Miller Tower Testing

8/3/2019 Miller Tower Testing

1/52

Howtospecify,workwithvendorandoverseetestingMikeMiller, PE

1


2/52

ManufacturingPlantLocationsHeadquarters Houston,Texas

Monterrey, Mexico

BeloHorizonte, Brazil


3/52

Whyis

testing

needed?

Isit

worth

the

expense?

Cantweverifydesigns

withour

computer

tools?


4/52

ThePurposeofFullScaleTowerTesting

Thearethreetypesoffullscaletesting.

Goal:ToVerify:Design

Detailing

ProductionDesignTestsCost:

$

ProductionVerif NewConce ts

Goals:ResearchTestsCost:

$$$

Design,Detailing

NewDesign

Concepts

ToVerify:

ExaminemodeofFailureFailureInvestigations

Goal:Cost:$$ Fabrication

ProductionQA/QC

Extremelyrarebecause

hassignificantschedule


5/52

Whyistestingneeded?

LatticeTowersarecomplexsystems

Indeterminate

3D

Structures

TrussDeflectedShape

Compression/TensionIn3D


6/52

Whyistestingneeded?


Modeled

as

2

force

(tension/compression)

systems


7/52

Whyistestingneeded?


8/52

Whyistestingneeded?


Connection

eccentricity


9/52

Whyistestingneeded?


Connection

eccentricity

TopView


10/52

Whyistestingneeded?

Inaddition:

Tover yt e es gnmet o sot at t sposs eto ave

confidencein

the

design

of

the

body

and

leg

extensions

thatarenottested.


11/52

Whyistestingneeded?

Testingvideos

ClickherefortheTowerTestingVideo


12/52

FullscaleTestingisalongstandingpractice


13/52

1948


14/52

1948


15/52

1965


16/52

Verticalloadapplication Deflectionmeasurements

1965


17/52

196080


18/52

TowerTesting:

HowtoSpecify


19/52

TowerTesting:HowtoSpecify

Specification: ASCE10 Specification: IEC60652


20/52


Familyof4to8towertypes:

ower

ype

e ec onTypically: 50100%offamilyistested.

Selectionoftowertypeisbasedon:

MostUsed

Tower

(Tangent

Suspension)

MostCriticalTower(DeadEnd)

,

treatdetailingindependently)

Note:

Ifany

two

towers

are

very

close

in

strength,

anditisexpectedthedetailingwillbeverysimilar,

thenadecisiontotestonlyoneofthetwotowersis

reasonable.

TowerTypesnottested:

RiverCrossingTowers OLF~1.2to1.5

o es


21/52

WhyareTubularSteelPolesTypically

NotTested?

StructurallyDeterminant

(LoadPathswellbehavedandpredictable)

Analysismatchespoleperformance

(Testinginthe70s)

Fabricationdependentonweldquality,e.g.AWS

D1.1(processesandprocedureswelldefined)

Largerdeflections applicationloadvectorsmay

change.


22/52

PoleTesting 1979


23/52

TowerTesting:HowtoSpecify 8to12LoadCasesaretypical:

oa

ase

e ec on e ec

oa

cases

a

con ro

e

es gno :

MainLegs

BodyCrossBracing

TowerArms

Shieldwirepeaks

Typically: on ro ngwea ercases

Construction/Stringing

BrokenWire/Torsional

LocateLoadcellsascloseto

loadapplicationpointaspossibletoavoid

u ey r c on osses

es ng anpro uce y

vendor/testingfacility


24/52


WindLoad

Application

napprox mat onon oa ng


25/52


TowerHeight

Selectioncanbebasedon:

BodyExt./Leg

Ext.

with

highest

reaction

. .

Tallesttowerconfiguration

Mostcommon yuse tower e g t


26/52


Genera

Test ngCr ter a

Themanufacturingprocessesusedinthefabricationofthe

pro o ype or es ngs a e o esamespec ca onsas oseused

during

the

fabrication

of

the

production.


27/52


Material

Itisimportanttousethesamematerialforthetestthatwillbe

producedduringthesupply.

Toughness

, . . .

ductility,A709,etc.,thenitmaynotbepossibletoobtainthismaterialin

smallquantitiesforthetest.

A572Gr.

50

with

Charpy notch

toughness

is

asupplemental

requirement

andthisexactsteelwillmostlikelynotbeavailableforthetest.

Yieldstrength

maylimit

test

material

availability.


28/52


IdenticalMaterialforTestingmaynotbeavailable.

ProcessforProduction towersupply:

.

2.Supplierpurchasesrawmaterial

2a. Mills:Productionquantities rollingschedules (materialisnotstockedinsufficient

productionquantities)

2b.Distributionwarehouses(ifstandardmaterialandinstock)

3. Fabrication

ProcessforTest towersupply:

1.TowerSupplier/TestingStationcontractaward

.

2b.Distribution

warehouses

(if

standard

material

and

in

stock)

3. Fabrication

Note: Nonstandardmaterialistypicallynotavailablefortesting.

Materialsubstitutionsmustbemadeutilizingthesameyieldstrength

andshall

be

representative

of

the

materials

used

in

the

production.


29/52


F n s ng

Blackvs.Galvanized

Structurallyperformthesame.

Blacksteeltendstobemoreaudiblethangalvanizedunderloading.

es e ower s o e e vere o ec en , en es nga

galvanizedtowermakespackagingandshippingsimpler.


30/52

TowerTesting: Scheduling

Howmuchtime

doesittake?

TowerDesign,DetailingandTestingProcess

TowerDesign

TestMaterialProcurement

48weeks

2weeks

TestingPlan

Prototype

Fabrication

Detailing(prototype)6to12

months

68weeks

2weeks

Assembly,Erection&Rigging

Testing

23weeks

12weeks

Detailingcorrection

(if

required)2weeks Test

Report

Note: 6to12monthsforcompletionofdesign/detail/testingincludingalldetaildrawings.


31/52

EquipmentusedinFullScaleTesting

oa

e sTemperaturecompensating,calibratedbeforeeverytowertest

MechanicalWinches(AutomaticorManual)

WireRopes

Sheaves

TheodolitesFormeasuringdeflections


32/52


LoadCells:

Calibratedbeforeeverytowertest


33/52


ControlRoomMonitoring

slowprocess


34/52


Reaction

Frame

Reaction

Frame

TestBed


35/52


VerticalCables(holddowns)for

Horizontalloadapplication

ControlRoom


36/52


How to Specify : Strain Gauges?


37/52

HowtoSpecify: StrainGauges?

StrainGauges Definition

StrainGaugesareanelectronicresistance

measuringdevicewhoseelectrical

ofstraininthedevice.

ClickhereforStrainGau eDemo

How to Specify : Strain Gauges?


38/52

HowtoSpecify: StrainGauges?

StrainGauges ExpectationsandLimitationsonuse

Notmandatory.

Notrecommendedforprooftestingnormallatticeconfigurations.

Doesincreasethecostoftestin . ~ 3 000 ermember au ed

Unlesscalibratedlikealoadcell,gaugesyieldapproximateresults.

Canbe

used

to

spot

check

load

flow

in

members.

Judiciousselectionoflocationsareimportant.

equa ewea erproo ng srequ re .

h d l d ?


39/52

WhatdoesTestingInclude?

es gno

a

x ure

Verificationofassemblyanderection

VerificationofRiggingthatsimulatesthedesignloading

DesignFitnessofPurposeunderloading.

mes,

es ng

a so

includes:

Failurerootcause

analysis

DetailingFitnessofPurposeunderloading.

RecordingDeflections

in

Longitudinal

and

Transverse

directions.

Structuralcorrection

Re

testing

Witnessingbyclient(orclientsrepresentative)

TestReport

AsInstalledinfield


40/52

Testingprocess Whattoexpectduringtesting

Loadingtowerto50%of1st loadcasetocheckrigging,etc.

LoadCaseloadingwith1min.holdsat50%,75%,90%,95%and5min.holdat100%.

, .

Videodocumentation.

LoadBalancingcanbe

pro ema c


41/52


ForRigid

lattice

towers:

Deflectionscanbe1.5to2timesthatpredictedbyanalysisduetoboltslippage.

Thereisadistinctionbetweendesignloaddeflectionsandserviceloaddeflections.


42/52


Ovalizationof

holes:

Generally,someovalization ofholesandpermanentdeformationofboltsare

.

Ifthe

test

tower

is

to

be

used

for

production,

these

deformations

must

be

replacedtoconformtheproductionspecification. Boltsusedintowertestsshould

notbereused.

Mechanical ClearanceIssues:


43/52

TestingCosts

~1.5 2.0xcostofproductionruntowers.

Test Basic: ~$150,000to$250,000

Failurecost: dependingonextentoffailureandavailabilityofmaterial,

downtimeofstation.

Strain

Gauges

(if

used)

~

3,000

per

gauge. Additionalloadcasesbeyondthestandard812.

Destruction Test: ~ 20 000 to 50 000

CostforFailureafterenergization $$$$$


44/52

DestructionTests

increasingthetransverseandverticalloadsinorder

togaininformationonactualvs.predictedbehavior. Thefailureloadcouldthenberelatedto

aincreaseinspanutilization.

~114%


45/52

FailureCaseStudy1

500kVSingleCircuit

TangentSuspension

Failure

at

95%

b /f


46/52

Freebodydiagram/force

summationcheck

Anglesdonothaveequal

flangestressdistributions


47/52

FailureCaseStudy2

500kVSingleCircuit

DeadEnd

Loading evaluation


48/52

Loadingevaluation

Solution


49/52

WhatsnewintheIndustryregardingTowerTesting?

TowerTesting: Newinformation (Cigre Report399)


50/52

oa

ase

e ec on

Deterministicvs.Probabilisticdesignloads

e%exc usion imits,w eree=5%,10%

Howtocorrelateittothestrengthfactor?

Return eriod

DestructionTestsRecommended

forcalibrationoftowerstrength

curves.

TowerTesting: Newinformation (Cigre Report399)


51/52

p t

test ng

snot

necessary

BayesianApproach

probabilitybasedanalysis.

FrictionLossesdueto ulle s stemsMethodofdeterminationbytest

Acceptable

tolerances

on

applied

loads


52/52

Whiletestingyoucan

Makenewfriends

The

End

Miller Tower Testing

Documents