Testing ship Ramanathan Krishnan

8/13/2019 Testing ship Ramanathan Krishnan

1/139


2/139


3/139


4/139


5/139

Three Dimensional Wake Survey Behind a Ship Model

UsingHot-Film Anemometers

Krishnan Ramanathan B Eng

A Thesis Submitted to the School Graduate Studies in Partial Fulfillment theRequirements for the Degree Master Engineering

Faculty Engineering and Applied ScienceMemorial University Newfoundland

S John s Newfoundland

OCtober 1995

St John s Newfoundland Canada


6/139

NaliQnalUbraryof CanadaAcquisilionsandibliographic Services ranch395Wellu lllO : Sl ee1

~ ~ ~ 4 l a n o

~ ~ b ~ ~ n a t i o o a l eOi ection des a c g u i s i t i o n ~ eldes services b i b h o g r a p h i q ~ s

3 9 S . . ~ l o o g 1 o nOllawa CWanolK ~

The author has granted anirrevocable nonexclusive licenceallowing the National Library ofCanada to reproduce loandistribute or sell copies ofh s her thesis any means andin any form or format makingthis thosls available to Interestedpersons.

The author retains ownership ofthe copyright in h is/ her thesis.Neither the thesis nor substantialextracts from it may be printed orotherwise reproduced withouthis her permission.

L auteur a accorde une licenceirrevocable et non exclusivepermettant ii la Bibliothequenationale du Canada dereproduire. preter dlstribuer ouvendre des copies de sa thesede quelque manlere et sousquelque forme que ce soit pourmettre des exemplaires de cettethese ii la disposition despersonnes i n b ~ r e s s e e s

L auteur conserve la p r o p r h ~ t e dudroit d auteur qUi protege sathese. Ni la these ni des extraitssubstantiels de celle-ci nedolvent etre imprimes ouButrement reproduits sans sonautorisatlon.

IS N 0-612-13941-7


7/139

Abstract

This thesis presents results of investigative experiments using constant temperatureanemometry to measure three dimensional flows in a lowing tarik.

A three dimensional probe was configured from three single 45 hot film probesmounted with thC ir sensor plane at 120 to each other. Presented are the calibration datafor each the single sensors for variations in velocity yaw and pitch angles well asthe performance of the three probes when mounted in close proximity to each other. Thecalibration dala each sensor was fitted by a two parameter non-linear polynomial invelocity and yaw angle. The probes ll P.re insensitive 10 variations in angle pilch.

A nomin l w k survey w s done with the three dimensional probe on theInstitute Marine Dynamics ship model No. M445 the model a modern fishingvessel form. The results from this wake survey are compared with the results from anexisting pitot tube wake survey done for this model. The spectrum turbulence in thewake is presented.


8/139

Acknowledgements

First and oremost, I thank my supervisor, Dr. Neil Bose, Professor, Faculty ofEngineering and Applied Science for his unceasing encouragement and technicalguidance throughout the course of this research work. I cannot adequately express myappreciation for his support

I thank Dr . Shukai Wu, Research Officer, IM for assista. ,ce with probecalibrations and Mr Carl Harris, Research Officer, for assistance with the model andpitot tube wake survey data Thanks are due to Mr. Andrew Kuczora, Tank Technician,for his help during the experimental work.

I thank the School of Graduate Studies and the Faculty of Engineering andApplied Science for their financiaJ support throughout the course of my study. Inaddition, I thank the lnstitute for Marine Dynamics for their partiallinancial support andfor letting me use their wake rake and calibration rig. Last but not the least, I also thankmy fellow colleagues for their helpful advice and constructive criticism.


9/139

Contents

Abstract cknowledgementsConlentsList of FiguresList of TablesNomenclature

Introduction1 1 Objectives of Study

2 Velocimetry Techniques2 1 Pilot Tubes2 2 LDV2 3 CTA

3 Sensor Calibration

II

IIl......................... V

VII VIII

1 .... ,,3

,,5,5

6

9 ,12

33 23 3

calibration ProcedureCalibration ModelsCalibration Results

12 6

9


10/139

4 Wake Survey us n er4 Experimental Procedure4 2 Wake Survey Results4 3 Comparison of Results4 4 Level and Frequency Turbulence

Discussion and Conclusions ReferencesBibliographyAppendix

IV

47 48

668

7


11/139

List of Figures

3

17 Yaw calibration 538 Yaw calibration 55

234567

891112

1345

6

Calibration setupModified calibration rigHolder for 3-D configuration of probesEffect of varied overheat ratio 55Repeatability of normal calibration 51Repeatability of normal calibration 52Repeatability of normal calibration 53Repeatability of normal calibration 55Effect ofvaricd pitch angle 51Effect of varied pilch angle 52Effect of varied pilch angle 53Effect of varied pilch angle 53Effect of varied pilch angle 55Yaw-pitch calibration 53Yaw calibration 51Yaw calibration 52

4 8

23

25 27

29333

35 37

39 4

43

4445

46


12/139

List of Tables

Normal calibration wilh varied overheat r r tio2. Normal calibration St3. Normal calibration 524. Normal calibration 535. Normal calibration 556. Pitch calibration 517. Pitch calibration 528. Pitch calibration 539. Pitch calibration 5310. Pitch calibration 5511. Yawpitch calibration 5312. Model particulars13. Pilot tube wake survey data radius 5.08 em

4 PilOI tube wake survey dala radius - 6.98 em15. Pilot tube wake survey data radius - 8.89 em16. Pitot tube wake survey data radius 10.67 em

.. 22 24

6............... 28

... 30 32

... 34 36............ 38

40.. 42

49

56 57

58 59

17. Pitot tube wake survey data radius 12.19 em 60

VII


13/139

Nomenclature

CTA Constant temperature anemometerIMD Institute for Marine DynamicsMUN Memorial University of NewfoundlandU. Normal speed in x directionUJ Normal speed in y directionU. Normal speed in z directionCltl Yaw angle for senso r I Yaw angle for sensor 2Cltj Yaw angle for sensor 3CIS Yaw angle for sensor 58 Sensor 52 Sensor 25 Sensor 3 Sensor 5

Order polynomial Overheat ratio Exponent an experimental constant Index of terms of a polynomial k = 0 1 2 3 m

VlII


14/139

V Curve fitled voltageCW Clockwise CW AnticlockwiseU ssumed velocity in test data analysisU 1 Calculated velocity in test data analysis

IX


15/139

CHAPTER ONE

ntrodu tion

Constant temperature anemomelry ~ first employed V. King in the earlypart of this century Perry 1982)]. Though advances have been made over the years,use of the instrument has predominantly been for measurements in air. Application ofconstant temperature anemometry for measurements in water which requires use offilm sensors with thick quartz coatings, has received less allcntion. Most of thepublished experiments in water were done in water channels under controlled conditions.Measurements in oilier fluid medium such as polymer solutions, mercury, blood,glycerine and oil are possible with special probes designed for such specific applications[, mas 1986 ,

The most commonly reported problems encountered with use of hot-film sensorsin water include sensor contamination, air bubbles on the sensor at high overheat ratio,thermal instability during the burn in period and uncontrolled environment varyingtemperatures). calibration of sensors at high overheat ratios results in bubble formationwhich is detrimental to the sensors and affect their response [WU and Bose 1992)].Sufficient time has be allowed for sensors to become stable before any measurements


16/139

can be done [ and Bose 1991. In addition, sensors arc sensitive to changes inwater temperature in the towing tank. Variation of nuid temperature during thecalibration and testing process has a more pronounced effect in water than in air; thereason being that low overheat ratios are used in water to prevent bubble formation andboiling. A change of C in water temperature causes an error of 6 in measuredvelocity [Resch 1969. In such uncontrolled conditions. normal calibrations have to berepeated before any test is carried out. Alternatively. by closely monitoring the watertemperature and applying an appropriate correction to the calibration, temperaturevariations can be accounted for. Various methods for temperature correction have beensuggested by researchers in the past. Teffiperature compensation can be obtained byevaluating the fluid propenies althe mean film temperature Rahman et al. 1987. Theconcept of varying overheat resistance provide results which represent hot-wire outputresponse to variations in fluid temperature [Hollasch and Gebhart l97I J.

Sensor contamination leads to a drift in the anemometer output voltage.However, the contamination does not affect the frequency response of the probes up to200 HZ since for this range the penetration depth of thermal nuctuations is much largerthan the thickness of any dirt layer [Jimenez et aI. 1981 ]. Tests at constant velocityshowed that, for cylindrical sensors drift was a function of dirt particle size to sensorradius [Morrow and Kline l974 J. A conicaHype probe has been used in water nowsbecause its contamination by dirt particles is less than that for other design [Okuno1988 ].


17/139

Successful calioration of the quartz-coated sensors in water ha s led to the use oftwo sensors X-configuration and vee sensors) and triple sensors for studying turbulentnow [Johnson and Eckelmann I984)J. The effects of mechanical interference betweensensors, as in th e case of X-probes, could be severe, depending on th e size of the probes[Fernando et al. 1988)]. A multi-sensor hot wire probe can be u ~ e to measure vorticityand velocity in turbulent flows [Vukoslavcenic et a 1989)].

Directional sensitivity, alignment of the sensor perpendicular to the axial velocitycomponent, is a key issue in the calibration of hot film probes. Errors in calibration,velocity and turbulence measurement using a hot wire/film probe depend on thedirectional sensitivity of the probe [Okuno 1988)]. Measures have been taken byresearchers to overcome the forwardreverse ambiguity in highly turbulent flows. Hotwire probes have been us ed to obtain the thl components of mean velocities and sixcomponents of Reynolds stresses [Lakshminarayana et al. 1982)]. However, such workhas so far been limited mainly to measurements in air.

Ohjectives of Study

The ai m of this work was to investigate the us e of constant temperatureanemometry :., measure three dimensional floWJ; in a lowing tank. The work included thefollowing pans.

Calibrate hot-film sensors using constant temperature anemometers and invesiigate


18/139

the effect of changes in velocity yaw and pitch angles on sensor outputInvestigate the effectiveness use of three 45 hot film sensors to measure 3 Dflow fieldsConduct nominal w k test round bilge fishing vessel modelCompare the results with data previously obtained from a pitot tube wake surveydone on the same model at a sp of 0 54 mlsStudy the turbulence spectrum in the wake of the ship model


19/139

CHAPTER W

Velocimetry Techniques

This chapter presents an overview of the various methods used to determine themagnitude and direction of nuid velocity Emphasis is put on those techniques whichhave found application in hydrodynamic testing facilities such as towing tanks andcavitation tunnels The use of constant temperature anemometry being the primeobjective of this work has been discussed in detail

2 1 Pitot Tubes

In spile of the development of constant temperature anemometry and laser dopplervelocimetry pitot lubes are still used extensively for fluid flow measurements Aconventional pitot lube in conjunction with water tube manometers can be used forpressure measurements which are later translated into velocity The development ofelectronic pressure transducers and the reduction in the size of the pitot tube whichotherwise affect the true flow pattem have helped researchers to reduce some of theerrors in measurements


20/139

Pitot tubes, apart from being relatively cheap, rc mechaniC


21/139

provide information on the turbulent nature of the flow which cannot be obtained withpitot tubes or other mechanical devices.

A laser doppler velocimeter has high spatial and temporal resolution dependingon the quality and quantity of the scattered particles. The density of seeding particleconcentration is a critical factor for LOA operation and effectively determines the datasampling rate. In some instances, the dirt particles inherent in the flow are sufficient,while in other cases artificial s particles are introduced in the flow These dirtparticles, also referred to as scattering centres, may vary in size irom a fraction of a p mto several tens of for efficient operation. Particles which are too small will notproduce enough scattered power to exceed the noise level of the detection system.Particles which are too large, on the other hand will tend to block the light completelyand thus render the instrument inoperative. In effect, sampling rates for laser dopplervelocimeter and the accuracy of measurement depend on the optical arrangements,detector, particle statistics, fluid medium, test speed, anticipated levels of turbulence andfinally on the approach to electronic signal processing [Durrani et. l 1977 .

A major disadvantage of laser doppler velocimeter is that the measurement is notcontinuous, unlike that from constant temperature anemometers. gives measurementsat various instants of time. Also. measurements are not made directly of velocity.Instead, these instruments measure velocities of particles in fluids, i.e. seed or dirtparticles, which m yor may not follow t the fluid velocity. The discrepancy between


22/139

particle velocities in fluids and fluid velocities may significant for large particlesespecially in turbulent flows Durrani et. al. 1977 . Only panicles as small as 0.1 p mcan faithfully follow turbulent flow in liquids such as water [Watrasiewicz et. al. 1976 ].

Laser doppler velocimetry has three fundamental advantages over constanttemperature anemometry. Firstly, there is no flow interference since the probe s purelyoptical. Secondly, the system docs not require calibration as fluid velocity is measureddirectly from the doppler shift in the frequency of coherent light reflected from scallering

centres. Thirdly, the instrument measures the component of velocity in a specifieddirection, the output being a linear function of this velocity component.

In principle, laser doppler velocimeter measurements are non-intrusive in natureexcept fo r the addition of seed particles to increase data acquisition rates to an acceptablelevel. However, implementation 7a laser doppler velocimeter system may become veryintrusive depending on the test facility and arrangement of the optical system, ormeasurements in cavitation tunnels and small open water channels, the optical systcm can extemalto the flow. or towing tank applications, this equipment has either beplaced directly in the flow or the laser beam has to be routed through an optical conduitimmersed in the flow, In the latter case, the optical conduit will absorb SOffie of thelaser light energy making less available to the interrogation beams. A laser dopplervelocimeter system costs as much as 10 times that of a constant temperature anemometersystem and because of this their use has been restricted.


23/139

2.3 Constant Temperature Anemornetry

Hot film atlemometry and hot wire anemometry work on the same principleexcept that the geomelry of the sensors is different The hot film anemometer isbasically a thermal transducer. The principle of operation is as follows: an electriccurrent is passed through a fine filament which s exposed to hc flow As the flow ratevaries the heat transfer from the filament varies. This lum causes a variation in theheat balance of the filament The filament is made from a material which possesses atemperature coefficient of resistance. The variation of resistance is monitoreJ by variouselectronic methods which give signals related to the variations in flow velocity or flowtemperature

There are three modes of operation of a hot film system. irst is the constanttemperature mode where the filament is placed in a feedback circuit which tends 1m int in the film at constant resistance and hence con l nt temperature_ The second isthe constant current mode where the current in the wire is kept constant and variationsin wire resistance caused by the flow are measured by monitOring the voltage drop acrossthe filament. A third type the pulsed wire anemometer. measures velocity bymomentarily heating a film to heat the Ouid around it This spot heated fluid isconvected downstream 1 a second film that acts as a temperature sensor. The time offlight of the hot spot is inversely proportional to the fluid velocity. Of the three modesdescribed above. the constant temperature mode is the most frequently used The reason


24/139

for this is that the feedback mechanism/servo mechanism responds rapidly and the sensortemperature remains virtually constant. The voltage difference across the bridge isrelated the fluid velocily.

Heat transfer/cooling rate can e attributed all of the following.Heat loss from a hot film sensor by convection to the fluid.

2. Heat loss from a hot film sensor by conduction the support needles.3. Heat loss from the hoI film sensor by radiation to cooler surroundings.4. Storage of heat in the hot film sensor.

King s law by far the most well known of the heat transfer laws used in hot filmanemometry is written as

V . , A+BU Iwhere V is the anemomeler output voltage t ken across the Wheatstone bridge in theelectronics package U is the fluid velocity is an experimentally determined constantand A B are experimental constants. Various modelling techniques have beendeveloped over the years which have been discussed in chapter 3.

Hot wire anemometry was originally developed for use in gaseous flow fields butthe delicacy of wire probes prohibited their use in harsh or conductive media. Howeverthe development of film sensors heavily coated with quartz have been used successfullyin fluid applications. The quartz coating protects the sensors from contaminants


25/139

electrical and thermal damage. The dimensions of the sensors are of the order of micronswhich hardly affects the flow pattern

A distinct advantage of hot-film ncmomeuy is that the output from hesensoris continlJOllS with an cxuemely low response time and this reveals information on theturbulence chasacteristics of the flow. It is possible to determine a mean flow componentfrom the sensor output, but a measurement c n lso be obtained of the level andfrequency of turbulence present in the flow. Sensors, once put through the buminperiod c n be calibrated for varied angles of pitch and yaw, is pertinent to c rry outIhe entire calibration process at the same overheat ratio.

For the following rtaSO \S constant temperature anemometry has advantages overother techniques:

1. Hot film probes are less intrusive to the llow than most pitot tubeassemblies.

2 Hi h spatial resolution is accomplished.3 h short response time permits investigations of high frequency flow

fluctuations.4 he high flow sensitivity permits detection and measurement of very low

flow velocities.5 Low cost compared to laser doppler velocimctry.6 Continuous response to the flow characteristics


26/139


27/139

type 55MlO CTA and five 45 hot fUm sensors type 55R12, manufactured by DantecElectronics Inc., were available for this research work. Description the CTAs, hotftlm sensors and the electronic circuitry associated with constant temperature anemometryhave been detailed in the manual published by Dantee Electronics. Each 45 hot-ftlmsensor was held individually at the end of a rod connected to the calibration rig. At theend the rod was a hole drilled at 45, to accommodate the sensor. This was done order to calibrate each sensor in a direction normal to the sensor film. Scales attachedto the rod allowed changes to be made in pitch and yaw angles. Perpendicularity therod was ensured by using a plumb line and by calibrating at slight changes yaw angleon either side lhe zero reading on the protractor. Figures I and 2 show the calibrationarrangement.

Fig 1 : Calibration Setup

3


28/139

~ ~ A n g l

Fig_ 2 : Modified Calibration Rig

hetowing carriage was operated at known speeds and the voltage outpUts fromthe sensors were recorded. Data of carriage spcc d and sensor output were recorded ona micro computer throulh a Keithley 16 bit AID convertor. Two channels wereused, one for carriage velocity and the other for sensor outpUt voltqc.

The sensors were set at an overbcat ratio of 0.4 during the entm calibnltionprocess. The overheat ratio was calculated based on the temperature of the fluid mediumTw) and the temperature of the sensor Ts). Ovettlc:at ratio, can expmsed as

a Ts-Tw) ITs

14

2


29/139

A constant decade resistane.e setting in the anemometer helped maintain a oonstantoverheat ratio Water temperaturewas recorded before commencementofcalibration andthe sensor temperature thereby obtained was used to set the decade resistance Theformula used to set the decade resistance is

R s x Rs Ts-Tw ] 3where is the decade resistance is the total resistance of the sensor including thelead resiSlance a is the temperature coefficient of resistivity and R is the sensorresistance

Firstly sensor 3 was calibrated for pitch and yaw angles During the initialstages of calibration it was found that this sensor was sensitive to pilch; which afterseveral runs disappeared t is possible that the sensor though put through the bum inperiod was not burnt in adequately on the sides These hot film sensors have a verysmall rectangular cross section nickel film and are coated with a 2 lm layer of quartzFollowing an adequate bum in perio the sensor was found to be insensitive to changesin pitch angle and yaw calibrations were done As the sensor output depends on the rateof heat transfer the effective area normal to the flow is reduced as the yaw angle isincreased Once the behaviour of the sensor was known calibrations of sensor 2 andsensor I were done and these showed similar trends to senwr 3 During the bum inprocess sensor 4 was accidently burnt out Three sensors were needed for the wakesurvey Sensor was calibrated as a spare Normal pitch and yaw calibrations for allfour sensors were done in November and December of 1993

IS


30/139

3.2 libr tion Models

King s law is usually expressed as .. A + BU 4)

where V is the anemometer output voltage, U is the fluid velocity flowing past the sensor,and A and B are experimental constants. King, based on his experiments, found themagnitude of the exponent n to be 0.5 [King 1914)).

A general fonn of modified King s law is written as = Akl1 5)

where V is the voltage, is the velocity of the fluid, m is the order of the polynomial,Ak and exponent n are experimental constants [Wu and Bose 1992)]. This calibrationmodel was used to analyze pitch and nonnal calibration data for all sensors.

In the case of a D flow. the sensor is cooled by the velocity components in alldirections, but it s sensitivity to each component differs. The heat loss in this casedepends on the effective cooling velocity, which should replace U in the aboveequation. For such cases, the effective cooling velocity is expressed as

6)where and kb are the yaw and pitch factors respectively [Champagne et al. 1967)).This method assumes t he yaw factor to be constant over the entire velocity range.

16


31/139

However, it has been found that the yaw factor showed strong velocity dependence at lowvelocities [Wu and :Jose 1994 J,

A two-variable non-linear polynomial o fonnV2 kjt a\j 7

was proposed as a better calibration model [Wu and Bose 1993 J. In the above equation,a is the yaw angle, m is the order o the polynomial and k is subscript o polynomial omth order. Yaw calibration data for each sensor was fitted using a least squares fitprogram based on the above equation. Coefficients obtained, for a polynomial o order3 m=3 , were used to analyze test data. Polynomial o order 2 gave an error which waslarger than that obtained using a polynomial o 3rd order.

Water temperature changed significantly during the calibration process. Initially,the anemometer was set such that the sensors worked at an overheat ratio o 0.4. Thedecade resistance setting on the r was left untouched in spite o temperaturefluctuations. Temperaturewas recorded during every calibration process. Effectively, thesensors were allowed to operate at slightly varying overheat ratios. In order to accuratelydetennine the characteristics o the sensor for varied overheat ratio, a series o runs wasmade using sensor operated at overheat ratios ranging from 0.2 to 0.4. Calibrationvoltage, in addition to curve-fitted voltage and percentage error t each point, is given in

t ~ l I. A plot o the nonnal calibrations at various overheat ratios shows that the voltagelevel from the sensor at a given speed increased with an increase in overheat ratios and

7


32/139

vice-versa fig. 4).

For simultaneous calibration of the three sensors, oriented al 120 apart, a holderwas made up which could also be used for the wake survey fig. 3). was decided not10 disturb the orientation of the sensors after calibration; once mounted in the holder, thesame sensor and holder assembly was used for both Ihe calibration and tests. Sensors 1 3and 5 were chosen for wake lests due to their steady perfonnance during the calibrationprocess. Adequate yaw calibration data was not available for sensor 2

120Fig. 3 : Holder for configuration of probes

18


33/139

3.3 alibration esults

Normal calibrations were repeated on different days to check the sensorresponses were stable at a given overheat ratio. Modelling curve-fitting o nonnalcalibration data was based on the modified fonn o King s law eqn. 5 A third orderpolynomial was chosen over a second order rolynomial for the reason that a beller curvefit with minimum curve fitting crror was achicvcd.

Figures 5,6,7 and 8 arc indicative o repealability levels ofnonnal calibration forsensors 1,2,3 and respectively. Tables 2,3,4 and 5 contain corresponding calibrationdata points, curvefitted voltage Vc and the percentage error Error) in curve fit atevery point. The response o each sensor was almost the same for calibrations done eithertwice on the same day or over a span o 2 to 3 days.

Figures 9,10,12 and 13 are indicative o the insensitivity o the four sensors tovariation in pitch angles. Tables 6,7,9 and 10 contain their corresponding calibration dataand curve fitting error at each point. Also, the nonnal calibration for each sensor hadgood correspondence with the pitch calibration. Pitch calibration data were modelled

based on a third order polynomial o the modified King s law eqn. 5 . Curve fittingeTTors were negligible and hence the modified fonn o King s law was adequate to checkfor repeatability o nonnal calibration. Though there were slight variations in sensor

19


34/139

response during pitch calibration. this was n as significant as in yaw cilibralions.Hence. was assumed that the sensors were insensitive to changes in pitch angle. Figure indicates the levels o sensitivity o sensor 3 to changes in angles o pitch observedin the earlier stages o th burn in period; these reduced ..ner funher bum in Table 8corresponds to this initial pitch calibration plot.

Yawc:alibration was modelled based on the two-variable non-linear polynomial oorder 3 eqn. 7). for the four sensors. The method o yaw factor modelling eqn. 6assumes that the yaw factor is constant for any given angle, over a range o velocity. Asdiscussed earlier, this is not necessarily true at lower speeds. Figures 15.16,17 and 18show the yaw characteristics o the sensors. Maximum heat transfer occurs when thesensor plane is at 900 to the flow direction. i.e. the nonnal projected area to the flow ismaximum. Also. heat transfer is minimum when the sensor plane is parallel to thedirection o now, i.e. the normal projected area to the flow is minimum. The yawcalibration data. coefficients o the polynomials and the percentage error at each point canbe found in the appendix. along with the computer program used for analysis. tcan seen that the curve-fitting error based on this model. at most points in the region ointerest 0.54 mls), was under 0 7 .

20


35/139

10a_0.4a_0.35a.0.3

a_0.25

a_0.2. >

20L--O: .2::----:0..., O: :----:O. : - - - 1 ~ - - - - : I . : . 2 - - - - 1 . 4 : : - - - - - : 1 . :----,-,. :-.---:V.loclly mI.)

Fig. 4 : Effect of Varied Overheat Ratio (55)

21


36/139

a 020 025Velocltv Voltage V Error O O 2 5261 2 5331 0 27440 1216 4 3993 3705 0 65180 2471 4 8147 4 B 84 0 13040 4996 5 2363 5 3070 1 34940 7511 5 6072 5 6157 0 15131 0034 5 9135 5 8370 1 29391 2554 6 0040 6 0049 0 01551 5064 6 1099 6 1364 0 4338

Valoclty Vo I aga V Error O O 2 9036 2 9059 0 08010 1216 5 1570 5 1511 1146 2479 5 6849 5 6637 0 37290 4988 6 1465 1 30370 7506 6 6277 6 5676 0 90621 0030 344 6 8041 0 .1 2546 9268 6 9769 0 72271 B e2 7 1206 7 1070 0 1910

035Veloctty Voltage V Error O O 3 5427 3 5487 0 18790 1224 6 3709 6 3442 0 41790 2478 9622 6 9589 0 04660 4994 7 5588 7 6342 0 99740 7514 8 0679 8 0367 0 38761 0027 8 3437 8 3107 0 39851 2546 8 5166 8 5089 0 09071 5056 6327 8 6548 0 2561

030=Veloc Volta V Error O O 3 1735 3 1767 0 09550 1213 5 6448 5 6093 0 62190 2471 6 2147 6 2854 1 13650 4994 7 0239 0 81870 7510 7 3504 7 3373 0 17851 0029 7 5103 7 5876 1 03121 2542 7 8413 7 7837 0 73341 5060 7 9447 7 9500 0 1737

a QVelo Vo V Error0 0000 3 8870 3 897 0 06890 1212 6 9107 6 8920 0 27090 2481 7 6363 7 5525 0 21470 4998 8 8 2697 0 15670 7612 8 7035 8 7 3 ~1 0046 9 0154 8 0002 O td851 2561 9 2221 9 2187 0 03691 5071 9 3745 9 3837 0 0979

Table I : Normal Calibration wilh Varied Overheat Ratio SS

22


37/139

1 0 r ~ ~ _ . . ~ _

~ . _ . . ~ ~ . . w

~ ~ . .t

i 65: run onl

ru tw

3 0 ~ - - - - : O : . 5 - - - - ~ - - - I : . 5 - - - - : - - - - - : 2 . 5vlllocity Im l

Fig. Repeatability of Nonna Calibration 51)

3


38/139

nVela V lta Vc rror

O O 3 6193 3 6192 16 122 6 4724 6 4762 583 2471 7 9 6 7 825 1141 4993 7 7 61 7 7 92 4 4 75 2 8 618 8 885 8341 1 44 8 3991 8 3963 3321 4556 8 6293 8 6158 15661 8 72 8 76 1 8 7754 17442 26 7 8 9298 8 9254 491

Run VelocItY VClttaae Vc rror

O O 3 6475 3 6478 69 1217 6 4822 6 4817 69 2468 7 544 7 544 5 4991 7 6886 7 8644 549 75 8 8 542 8 631 111 2545 8 5312 8 5 6 5881 7578 8 8 9 8 6 2 762 2813 8 9724 8 9737 148

Table 2 : Nonnal Calibration SI

4


39/139

1 O ~ ~ ~

run onerun two

5 VelocIty m/s

Fig. 6 : Repeatability of Nonna Calibration S2)

2S

1


40/139

A 1vel itY VoltWie a rror

O O 3 5292 3 53 9 0 04730 1218 6 6103 6 5958 0 2188 2473 7 2211 7 2363 21 50 4998 7 9349 7 9459 0 13930 7506 8 3992 8 3746 0 17581 45 8 6864 8 68 731 2554 B ~ ~ q 8 9 68 892

R nVela Votta a rror

O O 3 5973 3 5997 6490 1214 6 8239 6 6 54 28210 2473 7 2 78 7 2243 23 40 4999 7 9 49 7 9138 0 1104 7512 8 3154 8 n 26881 34 B 7065 8 6435 0 72281 2547 8 8458 8.B7B1 3851

Table 3 : Nonna caubl 1.tion (52)


41/139

. r ~ _ _ r _ ~ _ . .

6

run on.run rwo

2. -----:. ': - - - - - ~ - - - - : : . 6 - - - - ~ - - - - - : 2 .Vlloclty mI.

Fig. 7 : Repeatability of Normal Calibration 53

27


42/139

R VelOcltV o Vo Error0 0000 3 3487 3 3588 29830 12 3 5 6040 5 5504 9559 2478 6 1 85 6 1444 5882 4992 6 7717 6 8242 0 n570 7511 7 2852 7 2511 46761 33 7 5659 7 5552 14 71 2546 7 7999 7 78 3 0 21981 5 66 7 9481 7 9599 0 13411 7581 8 0850 8 956 0 13082 0090 8 2043 8 2017 3 7

R nVelocltv Voltaaa Vo Error0 0000 3 3558 3 3835 0 2341 1212 5 8406 5 5994 92580 2474 6 1217 6 1759 8851 49 8 6 8338 6 8479 0 2061 7524 7 2853 7 2659 2671 0020 7 5820 7 5585 0 04841 2536 7 7894 7 7794 12871 5063 7 9424 7 9496 0 0899t 75 9 8 0n5 8 794 0 02322 0096 8 1808 8 1822 172

Table 4 : Normal Calibration 53)

28


43/139

1 0 r ~ ~ _ _

run oru tworu t r

3 0 ~ - - - - - - - : O : . 5 - - - - - - - C - - - - - - - . . J . eloc tymIa

Fig. : Repeatability of Normal Calibration 55

9


44/139

RunVelocltv olto o Error O O 3 6314 3 6339 685 1215 6 72 2 6 6995 3123 2478 7 3662 7 3648 2511 4995 6 15 4 8 1731 O zn4 75 9 6 6999 6 662 19251 28 9 872 9 63 26651 2546 9 3486 9 3674 2 13

RunVeloCItY Voltaae Error

O O 3 5496 3 5514 788 1222 6.7131 6 89 9 3317 2476 7 356 7 3745 2518 4988 6 1376 8 1817 95

75 4 8 6798 aan 13831 29 9 763 9 44S 37211 255 9 32 1 9 3432 2483

Runol O O

1223 2473 4998 75151 181 255

Votta e Vc3 5887 3 57 66 7266 6 7 997 3691 7 36188 1498 8 1698

8 68129 6499 3782

Error 538 2446 1725 2456 62

3729 23

Table S : Normal Calibration 55)


45/139

1 O r ~ _ r _ ~ ~ _ _ _

normal1 dllg30dllg.

0 1 6V,locily m/.Fig. 9 : Effect of Varied Pitch Angle 51

31

2


46/139

ormalVeloc tY Vottage rror O O 3 5939 3 5961 6 1 1219 6 5749 6 5424 0 4947 2475 7 887 7 14 1 7253 5 1 7 6063 7 7971 1168 7516 6 2096 8 1923 0 21361 2551 6 6819 8 6773 0 05331 7595 8 9479 8 9668 21112262 9 1592 9 1511 0 0685

d eeel Volta e ~ E r r O f

O O 3 6295 3 6284 4280 1217 6 5282 6 5434 0 2461 248 7 2349 7 2069 0 3868 7 6602 7 8733 0 16028 23 1 6 2409 12521 2549 6 6864 8 6869 0 19791 7569 6 9252 6 9346 11 52 2622 9 1462 9 1444 229

3elo ltv Voltage rror

O O 3 6399 3 6445 0 12600 1208 6 5306 6 5163 0 1668 24n 7 1232 7 1 58 24380 5004 7 7222 7 7478 331

7518 6 0960 8 1396 0 52541 2539 6 6871 8 6261 68 11 7575 6 9262 8 9322 6672 2612 9 1257 9 1368 0.1221

Table 6 : Pitch Calibration 51

32


47/139

1 O r r ~ r _ . . .

deg,20 deg30 degdell3 0 ~ = O : . 5 : c 7 t . 5

Vtloelly m/s

Fig_ t ffect of Varied Pitch Angle 82

33


48/139

10 rV V Vo Error

O O 3 5947 3 5963 0 04280 222 6 5970 6 6806 24 80 2477 7 1713 7 189 0 24700 6000 7 8683 7 8756 0 11720 75 4 6 3134 6 3044 0 0761 0037 6 B33B 6 6 80 0 18241 2550 6 8495 8 8619 0 1398

SOd rl SV V Vo EnOl

O O 3 5720 3 5723 0 00850 1215 6 5628 6 5545 0 12360 2484 7 0791 7 874 0 11750 4996 7 7335 7 7377 0 05420 7510 6 679 8 905 0 27601 31 6 6066 8 5506 0 64951 2552 6 B254 6 8S43 0 3280

20 oarV Volt Vo ErrOf

O O 3 6043 3 6052 2330 1218 6 5899 6 5827 0 09470 246 7 1681 7 1672 10 4991 7 6086 7 8278 0 27210 7516 6 2573 6 2536 0 0450 0039 6 599B 8 5707 0 3382 2549 6 6053 8 6232 0 202lI

4 daareesV V Vo Error O O 3 6081 3 6095 0 0377 1222 6 6124 6 9 8 O 4n 2481 7 1513 7 1628 0 15850 4994 7 78 5 7 6125 0 398ll0 7518 6 2530 B 2362 0 20451 0038 8 5924 8 5553 0 4325 2553 8 7883 6 6 33 3 73

Table 7 : Pitch Calibration (S2)

34


49/139

lOr . r ..Initial Observation

::: : :: :

10deg. cw20 dig. ew30 dig. CW40 de;. ew)10 dIg. aew20 dlg.faew3 0 : - - - - - - ; O c : . - - - - : - - - - - - : c : . : - - - - ~ - - - - : 2

Velocity 11 1/1

Fig. 11 : Effect of Varied PilCh Angle 53)

Initial observation during the bum period.35


50/139

10_ /C Mv v V, .0000 3 3165 3.3 0.00111iJ 1219 5.9961 0.13380 2478 sn U3D5 0 4 070.4993 7.2570 7.2093 0.57910 751 7 7 e922 0.0815I 7._ 7 0.421-41.2543 ~ 00061 usn .3022 0 18261.0092 8.4163 8.4357 0.2305

~ 7 8 5210 as03 0._30 dear , ICWI

V.1oollY VoIItQe V, Em>

odeg.S deg.10 dog. 5 5 elocl tymI. 5

Fig. 12 : Effect Varit d it h Angle (53)

fter the sensor h d stabilized.


52/139

orVel ~ Vc Erra

O O 3 4336 3 4349 0 03730.1211 6 5 5 9956 6150 2478 6 697 6.5124 4 60.7515 7 6 78 7 6007 9331 32 7,8514 7 8760 0.31331 5 62 8 27 2 8 2435 32282 88 8 4614 8 4718 1234

de reesVeloCItV Voltage Vc Error

O O 3 4336 3 4413 2218 1113 6 74 5 9633 7334 246 6.5713 6.6146 0 6590.7512 7 8223 7 6488 0 34711,0026 7 9547 7 92 8 0 42621 5 61 8 2854 8 2754 1 892 94 8 4761 8 487 0 1305

lOde reesVelocItY Voltage Vc Error O O 3 4336 3 4355 545 1213 5 9654 5 9678 415 2472 6 5792 8 5841 23 2 7518 7 6 54 7 63 8 33171 32 7 9 69 7 91 1 4 11 5 65 8 3 44 8 2727 38152 87 8 4882 8 4828 1723

Table 9 : Pitch Calibration 53)

fterthe sensor h stabilized

8


53/139

1 0 r ~ ~ _ _ _ _

normal1 d g3 d g acwi3 d g ew3 0 - - - - - - - - : o . ; s - - - - - - - - ; ~ - - - - - - 7

V.locHy mI.

Fig 13: Effect of Varied Pitch ngle5S

39


54/139

v0 0000 12150 24850 50050 75181 00401.2555

1 de rees AVolta e Vc3 6173 3 61766.8015 6.80137.4584 7 45148.2038 8.21788 7213 8 73089.1617 9.12879.4393 9.4560

Error0 00680 00280.09390.17000 10850 3604 1768

30 degrees {ACWlVelocItY Voltaae ErrOf0 0000 3 59 4 3 5909 0 01340 1228 6 8575 8 8550 0 03570 2485 7 5140 7 5090 0 06680 5003 8 2517 8 2735 0 2


55/139

. r r r ~ _ r _ ~ _ _ _ _ _ _yaw anlll lo dIg.

10 dig.20 d ...0 d.g.g . L 3 ~ O . .... O ~ . ~ 6 0 .... 0 . ~ . . . J

Velocity mIl

Fig. 14: Yaw-Pitch Calibration 53)

Fixed yaw angle of 10 and varied pitch angle.

1


56/139

de reesVeloci ~0 4993 7 01630.7517 7.3044

de reesVeloc Voltage 4995 6 99810 7509 7 3001

20dVelo0 49860 7512

reasVolt e7 03817 3043

Table : Yaw Pitch Calibration 53)

Fixed yaw angle of l00 and varied pitch angle

4


57/139

1 0 r ~ ~ ~ ~ r _ _

polyfitr w t

3 0 L - - - - : O : . S ~ - - - - - - - - : , : . S - - - - : - - - - - 7 2 . Selo ity mf.

Fig. is: Yaw Calibration SIr

Yaw calibration was also done al8O 7rr W. 30 20 and 10 .

43


58/139

1 0 . . ~

pofylllt w ll

O ; - - - - - - - - - : .; ; . , ; - - - - - - - -7 - - - - - - - - , , JVelocity mIl

Fig. 16 : Yaw Calibration 52)

Yaw calibration was also done at 80 and f ff


59/139

1 0 r ~ ~ ~ ~ _ _ _

i 76>

0 5 1 5V.loal) mI.)Fig. 17 : Yaw Calibration S3)

r

polyfltraw dala

2 5

Yaw calibration was also done at 80 , 70 , 6lr 300 2 and 10

45


60/139

1 O r ~ ~ . . .

polylllt w t

0.5V l o c ~ y m / l

Fig. 18 : Yaw Calibration 55r

Yaw calibration was also done at 80 60 , 20 and 10 .

5


61/139

H FfER FOUR

Wake Survey Using TA

Wake survey in general is conducted on ship models to clearly quantify the flowpattern in the propeller plane. The nature of the flow in the wake governs the design ofan efficient propulsion and steering system Also. the pattern of flow behind the hullwill be influenced the basic hull form and will determine the selection and placementof propulsors. The wake behind a ship model is complex and three-dimensional. Atheoretical approach its determination is nOI generally pursued. However with theadvent of supercomputers a computational approach could become feasible in the nearfuture. An analytical approach requires the clear interpretation of the flow in addition solving the Navier Stokes equations in three dimensions at high Reynolds numbers.The most practical method for wake determination which has been followed to date isto cany out experiments on a model and extrapolate the results 10 that of a full scaleship. Tests on a full scale ship are expensive and not generally feasible.

Model nominal wake data for ships are presented either as a contour plot or ascurves of velocity versus angular position and radii in the propeller plane.

47


62/139

4.1 Experimental Procedure

A model of a round bi1ee fishing vessel M44 ) WLl used for tltis study. Thismodel is similar recent designs of lon Iiners in Newfoundland. Figure 19 is a bodyplan of the model and table 12 gtvt:S the model particulars. Pitot tube wake surveyresults at speeds of 54mlJ and 8mls t available for this model from previoustests done at the Institute for Marine Dynamics. These speeds were equivalent to fullscale ship speeds of 3 knots and 9 knots respectively. The wake survey using hot filmanemometry in this study was done at a speed O S m s

Fig. 19 : Body Plan of Round Bille Model


63/139

ModeVShipScaleLength BPBreadth MaxDraft FPDraft APTrimDisplacement

:2 12Sm0 838m0 396 m0 431mO Q3Sm0 438Mt

Table 12 : Model Paroculats

this work modifications were made to the wake r that was used for thepilot tube wake survey in order to accomodate the J D hot film sensor Figure 3 showsthe configuration of the three sensors in a holder which was held in the wake rake herake unit and the associated sensor holder for wake survey is shown in figure 20

l UeSuppon Ship Hull

Fia 20 : Wake Survey Setup

49


64/139

The rake unit, which could be rotated about the propeller shaft axis, accommodatedthe sensors at five different radii; each radius was on the propeller plane. The wake rakewas rotated by a digitally contrclled stepping motor fixed on the inboard end the shaft,The top portion of the skeg near the region the hull represented 0 and the boltomportion of the skeg in the region the keel represented 180, The entire wake rake unitwas rotated in the clockwise direction looking from the stem ponion which covered halfthe propeller plane, Each sensor, in the 3-D setup, connecled by standard cables tothree anemometers which helped maintain a constant overheat ratio 0.4 throughout thetest. The response of the three sensors in the wake of the model was sampled at 300 Hzfor a sampling period 20 seconds, Data of carriage speed and sensor output wererecorded on a computer through a Keithley 16 bit AID convertor. Tests were doneover a period five days one radius per day , Calibration sensors were done everymorning and the actual tests were done in the afternoon.

4.2 Wake Survey Results

Test voltages from the three sensors contained velocities and directions, As aresult. an iterative approach was adopted to determine the velocity and direction of flowwith respect to each sensor . The coefficients of the twovariable non-linear polynomialone set for each sensor and test voltages were put through an iterative program to solvefor axial, tangential and radial components of velocities. The routine the program was


65/139


66/139

to be O.SC. In Ihe region of 50 150 the sensors were exposed to the free streamflow and as a result should measure the free stream velocit) . The reduction and variationin the axial component in this range fig. 24) shows that the temperature correction basedon calibrations done on the respective days the test needs to be improved. Based onthe calibration done every da) before the test, the test data w as approximated Iinearl)with 0 55 mJs as the reference, a multiplication factor was established). Figures 25 and

26 represent the tangential and radial component velocities. The pattern is comparable10 that the pitot tube wake survey results. The amount reduction in the axialcomponent ofvelocit) tends to augment the tangential and radial velocity components dueto the geometry used in the iteration process.

Both wake surveys confinn a relatively unifonn distribution water flow in thepropeller plane for ...ngles ranging from 50 to ISO , except for the region skeg. Thedrop in the axial velocity component in the region skeg is due to the flow retardationgenerated by the presence the skeg.

S2


67/139


68/139

0 5

0 5

radii cm5 088 988 8910 6712 19

50 100 150 200 250Angle Theta degrees 300 350 400

Fig : Pilot Tube Wake Survey tangential velocity

54


69/139

0.4radii cm

0.3 5.08 6.98

8.8910.670.2 12.19

1 0. x 0::> . t0.1 x x

it~1 0.3 o tS 80.4 50 100 150 200 250 300 350 400Angle Theta degrees

Fig. 23 : Pitot Tube Wake Survey radial velocity

55


70/139

dea. X t r3 0.81051 0.02297 0 066896 0.78138 0.06613 0.071369 0.73743 0.08159 0.0503415 0 72309 0 16840 0 00966

20 0.78707 0.23801 -0.06155 s 1 28862 I 0 11024 fs I 0 18527 .26940 I 0 096790.00890 ;it 1.01502 -0.21902 0 09B02135 0.99415 0.14480 14415155 0 95690 0.05742 0 16632160 0.95119 0.03785 17 12165 0.94132 0 00290 173597 O 8 7 ~ J 7 0.09595 0.16127183 0.53912 0.25827 0.07403186 0.82341 0.12403 0.17563189 0.91676 0.07252 0.19446195 0 94892 O O 2 0.19244200 0.95632 0.03522 0.18783205 0.95460 0.06299 0.18240225 1 158 16831 0.15683245 1 00897 0.23906 0.10954270 0 99048 0 29843 0.01954295 0.97812 0.31441 a OlSlO315 0.96706 0.29843 -0 15379335 0.96982 29311 -0.18682340 0 90235 0.29109 -0.13063345 0.88442 0.28701 J.114905 0 79888 0.24150 -0.05991360 0.83349 -0.02708 0.07032

Table 13 : Pitot Tube Wake Survey Data radius 5.08 em

56


71/139

Xl l Ur/U3 0 76378


72/139

de Ux U UtJU Ur U3 0.79088 0 10025 0 241626 0 88392 0 10695 0.278239 0 91180 0 11430 0 2745915 0.92633 0 15530 0 25086

20 0.94332 0.18899 0.2464525 0 94707 0 21386 0 2262545 0 97085 0 27191 0 1512965 0 97702 Q 28546 0 0586690 0.98664 Q 26463 0 04740

11 5 0.99390 0 21135 0 1276413 5 0 99846 0 14806 0 169405 5 0 98812 0 06561 0 1866360 0 98215 Q 04476 0.19083165 0.97652 0.02033 0 18729 7 0 97360 0 01218 0 18820183 0.85926 0 15564 0.20825186 0 97060 0 07602 0 2002118 9 0 97819 0 04282 0 2025819 5 0.98133 0.00985 0 209322 0.98499 0 01739 0 213092 5 0 98248 0 04490 0 2095122 5 1 00301 0 13594 0 1S77024 5 0 99776 0 19948 0 1425527 0 98095 0 26123 0 0595929 5 0 96404 0 28487 ..Q.04938315 0 95240 0 27051 0 13772335 0 96232 0 24129 0 2183634 0 94545 19nS 0 24545345 0.94038 0.10028 0 25213351 0 92993 0 15621 0 2708436 0 79561 0 10650 0 25190

Table : pitot Tube Wake Survey Data radius 8 89 em

58


73/139

eg UxIU Ut U Ur U3 0 68804 0 02877 0 253596 0 78367 0 05356 0 329489 0 75911 0 06275 0 3144515 0 77536 Q 10840 0 28693

0 81179 0 13861 0 267705 0 83668 0 16483 0 2485045 0 97284 0 26666 0 1654265 0 98914 0 27832 0 072789 0 98993 0 25575 0 03199115 0 99123 0 20207 0 11133

135 0 98559 0 14919 0 15095155 0 99289 0 10514 0 17645160 0 98987 -0.06624 0 16676165 0 98239 0 06036 0 18781171 0 97244 0 02118 0 19297183 0 96132 0 14252 0 12487186 0 94834 0 11485 0 17400189 0 94086 0 08727 0 20662195 0 95548 0 02498 0 21571200 0 97684 0 01140 0 21474205 0 97808 0 03947 0 20954225 0 97779 0 11804 0 18264245 0 98637 0 18415 0 14889270 0 98099 0 24534 0 07346295 0 98461 0 28172 Q 03576315 0 97794 0 27970 Q 13434335 0 85417 0 19483 Q 22867340 0 81571 0 16917 0 24705345 0 77477 0 15190 Q 2633735 0 75534 0 11685 Q 28655360 0 69842 0 08748 Q 24872

Table : Pitot Tube Wa ke Survey ataradius 10.67 em

59


74/139

de X rtu 0.01750 0.09561 0 00858 0.22262 0.00143 0.23957 0.07558 0.23529 0.11973 0.2597525 0.68395 0.14203 0.2489945 0.92452 0.27783 0.1814165 0.98795 0.28141 0 0700990 0.99573 0.24942 0.04753

115 1.01054 0.19242 0.11730135 1.01823 0.13867 0.15156155 1.00901 0.07294 0.17136160 1.00520 0.05781 0.17947165 1.00164 0.03285 0.191387 0.99131 0.02700 0.22535183 0.69852 0.02575 0.11293186 0.61364 0.17661 0.05423189 0.98377 0.08569 0.28038195 1.00671 0.00160 0.23110200 1.01138 0.02664 0.20735205 1.00621 4 ~ 9 9 0.19438225 1.02249 0.12328 0.16454245 1.01362 0.17877 0.12903270 0.99409 0.24168 0.05663295 0.97815 0.27645 0.05829315 0.90512 0.27739 0.16985335 n343 0.18318 0.23936340 0.67237 0.14001 0.25067345 0.63198 0.12169 0.2572635 0.55130 0 09599 ..Q.24065360 0_19839 0.01658 ..Q.10850

Table 17 : Pilot Tube Wake Survey Data radius 12.19 em

6


75/139

2 0

1 5

1 0

0 5

0 0

5 0116 988 891 67 2 9

0 5 - - - - - - - - - - - - - - - - - - - - - - - - - - 150 100 150 200 250 300 350 400

Angle Theta degreesFig. 24 : erA ke Survey axial velocity


76/139

0 8 5 81; 988 891 6712 19

0 4

00 0 0 V : 0

0 4

0 8 . . . L ~ ~ ~ ~ . . ~ ~ i50 100 150 250 300 5 4

Angle Theta degrees

Fig. 25 CfA Wake Survey tangential velocity

6


77/139

0.8 ..... =5 86 988 8 9

1 67 2 9

0.4

i> 0.0

-0.4

-0.8 . . . L ~ r _ _ _ _ ~ . _ r _ _ _ r i50 100 150 200 250 5 4

AngleTheta degreesFig. : CTA Wake Survey radial velocity

63


78/139

4.3 omparison esults

Figures 27, 28 and 29 represent a comparison between wake SUTVey data obtainedwing constant temperature anemometers and pitottubes. The wake pattern at a radius of12.19 emalong the propeller plane, established by these twomethods, has been compared.

Wake survey done using both r and pitot tubes indicate a relatively unifonnflow in the region of 6 to 17 (fig, 27). The lowcr values of the axial component ofvelocity in this region, using CTA, can be attributed to problems associated withtemperature concetion. In this range, the sensors were exposed to the free stream nowand hence should have predicted the test speed of 0,55 mi In the region of the bottomponion of the skeg (around 180 ), there is a drop in the axial velocity component, asobserved by both methods. Turbulence in this region resulted in an increase in thetangential velocity component (fig. 28). In the range of f to 3 (skeg ponion near thehull), the crA method indicates more or less a smooth now as compared to pitot tubewake SUTVey; which is due to an ambiguity in the waler line length and the modeldisplacement indicated in the pitot tube wake survey report [Nordco Ltd. (1991)].Another indication of CTA measuring more axial and unifonn flow in that region can beseen in figure 28; where the tangential velocity components arc smaller in magnitude ascompared to pitot tube wake survey. A similar trend is also obseTVed in figure 29; whichrepresents radial velocity component in the range of to 180 ,

64


79/139

2 .0

5

1 0 .

0 5

0 0 CfAdalaPitollubedata

40 80 2 160 240 280 3 36 400ngle heta degrees

Fig. 27: Comparison xi l Velocity Component radius 12 19 em

0 5 f . . . r 140

65


80/139

0 8

0 4

0 0

0 4

:

CTAdatailollubcd t

40 80 2 6 200 240 28 320 360 4000 8 l . . c r j

40Angle Theta degrees

Fig 28: Comparison ofTangential Velocity Component radius 12.19 em

66


81/139

4 80 120 16 200 240 280 32 380 400

O B

4

0 4

CTAdataPilOllllbedata

O B t r r .40

Angle Theta degreesFig 29 : Comparison of Radial Velocity Component radius 12 19 em


82/139

4.4 Level and Frequency of Turbulence

The intensity of turbulence can represented by ils energy spectrum over theturbulence frequency range. To analyze the frequency components of turbulence, a fastFourier transform FFf was performed on the time-domain recording of convertedvelocities values obtained after iteration . During the iteration process, some of the testdata in the close vicinity of the skeg showed signs of flow reversal. This is indicative

of the turbulent nature of now in the propeller plane. For practical purposes, turbulencecan be regarded as having frequencies below 45 SOHz. The sampling frequency isnormally ten times that of the maximum turbulence frequency of interest. By consideringthe turbulence to mostly under 30Hz from the present results, a sampling rate of300Hz was chosen.

Typical energy spectra of the recording at various radii and angular positions areshown in figures 3 31 and 32. Figure 3 indicates the presence of maximumturbulence in the region of skeg near the keel , at a dominant frequency ofapproximately 13Hz. Figure 31 represents the energy content at radii of 5.08 cm and10.67 cm along the radial line of 90 . It can seen that the dominant frequency occursat a very low frequency of approximately 2Hz and the lower energy content clearlyindicates lower levels of turbulence far away from the skeg. Figure 32 representsturbulence in the region of the skeg near the hull which appears to have decreasing

68


83/139

1 I ro w M Frequency Hz

: : : : : : :I 10 W 5 M 35 4Frequency Hz

Fig 3 : Spectral Analysis 16 radial line

69


84/139

~ ~ = JW W requency Hz=-: I

W W requen y Hz

Fig. 31 Spectral Analysis 9lr radial1ine


85/139

~ : : . : . : I 5 5 2 25 3 5 4 45 5 55 Frequency HzI c W_: I

2 3 5 4 5 55requen y Hz

Fig 32 : Spectral nalysis S radial line


86/139

levels of turbulence t higher radii. Due to varying levels of turbulence, different scaleshave been used in these plots in order to identify the dominant frequency t one angularposition and different radii.

The level of turbulence at the inner radius near the region of the skeg alongradial lines of 160 and 15 is much higher than that for the outer radius. In the regionof skeg near the keel), a decrease in the axial component of velocity fig. 24) andsubsequent increase in tangential and radial components of velocity fig. 25 and 26),indicates p r s ~ of turbulence in the region of skeg ne r the keel). Along the radialdirection of 90 , the energy content at radius of 6.98 cm is larger when compared to thatat radius of 10.67 cm; which was mostly exposed to the free stream now. Due to arelatively uniform distribution of water flowing into the propeller plane except the regionof skeg), the amount of energy in the spectrum is not significant and indicates that thelevel of turbulence is minimal.

72


87/139

Chapter Five

iscussion and onclusions

All sensors were calibrated over a range speeds. Prior to calibration eachsensor, the constant temperature anemometers were warmed up at the beginning theday) for nearly 90 minutes This ensured thermal stability the resistances in theelectronic circuitry which otherwise would induce an error in measurements. Effects humidity on the Wheatstone bridge balance were ignored

Each sensor had to be put through a bum in period approximately 70 - 80hours before calibrations were done Repeatability levels normal calibrations werechecked for individual sensors, which when curve fined using the modified form King s law showed their output to be stable pilch calibration was attempted on sensor3; which during the initial stages appeared to be sensitive to changes in angles pilch.However, sensor 3, once put through a further bum in period demonstrated insensitivityto changes in angles of pitch. Pitch calibration was done for all sensors and data werecurve filted based on the modified King s law was concluded that the sensors wereinsensilive 10 changes in angles of pilch. A polynomial 3rd order gave the best fit forboth normal and pitch calibration data.


88/139

The directional sensitivity calibration, also commonly referred to as yawcalibration, covered yaw angles ranging from 0 to 90 . Yaw calibration data were curvefitted based on a two-variable non-linear polynomial of 3rd order. Yaw characteristicsaround 45 were interpolated based on the two-variable non-linear polynomial. In theregion interest 0.55 m/s curve fitting error varied from 0.03 to 1.5 .Percentage error at each yaw calibration point is available in the Appendix.

The calibration velocity was measured with an accuracy about 0.005 t[Bose and Luznik 1995)]. The entire calibration processes all sensors were done atan overheat ratio 0.4, which to a large extent eliminated the need for temperaturecorrection. However, thermal currents were present in the wave tank which affectedsensor response during the calibration process. The level variation in temperaturealong the horizontal direction in the wave tank could not be measured continuously.Normal and pitch calibrations were done at a constant depth 0.3 m measured fromthe top surface the water in the tank). However, yaw calibration was done at varyingdepths in the top 0.3 m water column) in the wave tank. On certain days, during theyaw calibration process. a temperature gradients of approximately a.soc wereobserved in the top 0.3 m of water column. As this phenomenon was not observedcontinuously during the yaw calibration process, temperature correction was not appliedto yaw calibration data. The complete effects horizontal and venicaJ temperaturegradients on sensor characteristics could not be accounted for by an appropriate

74


89/139

com:ction to the output of the sensors. Also, the thermal c:oefficient of expansion wasassumed to be conslant while calculating the decade resistance of the anemometer.

From the calibration plots it can be seen that, at high velocities, the gradient ofthe calibration curve is reduced compmd with that at lower velocities. The calibrationcurve was flat at higher velocities, which indicated that the heat loss rate or theconvection rateof the sensor had reached a limit and that the anemometer output voltage,V, became v y n ~ s t v to any further increase in velocity, U. Any test conductedin this range yields a large error in measured velocity, for the reason that any slightchange in voltage, V, will correspond to a large change in velocity, U Calibration at0.S5 mIs which was our prime concern, fell in the highly non-linear range where thegradient of voltage against velocity was relatively large.

Results obtained from the pilot tube wake survey and this study indicate arelatively uniform u l distribution of flow in the propeller plane for angles fromapproximately 5to 165 i.e. with the exception of the sleeg area . There is a lar eapparent scatter in the data obtained by this study which is due to the fact thai thetemperature varied within a range of a soc during the test period five days . Thepattern of flow was the same for all radii except for either a drop in velocity or anincrease in velocity relative to one anolher different radii . An improved temperaturecorrection procedure will help eliminate the errors introduced by linear approximationwhich was employed in Ihis study. Also, there s a possibi:ity that the sensors


90/139


91/139

continuously measure water temperature during a time series record andcorrect each data sample for the appropriate temperature Herecorrections were made based on calibrations done at the beginning eachday of tests. This overall correction for the general trend in thetemperature variation in the tank was insufficient to maintain acceptableaccuracy in the records

2 The velocity components axial. radial and tangential were resolved fromthe three outputs from three single sensors. Any small variation in theaccuracy of a signal say that arising from an error in temperaturecorrection affects all three components of velocity.

To be a useful instrument for nominal wake survey both these sources of errorneed to e addressed. This might be done by more reliable means for temperaturecorrection or it might be alleviated to some extent by use of a bank of sensors such thattime changes between measurements are redUl:ed. The latter would not account forspatial variation in temperature over the extent of the tank. The sensors do allow forrelatively high spatial resolution in the measurement of velocity and for measurement ofthe energy levels contained in the turbulence. They are the only type of sensor whichallow n effectively continuous signal to be made of velocity a high frequencyresponse.

77


92/139

ReferencesBose N and Luznik L., ~ n c e n a i n t y analysis in propeller open water tesls,Submitted to International Shipbudding Progress 1995.

Champagne, F. H., Sleicher, C. A. and Wehrmann, O. H., Turbulencemeasurements with inclined hot wire. Part I heat transfer experiments withinclined hot wire. , Journal of Fluid Mech . vo1.28 part I ppI53.175. 1967.

Disa Eleklooniks, Inc., Instruction and service manual for type SSMIO eTADisa Elektronik lS

Durrani, S. Tariq, and Greated, A. Clive, Laser Systems in FlowMeasurement Plenum Press New york. 9n

Eric W Nelson and John, A. Borgos Dynamic response of conical and wedgetype hot ilms Comparison of experimental and theoretical results TSIIncorporated, Vol. IX Issue I, January-March 1983.

Fernando, E. M., Donovan J. F. and Smits A. J., The calibration andoperation ofaconstant-temperature crossed-wire probe in supersonic flow , Dept.of Mech. and Aero. Engg., Princeton Univ., New Jersey 1988.

Freymuth, Peter, Review : A Bibliography of Thermal Anemometry ,Transactions of the ASME Vol. 102 June 1980.Harris, C. A particle tracking system for wake survey\ M. Engg. Thesis,Memorial University ofNewfoundland SI. John s, Newfoundland Canada 1992.

78


93/139

Hollasch, K. and Gebhart, B., Calibration of constant-temperature hot-wireanemometers at low velocities in water with variable fluid temperature , ASMEAIChE Heat Transfer Conference, 1971.

Jimenez, J., Martinez-Val, R. and Rebollo, M., Hot-film sensors calibration. drift in water , J. Phys. E. Sci. Instromen., Vol. 14, 1981.

Johnson, F. D. and Eckelmann, H. A variable angle method of calibration forX-probes applied to wall-bounded turbulent shear flow , Experiments in Fluids2, pp121-130, 1984.

King, L. V., On the convection of heat from small cylinders in a stream offluid phil. trans. Roy. Soc. London, Sec. A Vol. 214, pp373-432, 1914.

Lakshminarayana, B., Three sensor hot wire/film technique for threedimensional mean and turbulence flow field measurement , Prepared for TSIIncorporated, Vol. VIII, Issue I Ianuary-March 1982.

Lomas, G. Charles, Fundamentals of Hot Wire Anemometry , CambridgeUniversity Press, London, 1986.

Morrow, T. B. and Kline, S. I TIle performance of hot-wire and o t ~ i l manemometers uSlXt in water, Instrument Society of America, ppSSS-S62, 1974.

Mulhearn, P. and Finigan, I A device for dynamic testing of X-configurationhot-wire anemometer probes . J. Ph) s. E. Sci. rnstrum., Vol. 11, 1978.

Nerdco Limited, An investigation of the resistance, self-propul.sion andseakeeping characteristics of two 6S fishing vessels , Report on work contractedby the Institute for Marine Dynamics. Project No. 143-900, March 1991.


94/139


95/139

Wu S. and Bose N. ~ a l i b r a t i o n of a wedge shaped vee hot film probe in atowing tank Measurement Science and Technology Vol. 4 pplOl 108. 1993.

Wu S. and Bose N. Axial wake survey behind a fishing vessel model usinga wedge shaped hOI film probe\ Kansai Soc. of Naval Arch. Japan May 1992.

WU S. and Bose N. An extended power law model for the calibration of halwire/hot film constant temperature p r o b e s ~ International Journal of Heat andMass Transfer Vol. 37 No 3 pp437 442 1994.

81


96/139

BibliographyArndt R. E. A. Stefan H. G. Farell. C. and Peterson S M. Advancementsin Aerodynamics Fluid Mechanics and Hydraulics . ASCE. 1986.

2. Durst F Melling A. and Whitelaw H Principles and Practice of LaserDoppler Anemometry Academic Press Inc. London) Ltd . 1976.

3. Dreyer G. F Calibration of HotFilm Sensors in a Towing Tank andApplication to Quantitative Turbulence Measurements , ASludent paper presentedto the Chesapeake Section of the Society of Naval Architects and MarineEngineers. U. S. Naval Academy Annapolis Maryland March 1967.

4. Fox J. A Introduction to Fluid Mechanics McGraw-Hili Book Co 1974.

5. Harvald SV. AA. Resistance and Propulsion of Ships John Wiley and SonsInc. 1983.

6. Jackson, Herring and James, C. McWilliams, Lecture Notes on Turbulence .World Scientific Publishing Company Pte. Ltd., 1989.

7. Kreyszig, E. -Advanced Engineering M a t h e m a t i c s ~ Fourth edition, John Wileyand Sons Inc., 1979.

8. Massey, 8. S. -Mechanics of Fluids Van Nostrand Co. Ltd., London, 1968.9. Melnik. W. and Weske, J. R. ~ d v a n c e s i n Hot-Wire Anemometry . AFOSR

final report no. 68-1492. Dept. of Aero. Engg, Univ. of Maryland, July 1968.82


97/139

10. Peerless, S. J MBasic Fluid Mechanics , Pergamon Press Ltd., 1967.

t Robert, P. Benedict, Fundamentals of Temperature, Pressure and FlowMeasurements , John Wiley and Sons Inc., 1969.

12. Rosenhead, L MLaminar Boundary Layers , x ~ o r University Press, 1963.

3 Saad, A M., MCompressible Fluid FlowM, Prentice-Hall Inc., New Jersey, 1985.

14. Smol yakov, A V and Tkachenko, V M., The measurement of TurbulentFluctuations - An Introduction to Hot-Wire Anemometry and RelatedTransducersM, Springer-Veslag, New york, 1983.

IS Wu, S. and Bose, N., Adoption of the DlSA SDOS CTA for WakeMeasurements in the TowingTank ,OERC91-WIT-TROO2, Memorial Universityof Newfoundland, Canada, June 1991.

6 Wu, S. and Bose, N., MMethods of Fitling Calibration Curves to Hot Film/HotWire Constant Temperature Anemometer Data., OERC91-WIT-TROO3,Memorial University of Newfoundland, Canada, June 1991.

17. Wu, S. and Bose, N Wedge-Shaped Vee Hot-Film Probe Calibrated forTowing Tank MeasurementsM, OERC91-WITTROO4. Memorial University ofNewfoundland, Canada, June 1991.

18. Wu, S. and Bose. N., Methods of Calibrating X-Wire Signals from ConstantTemperature Anemometers , OERC91-WIT-TROOS, Memorial University ofNewfoundland, Canada, June 1991.

83


98/139

Appendix

Program to examine repeatability normal and pitch calibration of sensors basedon the modified Conn of King s law.Program to evaluate the coefficients of two-variable non-linear polynomial basedon yaw calibration data.Input and output files yaw calibration from the above referred program .Program iterate values velocity a nd yaw a ng le for a given voltage used t oanalyze wake survey data

84


99/139

Program ensure repeatability normal and Ritch ClIlibI l tjpD Of sensors

Based on the modified form King s lawA 3rd order polynomial was used in this studyNegligible curve fitting errors for repeated normal and pitch calibrations presented in the body this thesis) confirmed stability the sensors.Subsequently, yaw calibrations were done.

85


100/139

THIS PROGRAM CURVE FITS DATA BASED ON THE VALUE OF M.KINO S LAW WHEN M - I AND OTHERWISE BECOMES MODIFIEDFORM OF KINO S L Wproaram dpfilimplicil real+S (l-h,o,:r.)panmelel (numu-IOO,namax I Idimension u n u m u ) , v n u m u ) , l I . n u l l l l l l . ) , ) { l I u m u ) , ~ r l I . y n u l l l l l l . )dilIlelUion l n.tmU ,b nlmu ,coel{nlllllll.,nlmu ,tt1 n.mucomnlOn/acomlllll.,ux IOO ,vlI. IOOCOMMON IboutllllOtdcham::ter+ISfilmClI.temalll.yfil.lran5.fabn.dif.npoly.cylwkt

o p e l 1 2 , f i l ~ _ i n f l t l t . d a t , 5 t a l u l l _ o l d )o p e l 1 9 , f i l ~ _ Olilput.dat ,status_ DeW )INPUT DATA IS FROM INPUT.OATOUTPUT IS STORED IN OUTPUT.DATORDERING OF U.NX - NUMBER OF CALIBRATION POINTSread(2,+)1lJI.READ IN VELOCmESrcad(2,+) (ux(it),it_I,IlJI.)RE D IN VOLTAGESread(2,+) (vx(it),it_I,IlJI.)l ip-TIlldoS it_l,numlllu(it)_Ul{it)v(il) vll.{it)WRlTE{S,+ (VX(lP),IP-I,NP)WRlTE{o5, + (V(IP),IP-I,NP)DO 120K-I,NP-1DO 130 I-K+I,NPIF(u{I).GE.u(K GOTO 130trl-u Ku(K)_u(l)u I _trltrl_v Kv K _v lv l -tr l130 CONTINUE120 CONTINUEw r i t ~ o 5 , l l )

11 f o r m a t l l , i n p u l m u d ~ f t n e d in Eq. 05) )read 05,+ mordmord-mord+l


101/139

r.. lInpoly(c,u,v,np,c:n,Iloe,lseod,llIO,uxy)write(9,SI)np,(u{i),i_l,op)write(9,61) (v(i),i_l,np)5\ formal(l1,'dala file ERRXY,DAT-output from STP lT

I f ll, numberofealibnlionpoinlsnp_ ,iSI fIIx,'calibralionveloc:ities:',17(/lx,6flO.461 formal(lx.'calibra ionvollageoutpul;',17{llx,6flO.4)ltIOrd_mordJwrite(9,21)mord,c.(cn(il),il_l,mo)write(9,31)l5eod,tse

21 fonnat(b,'order m- ,i41 Ib, exponcnln= ,f10.41 lb, coef. Ak .. ',SfI3.6/1101,5fl3.6)31 fonoat(b, error function E I(V). ,eI2.3I /Ix, errorfunclionE 2(V) ',eI2.311)wrile(9,41) (eny(ik).ik-;'I,np)41 fomut(lx,'l'iterror aleach poinlin ~ , 1 7 l l x , 6 f 1 0 . 4

wrile(5,) needloealculalelbeyawfaclor /(I)read(S,)nyesif(nyeu:q.O)gotomwrite(5,) (cn(il),it_l,mo)callcyawkt(cn,mo,c)999 continue psubroutine npoly(c,u,v,np,cn,tse,tseod,mo,cny)implicit real.S (ah,o-z)parameter(nurnax 'IOO,namu-IO)dimensionu{np), (np).l(nulI1I1),y{numu),cn(narnax+l)dimemion a(namax+ l).coef(nlmax+ I,namax + I),erxy{nurnax)CO-O.2ood-


102/139

call fabn(tc,u,v,np,a,tse,r-xI,mo,Crl;y}if{lSe,gC.tseO)golo40c_lcdo 45 iI_I ,rna

45 cn(it)-a{it)40 continuetc . .cc.().OO'icall fabn(tc.u,v,np,a,lse,r-xI,mO,crl;y)if(tse.ge.tseO)goCo30,atseO_lsIIdo 35 i t -I ,mo35 cn it -a it

30 continuedo50i_I,IO1r._cc+O.OOOIical1fabn(lc,u,v,np,a,lse,tseod,mo,Crl;y)if(tse.gc.tseO)golo60C_IC -do55it I DO

55 cn ill-a it60 continue

Ic cc-o.OOOIicaUfabn(lc,u,v,np,a,tse,15cod,IDO,crl;y)if lU.gc.tseO goto50C_lcIseO_tsedo 65 i t -I ,mo

65 cn{it -a il50 continucreturn,,5 broutinc fabn(cll,u,v,np,a,tsc,beOd,mo,en:y)

i m p l i c i l ~ I 8 . h , o J ;parametcr(numu.-1OO,1I11I1 ll_IO)dimcn5ionu(np),V(np),ll(Ullmu),y(nulDIll)dimension a(namaxt),c(namax + I,llamax + I),erl;y(numall)call1lllf1s(ex,u,v,np,ll,y,mo)call llyfil{Jl,y,llp,mo,a,c:,tsc,var,cIl.Y)lseod_O.do IOi-l ,npvfil-dif(l,Iuo,Jl(i),O.)if(vfit.lc:.O.)vfil_O,\seod .. Iscod+(v(i)-sqrt(vfit' 2erxy(i)_abs(v(i)5qrf{vfil/v(i)IOO.

88


103/139

10 conlinueC gl . . tseC lsc_tscodC lseod-t:l msubroutine t T l \ s ( e ~ , u , v , l l p , l , ) , m o )illlplicilre.lllS(I-h,o-z)oommoolboullmord

dilIlCllSioou{np),v(np),l{llp),Y(IlJI)DO momdo IOJ-I,npl(i)-u(i)ely(i)_v(i)* 2

10 conlinuemNmcSUBROUTINE XYFIT(X,Y,NP,MO,A,C,TSE,VAR,ERXY)implicil realS (I-h,o-z)plflIme er(nlmu-IO)DIMENSION X(NP),Y(NP},ERXY(Nf ),A(MO),C(MO,MO),B{nllfW: + I)NP:No.ofdalipointsM o I : ordcroflbcpolynomill (MO


104/139

DO 300 KO-I,MOK_KO_IXB .. OD0400I-NI.N2XPART-IIF(ABS(X(I))+l+ K.GT.O) XPART-X{I) {K+l)400 XB-XB+XPART

3 C(JO.KO}-XDINP1 CONTINUE

CALL DLSARG(MO.C,MO,B,I,A)TSE-O.VAR . . OD0500I_NI,N2SXY-O.DO 600 K.. I.MOXPART_IIF(ABS(X(I))+KI.GT.O) XPART-X(I)(K-I)

6 SXY . .SXY+A{K) XPARTC error inERXY{I)-ABS(Y(I)-SXY)IABS(Y(1)) IOO.VAR-VAR+(Y(I)-YM) 2500 TSE-TSE+(Y(I)-SXY) 2RETURNENDFUNCTION DIF(A,M,X.YO)implicilrt.llS(a-h,o-z)DIMENSiON A(M)function DIF-F(I) Y F(I) is a polynomial f (M-Ithordcr)A(M) are tbecoefficienuY_O.DO 1 I . . I,M10-11Jpart-Iif(abs(I)+io.gt.O) lpart_Juio

1 Y_Y+A(I)lpartDIF .. YYORETURNENDcaleulation oflhc yaw factorsubroulinccyawkt(aco.mo,CJl)implicitrCII:S(a-h,o-z)paramcter(nlmax-ll.numu_IOO,pi_3.1415927/18O.)dimension aco(mo).yawf(numu,I\.llIIU).ylwkl(lIumu.namu)dimension ~ m e a n y n l m . ~ n n s y ( n . m u . ) . y t l \ l l ( l l. . . . . .),.ymi(llImax),

cmeank(nanlU).cnnsk(n.mal),dJna{namn).Jkrni(ll mu.)dimension u n u m u , n a m a x ) , v { n u m a x . D a m u ) , l I u n a m u ) . ~ D a m u )


105/139

dimensiOIl crf(numn,namu),Vl:(numax,n. lRIlt),suber (n rlIu),5ubrms(n.mu)

comlDOnlacomlnx,llIlorm(IOO),vnorm(IOO)ehaTllcler15fi1en ll u,v122oo+o./slItcmenl fuDctioooftbccffcclivevclocitydU:lhcc.librationvelocilyda: yaw lllI le (in radillll)dkt:y.wf.ctortn.(da)-sin(da)ylTll(da)_.1874269+3.222723c3+tTll(cla)+I.02656227+

1 tra(da)U2..Q.22411871+1n.(da)+3ylra(da) . . .207606+.313142+111l(da)+O.994339+1r.(da)UZ

1 ..Q.526574tTll{da)3ytTll(da)_O.150464+.118694+tra(da)+O.7190691ra(da) 21 +.:)24345cZII1l(da) 3ytTll(da)_O.Z07JI9.06205221ra(da)+Z.0799111l(da)UZI -1.21912tn.(dalu Jcuef(du,da,dkt)=(duytl1l(daUuC cucf(du,tIa,dkt)-(dusqrtJin(da)Z+C 1 dklu2 '{cos{da 2+u

C iopuldata fromcllannel2C input data file should e ilCl up in thc fo:lowing formal:C 1st row: na no. of yaw anglc calibr.tiolUC 2nd row: v.ll1CSofthose w angluC 3rd row: nuv)-no. ofvclocity calibratioll.t eachC cOlTeSpOnding y.w angleC 41h row: i v.luesofthcClllibratiollvclociliesC - allsly,wlll lcC 51h row: v .lues of lhc lllIcmomctcr oulput vollal:cC .ltbClC ls i y.wClllibrationvclocilicsC 6th row: u i-v.luesofthcClllibrationvcloc:itiesC - 1 nd y.w 1lI8leC 7th row; Y_i v .lues of thc anemometer output yoltalcC 1 tbcsc Znd y.wClIlibmlion velocitiesC .... lllIdSOOll

wrilc(5, ) (aco(it),it-I,mo),cxwritc(5,*) enter inpul data filcnlmc'IUd(5, ) filenopen(Z,fi]c_fiIen,slatus. . 'old')rcad in no. of yaw IlIglcs of calibrationread(2,)n.write(5,)n.read in lbcvalues oflhosey,wllIglcsrc.ad(2, ) (.le(i),i-I,n.)write(5,) (.te(i),i-\ '1l')read in the 1lO of velocity calibTllions It each oflhcg y,wllIlllesread(2,) (nu{i),i=l,na)


106/139

wrile(S,) (nu{i),i-I,IUl)read inlhosevelocides and output vol ..ges . teach of thoseyaw anglesdo 10i - l ,nlread(Z,) (u(k,i),k:-I,nu(iread(2,) (v(k,i),lt-I,nu(i1 continueC (lulputoflheinpulclatldo i_I,na

w r i t e 9 , 1 ~ ) i,llc(i),nu(i)fonnal{lx, Dal.l input IS follows:1 Ih,iZ, thYlwllllglevlllle_ ,n.2,I h , no. ofcalibra tionveloci ties_ ,iZ, theYll e: 1)wrile{9,2S) (lI( j,i)J- I,nu(i)25 fonnat(Ilr. , U_: ,IOnA

w r i l e { 9 , 3 ~ ) v j , i ) J - I , n l l i fonnat(b, V-: ,IOn.420 continue

mont-mo-Iif(mord.1.3) 810999(1gOlo (100,200,3OO),mordC m I-King sJlw100 continuedo I l0 i I n lIf-ate(i)*pido 120m_I,nu(i)effeclive velocity of the powerofnudn-(v(m,i) Z-aco(l/lco(Z)if(uefn.Il.0.)uefn_O,effeclivevelocityuef_uefn(I. t.l.)Ylwfunclionylwf{m,i)-uef/lI(m,i)alt-(yawf{m,i)j Z-(sio(If) 2j f a l t . ~ . 0 . ) Y l w l t t m , i ) _ l I q r t { a k { c o s a l } ) 2 )if(ak.h.O.)yawl:l(m,i)--sqrt(-ak (cos(afW 2)C neallive values of ylwkl represents iINginl? values120 coolinue130 conlinW

110 continue810400200 continueC m modified KinS s lawdaZIOi-l ,nl

a f ~ ..te(i)3.14IS51Z7fl801do 220m-l,nll(i)uefn-(-aco(2)+6Ij1t{ac0(2)aco(2)-4lCo(3)(IC1l(I)-I v(m,i)v{m,i))))/(Z lco(3))


107/139

u e r - u e f n I . / e ~ lwrite(S.) . e o i t . i l - 1 . l I l O . e ~y.wf(m,il-uerlu(m.i)write(S.lu(lll.i),uefu,uef,y.wr(m,i)

u - u e f / u m , i ) l ~ - s i n a f l ) ~if(ak.ge.O.) 8wkt{m,i)-sqrt(akllabs(eos(af)if(Ik.II.O.lyawkt(m,il--sqT1(-ak)labs{eos{al)l220 continue210 continuegoto400300 continue. -.co(4)b-.co(3)c-.c0(2)delll b/(3.a)p-(3.. cbb)/(3...)I3.do310i-l ,n.af . .ate{i)3.14IS927/1801do3IOm-l.nu(i)d-aco(ll-(v(m,iW 2q a b 3+27...


108/139

num .. Onnsy-O.nnsk-O.do430m_l,nu i)ir(y.wkt(m,i).Ie.O) 1 1 43num_num+1nnsy_nnsy+{cmeany(i)y.wf(m.i))-OO2muk_ nIlSk+ (cmeank(i)y.wJ:t(m, i OO243 OOlIlinuecnnsy{i)-Ocnnsk(i) . . Oif(numII.O)crmsy(i)_sqrt(nnsyfnum)if(numlt.O)crmsk(i)_sqr1(rmskfnum)410 continuewrite(9.405) mord4 5 format(flh,'rf'sul15 from m- ,il

I JI1.'neg.li...e ....luesofkt represenlSimaginary ....lues'I 111. ofthey.wfaclorastbey.wfunclionislessI 11l, lhanlhesineoflhey.waDlle)do 44 i-I,nawrite(9,415)i,.le(i),nu(i)4lS formal (h. i2, lhy.wangle ....lue- ,f7.2,I Ih, no.ofc.libralion ...elocities-',i2.I fh,y.w function y_ ; yaw faclor k. . below:')write(9,42S) (y.wf(k,i),k_l ,nu{i425 fonnal(ll.'y_:'.10e10.2)wrile(9.43S)cmeany(i).crmsy(i)435 formal(b.'the mean of those yaw function v.lues .bove th.tI fll, .re gre lerthan lite sine ofYlw Ingle is: ,n.3

I Ih, lheir eorrespollding rool-mean'square is : ,17.31write(9,445) (y.wkl(k,i),k_l,nu(i445 format la, k-: .IOn.3)write(9,455) cmeank(i),crmsk(i)455 fonnal(h, 1 16 mean of the y. w factor ....Iues abo ...e zero is:'I ,17,3I Ih,'theircorrespondillgroot-mean.squareis : n 31440 continuewrite(9,465) ate i),i-I.n.)465 fonnal(h, a.. ,619.2)C do 4SOi l,nawrile(9,475) (Yawf(j.i)J .. l.nu(i475 formal{Il.6l'9.3)450 cOlltinue

lot05009990 cOlltinue


109/139

write(9,505jmord505 fonmt(h, '00 ulcul.tion routine s aVlillble for m>3Ih, nooutputislvail.bleform= ,i1

1 1 ~ l r y another value m omon than 3 )~

500 continuenonn nl+1D U { D o r m - n ~ate(norm)-90.do 510 i _ I n ~u(i,norm)_unorm{i)v(i,norm)-vnorm{i)510 ~ n t i n u e

Cdo910i_l,norm910 write(9,905)lte{ij,(u(m,i),m_l,nu{ido 92Gi-l,norm

920 write(9,90S)ate(i),(v(m,i),m. . l,nu{i905 fonnat(h,f6.1,14fB.3)C do 810i_l,nl810 write(9,905) ate{ij,(yawf(m,i),m-l,nu(i),emtlllly(i),cl'lI1.'Iy(ijdo 820i- l ,nl820 write{9,905) Ite{i),(ylwkt(m.i),m .. l,nll(ij),cmcank(i),cnnskji)600 continuewrite( l,) inputlvaluefortbeyawfactOfread{5,)cktnum_O.ersum-Odo 520 i-I,normar 'lte{ij pinsub_Oersub_O.do 5 m_l,nu(i)vc{m,ij_sqrt(dif(lco,mo,cuef(u(m,ij,af,ckt),O.))erf(m,i)-abs(vc(m,i)-v(m,i/v(m,i)IOO.ersum .. ersum+erf(m,i)ersub_crsub+erf(m,i)Dsub_nsub+lnum-num+1530 continuesuber(ij-crsub/nsub520 continueersum_ersumlDumrool-rnean-sqUlllllofthcerrors

os


110/139

ntI l-O.do540i-l,nonnnsuh_Onnssub-O.do5S0m_l,nu(i)ntI l-nns+(erlium-erf(m,i 2nnssuh-nnssuh+(subel(i)-erf(m,i' 2nsuh_nwb+1550 cOlItinuesubrms(i)_sqr1(rmssuh/n5llb)540 continuerms sqr1(rmslnum)write(9,515)ckl515 fontllt(h:,'the input yawfllClOr value - ,n,3fdo 560 i_l,nonnwrite(9,525)i,lle(i)525 fontlll(IJt,i2,'thyawangle-',h,2I Ib,' calculated volt-ges (Vc)'

I 1 J ~ p e r c e n l a r e fiuingerrors e( ) are'/)write(9,SJS) (vc(m,i),m_l,nll(i fontllt(Il, 'Vc-',IOn.4)write(9,S4S) (erf(m,i),m-I,nu(i)) fontlll(lJt,'e( )-',lOn.2)write(9,555)suber(i),subrms(i) fonnal(IJt,'wlll a mean percentage error of ',19.4

Ib, 'and roolmeaftsquare of the mean error or,f9.4/1)560 conlinuewrile(9,SIS)ckldo 930 i-I,norm

93 wrile(9,90S) ate(i),(vc(m,i),m_1 ,nll(i))do94 i_t nonn940 wtite(9,91S)ate(j),(erf(m,i),m l,nu(i,luber(i),subrnl$(i)915 fOrmlt(11,f6.1,14fS.2)wrile(9,S6S) ersum,rms

6 fonnal(ll,'tnean value of the total percentageI fillinserron-',fS,21 ,'wilharoot-mean-squarevalueor,fS.2)wrile(5,) 'try anolher k tvalue?(l)'read(S,)nrel -if(nret.eq,l) golo600return d


111/139

Program t m @ e the cosfficj;n't gf two variable ngo. jDQ[ pgIyngmii

sed on yaw calibration dataA 3rd or r polynomial w u used in this studyUsed for all foor sensorsFollowing the program, input files and output files (voltage, velocity, curvefittingerror in percentage at each point, corrected voltage, coefficients of two-variablenon-linear polynomial) have been attached

97


112/139

THIS PROGRAM CURVE FITS DATA BASED ONTWOVARIABLE NONLlNEAR EQUATION AS DESCRIBEDIN THE REPORT. THIS PROGRAM WAS USED TO CURVE FITYAW CALIBRATION DATA. A THIRD ORDER POLYNOMIAL WASUSED IN THIS WORK.NX - NUMBER OF CAlIBRATION ANGLES SHOULD BE SET INITIALLYBEFORE COMPILING THE PROGRAMMO ORDER OF POLYNOMIAL + SHOULD ALSO BE SET INITIALLYIMPLICIT REAL.S (AH,OZ)PARAMETER (NX S,MO-4,API 3.J4IS926S36fl80.)PARAMETER (MS MO+(MO+t)12)PARAMETER (MT-MS+4S)DIMENSION X(NX),Y(NX,30).Z(NX,30),A(MS,MS),C(MT)DIMENSION NY(30),ERXYZ(30,NX).YT(NX.30).ZT{NX,30jDIMENSION D3XYZ(30,NX),D3CON(NX,30),g(30,NX),H(JO,NX)DATA Y,Z,G,HI96O O.fEXTERNAL ZXYFJTOPEN(2,FILE- 7y.w.DAT ,STATUS_ OLD )OPEN(3,FILE- 7YAWO.DAT ,STATUS , NEW )OPEN(4,FILE- 7YAWU.DAT .STATUS_ NEW )OPEN(7,FILE- 7YAWV.DAT ,STATUS,. NEW )OPEN(8.FILE,,7YAWVC.DAT ,STATUS_ NEW )YAWO.DAT ASSORTED OUTPUTYAWU.DAT CALIBRATION VELOCITYYAWV.DAT CALIBRATiON VOLTAGEYAWVC.DAT FmEO VOLTAGEdo 998 irel_I,1READ(2,) (X(I),I-I,NXjREAD(2,) (NY(I),I-I,NX)DOIOI-I ,NXREAD{2,) (Y(I,I),J -I,NY(1))

1 READ(2,) (Z(I.J),J-I.NY(I)do9i_l ,nxx{i)a ll.(i)apil npoly(ex,y.z,Yl,zl,lU,ny,a,mo,ms,a,c, l5e,tseod,enyz)ealllrans(ca,y,..,tu:,lly,yl,Zl)CALL ZXYFIT(X,NX,YT,NY,ZT,A,C.MO,MS,TSE,ERXYZ)60 FORMAT(IX:. ._ ,SE1S.7)wrilc(3,60) l5e,cawrilc(3,60) (e(i),i-I,ms)ealelll.tionorthcsumorl.bcsqUlrcdcrrorsincquatioll(lO)T5EOO . .0.DO 100 I-I ,NXDO 200 J-I,NY(I)yt(iJ)-(j O.IS}+ u

98


113/139

SUBT_O.NUVS_ODO 1l01U_I,MOJU-IUID0210IV_I,1UJV_IV_IXPART-]IF(ABS(X{I))+JV.GT.O} XPART .. X lluJVYPART. . IIF(ABS{YT(I,J))+ABS(JU-JV).GT.O) YPART..YT(I,Jj (JU-JV)NUVS-NUVS+1SUBT-SUBT+C{NUVS)XPART-VPART

210 CONTINUE110 CONTINUE

TSEOD-(Z(I,J)SQRT{SUB1) 2+TSEODERXYZ(J,I)-ABS(Z(I,J)oSQRT(SUB1)IZ(I,Jj IOO.d3xyz(j,I)_sqr1(subt)

200 CONTINUE100 CONTINUEWRITE(3,60) TSEOD,EX

WRITE(3,SI) ERXYZ(I,J),J .. I,NX),I .. 1,14)5 FORMAT(lX,7F7.4)DO 3 I_I,NX

00301 J_I,14H(1,I) Z(I,J)301 G J,I _Y I,JWRlTE(4,71) G I,J ,J_I,NX ,I_I,14WRlTE(7,71) H I,J ,J_t,NX ,I_I,14WRlTE{S,71) D3XYZ{I,J ,J-I,NX ,I_I,14

7 FORMAT(lX,7F7.4)998 continue

STOPENsubroUline npoly{cx,u,v,yl,Zl,nll,ny,x,lIlO,ms,eocf,cn,tse,lSeOd,crxy:)implicitrealS{a-b,o-z)dimellSion u(nx,30),v(nx,30),yt(nx,30),ll{nx,30),x{nx)dimension n y 3 0 , c r x Y : D J I . , 3 0 , c o e f n I l I , I l 1 S , . 4 ~ , c n m sex-O.3nO_exIe_excall (abn(tc,u,v,yt,Zl,nx,ny,x,mo,ms,coef,cn,tseO,l5eodO,erxyz)do 20 i_I,30lc=exO+O.OI*icall fabn(tc,u,v,yt,Zl,nx,ny,x,mo,ms,cocf,a,tsc,IsCOd,erxy:)if(tsc.se.lseO) soloWeX_I

-


114/139

doZS il_l,msIS en h)-. iI)20 continue

do 30 i_I,IOte_cc+0.OO1ioll f.bn(lc,u,v.yl,zt,lU,ny.lI,mo,ms,eoef,.,lsc,tseod,cnyz)i f { ~ . g c , t s d J g o l o 4 0ell-IetseO=ISCdo jt l,ms

en il)-. it)4 >nlinucte_cc-o,OOI-iClllf.bn{lc,lI,v,yl,Z1,Illt,ny,:r.,lIIo,llIS,cocf,.,tsc.tseod,enyz)

if tse,gc,tsdJ)goto30UafetseO_lsedo 35 il-l,llIS35 en{it)_.(it)

3 continuedo50i-I ,10te-cc+O,OOOIiClllf.bn(tc,ll,v.yt.Z1,Illl,ny.lI,mo,ln.'I.eoef ,lse,tseod.enyz)if(tse.ge,tscO)a:oI060ell-Ie

~ _ I s edo 55 il l,ms55 eD(it)-=.(il)

60 continuele-cc.o.OOO1ic.lI f.bn(lc,u,v,yt.zt,nll,ny,:r.,tnO,ms,coer,.,

Documents

Testing ship Ramanathan Krishnan