NOTE - Memorial University of Newfoundlandcollections.mun.ca/PDFs/theses/Tucker_JohnR.pdf · NOTE TO USERS The original manuscript received by UMI contains indistinct and light print.All

NOTE TO USERS

The original manuscript received by UMI contains indistinct and light print. All efforts were made to acquire the highest quality manuscript from the author or school. Microfilmed

as received.

This reproduction is the best copy available

The Design, Manufacture and Verifcation of a High Capacity

Force and Moment Measurement System

John R. Tucker. P-Eng.

A thesis submined to the school of Graduate

Studies in partial fulfillrnenr of the

requirements for rhe de- of

-Vaster of Engineering

Faculty of Engineering and Xpplied Science

Mernorial University of Sewfoundland

.Au,yst 1998

St. John's

National Library BiMiottiéque nationale du Canada

Acquisitions and Acquisitions et Bibliographie Services senrices bibliographiques

395 We4mgîon Street 395. rue Wdingtm OrtawaON KIAON9 OaawaON K I A W Canada canada

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant à la National Libmy of Canada to Bibliothèque nationale du Canada de reproduce, loan, distnbute or sell reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/61m, de

reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts fkom it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author's ou autrement reproduits sans son permission. autorisation.

A series of tests was conducted to study the interaction of a multifaceted conical structure

with multi-yar rïdges. The tests were conducted in three phases. one in Calgary. Alberta.

one in Ottawa. Ontario. and the third at the Institute for Marine Dpamics in St. John's.

Newfoundland. ïhe St. John's phase of the program tested some of the largest ice features

ever attempted at this facility, including an attempt at a one in one hundred yearconsolidated

multi-year ridge. This thesis documents the design. manufacture and testing of the force and

moment measurement system developed for use in the St. John's tests. A series of

calibration tests were conducted in the Structures Laboratory of The Faculty of Engineering

in which the global force measurement system was secured to the floor. and Ioads of known

magnitude and direction were applied using a hydnulic rarn with an in-line force transducer

instailed. Founeen different orientations were tested and a fifteenth test was conducted in

which one half of the loading system was chilled using ice to simulate the condition of

having haif of the force measurement system subrnerged in the iMD ice tank. Following this

series of calibration tests, the equipment was taken to the hstitute for Marine Dynamics and

installed. Calibratîon tests were conducted there to venfy the integrity of the force and

moment measurement system, and a series of dynarnic 'pluck' tests were canied out CO

determine the natural frequencies of the towing system and model. In this thesis. the results

of these pluck tests are compared to the frequency of the forces observed during an ice test

to ensure that resonance did not occur during testing, and that the data collected are sound.

1

ACKNOllrLEDGrnNTS

This thesis was compiled from research and work conducted at Mernorial University of

Newfoundland and the National Researc h Council's Insti tute for ~EiIarine D ynarnics. w i th

joint univenityhdustry funding from the National Science and Engineering Research

Council, the National Research Council. Esso Resources Canada Limited and Mernoriai

University of Newfoundland. This, in addition to the deparîmentai suppon within the

Faculty of Engineering and Applied Science were essentiai to the success of this study.

1 would like to sincerely thank my supervisor. Dr. D.B. Muggeridge. Dean of Science at

Okanagan University College. B .Cs. whose support. encouragement. guidance and patience

made this endeavor a reality. 1 would also like to thank Dr. A.S.J. Swamidas and Dr. L Lye.

Professon of Engineering, .Memonal University of Newfoundland the many individuals

working in the laboratones of this faculty as well as the IMD for their assistance and advice.

The manufacture of the equipment outlined in this thesis and ail of the equipment used in this

test program including the models was conducted at Technical Services within the Faculty

of Engineering and Applied Science. It is a testament to their superb craftsmanship, and 1

wish to thank them also for their generosity, time and patience.

Finally, 1 would like to express my sincere gratitude to the members ofmy family, especially

my wife Cindy, for support, encouragement and affection provided during this uying period.

. . I l

TABLE OF CONTEXTS

TABLE OF CONTEINTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii

LISTOFTABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

XOMENCLATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii

1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.2 The Test Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

. . . . . . . . . . . . . . . . . . . . . 1.3 Objectives and Scope of the Ovenll P m p m 7 f . 1 Thesis Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 9

3.0 THE DESIGN . FABRICATION. C ~ R A T I O N AIW ASSELMBLY

. . . . . . . . . . . . . . . OF TKE FORCE MEASURE,aW SY STEM MOCK-IJP 18

3.1 Design And Fabrication Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 3.3 Global Load Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 4 3.3 Cdibration of Test Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 3.4 Test Setup and Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 3.5 Method of Cornparison of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 3 3 3.6 Discussion of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -41

4.1 The Mode1 Test Stnictm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 4.2 instrumentation and Data Acquisition System . . . . . . . . . . . . . . . . . . . . 48

4.21 Neck b a d Cells . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50

TABLE OF CONTENTS (continued)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2.2 AccelerometersandLVDT - 5 2 4.2.3 Data Acquisition Systern . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53

4.3 Test Program with Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 5 4

5.0 PERFORMANCE OF THE EQUIPMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . - 5 7

5.1 Load Ce11 Calibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 5 7 5.2 Natural Frequency of the Test Structure . . . . . . . . . . . . . . . . . . . . . . . . - 6 0 5.3 Test Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -64

6.0 CONCLUSIONS AND FiNDDJGS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

APPENDR A SPECIFICATION SHEETS . . . . . . . . . . . . . . . . . . . . . . . . . . . - 7 5

CALIBRATION TEST RESULTS . . . . . . . . . . . . . . . . . . . . . . . 80

Table

. . . . . . . . . . . 1.1 Test Parameters Varied in Test Facilities at ERCL. IME and IMD 3

3.1 System Capacities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -38

3 -7 MAPET and MMAPE' Factors for Force iMeasurernent During the CdibrationTests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39

3.3 MME' and MMAPE' Factors for Force Measurement During the CalibrationTests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 O

. . . . . . . . . 4.1 The values of Xi. Y. and 2, for the 125 and the 150 scale models -49

. . . . . . . . . . . - - - . . . - - . - . 4.3 Test Matrix for the Multifaceted Cone Tests IMD -56

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Summary of Ridge Test Results - 6 8

Fi oure Pace

3.1 Global Load .M easurement Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

3.2 AMn 6 Component Force Transducers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 0

7 3 3.3 Angular Contact Spherical Bearings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7 3 3.4 Angular Contact Sphericd Bearing and Fiangd Housing ...................

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Test Assernbly - 2 3

3.6 Orientation of load ce11 coordinate axes with respect to global coordinate axis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 7

. . . . . . . . 3.7 Orientation of global coordinate axis with respect to mode1 structure 28

3.8 Calibration and ,Mot k q equipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . - 2 9

3.9 LuadingArm . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

3.10 Loading Equipment for Calibration Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

3.1 1 Perfect Correlation Between .M easured and Xpplied Data . . . . . . . . . . . . . . . . . 34

4.1 1.25LargeNeck.M ode1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45

4.2 1:25 Large Neck Mode1 Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . -16

4.3 1 :25 Srnall Neck Mode1 Dimensions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4 150 Large Neck Modei Dimensions - 4 7

1.5 Neck ioad ceil arrangement for the 125 large neck mode1 . . . . . . . . . . . . . . . . . 50

4.6 Neck load cell arrangement for the 150 large neck . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . and 1 :35 smail neck models - 5 2

4.7 Schematic of Data Acquisition System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54

LIST OF FIGURES (continued)

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.1 Load Cell Calibration Setup - 5 8

5.2 Global Force Calibration Data from Test C o n e d 4 . . . . . . . . . . . . . . . . . . . . . 59

. . . . . 5.3 Setup for Dynamic 'Pluck" Test 555.1PSD Analysis of Cone Acceleration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Data in the X Direction -60

5.5 PSD Analysis of Cone Acceleration Data in the X Direction . . . . . . . . . . . . . . - 6 1

5.5 PSD Analysis of Cone Acceleration Data in the Y Direction . . . . . . . . . . . . . . - 6 1

5.6 PSD Analysis of Post Acceleration Data in the X Direction . . . . . . . . . . . . . . . 62

. . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 PSD Analysis of Ice Force Data X Direction - 6 2

. . . . . . . . . . . . . . . . . . . . . . . . . . 5.8 PSD Analysis of Ice Force Data Y Direction - 6 3

5.9 Time series trace of giobal forces in the X . Y and Z directions fortestMUNCONE7-007 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65

5.10 Time series trace of giobal moments about the X . Y and Z axes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . for test MUNCONE7-007 66

5.1 1 Time series trace of forces on the neck in the X. Y and Z directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . for test MUNCOE7-OO7 66

vii

a,

PSD

Total force in the X direction

Measured force in the X' mis direction on the i'th load cell

Total ovemming moment about the X axis

X location of the i'th giobal load ce11 with respect to the defined coordmate

system

Force applied to the load measurement system

Force recorded by the load rneasurement system

Sum of absolute prediction error

.Mean absolute percentap emr

Mean absolute percentage e m r desensitized to the applied load

The -VAPE' desensiuzed to the system capacity

Total nec k force in the X hrect~on

Yodel velocity during ice test

Ice sheet thichess

Fiexural strength of the ice sheet

Ridge thichess

Rexurai saen,@ of the ridge tested

Power spectrum density

Chapter 1

When conducting mode1 tests to study such events as icdstrucnire interaction. ir 1s rmponant

to perform rhe tests at a number of mode1 scales to determine preciseiy whar parameters

should be accounted for in the analysis. This is of panicular importance uhen studying

events involving cornplex. anisompic materials. -4 mode1 test series was conducted to study

icdstructure interaction berween multi-year ice nees and mulu-faceted conical stucnires.

with the ooal in mind of testing at as large a scale as possible to facilime the development

of an algorithm to predict forces experienced in such an interaction at full scale.

1.1 Background

In the past two decades extensive research has been conducred to examine the feasibiliry of

using smooth cones to protect offshore strucrures such as oil ngs and bridge pien from sheer

and rnulti-year ice. .A recent state-of-the-an rewrew of ice forces on conical ~tnicnires

Wessek md Kato. 1988) descnbes ice faiIure modes around conical stnicrures. and has

sumrnarized awiab le mode1 scde and Mi scaie rneasuremenE. These t s r s showci rhe

effectiveness of conical srnictures in ice defence. Houever. it was highly desirable h m 3

manufacturïng standpoint to replace the rounded conical surface wirh a ilat facemi one.

This pro-oram was developed ro studp the differenr aspects of ice loading on a mu10 -facececi

conicd stnicnire. The interaction loads with a muiti-~ear ndge was of parricular interest Io

designers as this would be the design ice conhtion for suucnues located in the Beaufort or

Chukchi Seas-

Another point of interest \vas rhe effecr of using a corncal s rn icm with a Iarger diameur

neck than previously iomidered This would have the obvious benefit of pemtting 3

smaller cone to prorect a 1-r piece of qupmenr juch ûs bridge pien or smicniral

members of offshore oil ri-.

Scde effects are dso considered with a total of four modei scales k i n g examned. The 1 : 10

and 120 scaie tests were conducred at ERCLi outdoor ba in in C a l w . a 150 mode1 was

utilized at the tests in ME in Ottawa and the 125 and 150 scale models were tested at the

MD'S facility in St. Johnk

The test parameters that were varied in each of the test facilities are indicated in Table 1.

The first two phases of the test program were completed and documented by .Verge and

Weiss. (1989). .Metge and Tucker. (1990). and irani et al. (1992). The third phase of the

program was carried out at MD and foms the bais of this thesis.

The test program conducted in IMD was a collaborative effort beween ~Mernorid LTnivenity

of Newfoundland (MUN) and the National Research Council (NRC) of Canada. The general

concept of the test series as well as the execution was done jointly. and several individuais

were involved in this project. The detailed design. fabrication and testing of the mode1 and

the load measurement assembl y were done by J. Tucker of under the supervision of

D.B. Wuggeridge. W. Lau of NRC had the responsibility of executing the test pian

Table 1.1: Test Parameters Varied in Test FaciIicies at ERCL, IME and IMD -- -

1 Variation Tested

1 Structure Orientation I --

Yes

1 Neck Diameter 1 No 1 Yes Yes

1 Ice Movernent Rate 1 No 1 Yes Yes I 1 Ice Floe Thickness 1 Yes 1 Yes 1 Yes 1 Ice Floe Strength

1 ~ i d g e ~trength I Yes I Yes I Yes I 1 ~ i d g e Orientation I Yes I NO I NO I

1 Yes

Yes -

Ridge Thickness

Yes -

Yes

Yes

Yes

developed jointly by D.B. Muggeridge and J. Tuckerof MUN, M. Lau of N'Etc. K. Croasdale

and LM. .Metge representing ERCL, A. Rodanovic representing Mobi 1 Oi l Research and

Development. J-C. Chao of Exxon Production Research Company, and G. Timco and S.

Jones also of the M C . B. Hi!l of XRC was responsible for developing the new ridge

rnodelling techniques and conducting the ice properties measurements. S. Bua of NRC

provided technical assistance in dl aspects of the tests. A number of students from the CO-

operative engineering program of MUN and 2. Wang, a graduate student at .MUN. also

provided assistance in executing the test program.

1.2 The Test Program

The scûles to be tested in the overall program were to approach 1: f O. which would mean that

the forces wouid tend to be closer to prototype values and the ice damage mechanisms would

be more realistic than the ones observed at much smaller scales.

The test series was part of a joint university-industry research program initiated to study the

interaction between a multi-faceted conical structure and a multi-year ndge frozen into a

multi-year ice pack. The program was the result of a research agreement between Mernorial

University of Newfoundand (MüN), the National Research Council of Canada (NRC), and

Esso Resources Canada Limited (ERCL), and was partly funded by the Naturai Sciences and

4

Engineering Research Council of Canada (NSERC). The program involved a total of four

test senes conducted in three phases. The fint phase consisted of tests conducted over two

winter seasons (scales 1: 10 and 1 :?O) at ERCL3 outdoor ice testing basin in Calgary. The

second phase consisted of model tests conducted at NRCS Institute for Mechanical

Engineering (ME) in Ottawa (scale 150): and the final phase was a series of tests conducted

at NRC's Institute for Marine Dynamics (IMD) in St. John's (scales 1 2 5 and 1 :SOI. The S t.

John's test matrix had several features incorporated into it:

- The 54 kPa ridge was built using the 'brick iayer' method whereas the lower strength

ridges (28 and 14 kPa) were constructed using the 'dump truck' method (Spencer et

al. 1990). These ridge building techniques were developed at the IMD in an attempt

to produce consolidated ridges of appropriace strength and dimension to simulate a

one in one hundred year, multi-year ridge.

- Tests from the phase conducted in Calgary with ERCL which were selected for

cornparison were incorporated into the matrix.

- A mode1 speed of 0.06 m/s was used in both the ERCL and ME tests senes and was

used as a standard speed for this series as well.

- A number of extreme ice conditions were dso simulated.

Smooth cones have been fairly widely used to protect bridge piea. caissons and offshore

structures, and it is generally accepted that forcing ice to break in bending results in less force

5

being experienced by the structure than if the ice were to fail in crushing. The manufacture

of large. smooth conical stnictures is inherently expensive due to the manufacture of rounded

plates and interna1 suppomng structures. An alternative to the srnooth conical des ig is to

approximate the surface with a multifaceted conical design. The interaction of ice features

with these faceted conical structures was the subject of investigations that consisted of a

combination of mode1 tests and analyticd studies. Loads associated with this interaction

could not be accurately predicted. this rquired that tests be cmied out on scaled models of

a generic design in an effort to produce an appropriate numerical algorithm.

At present several methods are available to measure the forces and moments exened by ice

ieatures. A newly developed method for the purposes of this study consisted of three six-

component force and moment transducen sandwiched between two plates. Concerns over

the inability to rigidly fix the three load transducen to two plates led to a design

enhancement in which the load cells were rigidly fixed to one plate using a bolted connection

while the connection to the other plate wûs achieved using flange-mounted spherical

bearings.

To test this transducer arrangement. and to alleviate concems about this new systems' ability

to measure forces and moments accurately, a series of rnock-up tests were performed on the

load ce11 arrangement in the Engineering Structures Lab in the S.J. Canw Building at

Memorial University of Newfoundland. The system was later used in the test pro- to

6

measure the ice forces expenenced by a six facetedconical structure interacting with a multi-

year ice pack consisting of embedded multi-year ridges.

13 Objectives and Scope of the Overall Program

The principle objectives of the program were:

(I) To understand how multi-year ice noes and ridges would intenct with a

multifaceted conical structure; and

(ii) To investigate rhe effects on ice interactions and forces of having the

diameter of the above-water vertical 'neck'of the structure be almost as large

as the waterline diameter,

Two scaled modeis, 125 and 150. of a prototype Beaufort Sea structure were used in the

tests, and two neck sizes were tested at the 125 scaie.

Two ridge targets were use& The first ridge target was one that wouid represent

approximately the yearly average multi-year ndge size and smngth tested at the 1 : 15 scde.

This was chosen because it is a more commonly encounteredice condition in which full scale

7

data might be available in the near future. Furthemore, this ice condition was also tested in

Essof '89-90 test series from which a direct cornparison was possible. The second target

was a 1 in 100 year multi-year ridge to be tested at the 150 scale. ïhis was tested as part of

the gant mandate.

The ridge construction techniques used in the tests conducted in Calgary and Ottawa were

employed in the present test series. This permitted examination of the effects of ridge

construction techniques on ice failure rnechanisms and Ioads. It also helped correlate the test

results from the three phases of the test program. Ridge strength and thickness as well as

sheet ice stren,@ and thickness were varied throughout the five weeks of tests conducted.

A transducer system had to be designed. manufactured and calibrated which would permit

the measurement of global (total overall) forces and moments on the model. Due to the

mode1 scale being used and the ice features being tested. the predicted global forces and

moments would be p a t e r than any previously measured in a test program at the MD. and

would be close to the capacity of the towing carriage. For this reason. use of a single six

component load transducer was ruled out.

1.4 Thesis Objectives

The major objectives of this thesis were:

(i) To outline a new design of a transducer system intended for use in the

multifaceted conical structure program outlined above.

(ii) To record a senes of calibration tests conducted at the S.J. Carew Building

in which a known load was applied to the load ce11 system and its outputs

recorded resolved into meaningful data and compared with the recorded

applied loads

(iii) To develop a meaningful method for comparing the results of these

calibration tests with the known applied load

(iv) To examine the performance of the test equipment in the IMD phase of the

multifaceted conical structure program.

The equipment designed, constructed and verified was an essential component to the success

of the program.

Chapter one is the introduction to the thesis. outlining the rationale behind the project and

the background behind the multifacetedconical structure project and chapter two is a review

of relevant lirerature. Chapter three discusses the design of of the force measurement system.

and details a set of calibration tests conducted in the Smctures Laboratory of the S-l. Carew

building. A set of two indices are also developed and presented which can be used to

indicate the degree of error in the calibration tests. Chapter four describes the entirety of the

test equipment and the instaihtion of it in the IMD ice testing facility. A discussion of the

multifaceted conical structure test program is also inciuded which descnbes the test

parameten varied throughout the series. Chapter five is an evaluation of the equipment's

performance in the test series. Calibration data gathered during the p r o k m is analysed and

discussed and a dynamic analysis is perfonned comparing the natural frequency of the test

structure with the frequency of the loading function during an ice test. Chapter six presents

an overview of the thesis as well as findings and suggested directions for future work.

Chapter 2

Physical testing. be it of model structures in a simulated environment or of full scde

structures in nature, requires the accurate rneasurernent of al1 important parameten. The

rnodeling of physical environments and objects for tesung purposes is necessary for the

maintenance of controlled environments. but the use of modelling and similitude laws for

scaling purposes then becomes necessary. In the ideal situation. one would test a full scale

structure in a controlled environment. This is of particular interest when developing

algorithms for predicting events on full scde structures based on model tests. The repeared

testing of the same conditions through a variety of scales would then optimize the algorithm,

and enhance our understanding of modeiing and similitude laws.

Ice-structure interaction testing has k e n carried our by researchers for decades with the

motive of reducing the risk to offshore structures from such threats. Strain gauged proving

rings were among the fint of the precise ice force measurement devices (Saeki. 1977). used

by a variety of researchen to mesure forces acting on a stmcture in a single axis. Devices

such as this were used in conjunction with plates to infer pressure or to permit the use of

multiple senson to measure multi-axis measuremenü, but their implementation and

vaiidation is difficult and cumbersome.

indeed, there have k e n tests conducted on actual structures where they have been

instrumented to measure the loads on them resulting from interaction with ice. In 1975. the

Kemi 1 lighthouse (Matanen. 1977) was constructed and instrumented for the measurement

of forces acting upon it from ice forces. The total ice load was rneasured by monitoring and

recording the bending deformation of the lighthouse (5.8 rn dimeter) undenvater structure.

Four 7.4 m long rods were mounted vertically inside shielded tubes so that they were free to

move relative to the concrete structure. Suain gauge transducen were used to measure the

relative rnovement of the rods. The notion was that output wouid be directly proportional

to the bending moment or the total ice load on the structure. It was stated by the author that

the total ice load measurement accuracy was reduced as a consequence of "mechanical signal

to noise ratio" and the vertical situation or location of the ice load.

The JZ20-2-1 platform was instrurnented in Liao Dong Bay (Fan and Jin. 1990)to measure

the total ice force on the jacket or platform in October of 1987. They used three types of

sensors, viz., sVain gauges. accelerometers and load panels which were used together to

12

determine the total ice force on the structure resulting from what was termed severe ice

conditions. The load panels measured the forces on the structure only at the location of ice-

structure interaction. and had to be mounted such that they could be adjusted to

accommodate the tide ievels.

A study was performed (Masroor and Zachary, 199 1) on the use of applying strain gages ro

structures. using the structure itself as the force measurement device for a physical test. In

this study. the authors demonstrated the viability of this technique. but cautioned that one

should be careh1 in locating the position of the strain gages. Emn may be propagated

through structural members. resulting in significant erron. They recomrnend that different

combinations of strain components could be snidied and evaluated on the basis of error-

propagation. and that the best among these could be selected for the placement of strain

gages. Their study. however. was restricted to linearly elastic structures with small strains

where the principle of superposition retained its validity.

In 1990. a joint Sino-German projecr to explore the correlation between mode1 and full scale

ice forces on a jacket piatform was conducted (Wessels and Jochmean. 199 1) in which

research was conducted in twci phases. In the fint stage, one of the four lep of a jacket

platforni in Bohai were instrumented using custom designed ice force measuring panels.

Five such panels were mounted on the surface of the leg and adjacent to each other so as to

cover the entire front sector ( 180') of the leg in the direction of ice impingement at rising

13

tide. For the scde model tests. a single force dynamometer was positioned at the top of a

frame io which the test structure was mounted and dnven through the model ice. in this case.

only sheet ice was tested and the vertical structure used tended to promote failure of the ice

in crushing. which resulted in very little rnovement of the line of action of the force

experienced by the test structure relative to the axis of measurement of the load transducer.

Structures which are bottom founded and which interact with ice in nature experience

extraordinady large forces. The ice failure on and about the structure could result in I q e

overtuming moments also king experienced during these types of interactions. This is

particularly the case where the walls of the structure have a slope, permitting the ice to slide

up or down the surface prior to failure. Such conical structures have been used to protect

offshore piatforms, lighthouses and bridge pien. These structures often have a sloped

surface at the waterline which would either iift the ice interacting with it up. or force it down

in order to cause the failure of the ice in bending as opposed to crushing. Uany of these

structures have more than one inclination of sloping surface. and a vertical wall or pillar

beyond the sloped surfaces. The net affect is that as the ice slides up the inclined surfaces.

the line of the action of the force relative to some danirn is constantly changïng. In their

book, "Ice interaction with Offshore Structures", Camrnaert and Muggeridge (1988) present

a number of dgorithms which may be used to predict the forces experienced by an inclined

structure experiencing ice. These dgorithms extend from two dimensional to three

dimensional analysis and include both elastic and plastic limit failure modes (Nevel. 1972

14

Ralston, 1977). The location of the forces and the resulting overtuming moment experienced

by the structure would be a consequence of the location of the centre of mass of the ice.

During the physical test of a structure in which the line of action of the force moves about

the location of the forcdmoment senson, it is not inconceivable for very large moments to

be experienced by the force transducers being used. One method used by researchers to

instrument structures which are to expenence this type of loading is to use single axis load

transducers and install them with couplings that are rigid in only the axis of measurement.

Using these, only forces in the load bearing axis of this coupling unit will be transferred to

the force sensor. Instrumentation such as this was used in the second phase of the

multifaceted conicd stmcture program (Irani. 1993). as well as a number of other tests

including a series of tests of impact forces on a flat jacket deck (Murray, 1997). This method

of instrumentation is effective. but does not lend itself to quick modification of the test

equipment due to the large number of components required for the setup.

h the book "Vibration and Testing - Theory and Practice" (McConnel. 1995). the issue of

the effects of bending moments on measured forces is addressed. Using an exarnple of the

measurement of the dynamic response of a steel bar r igidy attached to a force tmsducer and

stnick with an impulse hammer, ghost resonances appeared in the measured experimental

frequency response functions. This issue is pnncipally the one being addressed in this thesis,

and more specifically a method to reduce the moment king applied to the force transducers

15

used in an experimentd setup. He goes funher to state that force transducen are sensitive

to bending moments and shear forces, and there is little known of the definition of bending

moment sensitivities, let alone the prescription of methods of correction.

Nonlinearity in force transducers is defined as the arnount by which the uansducer output

deviates from between a svaight line zero Ioad and the nted load outputs (Antkowiak and

Rencis, 1994). This percentage of the rated output can be used to classify the performance

of a transducer (setup). In their paper, they evaluate the lineari ty of a small strain gage force

transducer and compare their experimentai calibration results with a finite element analysis

of the same structure. In their study, they found that small discrepancies occured between

the FE and test results. ïhey theorized that these could be the result of nonlinear material

behaviour, machining tolerances or residual stresses due to machining. Such issues as these

would certainly be compounded in a compiex transducer setup like that examined in ihis

thesis.

For accurate force measurements and to simplify analysis of the forces being measured it is

beneficial to attach the transducer systems base to m ideal or 'riad' foundation (McConnell,

1995). In such a case, the mass of the foundation is considered infinite, and the

transducer/model configuration acts like a single degree of freedom system where the

transducers may be considered as very stiff springs.

A iwo dimensional mathematical mode1 for simulating the behaviour of a force transducer

when subjected to bending moments was developed in 1993 (McConnell and Varoto. 1993).

The analysis was verifted with experimental results and it was demonstnted that bending

moment sensitivity affected the transducer's overall sensitivity. and that this may cause large

measurement erron depending on how the transducer is employed in a test.

From the above review. it cm be seen that a force measurement systern should be: ( i ) rigid:

(ii) linear in response: (iii) shouid not respond dynarnically to the applied forces; (iv) have

low signal to noise ratios. and (v) be able to measure forces and moments in a repeatable

manner. The following pages outline the design, fabrication and verification of a system to

measure the ice loads exerted on a faceted conicai structure.

Chapter 3

3.0 THE DESIGN, FABRICATION, CALIBRATION AND ASSEMBLY OF

THE FORCE MEASUREMENT SYSTEM MOCK-UP

To measure the high forces and moments experienced by the test structure and maintain as

rigid a test structure as possible. three large capacity six-component dynamomecen were

utilized simultaneously and the global forces and moments were resolved from the recorded

data streams. The dynamometers were rigidly attached to an upper load celi plate using

spherical bearings in flanged housings (Figure 3.1). This was done for ease of assernbly and

to reduce the moment pre-load on the dynamometen due to imperfections in the

manufacture of the instrumentation assernbly. For the test series in the IMD. the upper load

ce11 plate was attached to a tow post and the mode1 was rigidly secured to the lower load ce11

plate. The load ce11 details and specifications are @en in Appendix A.

Rigidly attaching several multi-ais force tramducers together with rigid reinforced plate

system. For the installation to be successful in the case where three load cells would be

sttached between two plates. the six surfaces on the plates which will contact the load cells

must al1 be perfectly paraIlel. and in the same plane. This is a difficult construct t o

manufacture given the size of the plates in question. If any irregularities exist between two

mating surfaces or in one or another of the plates. the result upon installation is that the force

Tow Post

Upper Load Ce11 Plate

Momentless Connection

Lower Load Ce11 Plate

Figure 3.1: Global Load Mesurement Assembly

transducer would be pre-loaded with a force equal to that required <O deform the plates such

that their surfaces mate with the end plates of the transducers.

Another issue is that any deformation in one or another of the plates would result in ri

moment being applied ta the load cells in the area of the setup. It has been found that large

moments applied to a multi-ais load ceIl will introduce erron in measurernents obtained

on other channels of that tmnsducer. To alleviate these concerns, a novel concept was

i ntroduced.

3.1 Design And Fabrication Details

The MD maintain

t -ve tri-axial force

transducers, three of

which were used in

the global force

measorement system

( F i g u r e 3 . 3 )

[ p r o d u c e d b y

A d v a n c e d Figure 3.2: AMTI 6 Component Force Transducen

Mechanical Technology hc. (AMTI)] of the MC8 series ( 1 X 89 kN capacity and 7 X 44.5

kN capacity). These are splashproof. six component load cells designed to measure forces

in each of the X, Y and Z directions as well as moments about each of these axes. If these

three force transducen were placed in series in the newly designed transducer setup. they

would have a combined force capacity in the Z direction of 178 kbJ and 89 hW in each of the

X and Y axes. These load cells are suain gage uansducen. structural members with strain

gages attached to them which, when loaded register a precise and consistent amount of svain

on any of a number of sets of strain gages installed on them. ïhe gages are positioned

strategically to avoid any strain king measured on <hem when off-axis loading is taking

place by fixing them on the neuval axes of the two non-sensing axes. A problem arises when

sufficient plastic or elastic deformation of the structural member takes place to alter the

geometry and shift the neutrai mis of the member. In this instance. errors will be registered

on the unloaded axes. This scenario most ofren occurs when large moments are being

applied to the force transducers. To aileviate this problem, a mechanism was devised to

minimize the moments experienced by the individual load cells during loading.

The load cells were attached to a rigd 25.4 mm thick steel plare with gussets applied to the

underside for additional tigidity. This mounting place was the one to be used in the acrud

test program (Lower Plate in Figure 3.4). Plate thicknesses and appropriate reinforcement

was determined using plate bending theory and forces anticipated from the maximum

forecasted testing conditions. Al1 plates and mounting fixtures were manufacturcd by the

21

w e l d i n g a n d

m a c h i n i n g

technic ians of

Technical Services in

the S . Crirew

buiiding. at -MUN.

The load transducers

were boIted to

seating pads on this Figure 3.3: Angular Contact Spherical Bearings

plate (called the 'lower load ce11 plate') which had k e n ground and polished fiat. The

purpose of the polishing was to reduce the chance of any

surface irreguiarities causing a deformation in the mounting

plate of the load cell. resulting in a moment applied to the

load cells during assembl y in the system.

T w o t e f l o n I i n e d ,

angular contact sphencal plain bearings (Figure 3.3) were

mounted in each of three flanged cups and separated by a

spacer. The spacer was designed such that the centre of

rotation of each of the sphericd bearïngs would coincide.

resulting in a 'bal1 joint' type arrangement. The flanged

Figure 3.4: Angular Contact Spherical Bearing and Flangeed Housing

cup was bolted to its rnating fiange, and

a 1.0 mm spacer was installed to permit

preloading of each of the bearing

assemblies preventing vibration and any

movement other than rotation. The

flanged assemblies were attached to the

second of the two rnounting plates

(Figure 3 3 , and provided the coupling

mechanism for attachment of the load

cells.

One of the two piates were to be Figure 35: Test Assembly

attached to a ngjd surface (the floor of the Structures Laboratory in the case of the

calibration, and the tow post in the case of the mode1 tests). white the load was applied to the

second plate. Assuming that the bearings were ideal. the effect of this arrangement was to

rernove any moment application to the load ce11 by the mounting furtures. The only moments

experienced by the individual load transducers would be a result of the axial forces applied

at each of the respective sphencal bearings. Physical limitations of the bearings resulted in

friction dong the sliding surfaces and a net moment being observed at the centre of

rotatiodconnection points between the load cells and the bearings. The arnount of friction

observed and the amount of moment applied to each load ce11 was directly propodonai to

33

the amount of axial load applied to bearing at that instant in time. Subsequently. an effort

was made in the design to minimize the distance berween the centres of the bearings and the

mounting plate of the load cell.

3.2 Global Load Celk

One AiMTI model SRMCS-6-20000 and two ~~ model SRMC8-6-lûûûû six component

load cells. manufactured by Advanced Mechanical Technology lncorporated (XiMiI), were

used in this configuration. The mesurement axes (X', Y ', 2') for the individual load cells

(Nos 1. 2 and 3) were oriented as shown in Figure 3.5. and the forces and moments were

resolved to a global X, Y. Z coordinate system. The origïn of the global coordinate system

was Iocated dong the centerline of the cone at the water level. The X-axis was positive in

the direction of ice motion. the positive 2-axis was directed venicaily upwards. and the

direction of the Y-axis was such that X, Y. Z form a right handed orthogonal coordinate

system.

The global forces in the X, Y and Z directions were obtained using a simple algebraic

sumation approach:

where:

Total force in the X direction

Total force in the Y direction

Total force in the Z direction

Measured force in the X' aris direction on the i th &bai load ceil

Measured force in the Y' a i s direction on the i i h global load cell

Meuured force in the Z* axis direction on the i i h dobal load ceil

Global ovemiming moment induced on the test structure was the result of the F,,. Fyl

and Fzl forces O bserved and their respective application points and directions of action wirh

respect to the giobd axes.

Due to different water levels that changed widi test scales. the relative location of the

global origin to which al1 moments were resolved changed. Consequentl y. the moment arrns

changed with varying water level.

The global moments were calculated using the following equations:

w here:

MX - Total overtmning moment about the X mis

My - Total oveminiing moment about the Y a i s

M t - Total ovemirning moment about the Z axis

XI - X location of the i th giobal load cell with respect to the defined axis system

Yi - Y location of the i'th giobal load ce11 with respect to the defined mis system

Z - 2 location of the i'th giobal load cell with respect to the defined axis system

The moment arms for each of the three axes used for each of the cdibration test setup

26

variations are given in the respective spreadsheets and are dependent on the location of the

transducer setup as well as the location of the hydraulic rarn (or ice ridge/floe levei) used to

load the apparatus.

Figure 3.6: Orientation of load ce11 coordinate axes with respect to globd coordinate axis

Figure 3.7: Orientation of global coordinate axis with respect to mode1 structure

3.3 Calibration of Test Setup

Prior to the test series in the MD. a mock-up of the test structure and support system was

created in MUN5 Structurai Engineering Laboratory. An in-depth series of calibration

checks were performed on the force measurement systems.

For mock up and

calibration purposes, the

Rangelbearing assemblies

were bolted to a 25.4 mm

thick mild steel plate (test

plate)(Figure 3.3). A plate

thickness of 25.4 mm was

used to reduce the arnount

o f deflection during

loading to a minimum and

was additionally stiffened

with 50 mm x 250 mm low

carbon steel flat bars

ninning lengthwise and

laterally, and attached to

Load Application

Figure 3.8: Calibration and Mock-up equipment

the plate with stitch welds. This plate had 12 bolt hoies flame cut in it to permit fastening

of the entire transducer setup to the structures laboratory floor. the^ were a total of 3 sets

of 1 holes cut in a 600 mm by 600 mm square, each set of holes rotated 30" from the other

such that the setup could be fastened to the floor in a total of 12 configurations. This test

plate was separated from the floor by cylindrical spacers to pemiit a safe separation distance

from the flanges to the floor surface. The spacers were fashioned as bushings. aligning the

Roor bolts and the test plate. The entire transducer assembly was attached to the Structures

Laboratory 900 mm thick reinforced concrete floor using the floor bolts and bushings. and

restraining the assembly in al1 axes.

The loading arrn shown in Figure 3.8 was also manufactured from 25.4 mm thick steel plate

and had a bolt pattern rnatching that of the Iower cone model to be attached to the transducer

assembly. The mating sufaces of the loading a m and model plate were also milled to a fine

tolerance (d.05 mm) to ensure a proper fit when attached and reduce residual stresses during

the test.

The manufacturer's specified transducer cdibrations were verified pnor ro the start of the test

series and used to compute the calibration coefficients for the individual force channels

throughout the test program.

3.4 Test Setup and Assembly

The test apparatus

described above and the

procedure listed below

were designed to

compare the forces and

moments measured

about the X, Y and Z

axes by the transducer

Figure 3.9: Loading A m conf igura t ion and

compare these measurements to the actual loads applied to the plates. The objective of this

test was to prove conclusively that the three tri-axial or six component loüd cells sandwiched

rogether between two rigid plates with spherical bearings. to eliminate the moments. would

accurately measure large trizxial forces and moments .

Loads were transmitted to the test jig using a 50 hydraulic rarn. The ram was mounted

in a test f rme constructed for this program using structural steel memben available in the

Structures Laboratory and had a load ce11 attached in series wiîh the loading a m to monitor

the applied forces. The applied load was tensile in nature, and was delivered to the loading

arm through an eye boit located near its front. using a braided steel cable and steel shackles.

3 1

To change the mgle of loading with respect to the Z axis. there were three options: < i l

Change the lateral position of the hydnulic ram: (ii) Change the orientation of the test plates

( L2 variations possible): or ( i i i ) Change the position of the loadin; arm on the front of the

mode1 plate (3 positions). In this test program. variations of al1 3 options were used.

The test consisted of two phases. the first was the application of known loads at

known distances from the load ceIl defined onpin after the transducer setup was zeroed. The

resuitant loads and moments on the loadcell plate were calculated andcornpared to the loads

and moments measured by the three six-component load transducen. Forces were applied

at each of three points (Figure 3.7) and varïed through a range of O to 13 W.

The second phase of the test consisted of electronicalty zeroing the transducer setup.

and then reducing the temperature of one of the load cell plates by a known amount. This

was achieved by placing approximately fifty hlograms of ice on the upper plate and

monitoring the plate's ternperature. The shi ft in the zeroes on al1 the transducer charnels was

nored as a consequence of irregular thermal expansion and confraction. and the equipment

was rezeroed. The procedure outlined in phase one was then repeared. This test was referred

to as the ice test.

Al1 tests were conducted in the Structures Laboratory of the S.J. Carew Building. .Mernonal

University of Newfoundand In dl tests. the test frame was loaded and permitted to 'settle'

32

Figure 3.10: Loading Equipment for Caiibration Tests

or allow for any reination of the loa&ng system. and the channels of al1 transducers were

scanned using a Hewiett Packard 3497A Data Acquisition/Control Unit and recorded.

3.5 Method of Cornparison of Results

To detemine an evaluation process for the performance of this system. we should firsr

consider what features an ideal force measurement system would have. Plotting the observed

data from the tmnsducer setup against a known standard. perfect correlation between the new

33

system and the standard would be indicated by a perfect fit of this line with the linear

equation of

where a = 0, and

b = l

A perfect fit would be indicated by linear regression with an R' value of 1. and a dope of 45 a

as shown in Figure 3.10, assuming that intercept of this plotted line is zero.

4 1

2 /

4 - 1 Slope O ;' I

I I I

* I

I

Figure 3.11: Perfect Correlation Between Measured and Applied Data

O 2 4 6 Measured Force

The value of R' and the slope alone are not good indicaton of goodness of fit with the line

of perfect agreement. It is conceivable that the line of Applied Force vs. Measured Force

plotted could have an £2' value approaching one. but that its slope and intercept could vary

greatly from the values of one and zero. It has been recommended that the Surn of Absolute

Predict~on Emor (SAE) (Castille et al. 1997) would be a p o d goodness of fit cntenon. This

is given as:

where n = the number of samples

If the value of the S a for a given set of measurements is equal to zero. then the values of

a and b given in equation (3.7) are equd to zero and one respectively. Slight modification

of this measure desensitizes it to the number of samples, resulting in a criterion called the

Mean Absolute Percentage Error (MAPE):

The MAPE is a finite measure of the magnitude of e m r observed through a senes of

measurernents. To present the error as a percentage of the applied load, the MME' could

be presented as:

The MAPE* is a gwd rneasure of the performance of the system overall. but to examine and

compare the performance of the system between each axis of measurement. there cornes into

play the issue of resolution. The system being examined here has a 89 kN capacity in the X

ruid Y axes, and a 178 kN capacity in the Z axis of measurement. These electrical

transducen (as are al1 electncal transducers) are designed such that they produce a maximum

output of. typicaily, IO volts when they are subjected to the peak loads for which they are

designed with the rnanufacturers prescribed excitation voltage king appiied. The resolution

of the system is determined by the tools that are used to measure the voltage output of the

transducer. A measurement tool which has 12 bit resolution wouid take the range of voltage

that it is designed to measure within and divide it into 1" discrete units. registering a detected

change in voltage whenever a change in voltage of at l e s t one of these uni& has k e n

detec ted.

As an example. a system rneasuring 10 volts with 12 bit resolution would register that a

change in

Discrete Unit = 10/(212 - 1) = 0.002442 Volrs

the output from the sensor 0.002442 volts. If we have. as is the case for two of the three

36

load transducers used in the senip. a capacity of J4.5 kN for which we would have an output

of 10 volts at maximum load then we would require a load of 10.867 N to be applied to the

transducer before a change would be registered on the system. In the case of the 89 k! load

cell. a load of 2 1.734 N would be required before a change in load was detected.

Based on this example. it is obvious that the capacity of the measurement system cannot be

ignored when evaluating the erron of the system. Given two load transducers. one with a

high capacity and one with a low capacity. if both have equivaient outputs and are connected

CO similardata acquisition systems. it would be expected that emrs in measurements for low

loads would be geater for the larger capacity system. Dividing the A W E by the capacity

of the system and multipiying it by 100 will result in the W E k i n g presented as a

percenrage of the total capacity of the system in this a i s . This will not only desensitize i t

to the capacity of the system. but dso result in the -Modified - W E ' (MMAPE) being a

number that may be cmsidered more universally.

The larger the capacity of the system. the more conservative will be the value of the

W E . a direct result of the system capacity divisor in Equation 3.12. The global force

applied to the systern may be defined as the vectorially summed forces in the X. Y and Z

axes k i n g applied to the system. The global force capacity will Vary depending on the

37

direction of loading, but will be a maximum when a load is being applied dong the Z axis

of the system only. Subsequently, calculated values of the MMAPE for the global force

component use the sarne capacity as that of the Z axis. 178 k.. The system moment

capacities may be determined from the axial load capacities of the load cells multiplied by

the moment m s or the distance from bearing centre to bearing centre about the appropriate

axes. S ystem capacities are given in Table 3.1.

Table 3.1: System Capacities

- -.

The vaIues of the LMMAPE and MAPE factors obtained for this set of cdibration tests are

given in Tables 3.2 and 3.3 and graphical cornparison of the forces and moments foreach test

with the line of perfect agreement is given in Appendix B.

Component

F,

Capacity

89 hV

Component

-M,

Capacity

35 kN rn I

Table 3.2: MAPE' and W A P E Factors for Force Measurement During the Cali bration Tests

MAPE' MMAPE MMAPE MAPE'

1 Orientation 2

1 Orientation 3

1 Orientation 4

1 Orientation 5

( Orientation 6

1 Orientation 7

1 Orientation 8

1 ~rientation 9

1 Orientation 10

1 Orientation 13

1 Orientation 14

1 Ice Test

NOTE: 1. Due to the direction of loading, there is no error mesure possible for this direction.

Table 3.3: L W E ' and MMAPE Factors for Force Measumment Dunng the Cdibration Tests

Test

Orientation 1

Orientation 2

Orientation 3

Orientation 4

Orientation 5

Orientation 6

Onentation 7

Orientation 8

Orientation 9

Orientation 10

Orientation 11

Orientation 12

Orientation 13

Orientation 14

Ice Test

NOTE: 1. Due to the direction of loading, there is no error mesure possible for this direction.

3.6 Discussion of Results

For the entire series of tests. the system generally pedomed well. Global force measurement

erron had a MME' value of less than 2% in al1 cases. includin; the ice test. Orientation 1

had a direction of global loading such that the force was largely in the X and Y ( 1300 N and

8000 N. respectively) directions. and some loading in the Z direction (approximately 4000

N). The MAPE' values for al1 three measurement axes were less than 5% far both force and

moment measurements.

Orientation 2 was a pull in the X and Z directions only, theoretically resulting in zero force

dong the Y mis of measurement and no moment about the X or Z axes. For this reason. no

emr analysis is available for these channels of measurement. In actud fact. small forces and

moments were recorded dong these zero load axes which were typically less than 10% of the

recorded values dong and about the intended axes. Imperfections of the orientation of the

loading and measuring equipment during loading most likel y resulted in some off-axis

loading for this test. Global and axial erron MME' error values were all well below 2%.

Orientations 3 to 14 inclusive were a series of pulls dong a variety of directions. and al1 with

the exceptions of orientations of 4.5 and 6 resulted in good MAPE' values dong al1 axes of

measurement of less than 5.55%.

Orientation 4 was a pull in which the Y load was less than half of that recorded in either the

X or Z directions. The MAPE' index for the global force of this test was 1.37%. and the

MAPET for the X and Z directions were 0.54% and 3.4 1 % respectively. The MAPET for the

force in the Y direction, however, was 13.94%, more than three times the error recorded on

any other mis. Similarly, the moment about Z axis was unusually high, with a MAPE* value

of 15.16%. Since the global errors are quite small, and inspection of the loading curves

demonstrates that the percentage error is consistent throughout the loading range, it is

theorized that once again. the equipment ei ther shifted when k ing loaded or the direction

of the applied load was incorrectly recorded Similar observations may be made for tests

conducted dong Orientations 5 and 6. Orientation 5. in particular, was a loading scenario

which was intended as extreme. The line of action of the force resulted in very little force

being applied dong the Y mis of the system. but it was significantly off centre. resulting in

a very large moment about the Z axis. The transducer setup would never be subjected to this

type of extreme loading scenario under regular loading conditions in the upcoming test

program. Still. the global force measurernents for orientations 5 and 6 had M E ' values

of 1.49% and 1 4% respectively.

MMAPE values were consistent as well. Again. the error indices were higher for tests on

Orientations 4.5.6 and 7, as were the MAPE' indices. but this is expected. The global force

values of the MMAPE were consistently very low (0.0004 - 0.0019). and the forces in the

X and Z axes were also quite good The MMAPE index for the force in the Y direction.

42

however, was the highest of the four calculated indexes in most cases (9 out of 14 tests). in

al1 of these cases. the loading frarne was positioned relative to the transducer setup such thar

loads being applied to it were mostly in the X and Z axes. resuiting in small loads being

measured with a high capacity system and consequently larger emon being recorded than on

the other sensing axes. This example is demonstrated in the extreme by examining the

ioading direction and subsequent erron for Orientation 5. in which the transducer setup is

loaded aimost exclusively in the X and Z axes. but off-centre of the setup.

The Ice Test conducted was intended to simulate the effect on the transducer setup of having

one plate submerged in ice water during the test program. while the other was in the open air

dunng testing. With one of the loading plates chilled to approximately two degrees Celsius

and the other at room temperature. the equipment was again loaded. The transducer setup

perfonned well within specifications for this [oading scenario. with 'MAPE' values for the

global forces and the forces in the X, Y and Z directions of 1.59%. 0.93%. 1 .OS% and4 16%.

respectively, but the MAPE' values for the moment measurement axes were a lirtle higher

with values for the measurements about the X, Y and Z axes king 8.45%. 1.70% and 7.108

respectively.

Chapter 4

The objective of the test program was to perform a series of ice structure interaction tests at

a variety of scales. The equipment used in the program had to be structurally sound and

simple enough that it would be possible to switch from one modei to another in minimal

rime. A finite window of opportunity was available when equipment modifications could

take place. This typicaily occumd after the ice sheet had been g o w n and the ridges

manufactured, and while the ice was king tempered or its strength lowered. Working on the

equipment was not possible during ice growth as it would have an effect on the crystal

structure of the ice to be teste4 affecting its material properties at the time of testing-

4.1 The Mode1 Test Structure

Two model scales. 125 and 150 with three configurations were tested in this pro-.

Figure 4. L is a photograph of the 1 : 25 scale large neck model rnounted in the lMD tank. Due

to the nature of

the geometry of

the cone, it was

p o s s i b l e to

manufacture a

model that could

b e converted

frorn one scale to

the other. This

was done by

manufac turing Figure 4.1: 1 :25 Large Neck Mode1

various sizes of

necks and collars which could be attached to a universal lower cone. By changing the neck

and collar of the model as well as the waterline diameter of the lower cone, the scale of the

model structure could be altered Dimensions of the three model configurations are shown

in Figures 4.2,4.3 and 4.4.

The mode1 was constructed of 6.35 mm thick marine p d e aluminum plates welded to a

rigid frame of 50 mm x 200 mm aluminum channel. It was designed to be as rigid as

possible and of a modular construction to speed the change over of the model when testing

two different neck sizes and two scales in a limited time period The main componenr of the

mode1 was the lower cone structure to which various necks and collars could be attached to

facilitate these changes.

Due to rigidity considerations. it was decideci to attach the model direcrly to the IMD carnage

as opposed to a mode1 test frame. The model was attached to the undenide of the M D

cômage

Y

taro

Noie: Al1 diameten are corner to corner; Ail r l o p u are ot the facct centen and given as a mtio of vertic*l:horizont.l: and Al1 dimensiorm u e in miiiimetcn.

Figure 4.2: 125 Large Neck Mode1 Dimensions

Note: Ai l diameters are corner to corner: ~ i 1 s l o p are of the lacet centers and giwen s~ a ratio of vertical:horizontak and âii dimeluions are in rnillimetcrs-

figure 4.3: 125 Small Seck Model Dimensions

Note: AU diameters are corner Co cornec AU dom art of the h c c t centers and &?en as a ratio of vertica.khorizoa~ urd AU dimeruions are in millimetut-

Figure 4.4: 1 5 0 Large Xeck Model Dimensions 17

using a specially designed towing p s t constmcted from 300 mm x 300 mm x 12.7 mm steel

box beam (Figure 3. L ). and was instrumented to mesure the global forces and moments on

the entire structure as well as the forces on the verticai neck portron of the model.

Some of the equipment used in the calibnùon and verification of the load ce11 setup at the

S J. C m w Building was reused in the final test senip for the program. The load ce11 plate

which had a truncated hexagonal bolt paneni was designed such that the model rnulti-faceted

conical structure would be able to be mounted on it. In the final test program. this was the

Iower load ce11 plate.

4.2 Instrumentation and Data Acquisition Spiem

The waterline was stationary relative to the scale model cone. but changed relative to the

force measurement system when the model was changed (Figures 1.1, -1.3 and 4.4.

Subsequently. the moment arms used for computations varied frorn one scale to another. The

values of y. Y, and Z for the 125 and the 150 scde models are given in Table 4.1.

The global force load cells located in the lower cone were partially submerged in one model

configuration and completely submerged in the others. The ioad cells were waterproofed by

the manufacturer and fined with a water proof rubber boot' prior to installation to ensure

48

Table 41: The values of X,, Yi and 2, for the 125 and the 1 :50 scale modek

I 1 : 35 Scale I

I 150 Scate I

z, (ml

-0.082 4

-0.082

-0.082

that no water leaked into the transducer housings.

Yi (m)

O

-0278

0.278

Load Ce11 I

1

7 - 3

Load Ceil 1

1

7 - 3

The force and moment instrumentation for the mode1 set up had to be designed and

manufactured especially for the program. No existing equipment had both the force and

moment capacities for the expected loading scenarios of this program.

X, (ml

-0.293

0.707

0.707

& (ml

-0.293

0.707

0.707

Yi (m)

O

-0.278

0.278

Z (m) -0.332

-0.332

-0.332

4.2.1 Neck Load Cells

Due to the t h e mode1 configurations tested. two vertical neck sections were utilized in this

test program. The large diameter neck was used in the L 2 5 large neck mode1 configuration

(Figure 4.5) and was equipped with two A m model SRMC6-6400 dynamometee of

capacity 17.8 kN (Nos. 4 and 5). A mounting unit for these load cells was rigidly attached

to the cone and the vertical neck was rigîdly attached to the dynamometen. The

rnanufacturers specification sheets of the load cells are given in Appendix A.

The local axes for the individual neck Ioad ce lis were oriented such that the Z' axes were

parallel to the global Z axis. and the X' axes were out of phase with the global X axes by r

30". and the Y' axes were

such that X', Y' and Z' formed

a nght hand coordinate

system (Fi,- 3.5).

Using the known geometry of

the system, the forces

Vertical Neck

MT1 MC6 Load Attachrnent to Lower Cone

of the model. The following equations resulted:

w here:

Total neck force in the X direction

Total neck force in the Y direction

Total neck force in the Z direction

~Veasured force in the X' a x i s direction on Ioad ce11 4

Measured force in the Y' axis direction on load ceil 4

Measured force in the 2' ruris direction on load ce11 4

Measured force in the X' axis direction on load ceil 5

Measured force in the Y' axis direction on load ce11 5

Measured force in the 2' a i s direction on load ce11 5

Moments on the neck portion of the test structure were not to be measured in this test series.

The full scale structure would be made of welded steel plates with moments king applied

to the neck portion of the structure only king an impossibility.

The srna11 diameter neck was used in both the 125 small neck and 150 large neck mode1

configurations. Due to size consuaints. the models were fitted with only one of the SRMC6-

6-4000 dynamometers. In a similar manner. the dynamometer was the connection point

between the neck of the mode1 and the lower cone (Figure 4.6).

The axis of the local coordinate system of the load ce11 used in the small diameter neck of

the mode1 was oriented such that the X', Y' and Z' axes were oriented in the sarne direction

as the X, Y and Z axes of the global coordinate system. Consequently. the forces on the

small diameter neck could be

read directly from the

transducer output.

4.2.2 Accelerorneters and

LVDT

'- verti , Xeck '- XMTI

Load

c a 1

MC6 Cell

Attachrnent to Lower Cone

Accelerations of the mode1 in Figure 4.6: Neck load ce11 arrangement for the 150

the three principal axes were large neck and 125 small neck models

measured using three Systron Donner accelerometers. The deflection of the tow pole and the

mode1 were measu~d by two Schaevitz linear voltage displacement transducers (LVDT).

The particulars of the accelerometers and displacement transducen. along with their

specification date are aven in Appendix A.

4.2.3 Data Acquisition System

A schematic arrangement of the data acquisition system is aven in Figure 1.7. The data

acquisition hardware was mounted in the towing carriage operators'room and was connected

to appropriate transducen mounted on the mode1 via cables passing under the floor.

Excitation for the transducers was provided by a NEFF System 620 Series 300 signal

conditioner available in the Ice Tank towing caniage. The tnnsducer outputs from the load

cells and the LVDTS were filtered by a 10 Hz anaiog Iow p a s filter and digitized at a rate

of 50 Hz whereas the accelerometer outputs were filtered by a 100 Hz analogue low pass

filter and digitized at a rate of 200 Hz by a NEFF system 620 Series 100

amplifier/multiplexer and stored on a Vax i in50 cornputer for analysis. The analog outputs

of the transducer were recorded by a KYOWA RTP-6OOB 14 channel tape recorder as

backup.

Data collected were down-loaded to the faculty's VAX cluster via ETERNET for

M e r analysis.

1.3 Test Program with Results

The test matrix. with details of the test program. is given in Table 4.2. It was developed to

accommodate the testing of two scales (125 and 150) of model. two sizes of neck at one

scde (1:15), and a variety of ridge and sheet ice strengths and thicknesses over a five week

period. The models were tested in the face on orientation.

Five tests were performed on 14 model ridges in 5 ice sheets. In each test. level ice tests

were perforrned at model velocities of 0.01 m/s, 0.04 mls and 0.06 d s to assess the effect

54

of different interaction rates. A11 ndge tests were performed at a velocity of 0.W mls.

Two types of ridge construction techniques were developed for this test proagam. The Dump

Truck (DT) ridge was constructed by dropping slabs of sheet ice in a gïven location and

breaking it up into srnaller pieces. The result was a ridge made of sheet ice rubble with

randomly oriented crystals. The Split Layer (SL) ridge was constructed by gently placing one

layer of sheet ice atop another unti 1 the desired thickness of ridge was achieved. The material

properties of the ice were adjusted by allowing them to temper by placing thermal blankets

over the top of them and allowing them to warm while the sheet ice continued to freeze in

the cold -20" C air. Typically, L DT ridge and 2 SL ridges were tested per ice sheet.

AI1 DT ridges tested in this program simulated the I in 100 year multi-year ridge ar the 150

scale (Le. 54 cm thick and 14 kPa flexural strenagh): while. the SL ridges tested were

designed to model conditions comparable to yearly average ice conditions found in the

Canadian Arctic. It should be noted that the DT ridge tested with the model at the 150 scale

(i.e. Test MUNCONE7-007, ridge thickness - 50 cm. ridge flexurai strength - 23.8 kPa)

specificaily simulated the design loading condition for the Beaufort Sea.

The degree of variation in the model ice paramerers throughout the St. John's test program

permits the extrapolation of the test data to other scdes by modifymg the geometric scaie of

the mode1 structure.

1 Test Modef: MOL: Sheet No. 5 1

Table 4.2: Test Matrix for the Multifaceted Cone Tests - M D

NOTES: 1. L - level ice test; R - ridge ice test 2. Ice flexural smnath - Bottom in tension 3. SL - Split layer ridge; DT - Dump truck ridge 4. Al1 tests run in face-on orientation.

a,' kPa

Test Test Type

V (cm/s)

H (cm)

I

h (cm)

Test ~Modet: 125s; Sheet No. 1

MJNCONE6-O03 L 4 13.4 12.5 l

lMUNCONE6-O03 L 1 1 23.5 LMUNCONE6-O04 L 6 12.4 22.5 iMUNCONE6-O05 R 4 1 22.5 50 12 DT

on.' kPa

tviU'NCONE3-00 1 MUNCONE3-O02 1MUNCONE3-003 -;MUNCONE3004

Ridge T-ype

1

SL SL

DT DT

L L

L+R R

44.4 14.1 43.6 42.5 39.4 29.3 38-9

Test Modei: M5L; Sheet No. 3

LMUNCONE~-005 1 L

1 6 4 4

42.0 32.0

50 50

L'MUNCONE~-OO I MUNCONE~-002 MUNCONE4-003 MUNCONE4-004 MUNCONE4-O05 MUNCONE4-O06 MüT\JCONE4_007

15.8 15.8 15.8 15.8

4

r

99.3 ,

48.7

162 33.5

41.1 40.6 10.4 40.2 39.7 19.7 19.6

14.8

Test Modei: 13L; Sheet No. 3

14.8 14.8

MUNCONE~-006 1MUNCONE3-007

43.5 32.7

50

16.0 16.0 16.0 16.0 16.0 16.4 16.3

L L L R R L R

R 1 4 R 1 4

1 6 4 4 4 4 4

.MUNCONES-O0 1 mNC0NE5-002 MUNCONES-O03 LMUNCONES-O04

112.2 69.2

26.5

Test Mode1: 1:2SL; Sheet No. 4 MUNCONE6-00 1 R 4 12.4 23.5 33.4 65.5 SL

SL SL

DT

80.5 1 SL 10.9 1 DT

30.7 30.3 29.9 37.3

L L

L+R R

36.8 50

1 1 9.5 6 4 4

9.5 9.5 9.5

Chapter 5

5.0 PERFORMANCE OF THE EQüIPMENT

The test equipment was installed in the M D ice testing facility and verified pnor to actual

testing. Modifications. adjustments and checks of the setup had to be completed prior to the

start of growth of the ice sheet. .4ny disturbance of the water during the formation of the ice

crystals would have disturbed the ice crysral structure and significantly affected the sheets

physicd propenies.

5.1 Load Cell Calibration

The manufacturer's specified transducer calibrations were again verified prior to the start of

the test series by applying

known loads to the individuai

load cells. and used to compute

the calibration coefficients for

the individual force channels

throughout the test pro-.

In-situ calibrations of the load

cells along al1 three principle

axes were conducted both

i mmediatel y after each mode1

was instailed and pnor to each

test to ensure that ail force

rneasurement systems were

working properly. The ongin (

place for the model c o n f i p t i

Figure 5.1: Load Cell Cdibration Setup

~f the coordinate system used was located in the appropriate

!on as presented in Chapter 4.

Al1 force measurement systems were calibrated at the same time usine the method s h o w in

Figure 5.1. h steel cable was anchored to the IMD tow tank and wrapped around the neck

portion of the model. Pads were placed between the cable and the mode1 to protect the paint

finish, essential to the friction characteristics of the model, A hand winch and calibration

58

O 1000 2000 3000 4000 5000 6000 Measured Force [NI

Measured Force

- Line of Perfect Agreement

Figure 53: Globai Force Calibration Data from Test Conecd 4

load ce11 were placed in series between the point where the cable was anchored and the

model. Varying loads were applied to the structure using the hand winch and the reference

force for cornparison with the test equipment outputs was obrained from the in-line load cell.

Measurement of the precise line of action of the applied force was difficult in the tow tank

with the equipment installed For this reason. the forces in the X, Y and Z directions

resoived from the test setup were vectorially summed and compared to the applied force as

measured from the in-line load cell. A sample of the data from one of rhese calibmion tests

59

is given in Figure 5.2. The test shown has a i W E ' of 4.40% and a A W E of 0.0049.

5.2 Naturai Frequency of the Test Structure

For model testing purposes. it was essential that the naturai frequency of the test structure

be significantly higher than the ice breaking frequency during testing. This was to ensure

that resonance in the model structure and force measurement equipment did not occur and

interfere with the recorded force data-

In-situ f r e e

vibration tests

were perfomed

on each model

configuration

(Figure 5.3). ui

each test. the

modei was loaded

up to 500 kg in

the along-tank

direction. and the

Cable Cutters

Assemblies

loading cable was Figure 53: Senip for Dynamic 'Pluck" Test

60

5 10 15 Frequency [Hz]

Figure 5.4: PSD analysis of acceleration data in the X direction

Figure 5.5:


PSD analysis of cone acceleration in the Y direction

61

abruptly cut. ,411

instrumentation and

data acquisition

systems. as well as

r e a l - t i m e

monitoring was

operated during the

t e s t . P o w e r

Spectrum Density

(PSD) analysis of

the accelerometer

signals (Figure 5.4.

5 . 5 and 5 .6 )

showed a dominant

frequency of the

cone in the X

direction at about

11 Hz, and two

r e s o n a n t

frequencies of the

cone in the Y


Figure 5.6: PSD analysis of post acceleration data in the X direction

O 1 2 3 4 5 Frequency [Hz]

Figure 5.7: PSD analysis of ice force data in X direction

d i r e c t i o n a t

ripproximately 8

and 1 3 Hz. The

post alone was

instrumented dong

the X axis only, and

d e m o n s t r a t e d

r e s o n a n t

frequencies at 8. 12

and 17 Hz. with a

resonance afso

being observed at

slightiy greater than

one Hz as a

consequence of

resonance of the

towing carriage

i tself. Acceleration

data were also

collected dong the

Z axis for the cone,

O 1 2 3 4 5 Frequency [Hz]

Figure 5.8: PSD analysis of ice force data in Y direction

but no distinct low

f r e q u e n c y

resonance was

o b v i o u s .

demonstrating the

high stiffness of the

towing systern in

this axis. The 8 Hz

resonant tiequency

observed was due

to the rigid and

relatively heavy towing pole and model structure undergoing a rotational motion about the

connection to the c d a g e while the higher frequencies were due to the model structure

undergoing rotational motion about the centre of stiffness of the global load ce11 assembly.

The ice breakmg Iength for the MD'S 125 scde model has been analyzed and was found to

be approximaieiy 0.1 characteristic length or 1 times the ice thickness. This is the case for

test thicknesses ranging from 0.09 to 0.16 m independent of test velocity (from 0.0 1 to 0.06

rn/s). The corresponding ice breaking frequency for the test velocity range (0.01 to 0.06) is

less than one Hz for dl tests with the 125 scaie model. (i.e. 0.06m/s / 0.09 m = 0.666 m is

the estimated higher Iimi t.)

PSD analysis of the force data collected (Figures 5.7 and 5.8) demonstrated that only very

low forcing frequencies on the order of 0.1 Hz were observed. The significant differences

between the frequency of the forces on the mode1 and the natural frequencies of the structure

and towing system validate the effectiveness of the system for this test series.

5.3 Test Results

Test data were acquired using the equipment outlined in section 4.2.3 and were digtally

filtered with an upper cutoff frequency of 2-75 Hz before plotting. The data for

MLNCONE7-O07 are presented in Figures 5.9. 5.10 and 5.11. This test was of the most

extreme ice feature tested in this program. the one in one hundred year ridg. and resulted

in the highest forces recorded as well.

For the test program, analysis of the force and moment data were broken into two sers. The

level ice tests were studied separately from the ridge test data in an attempt to isolate the

extreme ice feature tests. Peak forces and moments acquired during the ridge tests of the

program are presented in Table 5.1.

The ridge test data had significantly higher peak loads. Forces in the X direction had a

maximum value of approximately 25 icN and forces in the Z axis peaked at approximately

13 kN (MIJNCONEJ) while Y forces had a maximum value of approximately 1.5 kN

(MUNCONE7). Y forces would only be significant in the cases where a large piece of ice

cleared around one side or the other of the cone. For this same reason, moments about the

Y axis were significantly geater than moments about the Z or X axes. Y moment had a peak

value (MLNCONU) of approximately 27 W-m. while moments about the X and Z axes had

a peak value of approximately 5.5 kMm.

F O R C E O N STRUCTURE IN X, Y , Z D l R E C T f O N S MUNCONE7 - 0 0 7

A V I - 4 - 0 6 k S MAX - O . L 4 k8

Y i M - 1 - 2 7 k l l

A V R - -3.60 k l l M A X - - 0 . 4 2 k l l M I S - -7.maiCll

T l M E (s)

WBtlE 1 1 1 8 -- I Y A L L I C C TtllCKtIS.8 - 10.3 C m n i n a 8 T U ~ C X W S . ~ - SO.O wm ? R l C T l o u -- 0.Oa ics D r l l 8 i T r - o l o . ec/ra-3 l i O I 8 T l D t U - 100 a m

O t l I C ? l O l l -- B R O I O QU ? L K X i T R (down) - 41 .0 k ? a rLCx S t l ( d o r a ) - 1s- IO Y?.

a r t 8 0 - 6 . c a r / r ILKX J T E ( u p ) - 18.4 k C a t ~ x UT R (UQ) - 11.70 k C a

M U L T I F A C E T E D C O N E T E S T S A = 5 0 . 0 0 , N R C / I M D

Figure 5.9: Time series trace of global forces in the X, Y and Z directions for test MUNCONE7-007

t

MOMENT O N STRUCTURE A B O U T X. Y. Z AXES MUNCONB7-0 0 7

- .. -

Figure 5.10: Time senes trace of global moments about the X. Y and Z axes for test ~MUNC0r\lE7_007

- -

F O R C E O N NeCK IN X. Y . 2 D1RGCTIONS MuNCONE7-0 0 7

2 a s O 7 s L O O 130 17s

T l M E (s)

M U L T I F A C E T E D C O N E TESTS

Figure 5.11: Time senes trace of forces on the neck in the X. Y and Z directions for test MUNCONE7-007

X = 5 0 . 0 0 . N R C / I M D !

Neck loads for the entire program were significantly less than those experienced globally as

a result of al1 forces being applied to the neck being a consequence of already failed ice

riding up the sioped side of the cone and clearing about the neck. Peak loads for the neck

occurred in the X direction during ridge tesring and resulted in peak forces of approximately

4 kN in the X direction (MüNCONE7). 2.6 kN in the Y direction (MUNCONE7). and 0.6

kN in the Z direction.

The data from the diflerent scales of the program were subsequently analyzed et al.

1993; Croasedale et aI. 1993). and it was determined that the data scaled as expected.

verifying the integrity of the force rneasurement system.

Table 5.1: Summary of Ridge Test Results

1 1 Maximum Global Forces 1 Maximum Global Moments 1 Maximum Neck Forces

Note: 1 . Absolute maximum 2. Vcnical downwurd 3. Moment, My, about ncgative Y üxis

Chapter 6

6.0 CONCLUSIONS AND FINDINGS

Two sets of experiments were actually used to gauge the performance of this system. The

fint was a set of tests conducted in the Structures Laboratory of the S.J. Carew Building to

study the performance of the system under a variety of loading conditions and two

environmentai conditions. The second was when the equipmenr was installed in the MD

for its intended purpose. Examination of the transducer data from the variety of calibration

tests conducted permits an evaluation of the system for its intended purpose.

A total of fifteen calibntion tests were performed on the giobal force transducer setup with

14 variations in the force line of action and one test in which one of the connection plates

were chilled using ice. resulùng in thermal contraciion of the plate and loading of the

individual load transducers-

Two measures were developed and used to gaauge the performance of the uansducen. The

69

fint is the *MAPET. which gives a good overall indication of the performance of the system

along the indicated axis of measurement in the form of a percentage error. The second is the

-MMAPE. which is a version of the MAPE' desensitized to the capacity of the systern dong

a specific measurement ais. This tool is a difficult one to use when gauging the

performance overall. and is useful only for relative performance indices between axes where

the same type of parameter (force. moment etc..) is k ing measurea.

A number of finding for this thesis are presented below:

1. The system generally performed weil. The global force error index (.LWE') did not

exceed 2% emr for any of the tests perfonned, including the ice test which was

intended to simulate the thermal conditions of the ice tank. Force rneasurements

along each axis resuited in larger but stiii acceptable erron generally less rhan 5%

with the exception of orientations 4 .5 and 6. in these cases. the errors recorded in

the measuremenü of the forces along the Y a i s were significantly higher. The

direction of the loading in these cases was such that the force was most1 y in the X

and Z directions. and the minimum force was registered dong the Y axis. These

small forces measured on a system whose capacity is quite large resulted in larger

erron between the measured and applied loads.

3 . In addition. the correlation between measured and applied loads dong each mis

relied on determining the precise line of action of the load in three dimensions. Error

in determining the anchor point for the hydraulic acniator used to load the system. or

locating the system itself. or possibly the system shifting during the course of the test

are al1 possible sources of error that would account for the discrepancy between the

excellent results of the global forces and the larger errors of the resolved forces.

3. Once installed in the MD for the performance of the final phase of the multifaceted

conical structure tests. the equipment once again performed well. Cdibrations

perforrned pior to each test sheet of ice and between each run verified the integrity

of the transducer setup throughout the program.

4- Power spectmrn density analysis of the ice failure &ta collected during testing

demonstrated that the loading frequency was of the order of 0.1 Hz. and the naturd

frequency of the structure was on the order of 10 Hz, ensuring that resonant action

of the test equipment did not corrupt the results.

In the future. more detailed calibrafions should be performed on this equipment using

methods in which the precise direction and magnitude of the loads king applied with respect

to the coordinate system of the force measurement system c m be determined Additiondly,

using a precise force vector for the applied loa& coefficients may be detennined for

7 1

correcting the output from this force measurement system.

The use of sphencal bearings in this rnanner to protect force uansducers from large moments

should be further studied.

Antkowiak, J.H., and Rencis, J.J.. (1994). "Geometric Nonlinearities in the Design of Force Transducers". Advances in Engineering Software, Barking, London. England. pp. 1 1-16.

Camrnaert , A.B., and Muggeridge, D.B.. (1988). "Ice hteraction with Offshore Structures". Van Nostrand Reinhold, New York, New York,

Castille. E., Losada, MA.. and Puig-Pey, J., (1997). "Probabilistic Analysis of the Nurnber of Waves and Their Influence hthe Design Wave Height of Marine Structures", Safety of Structures Under Dynarnic Loading, (Tapir, Norwegian institute of Technology). Trondheim. Norway. pp. 730-729.

Croasedale, KR.. and Muggeridge, D.B.. (1993), "A Collaborative Research P r o g m to hvestigate Ice Loads on Multifaceted Conical Stmctures", Proc. l î h Intl Conf. On Port and Ocean Engineering under Arctic Conditions. Vol. 3. Hamburg, pp. 475486.

Fan, L.C.. and Jin. Y.G.. (1990). "Ice Survey in Liao Dong Bay and Ice Force LMeasurement on JZ20-2- 1 Platform". SMM '90. International Shipping and Marine Technology Market with Congress. Proceedings, Hamburg, Germany. pp. 133-137.

h i . M.B., Timco. G.W.. and Muggeridge, D.B., (1992). "Ice Loadin; on a ~Multifaceted Conical Structure", Technical Report IME-CRE-TR-005. National Research Council of Canada, hstitute for Mechanical Engineering, Cold Regions Engineering, Ottawa, Ontario.

Lau, M., Jones. S., Tucker, J., and Muggeridge, D.B.. (1993). "Model [ce Forces on a Multifaceted Cone", Proc. 1" Inrl Conf. On Port and Ocean Engineering under Arctic Conditions, Vol. 3, Harnburg, pp. 537-546.

h u , M., Tucker, 5.. Jones, S., and Muggendge. D.B.. (1993), "Model Ice Forces on an Upward Breaiung Multifaceted Cone". NRC Canada Report. TR- 1993-07.

Maatanen. M.. (1977)' "Ice Force Measurements at the Gulf of Bothnia by the Instmrnented Kemi-1 Lighthouse", Proc. 4& h l . Conf. On Port and Ocean Engineering under Arctic Conditions, Vol. 2. St. John's, pp. 730-740.

Masroor. S.A.. and Zachary, L.W.. (199 1). "Designing an All-purpose Force Transducer". The 1990 SEM Spring Conference on Experimental Mechanics. pp. 33-35. Albuquerque. *M.

McConneIl, KG., (1995). "Vibration Testing". John Wiley and Sons, New York.

McConneIl, K.G., and Varoto, P.S.. (1993), "A Mode1 for Force Transducer Bending Moment Sensitivity and Response During Calibration", The 1 lh International Modal Analysis Conference, Vol. 1. pp. 364-368, fissimmee. Ronda.

Metge, M., and Tucker. J.R.. ( 1990). "Multifaceted Cone Tests - Year Two. 1989- 1990, Technical Report. Esso Resources Canada Limited, Calgary, Alberta

Metge, M., and Weiss. R.T.. (1989), "Multifaceted Cone Tests 19884989". Technical Report. Esso Resources Canada Limi ted, Calgary, Al berta.

Murray, J.J.. Winsor. F.N., (1997). "Impact Forces on a Jacket Deck in Regular Waves and Irreguiar Wave Groups", The 1997 Offshore Technology Conference. pp. 45-54, Houston. Texas.

Nevel, D.E., (1971), 'The Ultimate Failure of a Fioating Ice Sheet". Proc. Zd WHR Ice Symposium, Leningrad U.S .S .R.. pp. 23-17.

Raiston, T.D., (1977), "Plastic Limit Analysis of Sheet Ice Loads on Conical Suuctures". Proc. IUTAiM Symposium on Physics and Mechanics of Ice, Copenhagen. Ed. Tryde, P.. S pringer-Verlag. Berlin, 1980, pp. 189-308.

Saeki. K.. Hamanaka. K.. and Ozaki. A.. ( 1977). Experirnental Snidy on Ice Force on a Pile". 4& International Conference on Port and Ocean Under Arctic Conditions, St. John's, Newfoundland. Canada, pp. 695-706.

Spencer, D., HiII, B., Kirby, C.. andNeve1 D.. (199 1). "Properties of Multi-Year Ridges Built in MD'S Ice Tank", The 11' International Conference on Port and Ocean Engineering under Arctic Conditions, Vol. 2, pp. 635-648, St. John's, Newfoundland-

Wessels, E., and Jochmean. P.. (199 i), "ModeVfull scale correlation of ice forces on a jacket platfom in Bohai Bay", Roc. 11" Intl. Conf. On Port and Ocean Engineering under Arctic Conditions, St. John's. Canada, pp. 198-313.

Wessels, E., and Kato K., (1988), "Ice forces on fixed and floating conical strucnires", IAHR Ice Symposium. Roceedings, 9th, Sapporo. Japan. Vol.2, pp.666-69 1.

Angular contact spherical plain bearings A

Sliding contact surfaces: hard ctiromium/PTFE

?ui

G E S W - sliding material: PTFE fabnc

GE 45 SW GE SOSW GE 55SW

14 0.20 15 430 16 420

16 424 17 424

18 424

19 434 19 0 2 4

22 -0.30

22 - 0 3 22 -0.30 24 Q#

24 o.= 27 4.30 27 4.40

30 0.w 30 4-40 30 -0.4

33 au, 36 4.a 38 4.40

42 4 s 42 0.- 45 0 3

48 OS 54 0.50 61 0.m

61 4.60 66 4,60

-2-Mo

72 -0.60 83 4.70 03 4.70

% 4m

. ' U.S. Patent 114493220

Butirlin SAMC8-Qô7

HYSTERESIS F ~ . f y . h 0.2 0.2 %FSb"'

m E S Ç

43-8 87.6 105 21 .O

NON-UN- 020 0.20

ÇRMCB - X - XXXX

APPENDIIX B

CALIBRATION TEST RESULTS

Figure B-1: Direction of Load Application - Orientation 1

81

O 5000 10000 15000 20000 25000 Measured Global Force [NI

Line of perfect agreement

+ Measured Force

Figure B-2: Global Force Comparison - Orientation 1

MAPE' - 0.97

O 5000 f 0000 1 5000 2000C Measured Force in X Direction [NI

- - Une of perfect agreement

+ Measured Force

Figure 8-3: Comparison of Forces in X Direction - Orientation 1

83

-1 O000 -5000 O 5000 Measured Force in Y Direction [NI

Line of perfect agreement - Measured Force

Figure B-4: Comparison of Forces in Y Direction - Orientation 1

-1 2000 -8000 -4000 Measured Force in Z Direction [NI

- -- Line of perfect agreement

+ Measured Force

Figure B-5: Comparison of Forces in Z Direction - Orientation 1

83

-- Une of perfect agreement

+ Measured Moment

-10000 -7000 -4000 -1000 2000 5000 8000 Measured Moment about X Axis [N m]


+ Measured Moment

Fipre B-6: Comparison of Moments About X Axis - Orientation 1

O 5000 1 O000 15000 20000 Measured Moment about Y Axis [N ml

Figure B-7: Comparison of Moments about Y h i s - Orientation 1

84

- Q

2 -8000 -4000 O 4000 8000 Measured Moment about Z Axis [N m]


+ Measured Moment

Figure B-8: Cornparison of Moments about Z Axis - Orientation I


86


- Line of perfect agreement

+ Measured Force


< Measured Force in X Direction [NI


+ Measured Force

Figure B-12: Comparison of Forces in X Direction - Orientation 2

87

-1 2000 -8000 -4000 O Measured Force in Z Direction [NI

Une of perfect agreement

+ Measured Force

Figure 8-13: Comparison of Forces in Z Direction - Orientation 2

O 5000 1 O000 15000 20000 Measured Moment about Y Axis [N m]

-- Line of perfect agreement + Measured Moment

Figure B-14: Comparison of Moments about Y Axis - Orientation 2

88


89


- Line of perfect agreement - Measured Force


O 5000 1 O000 1 5000 20000 Measured Force in X Direction [NI

-- Line of perfect agreement

+ Measured Force

Figure 8-16: Comparison of Forces in X Direction - Orientation 3

90

-1 0000 -5000 O 5000 Measured Force in Y Direction [NI


+ Measured Force

Figure 8-17: Comparison of Forces in Y Direction - Orientation 3



+ Measured Force


91



+ Measured Moment

Figure B-19: Comparison of Moments About X Axis - Orientation 3

O 5000 10000 15000 20000 Measured Moment about Y Axis [N m]

--- Line of perfect agreement

+ Measured Moment


92

-8000 -4000 O 4000 8000 Measured Moment about Z Axis [N rn]


+ Measured Moment

Figure B-21: Cornparison of Moments About Z Axis - Orientation 3


25000

20000

1 SOOO

1 O000

5000

O T



+ Measured Force


O 5000 10000 1 5000 20000 Measured Force in X Direction [NI

- - * - Line of perfect agreement

-- Measured Force

Figure B-24: Cornparison of Forces in X Direction - Orientation 4

95

MAPE' - 13.94

. W N E - 0.0 156

R I

Measured Force in Y Direction [NI


+ Measured Force

Figure B-25: Cornparison of Forces in Y Direction - Orientation 4

- 12000 -8000 -4000 O Measured Force in Z Direction [NI

- - Line of perfect agreement

+ Measured Force

Figure B-26: Cornparison of Forces in 2 Direction - Orientation 4

96


Line of perfect agreement - Measured Moment


Q

2 O 5000 10000 1 5000 20000 Measured Moment about Y Axis [N m]

- Une of perfect agreement - Measured Moment

MAPE' - 1.14


97

u a .- - -8000 a 2 -8000 -4000 O 4000 8000

Measured Moment about Z Axis [N ml


+ Measured Moment


Figure 8-30: Direction of Load Application - Orientation 5



+ Measured Force



- .- Line of perfect agreement

+ Measured Force


100



+ Measured Force

Figure 8-33: Comparison of Forces in Y Direction - Orientation 5

- 1 2000 -8000 -4000 O Measured Force in Z Direction [NI


+ Measured Force


101

MAPE' - 9.65



+ Measured Moment

Figure B-35: Comparison of ~Voments About X Axis - Orientation 5

C

E 7

O 4000 8000 12000 16000 20000 Measured Moment about Y Axis [N ml

- - Line of perfect agreement - Measured Moment


102


+ Measured Moment

- f z

Figure B-37: Cornparison of -Moments About Z Axis - Orientation 5

V) .- 8000

a N d

4000 3 O n Cu c.

O t

E -4000 s u a .- - -8000 P

2 -8000 -4000 O 4000 8000 Measured Moment about Z Axis [N mj

t *

- ---

K

-

. L I S


Measured Global Force [NI


+ Measured Force


Measured Force in X Direction [NI

-- Line of perfect agreement - Measured Force

MAPE' - 1.73


105

Applied Force in Z Direction [NI Applied Force in Y Direction [NI

-10000 -7000 -4000 -1000 2000 5000 8000 Measured Moment about X Axis [N ml


+ Measured Moment

Figure B-43: Comparison of Moments About X Xxis - Orientation 6


- - Line of perfect agreement - Measured Moment

Figure B-44: Cornparison of Moments about Y Axis - Orientation 6

107


+ Measured Moment

8000

4000


I

-8000 , ' ! # L I . 1 1

-8000 -4000 O 4000 aooo Measured Moment about Z Axis [N m]

Measured Global Force [NI


+ Measured Force

Figure B-47: Global Force Cornparison - Orientation 7

O 5000 1 O000 15000 20000 Measured Force in X Direction [NI

-- - Line of perfect agreement

+ Measured Force

Figure B a : Cornparison of Forces in X Direction - Orientation 7

110

-1 0000 -5000 O 5000 Measured Force in Y Direction [NI

MAPE' - 1 1 .O0


+ Measured Force



o


+ Measured Force

.- u

O O 2 E -4000 N C .- al O -8000 8

L L U al -12000


111

m

r

MAPE' - 1.68

MMAPE - 0.0009

I P l

-1 0000-7000 -4000 -1 000 2000 5000 8000 Measured Moment about X Axis [N ml

Figure B-51: Cornparison of Moments About X Axis - Orientation 7

W E ' - 1 5000 1.28

W E - 0.0014

O 5000 t O000 15000 20000 Measured Moment about Y Axis [N ml

- --- Line of perfect agreement

+ Measured Moment


112

-8000 -4000 O 4000 8000 Measured Moment about Z Axis [N m]


+ Measured Moment

Figure 8-53: Cornparison of Moments About Z Axis - Orientation 7




+ Measured Force

Figure B-55': Global Force Comparison - Onentauon 8

Measured Force in X Direction [NI


+ Measured Force


i 1s

- 1 0000 -5000 O 5000 Measured Force in Y Direction [NI

Line of perfect agreement + Measured Force


- 7 ._ -

- 1 2000 -8000 -4000 O Measured Force in Z Direction [NI


+ Measured Force


116


tine of perfect agreement

+ Measured Moment


O 5000 10000 1 5000 20000 Measured Moment about Y Axis [N m]


+ Measured Moment


117


+ Measured Moment


-8000 4 0 0 O 4000 8000 Measured Moment about Z Axis EN m]

I 1

---

I i 1 I

8000

4000

1

,

O

-4000

-

-8000 .

Figure B-62: Direction of Laad Application - Orientation 9

25000 T

W E ' - 20000 1.69

15000 MMAPE 0.0019

4 -



+ Measured Force


O 5000 10000 15000 20000 Measured Force in X Direction [NI

-- Line of perfect agreement + Measured Force

MAPE' - 5.55

Figure B-64: Cornparison of Forces in X Direction - Orientation 9

120



+ Measured Force

Figure 8-65 Comparison of Forces in Y Direction - Orientation 9


+ Measured Force



121



+ Measured Moment

B-67: Comparison of Moments About X Axis - Orientation 9


- , - Line of perfect agreement - Measured Moment


122



+ Measured Moment

Figure B-69: Cornparison of Mornenu About Z Axis - Orientation 9




+ Measureâ Force


O 5 m 10000 15000 20000 Measured Force in X Direction [NI

- Une of perfect agreement - Measured Force

Figure B-72: Cornparison of Forces in X Direction - Orientaiion 10

125



+ Measured Force



---- Une of perfect agreement

+ Measured Force


126

-1 0000 -7000 -4000 -1000 2000 5000 8000 Measured Moment about X Axis [N ml


Figure B-75: Comparison of Moments About X A. i s - Orientation 10


+ Line of perfect agreement

+ Measured Moment


127

-8000 -4000 O 4000 8000 Measured Moment about Z Axis EN m]


Figure 8-77: Cornparison of Moments About Z Axis - Orientation 10


MAPE' - 20000 1 .O6

MMAPE - 0.00 12



+ Measured Force

Figure B-79: Global Force Comparison - Orientation I l



+ Measured Force

Figure B-80: Cornparison of Forces in X Direction - Orientation 1 1

130

- 1 0000 -5000 O 5000 Measured Force in Y Direction [NI

Line of perfect agreement - Measured Force

Figure B-81: Cornparison of Forces in Y Direction - Orientation 1 1



+ Measured Force

MAPE' - 6.32

m A P E - 0.007 1

Figure B-82: Cornparison of Forces in Z Direction - Orientation I l

131



+ Measured Moment

Figure B-83: Comparison of Moments About X Axis - Orientation I I

L W E ' - 1 -40

m & w E - 0.00 16

O 5000 1 O000 15000 20000 Measured Moment about Y Axis [N ml


+ Measured Moment

Figure B-84: Comparison of Moments about Y Axis - Orientauon 1 I

132



+ Measured Moment


Figure B-86: Direction of Load Applicauon - Orientation 12

134



+ Measured Force

Figure B-87: Giobal Force Cornparison - Orientation 12

0 5000 1 O000 1 5000 20000 Measured Force in X Direction [NI

-- Line of perfect agreement + Measured Force

Figure B-88: Compmison of Forces in X Direction - Orientation 12

135


C


+ Measured Force

O *- CI

5000 O L E O > C .- CD 2 -5000 O

L L 'c3 a .- -1 0000


-


- - Line of perfect agreement - Measured Force

t 1

MAPE' - 3.88


L 36

MAPE' - 5.23

I

.LIMAPE - 0.0059

r J

a Measured Moment about X Axis [N rn]


+ Measured Moment

Figure B-91: Cornparison of Moments About X hxis - Orientation 12

Q

2 O 5000 1 0000 1 5000 20000 Measured Moment about Y Axis [N m]

-.- Line of perfect agreement

+ Measured Moment

Figure B-92: Cornparison of Moments about Y X x i s - Orientation 12

137


Line of perfect agreement + Measured Moment


Figure 8-94: Direction of Load Application - Orientation 13

139



+ Measured Force

Figure 8-95: Global Force Comparison - Orientation 13

O 5000 10000 15000 20000 Measured Force in X Direction [NI

- Line of perfect agreement + Measured Force


140

a a -1 O000 -5000 O 5000 Measured Force in Y Direction [NI


-c- Measured Force


Measured Force in Z Direction [NI

- Line of perfect agreement - Measured Force


141

-1 O00 2000 5000 8000 Measured Moment about X Axis [N m]


+ Measured Moment


O 5000 1 0000 1 5000 20000 Measured Moment about Y Axis [N m]

.- Line of perfect agreement

+ Measured Moment

Figure 8-100: Comparison of Moments about Y Axis - Orientation 13

142

P

2 -8000 -4000 O 4000 8000 Measured Moment about Z Axis [N rnj


+ Measured Moment

MAPE' - 4-46

MMAPE - 0,0078



144



+ Measured Force



* - Line of perfect agreement

+ Measured Force


145

- 1 O000 -5000 O 5000 Measured Force in Y Direction [NI


+ Measured Force




+ Measured Force

MAPE' - 2.1 1

mMPE - 0.0033

Figure 8-106: Comparison of Forces in Z Direction - Orientation 14

146

-10000 -7000 -4000 -1000 2000 5000 8000 Measuted Moment about X Axis [N m]


+ Measured Moment

Figure B-107: Cornparison of Moments About X Axis - Orientation 14

O 5000 1 O000 1 5000 20000 Measured Moment about Y Axis [N m]


+ Measured Moment


147

a -8000 -4000 O 4000 8000 Measured Moment about 2 Axis [N ml


+ Measured Moment

Figure B-109: Cornparison of .Moments About Z Axis - Orientation 14

Figure B-110: Direction of Load Applicarion - L e Test

149



+ Measured Force

Figure B-111: Global Force Comparison - Ice Test

MAPE' - 1.59



+ Measured Force

Figure B-112: Comparison of Forces in X Direction - Ice Test

150



+ Measured Force

Figure B-113: Cornparison of Forces in Y Direction - Ice Test


- Une of perfect agreement

+ Measured Force

MAPE' - 1-05

.MMAPE 0.00 12

MAPE' - 4.16

W E - 0.0023

Figure B-114: Cornparison of Forces in Z Direction - Ice Test

151



+ Measured Moment

Figure B-115: Cornparison of Moments About X Axis - Ice Test

O 5000 1 O000 1 5000 20000 Measured Moment about Y Axis [N m]


+ Measured Moment

Figure B-116: Cornparison of Moments about Y Axis - Ice Test

152

-8000 -4000 O 4000 8000 Measured Moment about Z Axis [N ml


MAPE' - 7.10

Figure B-117: Cornparison of Moments About Z Axis - Ice Test

IMAGE EVALUATION TEST TARGET (QA-3)

APPLIED - & INIAGE . lnc = 1653 East Main Street - -. - - Rochester. NY 14609 USA -- -- .- - Phone: 71 6M82-0300 -- -- - - F a 716/28&5989

Documents

NOTE - Memorial University of Newfoundlandcollections.mun.ca/PDFs/theses/Tucker_JohnR.pdf · NOTE TO USERS The original manuscript received by UMI contains indistinct and light print.All