthe new design.

Finally,a user's manual waswritten in order to guidethe userthrough the wire onne tions,testsand the developed softwareinLabVIEW.

En transformatortestbänk avsedd för laborationer inom forskning o h utbildningharutve klats,därfokusharlagtspåsäkerheto hanvändbarhet, eftersomarbete medhöga spänningarmedför säkerhetsrisker.

En bänk modierades för att möjliggöra fastsättning av moduler samt banankontakter. Vidare anskaades elektroniska komponenter, bearbetades o h installerades. En design med en magnetisk swit h samt lättanvända kretsar föreslogs. Dessutom skrevs en kod i LabVIEW med ett brett användningsområde. Ett ertal tester utfördes på en enfastransformator o h en trefastransformator med syfte att säkerställa att komponenter-na fungerar väl ihop med den framtagna transformatortestbänken. En del komplikationer uppstod på grund av slumpmässigt, additivt brus. Detta gi k do k att undertry ka. Det slutliga resultatet visade att kom-ponenternafungeradevälihopmeddenframtagnatransformatortestbänken.

Slutligen skrevs en användarmanual med syfte att instruera använda-re av transformatortestbänken beträande kopplingarna, testerna samt mjukvaransomär skriven iLabVIEW.

1 Introdu tion 6

1.1 Ba kground . . . 6

1.2 Safetyrisks and omponent sensitivity . . . 6

1.3 Proje tgoal andobje tive . . . 7

1.4 Proje toutline . . . 7

1.5 Stru ture . . . 8

2 Theory 9 2.1 Transformersingeneral . . . 9

2.2 Typesof transformers . . . 11

2.2.1 Step-up&step-down transformers . . . 11

2.3 Single-phase andthree-phase transformers . . . 13

2.3.1 Wye-delta . . . 15 2.3.2 Delta-wye . . . 16 3 Transformer tests 17 3.1 Open- ir uittest . . . 17 3.2 Short- ir uittest . . . 18 3.3 Transformer losses . . . 19 3.3.1 Corelosses . . . 20 3.3.1.1 Hysteresisloss . . . 21 3.3.1.2 Eddy- urrent loss. . . 23 3.3.2 E ien y oftransformers . . . 25 4 Implementation 26 5 Alternative power sour e 36 6 LabVIEW 39 7 Results 39 7.1 Cal ulations . . . 39 7.1.1 Rated urrent . . . 40 7.1.2 Ratedvoltage . . . 40 7.1.3 Ratedpower . . . 41 7.2 Testresults . . . 41 7.2.1 Single-phase . . . 41 7.2.1.1 Open- ir uittest . . . 42

7.2.2 Three-phase . . . 45

7.2.2.1 Open- ir uit- delta . . . 45

7.2.2.2 Open- ir uit- wye . . . 48

7.2.2.3 Short- ir uit -delta-wye . . . 50

7.2.3 Loadlossand short- ir uitimpedan e . . . 51

7.3 Analysis . . . 52 8 Future work 52 9 User manual 54 9.1 Software . . . 54 9.2 Hardware . . . 63 9.3 Testinstru tions . . . 75 9.3.1 Single-phase transformer . . . 76 9.3.1.1 Open- ir uittest . . . 77 9.3.1.2 Short- ir uit test . . . 79 9.3.2 Three-phasetransformer . . . 80

9.3.2.1 Open- ir uittest - Delta. . . 81

9.3.2.2 Open- ir uittest - Wye . . . 84

9.3.2.3 Short- ir uit test . . . 86

The revolution in thete hnology behindtransformers started in the1830s, whenMi haelFaradayand Joseph Henry dis overed theproperty of indu -tion, independently. Faraday used an iron ring on whi h he wrapped two wireson oppositesides.He then pluggedone wireinto a galvanometer, and onne ted the other wire to a battery. This led to a transient urrent that o urred due to hange in magneti ux when the battery was onne ted andthendis onne ted, repeatedly[1℄.

About 50 years after the dis overy of indu tion, the rst transformer wasinvented. Nowadays, the transformeris a part of our everydaylife and omes invarious sizes andshapes.

1.1 Ba kground

Working with high voltages requires extreme aution. However, this is not ommon knowledgeto everyone. Thus, asafe environment is ane essityfor theworker, to prevent him/her fromgettinginjured.

This proje t was undertaken in order to solve the safety issues when working with high voltage equipment and sensitive ele troni omponents. The user must be prote ted from getting in dire t onta t with the high voltages,andtheele troni omponentsmustbeprote tedfromover urrent inorderto prevent them frombreaking.

1.2 Safety risks and omponent sensitivity

Whenworkingwithele tri alequipment,pre autionsmustbetaken.Ifnot, it an resultinserious injuriesor even death.Thefollowing table showsthe probable ee tsofgetting in onta t withACand DC urrents[2 ℄.

0-1 0-4 Per eption

1-4 4-15 Surprise

4-21 15-80 Reex a tion 21-40 80-160 Mus leinhibition 40-100 160-300 Respiratoryfailure Over100 Over300 Usually fatal

Table 1:AC/DC urrentsand theiree ts

It isimportant to notethat table 1 shows theprobable ee ts. Thismeans thatthe a tual ee tmay be dierent asitdependsonthe ondition ofthe environment, the ondition ofthebody,andtheareaof thebodywhere the urrent ows.However, physi al onta t withele troni equipment mustbe preventedwhen thepoweris on.

The transformers, and the other ele troni omponents should be pro-te ted as well. The transformer may break due to overload if exposed to a loadthatislargerthanwhatitisdesigned for[3℄.Theele tri al omponents thatwill be used in this proje t are very sensitive and expensive, and may break easily due to overvoltages. Thus, it is important to be autious and try not to overload the transformer, or expose the other omponents to overvoltages.

1.3 Proje t goal and obje tive

Thisthesisproje t onsistsof reatinga transformertest ben h,withfo us on usability, safety, and tidiness. The test ben h is supposed to be used in theETKdepartment at KTH,bystudentswho mayor maynot befamiliar withtransformersand possiblerelatedequipment.

1.4 Proje t outline

Theproje t isgoingto start owitha pre-study, inorder to graspthe fun-damentals oftransformers and relatedhardware. Next,a oupleof testsare goingtobeperformedonasingle-phase transformerfor purposeoflabelling its dierent properties, and to determine the losses. After that, a layout of thetransformertestben hisgoingtobe reated.On ethelayoutis reated, the transformer test ben h is going to be onstru ted and veried. Several tests are going to be performed on the transformers in order to verify the

in order to log the data that is a quired when performing the tests. On e the transformer test ben h is veried, a user manual is going to be writ-ten. The user manual is going to instru t the user of the transformer test ben h through the hardware, the software, and the tests. The user manual willdes ribethehardwareofthetransformertestben h.Additionally,itwill have instru tionsfor theuserwhenusingthesoftwareand performingtests.

1.5 Stru ture

The report onsistsof 8 se tions. Se tion 1 overs the history of the trans-former, the aim of the proje t and the outline of the proje t. In se tion 2, an analyti des ription of the transformer is presented for a deeper un-derstanding. Se tion 3 des ribes the transformer tests and al ulations of thetransformerlosses.Inse tion4,theimplementationisexplainedandthe layout isstudied.Analternative powersour eispresentedinse tion5.The softwareismentionedinse tion6,followedbyresultsandanalysisinse tion 7.Futureworkisdis ussed inse tion8.Finally,ausermanual ispresented.

Thisse tionprovidesinformationaboutsingle-phasetransformersand three-phasetransformersingeneral.Thevariableautotransformerwillbedes ribed inse tion4.

2.1 Transformers in general

Transformers are stati ele tromagneti devi es. They onsist of two or more oils where ele tri al power from one oil, at a given frequen y, is transformed to the other oil,at thesame frequen y [4℄.

A transformer an in rease or de rease the voltage, depending on the oils' turns ratio. This results in a de rease or an in rease of the urrent, respe tively [3℄.

The prin iple of mutual indu tion states that if two oils are indu tively oupled,andthe urrentinone oil hangesuniformly,thenanele tromotive for e(e.m.f.)willbe indu edintheother oil.Thus,ifthese ondary ir uit is a losed path, the indu ed e.m.f. will drive a urrent in the se ondary ir uit[4℄.An alternating urrent (AC)supplies theprimarywinding ofthe transformer,thusgeneratinganACmagneti uxinthe ore. Themagneti ux then generates an AC voltage in the se ondary winding, and thus the indu ede.m.f.inthese ondarywindingdrivesa urrentinbothwindings[3℄.

The oils of a transformer an ea h be wound either lo kwise or ounter- lo kwise,determining the voltage polarity[3 ℄. Figure1 shows thedierent polarities ofa transformer.

Ifboth windingsarewoundinthesamedire tion,thentheprimary andthe se ondary voltage will be in phase. However, ifthe windings are wound in oppositedire tions, thentheprimary and these ondaryvoltage will beout ofphase by180

◦

.

If two transformers are onne ted in parallel, the voltages and the polarities inboth theprimary andthe se ondary windingsmust beequal in both transformers inorderto avoid ir ulating urrentsintransformers.

For a series onne tion of the windings, the voltages will add up if the polarities are the same, and subtra ted if they are opposed to ea h other. Hen e thepolarityis ofgreat importan e when workingwith several transformers.

There are several types of transformers. However, this se tion only overs thebasi sofsingle-phase transformersand three-phase transformers,where ea hofthem anbeeitherastep-uptransformerorastep-downtransformer.

2.2.1 Step-up &step-down transformers

Thedieren e between a step-up transformer and a step-down transformer is the turns ratio, i.e. the ratio between the number of turns in the primary (supplied side) oil and the number of turns in the se ondary (output side) oil. The primary oil of a step-up transformer onsists of fewer turns of wires than the se ondary oil. This results in an in reased voltage, and thus a de reased urrent. The primary oil of a step-down transformer, however, onsists of more turns of wires than the se ondary oil, thus resulting in a de reased voltage and an in reased urrent. Fi-gures2and3showtherelationbetweentheprimaryandthese ondary oils.

Figure 3:Step-uptransformer

Theoutputvoltage isdetermined by

V

s

=

V

p

N

(1) and

N =

n

1

n

2

(2)

where

V

p

isthevoltageonthesuppliedside,

n

1

isthenumberofturnsinthe primary oil,

n

2

is the number of turns in these ondary oil,and N is the turnsratio[5℄.However,equations(1)and(2)applyonlyinideal onditions. In pra ti e, there is always a voltage drop. The transformers that are used inthisproje toperatewitha voltage drop of2-7%, seese tion7.2.3.

A single-phase transformer onsists of two or more windings that are used to transform an input voltage to an in reased or de reased output voltage. A three-phase transformer, however, onsists of at least six windings and works dierently in omparison to a single-phase transformer. The three-phase transformer an be onstru ted either by using three single-phase transformers onne ted to ea h other, or by using an E-I shaped ore with two windings (primary and se ondary oil) on ea h leg. The E-I ore an be used for smalltransformers. The orethat isused for larger three-phase transformersis alledathree-limb ore.Figure 4shows an E-Ishaped ore.

Figure4: E-Ishaped transformer ore

Thereareseveralwaysto onne tthewindingsofathree-phasetransformer. Themost ommonve torgroupsaredelta(

∆

),wye/star(Y), andgrounded wye/star (YN).The dieren e between theve tor groups Yand YN isthat the neutral point, denoted N in gure 5b, is grounded in a YN, while it is oating inY.

The hoi e of transformer ve tor group is usually based on the equip-ment that the transformer is supposed to work with. However, this report only overs the basi s of the wye-delta onguration (oating or grounded wye) and the delta-wye onguration (grounded wye), sin e these arevery ommon ongurations.

An ungrounded primary winding, su h as the delta or oating-wye, may ontribute to ferroresonan e (an overvoltage phenomenon). However,

A grounded primary winding may provide ground fault urrent, but ferroresonan e is not as likely to o ur with a grounded primary winding as it is with an ungrounded primary winding. Thus, there is a trade-o between ferroresonan e and ground fault urrent. Further, a grounded se ondary winding an handle loads with less on ern about unbalan es, in omparison to an ungrounded se ondary winding [6 ℄. However, there are more advantages and disadvantages of grounded and ungrounded onne tionsthan what arestated inthis report.

In gure 5 below, the delta and the wye onne tions are shown, re-spe tively [7℄.

a

b

N

N

gurations.However,thisthesiswillonly overthebasi softhewye-delta

(Y-∆

)andthedelta-wye(

∆

-Y) onne tions.Thefollowing subse tionsdes ribe the ir uitdiagrams.

2.3.1 Wye-delta

Ina wye-delta onne tion, theprimary (supplied) side is onne ted inwye, andthe se ondary(output) side is onne ted indelta,asshown ingure6. Theneutral point ofthe wye onne tionis either groundedor oating.

Figure6:Wye-delta onne tion

In a wye-delta onne tion, the primary winding, AN, and the se ondary winding,ab,areonthesameleg[3℄.Thisresultsin

V

AN

beinginphasewith

V

ab

,and

I

A

being inphase with

I

ba

,where:

• V

AN

,

V

BN

and

V

CN

arethe primary line-to-neutral voltages of phases A,BandC, respe tively

• V

an

,

V

bn

and

V

cn

arethese ondaryline-to-neutral voltagesofphasesa, band ,respe tively

• I

A

,

I

B

and

I

C

are the primary line urrents in phases A, B and C, respe tively

• I

a

,

I

b

and

I

c

are the se ondary line urrents in phases a, b and , re-spe tively

• V

AB

,

V

BC

and

V

CA

aretheprimaryline-to-linevoltagesbetweenphases A&B, B&Cand C&A,respe tively

• V

ab

,

V

bc

and

V

ca

arethese ondary line-to-linevoltages between phases a&b,b & and &a,respe tively

• I

AB

,

I

BC

and

I

CA

aretheprimaryline-to-line urrentsbetweenphases A&B, B&Cand C&A,respe tively

• I

ab

,

I

bc

and

I

ca

are the se ondary line-to-line urrents between phases a&b,b & and &a,respe tively

2.3.2 Delta-wye

In a wye-delta onne tion, the supplied side is onne ted in delta and the outputside is onne ted inwye,asshowningure 7.

Figure7:Delta-wye onne tion

In a delta-wye onne tion, the primary winding, AB, and the se ondary winding, an,areonthe same leg, whi hmeans thattheprimary line-to-line voltage,

V

AB

, and the se ondary line-to-neutral voltage,

V

an

, are in phase. Furthermore, the primary line-to-line urrent,

I

AB

, and the se ondary line urrent,

I

a

,areinphase.

Transformersneedto be testedinorder tomeasuretheire ien y, determi-netheirlosses anda quireseveralparameters tolabelthem.Inthisproje t, asingle-phasetransformerandathree-phasetransformerareused.Thetests that are performed on the transformers are alled open- ir uit test (or no-loadtest)and short- ir uit test (orimpedan e/load test). Usually,the low-voltagesideissuppliedinanopen- ir uittest,whilethehigh-voltage sideis suppliedinashort- ir uit test.Thefollowing subse tions explaintheir pur-pose, andthe usermanualinse tion9.3inthisreport ontains instru tions ofhowto makethe onne tionsfor the tests.

3.1 Open- ir uit test

The main purpose of this test is to determine the ore losses.A voltmeter, anammeter,andawattmeter are onne tedonthelow-voltagesideinorder to measurethese quantities,and to measurethe inputvoltage,

V

OC

.Figure 8belowshows the ir uit.

P

N LL

=

1

T

Z

T

0

u(t) ∗ i(t)dt

(3)

T is the period size. The no-load loss onsists of hysteresis losses, eddy- urrent ( lassi al eddy- urrent)lossesand ex ess-eddy urrent losses.These lossesare des ribed further inse tion3.3.1. However, theequation above is su ient for al ulatingthetotal no-load losses.

The no-load urrent is the amount of urrent that the transformer onsumes at rated voltage and frequen y, regardless of whether there is a load or not on the other side. The no-load urrent ontains information about the transformer ondition [8 ℄. It is very low, about 1 per ent of the full-load urrent for alarge transformer.

When performing an open- ir uit test, it is possible to supply the transformer from the high-voltage side. However, the no-load urrent may be too small to measure. Supplyingthe high-voltage side is only appli able to small transformers, and only with the spe i ir uit that is des ribed. Whi h side to supply depends on the voltage interval that is of interest to the user.

3.2 Short- ir uit test

The purpose of the short- ir uit test is to determine the short- ir uit im-pedan e of the transformer, and the opper loss, or load loss [4℄. This in-formation is important in order to verify the ondition of the transformer ore.Ina short- ir uittest,thelow-voltage windingisshort- ir uitedand a ertainlevelof voltageisapplied tothehigh-voltage sideinordertogetthe full-load urrent to ow in the high-voltage side and the low-voltage side. Thethree important parameters, whi hare urrent, voltage and power, are allmeasuredonthehigh-voltageside[3℄.Theshort- ir uittestisalsoknown asfullloadtest [9℄.The ir uitis showningure9.

Theabsolutevalueofthe impedan e is al ulatedbytakingtheratio ofthe voltage and urrent:

Z

SC

=

V

SC

I

SC

(7)

However, the short- ir uit impedan e is usually spe ied in per ent, using the following equation:

Z

SC

=

z

sc

100

U

2

m

S

m

(8) where

• z

sc

isthe per entage short- ir uitimpedan e [dimensionless℄,

• U

m

isthe ratedvoltage [V℄ and

• S

m

istherated power[VA℄. 3.3 Transformer losses

Thelosses in a transformerare dividedinto two parts: ore losses and load losses.The orelossesemergefromhysteresisandeddy- urrentsinthe ore,

lossisdetermined fromtheshort- ir uit (SC) test,being

Load loss = P

LL

= I

rated

2

∗ R

winding

(9)

where

R

winding

is the winding resistan e. Thisapproa h,however, an only be used for asmall transformer, where theeddy- urrents inthe opper an benegle ted.Forlargertransformers,theloadlossneedtobedetermined by the measurements, sin e it is di ult to derive the load losses analyti ally forlarger transformers.

Thewinding resistan eis al ulated using equation10.

R

winding

=

ρ

Cu

∗ ℓ

winding

A

cross

(10)

ρ

Cu

istheresistivityfor opper,

ℓ

winding

isthewindinglength,and

A

cross

is the ross-se tionalarea.Thewindinglength,

ℓ

winding

,dependsonthenumber ofturns.The windingsthat areusedinthisproje thave500 turns and100 turns.Table2shows the windinglengthand the ross-se tionalarea forthe two windings. The resulting load losses for ea h winding are presented in se tion7.2.3.

Winding: 100 turns 500 turns Lengthperside: 62

mm

75

mm

Total windinglength: 24.8

m

150

m

Cross-se tionalarea: 5.6

mm

2

1.12

mm

2

Table 2:Windingparameters

The resistivity depends on the temperature; for instan e, theresistivity at

350Kis

ρ

Cu

= 2.06 ∗ 10

−8

_Ωm

.

The ore loss, however, is determined from theopen- ir uit(OC) test. The ore lossis more ompli ated to derive than the load loss, and will thus be derived inthefollowing subse tion.

3.3.1 Core losses

Ferromagneti materials experien eenergy losseswhensubje ted to magne-ti uxes whi h varywithrespe t totime. Thisisnot the asefor magneti ir uits of devi es in whi h there isa onstant magneti ux. However, the

losses (oriron losses). The orelosses are dividedinto hysteresis lossesand eddy- urrent losses,seeequation11.

P

core

= P

h

+ P

e

(11)

3.3.1.1 Hysteresis loss

TheB-H hara teristi followsdierentpathswhenin reasingandde reasing the values of the magneti eld strength, H, as shown in gure 10. This resultsinhysteresisloss. CarryingHfrom

+H

max

to

−H

max

andthenba k to

+H

max

forms a hysteresisloop,asshown ingure10a [10℄.

to

H

max

result inanabsorbed energy equal to

W

1

=

Z

B

max

−B

r

H dB

(12)

Theenergyisrepresentedbythearea1-2-4ing10b.De reasingHba kto zero, asshowningure 10 ,will release energy

W

2

represented by thearea 2-3-4.

W

2

=

Z

B

r

B

max

H dB

(13)

De reasing H further, to

−H

max

, following the path 3-5, one a quires an absorbed energy bythe orerepresentedbythearea3-5-6,and

W

3

=

Z

B

r

−B

max

H dB

(14)

Sin e H and dBare both inthe third quadrant and thus negative, positive energy is a quired. When H is in reased ba k to zero, the ore givesup an amount ofenergy

W

4

represented bythearea 5-6-1.

W

4

=

Z

_−B

max

−B

r

H dB

(15)

Inequation15,Hisnegative anddBispositive,thustheabsorbedenergy is negative,meaningthatthe oregivesupenergy.Thesumofthesefourvalues representsthetotalenergyandequalstheareaofthehysteresisloopingure 10a.Thus,the energy lossfor a ore, duringone y le, witha volume, Vol, anda uniform uxdensity,B, is

W

h

= V ol

I

H dB

(16)

where thelineintegral representsthearea ofthehysteresisloop.

The hysteresis loss, in a volume, Vol, in whi h the ux density is uni-form and varies y li ally at a frequen y of f Hz, is expressed empiri ally as

P

h

= η ∗ V ol ∗ f ∗ B

max

n

(17)

where

η

andnarebothdeterminedbythenatureofthematerial [11 ℄. Equa-tion17 isnot validfor unsymmetri al loops.

Eddy- urrentsareele tri urrentsthatareindu edinalaminationofa on-du tingmaterial, whenthematerial issubje tedtoan alternating magneti eld[12 ℄.Figure 11shows a pie eof laminationwhere thethi kness, height andwidthare

τ

,hand

w

,respe tively.Bistheuxdensityandis onsidered to be uniform throughout the entire lamination. Considering that the ux varies over time, an e.m.f. is indu ed inthe path 1-2-3-4 (see gure 11) by the uxlinking thispath, andthe indu ed e.m.f. be omes

e =

dφ

dt

= A ∗

dB

dt

= 2hx

dB

dt

(18)

where A is the area of the entire path. The approximation A=hx is due to

τ

beingsmall in omparison toh.

variessinusoidally[10℄. Let

B = B

m

∗ sinωt

(19)

where

ω

=2

π

f, f being the frequen y in hertz. The voltage indu ed in the path be omes

e = AB

m

d

dt

sinωt

(20)

This produ es eddy- urrents in the path. Sin e the resistan e of the path isrelatively high, the eddy urrent isnearly inphasewith thevoltage. The resistan eis

R =

ρL

A

=

2hρ

wdx

(21)

where L=2h is the length of the urrent path, and A=

w

dx is the ross-se tional area of the urrent path. The length, L, is approximated to 2h be ause ofthethi kness

τ

being small.

Dividing equation 20 by equation 21, the urrent in the dierential path isobtained as

di =

ωB

m

wxcosωtdx

ρ

(22)

Theinstantaneous power onverted into heatis

dp

ce

= e di =

ω

2

B

2

m

hwx

2

cos

2

ωtdx

ρ

(23)

wheretheindex e standsfor lassi aleddy- urrents.Theinstantaneousloss inthe laminationbe omes

p

ce

= (

ω

2

_B

2

m

hwcos

2

ωt

ρ

)

Z

x=

τ

2

x=

−

τ

2

x

2

dx = (

ω

2

_B

2

m

hwcos

2

ωt

ρ

) ∗

τ

3

12

=

ω

2

B

2

m

hwτ

3

cos

2

ωt

12ρ

= {hwτ = V ol} =

V ol ∗ ω

2

B

2

m

τ

2

cos

2

ωt

12ρ

(24)

Equation 24 expresses the instantaneous power. However, the quantity of interestis theaverage power, whi his

P

ce

= (

1

π

)

Z

ωt=π

ωt=0

p

ce

dωt = (

V ol ∗ ω

2

B

2

m

τ

2

12ρπ

)

Z

ωt=π

ωt=0

cos

2

ωt dt =

(

V ol ∗ ω

2

B

2

m

τ

2

12ρπ

)h(

1

2

)

Z

ωt=π

ωt=0

(1 − cos2ωt)dωti = (

V ol ∗ ω

2

B

2

m

τ

2

12ρπ

) ∗

π

2

=

V ol ∗ π

2

f

2

B

2

m

τ

2

6ρ

(25)

A transformer ore onsist of thin sheets in order to de rease the eddy- urrents, and thus de reasetheloss.

The hysteresis losses and the ( lassi al) eddy- urrent losses do not a - ount for all the ele tromagneti losses in a ferromagneti material [13 ℄. Besides the lassi al eddy- urrent losses des ribed above, there are other dynami losses alled ex ess eddy- urrent losses. The ex ess eddy- urrent losseso ur due to imperfe tions inthe ore. The total eddy- urrent lossis given by

P

e

= P

ce

+ P

ee

(26) where

P

ee

= V ol ∗ 8

s

Gτ wH

0

ρ

B

3

/2

m

f

3

/2

(27)

with

G

=0.1356 and

H

0

[A/m℄ being a magneti eld whi h represents the distributionoftheinternal potentialthatisexperien edbythedomainwalls [14 ℄[15 ℄.

3.3.2 E ien y of transformers

Thepower e ien y ofa transformerisdetermined by

Ef f iciency =

Output

Input

=

Input − Losses

Input

= 1 −

Losses

Input

(28)

A ording to National Ele tri al Manufa turing Asso iation (NEMA) the e ien yofadrytype,low-voltagetransformerliesbetween97%-98.8%[16 ℄. Larger transformers are usually more e ient than smaller transformers, withane ien y up to 99.75% [17 ℄.

The transformertest ben h, whi h was usedbefore this thesis proje t, had a lot of aws; the ables were entangled, the devi es and omponents were hard to rea h and the environment was not safe for a user who might not be familiar withtransformers. Overall, itwasmessy.The gures 12 and 13 below shows thetransformertestben hbeforethenew design.

Figure12:Previous transformertest ben h

transformer and the three-phase transformer via a breadboard. The single-phase transformer and the three-phase transformer are further onne ted to a voltage divider (bla k box), whi h is in turn onne ted to a data a quisition (DAQ) ard. The DAQ ard is a module used to a quire data throughmeasurements, whi h isexplained further onpage 30.

Sin e the transformer test ben h is supposed to be used by students, whomayormaynot bepro ient engineersintheareaof ele tromagneti s, thesafety for the users is priority number one. This was, however, not the aseinthe previous state ofthetransformer testben h.

For the new transformer test ben h design, a ben h with the dimen-sions seen in gure 14 below was pur hased and modied. The plan is to utout two re tangular po kets, 26 m x 5 m, one at the left side and one atthe right sideof theben h.The purposeof thisisto atta h aframeof12 modules, one frame for ea h po ket. Ea h module will be onne ted, from underneath the ben h, to one of the banana plug onne tors. The banana plug onne tors are also atta hed to the ben h and an be seen as small white ir les in gure 14. These banana plugs areused as bridges between thetransformerwindings and the modules. Thus, thetransformerwindings are onne ted to the frame of modules underneath theben h, leaving more freespa efor theuser.

Beneath ea h frame of modules, there is a single module atta hed to theben h, seegure 14.Thesemodules aregrounded.

Below the top enter frame of modules, there are four other re tangu-lar po kets, 6 m x 4 m. The purpose of these po kets is the pla ement of urrent probes. The urrent probes will be pla ed halfway into the ben h (thewidthoftheupperpartoftheprobesarewiderthan6 m).The urrent probes' battery wires will be onne ted to the battery eliminator, and the oaxial ables willbe onne ted tothe ir uitboard,underneath theben h. Thelarge, white ir le, above one of thebla k re tangles, represents a hole whi h is usedfor this purpose. Thisway,there will bea lot offree spa eto use.

The grey boxes in gure 14 below represent the omponents that will be pla ed ontheben h.

A over, onsisting of a metal grid, is atta hed to the top of the ben h, as shown ingure 15.One part of a magneti swit h is pla ed on the bottom

onne tedtoarelay,ispla edonthetop enteroftheben h(arrowpointing up). The grid and the swit h are used for user prote tion. How the swit h workswill be explained later inthisse tion.

Figure 15:Transformer test ben h witha gridon top

Theequipment that is going to be usedtogether withthesingle-phase and the three-phase transformers onsistsof:

•

onevoltage divider,

•

four urrent probes, HamegHZ050

•

onebattery eliminator,

•

oneDAQ ard, NI-9205

•

one ontroller hassifor theDAQ ard, RIO-9074

•

onepowersupplyfor the ontroller,

•

one ir uit box,

•

onevariableautotransformer, and

redu ing the input voltage, making the output voltage only a fra tion of its input voltage. A urrent probe is an ele tri al devi e that measures the urrent ina ondu tor.

A data a quisition (DAQ) ard is a ard that a quires data by mea-suring ele tri al and physi al phenomenon su h as voltage, urrent, power and temperature [18 ℄. The a quireddata isfurther sent to the omputerto be analyzed and logged. Thus, theDAQ ard a ts asthe interfa e between the ir uitryonthe transformertestben h,anda omputer.Therearethree key omponentsofa DAQ ardusedfor measuringsignals.Thesearesignal onditioning ir uitry,analog-to-digital onverter(ADC)and omputerbus.

Signal onditioning ir uitry uses tools su h as ampli ation and at-tenuation to make a signal suitable for the ADC. This is ru ial sin e a signalfrom thetransformertest ben h ould be too noisyor toodangerous tomeasuredire tly,andmaybreak theDAQ ard. ManyDAQdevi es have built-in signal onditioning. If the DAQ devi e does not have a built-in signal onditioning, however, thenit an be applied externally by designing a ir uit to pro essthe signal, or byusingother devi es[19℄.

The ADC hip in the DAQ ard gives a digital presentation of an analog signal by periodi ally sampling the analog signal. The samples are then transferredtoa omputerthroughthe omputerbus,andthenre onstru ted.

The omputer bus works as an interfa e between the DAQ ard and the omputerto transferthemeasured data.

The ontroller system for the DAQ ard is alled Compa tRIO. Com-pa tRIO, or RIO, is a re ongurable embedded system. It ontains three key omponents: a pro essor that runs a real-time operating system (RTOS), a re ongurable eld-programmable gate array (FPGA), and inter hangeable I/Omodules(DAQ ards) [20 ℄.

The real-time ontroller is shown in gure 16. It ontains a pro essor that is able to exe ute LabVIEW real-time appli ations and oers servi es su hasmultirate ontrol and datalogging,among others.

Figure16: Real-time ontroller

An RTOS reliably exe utes programs for whi h time is of the essen e. The dieren e between a real-time OS and a general-purpose OS (su h as Windows) is that they are optimized for dierent purposes. A general-purpose OS runs many appli ations simultaneously and might delay the exe ution of a user program. This delay might ause system failure when running a measurement or a ontrol program whi h needs to run uninterrupted at a ertain rate. Thus, the general-purpose OS is not preferred when ontinuity is of importan e. A real-time OS, however, fo uses on a single program and performs the task in pre ise timing. The real-time OS is optimized in su h a way that it is possible to run multiple pro esses where the user an prioritize a ertain pro ess over the other pro esses. Real-time OS an be set up to run reliably for years. If the user program would stop running for some reason, then the wat hdog featurewhi hisin ludedinthesystem,automati allyrestartsthe omputer.

FPGAs are reprogrammable hips where dierent ir uitries an be programmed [21 ℄ [22 ℄. The FPGA hassis is shown in gure 17. Ea h moduleis onne teddire tlytotheFPGA,givingalmostno ontrol laten y for system response. The modules are dire tly onne ted to the FPGA, and this makes it possible for the programs on the real-time ontroller to a esstheI/O(Input/Output)withlessthan500nstime-lagsbetweenloops.

and pre ise timing. Sin e FPGAs are reprogrammable, it is possible to hange the program wheneverdesired.

Figure17:FPGA

A non- onta t safety interlo k swit h is a swit h onsisting of two pie es thatoperate magneti ally.Separatingthesetwopie esappliesastopsignal, thus removing the a tion of the a tuator from the swit h. This results in isolatingthepowerfrom thema hine/devi e [23℄.

An autotransformer has only one winding that links the primary side andthe se ondaryside together, ele tri ally and magneti ally[24 ℄ [25℄ [26 ℄. Thewindingistappedonseveralpointstoprovideseveralvoltageratios[24 ℄. A variableautotransformer (orvaria ), however, uses a xedAC supplyto produ evariablevoltages. The onstru tion isthesamefor thevaria asfor the autotransformer, with thedieren e that the varia doesnot have any predetermined tapping points. Instead,the se ondary voltage of the varia is tapped through a arbon brush that is able to slide along the primary winding. The arbon brush ontrols the length of the se ondary winding. Adjusting the length in reases/de reases the voltage applied. This is very useful,sin e the arbonbrushallows tuningoftheapplied voltage byvalues ofapproximately one volt. [26 ℄.

Figure18:Variableautotransformer

In some varia s, the supply voltage is tapped at some point along the primary winding. In these ases, it is possible to get a se ondary voltage thatisevenhigher than thesupplyvoltage [24℄.

The nal design of the transformer test ben h is shown in gure 19. There are several omponents that are pla ed on the ben h. The ompo-nents arelistedbelow.

1. RIO 2. DAQ ard 3. Powersupply 4. Cir uitboard 5. Batteryeliminator 6. Voltage divider 7. Current probes

8. Load (notne essarilyused)

1

2

3

4

5

6

7

8

9

Figure 19: T ransformer test b en h 35

By utilizing an amplier as a power sour e instead of a varia , it would be possible to ontrol and make adjustments to the waveform of the signal being sent [27℄.

With the use of an analog output module, it is possible to generate signals together with e.g. LabVIEW. With LabVIEW, it is possible to adjust the generated output signals. Analog output modules are not able to generate signals with high voltages or powers. This an however be over ome by onne ting the analog output module to a power amplier, whi h ouldbringe.g.a5Vgenerated signaluptoa100Vsignal,depending onthe hara teristi s ofthe power amplier.

An operational amplier is shown in gure 20. The operational amp-lier is ommonly denoted as op-amp. If the op-amp is used in ertain ir uits, the voltage will be amplied by a voltage gain, examples of these ir uits are provided on page 38.This implies that an input voltage V an be in reased or de reased by a multipli ative voltage gain and produ e a higher or lower output voltage. The minus and the plus sign identify the inverting andnon-inverting input, respe tively.However,ampliers analso beusedtoboosttheinputpower,the ele tri urrentortoboostfrequen ies available inthe inputsignal[28℄.

Figure20: OP-amp

E

+

isthe positive powersupplynode and

E

−

isthenegative power supply. They are used to bias the op-amp, whi h an be done to a hieve ertain

highinputresistan e andan outputresistan e equalto zero.

The input resistan e is lo ated between V- and V+ (see gure 20), and sin e it is innitely high, there will be no urrent owing between V-andV+.There isapossibilityto limitthe magnitudeoftheoutputvoltage, whi h ould be done byusing the powersupplynodes.

E

+

V

_out

V

+

- V

-

E

-Figure 21:Voltage inputversus voltage output

Figure 21 above indi ates that if there is a voltage dieren e between the inverting and non-inverting inputs, whi h is not equal to zero, the output voltage will be for ed to either the plus state or the minus state [30 ℄. The steeptransitionfrom

E

−

to

E

+

indi atesthatthevoltagegainisinnitehigh. The ideal voltage gain is not always the ase, and in order to a quire aspe i voltage gain, equation29 an be used.

A

v

=

V

out

V

in

(29)

A

v

isthe voltage gain.

To a quire a nite voltage gain, the onguration of the omponents onne ted to the amplier must be addressed; these omponents an be seeningure22 below, where

R

1

and

R

2

areresistors.

The op-amp an be used in dierent ir uits to a quire dierent on-ditions. Figure 22 below presents two dierent ongurations of how

op-ampwhi hwill produ ea negative voltagegain, wherethenon-inverting op-amp ingure 22bwill produ e a positive voltage gain[31 ℄.

Figure22:a) inverting op-amp ir uit b) non-inverting op-amp ir uit

Ingure23below, gurea)shows thetransfer hara teristi sforthe aseof annon-inverting op-amp ir uit and gureb) isfor the ase ofan inverting op-amp ir uit.

Figure23:Transfer fun tion hara teristi s ofa non-idealop-amp

To summarize: foran ideal op-amp,the onditions 30-32applies.

A

v

= ∞

(30)

R

out

= 0

(32) For the ir uits seen in gure 22, the sign of the voltage gain will vary depending onthe ir uitinuse.

A

v

= ±

V

out

V

in

(33)

The desiredoutput voltage an be a quired by onguring the op-amp ir- uit.Bywriting aprograminLabVIEWthatusestheDAQ ardtogenerate a spe i signal, and withtheuse ofan op-amp ir uit,it ispossible to ge-nerate anamplied voltage/ urrent signalwhi h isfed into thetransformer andtherebyrepla ethe varia .

6 LabVIEW

LabVIEW is a graphi al programming language that was usedin this the-sis proje t. A ready-to-use LabVIEW ode was downloaded from National Instruments' website, and modied to better meet the needs of the trans-former test ben h.In its default state,the ode olle ts data without being designedfor aspe i test.The odewasmodiedsu hthatit ouldhandle more hannels, and its default sampling frequen y was hanged. The user manualinse tion9 will give detailed instru tionsof how to usethe ode.

7 Results

Thisse tion presents the rated values of the transformers,and provide the resultsfrom the various testsexplainedthroughout thereport.

7.1 Cal ulations

It is important to determine the urrent, voltage, and power values that shouldnot be ex eeded. Thesevaluesare theratedvalues of atransformer. Thissubse tion presents therated voltage, therated urrent and the rated power.

Preferably a transformer's ux density, B, should be around 1.6 - 1.7 T. Setting B too high may ause overheating, and will thus damage the

a ompanies theuxdensityvalue1.7T.

B = 1.7 T

Theareaof ea hsteel sheetthat isusedinthis thesisis

A

sheet

= 0.27 ∗ 40 mm

2

Thenumberof sheetsthatareusedis 75.The angularfrequen y,

ω

,is

ω = 2π ∗ 50 rad/s

7.1.1 Rated urrent

Thegeneration of heatis what limitsthe urrent. Ex eeding therated ur-rent will ause the transformerto overheat,and thus de reaseits'e ien y while in reasing the losses. A rule of thumb is to keep the urrent density within2

[A/mm

2

]

.Thus,therated urrent an be al ulatedbymultiplying

2

[A/mm

2

_]

withthe rossse tional area ofthewinding, whi h is1.12

mm

2

forthe 500turn winding,and 5.6

mm

2

for the100turnwinding.

I

rated

= 2

A

mm

2

∗ 1.12 mm

2

= 2.24 ≈ 2 A (r.m.s.)

I

rated

= 2

A

mm

2

∗ 5.6 mm

2

= 11.20 ≈ 10 A (r.m.s.)

7.1.2 Rated voltage

The rated voltage values below were a quired by setting the ux B to 1.7 T, and is presented for the 500 turns winding and the 100 turns winding respe tively.

U

rated

=

B ∗ 500 ∗ A

√

sheet

∗ ω

2

= 152.95 ≈ 150 V (r.m.s.)

U

rated

=

B ∗ 100 ∗ A

√

sheet

∗ ω

Inorderto al ulatetherated power,therated urrent wasmultipliedwith theratedvoltage.

One phase

500turns : S

rated

= U

rated

∗ I

rated

= 300 V A

100turns : S

rated

= U

rated

∗ I

rated

= 300 V A

T hree phase

For the three-phase transformer,therated voltage hasto bemultiplied by a fa tor

√

3

.

500turns : S

rated

=

√

3 ∗ U

rated

∗ I

rated

= 519.62 ≈ 500 V A

100turns : S

rated

=

√

3 ∗ U

rated

∗ I

rated

= 519.62 ≈ 500 V A

7.2 Test results

Thisse tionprovidesresultsfromthe performedtests.Italsoprovidesplots relevantto therespe tivetest.Ase tion ontainingthe al ulation dataand thenananalysisareprovidedaftertheplots.Thesamplingfrequen yusedin thesetestswas2.5kHz.The al ulatedparameters,whi harea quiredfrom thesetestsare al ulatedoverlargerintervals,su hthatmultipleperiodsare a ounted for instead of just one, and then averaged. The larger intervals onsistsof 1000 data points. This provides better and more a urate data, nullifying mostof the noise.

7.2.1 Single-phase

The tests arried out for thesingle-phase transformer were an open- ir uit testand a short- ir uittest. Theout omes ofthetests arepresentedinthe following se tions.

For the open- ir uit tests, three dierent voltage levels were applied. The rated voltage isequal to 153.0 V (100%) in r.m.s. Two other voltage levels were also applied, whi h were 160.60 V (105%) and 168.20 V (110%), also in r.m.s. The 100% ase is presented in gure 24, whi h ontains both the voltageand the urrent.They-axisontheleftsideshowsthevoltage values andthe y-axison the right side shows the urrent values.

116.01

116.015

116.02

116.025

116.03

116.035

116.04

116.045

−250

−225

−200

−175

−150

−125

−100

−75

−50

−25

0

25

50

75

100

125

150

175

200

225

250

Voltage [ V ] OC 100 % Current [ A ]

Time [ s ]

116.01

116.015

116.02

116.025

116.03

116.035

116.04

116.045

−0.1

−0.09

−0.08

−0.07

−0.06

−0.05

−0.04

−0.03

−0.02

−0.01

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

Voltage

Current

Figure24:Open- ir uittest -100% ofrated voltage

Itispossibletonote,byobservinggure24,thatthebehaviourofthesignals isperiodi .

100.01

100.015

100.02

100.025

100.03

100.035

100.04

100.045

−250

−225

−200

−175

−150

−125

−100

−75

−50

−25

0

25

50

75

100

125

150

175

200

225

250

Voltage [ V ] OC 105 % Current [ A ]

Time [ s ]

100.01

100.015

100.02

100.025

100.03

100.035

100.04

100.045

−0.175

−0.15

−0.125

−0.1

−0.075

−0.05

−0.025

0

0.025

0.05

0.075

0.1

0.125

0.15

0.175

Voltage

Current

Figure25:Open- ir uittest -105% ofrated voltage

The urrent seenin gure25 isalmost twi eas large asthe urrent seenin gure24.

The last level of voltage that was applied was the 110% ase, whi h an be seeningure26.

38.01

38.015

38.02

38.025

38.03

38.035

38.04

38.045

−250

−225

−200

−175

−150

−125

−100

−75

−50

−25

0

25

50

75

100

125

150

175

200

225

250

Voltage [ V ] OC 110 % Current [ A ]

Time [ s ]

38.01

38.015

38.02

38.025

38.03

38.035

38.04

38.045

−0.5

−0.45

−0.4

−0.35

−0.3

−0.25

−0.2

−0.15

−0.1

−0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

Voltage

Current

feren e inthe amplitude for the urrents. The no load losses al ulated are shownintable 3.

100% ofrated voltage 3.07W 105% ofrated voltage 5.91W 110% ofrated voltage 10.87W

Table 3:No loadlosses

7.2.1.2 Short- ir uit test

Next,theshort- ir uittest wasappliedto thesingle-phase transformer. Ho-wever, therated urrent wasnot appliedintheshort- ir uittests thatwere performedinthis proje t.

68

68.01

68.02

68.03

68.04

68.05

68.06

−4

−3

−2

−1

0

1

2

3

4

5

Time [ s ]

[ V ] or [ A ] Short circuit

Voltage

Current

Figure27:Short- ir uit test - urrent and voltage

From the a quired test data, r.m.s. values ofthe voltage and urrent ould be al ulated,whi hareshowningure27 forboththevoltageand urrent.

Voltage r.m.s. 3.09V Current r.m.s. 0.71A Impedan e,Z 4.36

Ω

Table 4:Thea quired valuesfrom theshort- ir uittest

7.2.2 Three-phase

For the three-phase transformer, three dierent tests were performed. The teststhat were performed were two open- ir uit tests, where one ontained a delta onguration and the other ontained a wye onguration, and a short- ir uittest that ontained a ombination ofdelta onguration and a wye onguration. The no load losses for the delta and wye ongurations areexpe ted to be equal, sin e they arisefor thesame reason,regardless of the ongurationthatis used.

The tests were performed by using 4-6 hannels; 3 hannels for the urrents and 1-3 hannels for thevoltage(s). The voltages and urrents are denotedwithU,V, andW. Thenotation denes whi hwinding (1,2,or 3) thevoltage or the urrent wasmeasured in.

7.2.2.1 Open- ir uit - delta

The following plot ontains the urrents and voltages measured over six hannels. Three dierent voltages were applied in this ase as well, however, this time, the 100%, 105% and 110% rated voltages were divided by the square root of 3, whi h was done to a quire the phase-to-ground voltages instead of the phase-to-phase voltages. Thus, the applied voltage levels were 88.30 V, 92.70 V and 97.10 V, respe tively. The orresponding phase-to-phasevoltages are153.0V,160.60 V,and 168.20V.

Figure 28 below shows the voltages of all three windings at 110% of ratedvoltage. Thevoltages thatareshownarephase-to-ground voltages.

48

48.005

48.01

48.015

48.02

48.025

48.03

48.035

−140

−120

−100

−80

−60

−40

−20

0

20

40

60

80

100

120

140

Voltage [ V ] Delta 110 %

Time [ s ]

48

48.005

48.01

48.015

48.02

48.025

48.03

48.035

Voltage U

Voltage V

Voltage W

Figure 28:Open- ir uitdelta test - voltagesat 110%of ratedvoltage

As an be seen, allthree voltageshave approximately thesame amplitude.

The following plots, gures 29-31, shows the urrents for all three windings at dierent voltage levels. Voltage U is in luded in these plots to showhow urrent Ubehaves for dierent voltage levels.

82.4

82.405

82.41

82.415

82.42

82.425

82.43

82.435

−140

−120

−100

−80

−60

−40

−20

0

20

40

60

80

100

120

140

Voltage [ V ] 3 Currents Delta 100 % Current [ A ]

Time [ s ]

82.4

82.405

82.41

82.415

82.42

82.425

82.43

82.435

−0.14

−0.12

−0.1

−0.08

−0.06

−0.04

−0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

0.14

Voltage U

Current U

Current V

Current W

71.2

71.205

71.21

71.215

71.22

71.225

71.23

71.235

−160

−140

−120

−100

−80

−60

−40

−20

0

20

40

60

80

100

120

140

160

Voltage [ V ] 3 Currents Delta 105 % Current [ A ]

Time [ s ]

71.2

71.205

71.21

71.215

71.22

71.225

71.23

71.235

−0.3

−0.25

−0.2

−0.15

−0.1

−0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

Voltage U

Current U

Current V

Current W

Figure30: Open- ir uitdelta test - urrentsat 105% ofrated voltage

48

48.005

48.01

48.015

48.02

48.025

48.03

48.035

−160

−140

−120

−100

−80

−60

−40

−20

0

20

40

60

80

100

120

140

160

Voltage [ V ] 3 Currents Delta 110 % Current [ A ]

Time [ s ]

48

48.005

48.01

48.015

48.02

48.025

48.03

48.035

−0.9

−0.8

−0.7

−0.6

−0.5

−0.4

−0.3

−0.2

−0.1

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Voltage U

Current U

Current V

Current W

Figure31: Open- ir uitdelta test - urrentsat 110% ofrated voltage

The no load losses were al ulated for ea h of the voltage levels, and are presented intable5.

105% ofrated voltage 6.77W 110% ofrated voltage 9.29W

Table 5:No loadlosses

7.2.2.2 Open- ir uit - wye

This se tion ontains the plots and the results for the wye onguration for the open- ir uit test performed on the three-phase transformer. The hannel properties mentioned in the delta se tion apply in this ase as well. In the wye onguration, the voltage is already measured from phase-to-ground, thus there is no need to divide by the square root of 3. Thus the dierent levels of the rated voltages will be the same as for the single-phase open- ir uit setup. However, the orresponding phase-to-phase voltagesare264.90 V,278.20 V,and291.40 V.

Figure 32 shows the voltage and urrent for all three windings at 110%of ratedvoltage.

74

74.005

74.01

74.015

74.02

74.025

74.03

−250

−200

−150

−100

−50

0

50

100

150

200

250

Voltage [ V ] Wye 110 % Current [ A ]

Time [ s ]

74

74.005

74.01

74.015

74.02

74.025

74.03

−1.5

−1.25

−1

−0.75

−0.5

−0.25

0

0.25

0.5

0.75

1

1.25

Voltage U

Voltage V

Voltage W

Current U

Current V

Current W

Figure 32: Open- ir uit wye test - voltages and urrents at 110% of rated voltage

thezero rossingsof its orresponding voltage.

The following plots, gures 33-34, shows how the urrents behave at dierent voltage levels.

108

108.005

108.01

108.015

108.02

108.025

108.03

−250

−225

−200

−175

−150

−125

−100

−75

−50

−25

0

25

50

75

100

125

150

175

200

225

250

Voltage [ V ] 3 Currents Wye 100 % Current [ A ]

Time [ s ]

108

108.005

108.01

108.015

108.02

108.025

108.03

−0.12

−0.1

−0.08

−0.06

−0.04

−0.02

0

0.02

0.04

0.06

0.08

0.1

0.12

Voltage U

Current U

Current V

Current W

Figure 33:Open- ir uitwye test- urrentsat 100% ofrated voltage

92

92.005

92.01

92.015

92.02

92.025

92.03

92.035

−250

−225

−200

−175

−150

−125

−100

−75

−50

−25

0

25

50

75

100

125

150

175

200

225

250

Voltage [ V ] 3 Currents Wye 105 % Current [ A ]

Time [ s ]

92

92.005

92.01

92.015

92.02

92.025

92.03

92.035

−0.3

−0.25

−0.2

−0.15

−0.1

−0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

Voltage U

Current U

Current V

Current W

presented intable6.

100% ofrated voltage 5.10W 105% ofrated voltage 6.82W 110% ofrated voltage 9.73W

Table 6:No loadlosses

7.2.2.3 Short- ir uit - delta-wye

Thedelta-wye onguration wasapplied to thethree-phase transformer, to perform a short- ir uit test. Ideally, the voltage should be the same in all threewindings, thus we onlyplotted one voltage level, seegure35.

16

16.005

16.01

16.015

16.02

16.025

16.03

−6

−4

−2

0

2

4

6

8

Time [ s ]

[ V ] or [ A ]

_{Delta−Wye SC all Windings}

Voltage U

Current U

Current V

Current W

Figure35:Short- ir uit delta-wye test

Thea quired voltage and urrent valuesarepresentedintable 7.

Voltage Ur.m.s. 4.24V Current U r.m.s. 3.28A Current V r.m.s. 3.59A Current Wr.m.s. 3.85A

impedan es presentedintable 8.

Impedan eZ1 1.30

Ω

Impedan eZ2 1.18

Ω

Impedan eZ3 1.10

Ω

Table8:The orrespondingimpedan es

7.2.3 Load loss and short- ir uit impedan e

The load losses are onstant, but dier for dierent sizes of the windings. The sizes of the windings that was used in this thesis were 500 turns and 100 turns.Thus,the losses are

500turns : P

LL

= 2.24

2

[A

2

] ∗

2.06 ∗ 10

−8

_{[Ωm] ∗ 75[mm] ∗ 4 ∗ 500}

1.12[mm

2

_]

= 13.84 W

100turns : P

LL

= 11.2

2

[A

2

] ∗

2.06 ∗ 10

−8

_{[Ωm] ∗ 62[mm] ∗ 4 ∗ 100}

5.6[mm

2

_]

= 11.44 W

The load losses al ulated from the measurements lies between 14-16.3 W for the three-phase transformer. For the single-phase transformer, however, itisaslow as2.2W.

The short- ir uit impedan e per entage represents the voltage drop at full load. This is due to the winding resistan e and the indu tan e. This per entage also orresponds to the terminal voltage that is required to ir- ulatefull load urrent undershort- ir uit onditions[32 ℄. Theshort- ir uit impedan e forea h winding, inper ent,is:

Single − phase

z

sc

=

Z

SC

∗ S

rated

∗ 100

U

2

rated

=

4.3595 ∗ 342.6080 ∗ 100

152.95

2

= 6.39%

T hree − phase

z

sc1

=

Z

SC1

∗ S

rated

∗ 100

U

2

rated

=

1.30 ∗ 593.41 ∗ 100

152.95

2

= 3.30%

z

sc2

=

Z

SC2

∗ S

rated

∗ 100

U

2

rated

=

1.18 ∗ 593.41 ∗ 100

152.95

2

= 2.99%

z

sc3

=

Z

SC3

∗ S

rated

∗ 100

U

2

rated

=

1.10 ∗ 593.41 ∗ 100

152.95

2

= 2.79%

7.3 Analysis

In general, the results a quired from the various tests are good. The data samples were a quired with a 2.5 kHz sampling frequen y, whi h proved to be a su iently high enough sampling frequen y to avoid aliasing, and noiselikebehaviorinthesignals.

When applying the dierent per entages of the rated voltage in the various open- ir uit tests, one may a quire dierent data as it is not easy to tune the varia to spe i voltage values. This will lead to probable hanges in the a quired measured signals when re reating the tests. The same prin iple applies to the short- ir uit tests. Ideally, the voltages and urrents in ea h winding should be equal to one another in a three-phase short- ir uit test,i.e. the voltage and urrent in winding1 should be equal to the orresponding voltages and urrents available in winding 2 and 3, but that is not the ase. In the three-phase short- ir uit test, only one voltage was measured and then it was assumed that the same voltage was appliedtoallofthewindings,wheninfa titwasnot.Thus,theimpedan es al ulated willvarysomewhat between the windings.

8 Future work

Thetransformer oreusedinthisproje tis a ompound ofthinsteelsheets thatisput ontopofea hother. Thismakesiteasy to hangethesizeofthe ore. This method, however, is very time onsuming, and thus a problem.

to be worked on.

The LabVIEW ode used in this proje t is general and an be used for any test or measurement. However, it is not oded to do anything spe i ,other than simulating and logging data. Sin e there areonly a few tests that an be performed on a transformer, the LabVIEW ode should be further developed to meet some spe i requirements. Mathemati al operators ould be added to the main ode to make it al ulate the losses. Thiswould save some time for the user, whose only taskwouldbeto make surethatthe ir uitry and theinputdata is orre t.

The adapter, whi h was used to eliminate the use of batteries for the urrent probes introdu ed noise inthea quired data signals,thus batteries hadtobeusedinstead.However,itispreferredtousethebatteryeliminator. Hen e, the batteryeliminator should be worked on inorder to suppressthe noise.Ifthe urrent batteryeliminatorisused,itmaybepossibletoremove the noise by adding a lter. Alternatively, resear h should be done to nd anotherwayof eliminating theuseof batteries.

This user manual is designed to explain in detail how the onne tions for useof the transformertest ben h an be done. It introdu es instru tions of howoneoperatesthe softwareprogrammedforthetestben h,inawaythat enables modi ation and alternation in order to a hieve plausible goals. Themanual onsistsofahardwareandasoftwarepart,where thehardware se tion des ribes the omponents and how you onne t them. The last se tion of the user manual provides instru tions of what to do if an error o urs when runningthe ode.

A folder named Trafo  Finished Code is lo ated on the desktop of the omputer. Openthe folder andlo atethelenamed trafo.lvproj.This will open the nal produ t of a ode that fun tions with the transformer ben h. The following steps explains how to make use of the transformer ben h software. Themodied ode isnow assumedto be default.

9.1 Software

1. Lo ate andopenCont A q NSW under realtimeproje texplorer. You should nowbe viewinga newwindow, seegure36.

If you want to use the default setup, skip step 2 and 3 and ontinue on with step 4.

Thedefaultsetupusesthefollowing analoginput hannels,ofthedata a quisition ard, whi h an be seen in the ti k box to the left of the graphingure36.

2. To hange the input hannels that you wish to use, open SAR A q Main whi his lo atedundertheFPGA target intheproje t explo-rer. This is where the software is oded. A window will open up, as seeningure37.Theinputsthatdeterminewhi h hannelstopro ess anbeseenwithinthe ir le. The hannels anbe hangedby li king onthem.

3. Savethe odeandpress therunbuttonlo atedonthetop leftinorder to ompile the FPGA ode and load it onto the FPGA target, see gure38. Note; Thisstep will not work unlessthe ontroller hassi is poweredon.

4. Beforerunningthis ode,if youneedto hangethe save lo a-tion of the log, pro eed to step 5-7 and then performstep 4. Lo ate and open LogVI whi h is lo ated under My Computer in the proje t explorer in LabVIEW. A window will open up. Running this ode will both plot the data a quired and save the data on the omputer. This ode needsto be running simultaneouslywith the o-deinstep1inorderfortheprogramtofun tion,andbothofthe odes willstop runningifyou press the stopbutton.

5. You an hange the lo ation of where the logs will be saved. Press CTRL+Tto openup the blo kdiagram orresponding toLogVI

6. Right li kon the i onof Write to measurement le and sele t pro-perties.

7. Under the area Filename, see gure 42, you an hange the save lo ation andthe name thatthe logwill have.

8. The logs are saved as text les, whi h you an open with e.g. Ex el or Matlab. Figure 43 below shows an example of what a log may look like, where the measured data is saved in an own orresponding olumn for ea h hannel. The olumn X_Value denotes the time instan ea datapoint wasa quired.

Note that you need to hange the lename for the logs before runningthe ode,sin eifyou hange itafter, itwill applyforthenext log.

organizethedata

Thevoltagedividers alesthevoltagedownbyafa tor475,andthe urrent probes s ales the urrent down by a fa tor 10. Thus the measured voltage needsto be multipliedby475,and the urrent by10.

The last se tion of this manual provides instru tions on what to do if anerror o urswhen runningthe ode.

This se tion will des ribe how you onne t the omponents lo ated on the transformerben h.Before pro eeding tothe omponents ofthetransformer test ben h,make sure thattherelay onne ted to thevaria isnot plugged in,seegure 44.

Figure44:Relayino-state

Componentsthataredes ribedinthisse tionare:Three-phasetransformer, modules for banana onta ts, urrent probes, battery eliminators, voltage divider, data a quisition ard (NI-9205) and a ontroller hassi (NI RIO-9074)withits orresponding omponents.You anof ourseuseothermodels ofDAQ ards and ontroller hassis.TheDAQ ardis tobepluggedininto the hassi ontroller, and the ontroller hassi needs to be powered up in orderto ommuni ate withboth the omputerandthe DAQ.

1. Conne t thepower supply's outputs, seen in ir le 1, in gure 45, to the ontroller hassi, and onne tthepowersupply'sinputs in ir le 2 toa wall so ket.

2. Thewires in ir le1 ingure45 areto be onne ted to the ontroller hassi, seen in gure 46. Conne t the positive output to the one of the V terminals on the ontroller. Conne t the negative output to a terminal denotedasC.

Figure46:Controller for theDAQ ard

3. Conne tan Ethernet ableto one of theports,seegure 47.Conne t the otherend ofthe able to the omputer.

4. TheDAQ ard an now be pluggedinto the hassi ontroller.

Figure 48:a) Inputsof theDAQ ardb)output oftheDAQ ard

Slide the DAQ ard, seenin gure48a, into theslot on the ontroller hassi by aligning the output, seen in gure 50b, to the input of the ontroller hassi.Theinputofthe ontroller hassiisseeningure49.

5. Thenextstepis towiretheanaloginputs oftheDAQ ard. Youneed topayattentiontowhi hwireyou onne tintowhi hholeasyouhave tolog datafromthespe i hannel whenmaking measurements.

Figure 50:Wires onne ted totheDAQ ard

Conne tthenumberedwires ingure50to the orresponding numbe-redslot ingure 51.

AI0 isthe hannel that measures thedata over wireone and two, i.e. holeone isto be usedwithholetwo,hole threewithholefour, andso forth.The inputhole 33is to be onne ted toground.

Figure52:Cir uitboxwhere thewires are onne ted

6. The other end of the wires ingure 51 is onne ted to a ir uit box, showningure54.The ir uit boxisfurther onne ted tothevoltage