7.5-02-03-03.3

Home , Reynolds number

ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 1 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

Table of Contents 2.5.2 Higher Harmonic Components .... 8 1. PURPOSE OF THE PROCEDURE ...... 2.5.3 Uncertainty Consideration ...... 9 ...... 2 2.6 Other Essential Experimental 2. MODEL SCALE EXPERIMENTS ON Conditions...... 10 PROPELLER INDUCED UNSTEADY 2.6.1 Air Contents, Cavitation Nuclei PRESSURES ...... 2 and Stabilizing Model Cavitation 2.1 Test Set-Up ...... 2 ...... 10 2.1.1 Propeller Model ...... 2 2.6.2 Influence of Wake Simulation ... 10 2.1.2 Wake Simulation ...... 3 2.6.3 Facility Wall Effects ...... 11 2.1.2.1 Global Information ...... 3 2.6.4 Effect of Induced Vibration ...... 11 2.1.2.2 Dummy model ...... 3 2.6.5 Free Surface Effects ...... 12 2.1.2.3 Simulation by a Ship Model ...... 4 3. PARAMETERS ...... 12 2.1.3 Pressure Transducers ...... 4 3.1 Parameters to be Taken into 2.2 Test Conditions ...... 5 Account ...... 12 2.2.1 Propeller Loading Condition ...... 5 3.2 Recommendations of ITTC for 2.2.2 Corresponding Pressure Field ...... 5 Parameters ...... 15 2.3 Measurements and Instrumentation ...... 5 4. VALIDATION ...... 15 2.4 Calibration ...... 6 4.1 Uncertainty Analysis ...... 15 2.5 Data Collection and Analysis ...... 7 4.2 Benchmark Tests ...... 16

2.5.1 Global Data Collecting and Presentation of Results ...... 7

Updated Approved

Propulsion Committee of 27th ITTC 27th ITTC 2014

Date: 02/2014 Date 09/2014

ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 2 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

Cavitation Induced Pressure Fluctuations Model Scale Experiments

signals should be displayed. The model scale results must be analyzed for harmonic content and 1. PURPOSE OF THE PROCEDURE the pressure amplitudes scaled to ship scale for the prediction of dimensional unsteady hull sur- To provide guidelines and to ensure a best face pressures. possible quality of test results in terms of accurate data on the performance characteristics of Test facilities for this type of experiment are: ship propellers concerning cavitation induced variable pressure water tunnel, depressurised pressure fluctuations (blade rate and higher or- towing tank, and circulating water channel with der excitation). a free surface in the test section. First harmonic or blade rate means the blade frequency i.e. the number of propeller blades 2.1 Test Set-Up multiplied by the rate of revolutions of the propeller shaft. 2.1.1 Propeller Model

The present meaning of higher harmonics The size of model propeller should be deter- covers a frequency range between 5 to 10 or pos- mined, within the capacity constraint of the test sibly 20 multiples of the blade frequency, i.e. in facilities and within an acceptable range of test- practice causes high frequency vibrations and up section blockage (less than 20 %), to achieve the to what on passenger ships would be called low highest possible Reynolds number. The propel- or medium frequency noise. ler model must have sufficient ( 0.05 mm for a 250 mm diameter model propeller) and con- sistent geometry accuracy and therefore applica- 2. MODEL SCALE EXPERIMENTS tion of numerical controlled machining may be ON PROPELLER INDUCED UN- helpful to keep consistency in geometric accu- STEADY PRESSURES racy. Model propeller blades are popularly made of strong aluminium alloys or brass. Whether The basic pressure fluctuation test consists the surface of the model is hydrophobic or hy- of measuring transducer signals from flush- drophilic might have a remarkable influence on mounted pressure gauges located on a surface the cavitation characteristics and hull-pressure representing the ship hull shell, located adjacent measurements. A thrust to disc area loading of to a propeller operating in a non-uniform inflow about 70 kPa is a useful upper limit value for velocity field representing the flow behind the strength considerations. ship. Choice of number and placement of pressure gauges should cover the local peak of induced pressure amplitudes and the distributions of pressures both laterally and longitudinally. Example characteristic time series of pressure ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 3 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

2.1.2 Wake Simulation any case, all wake field simulations shall comply with ITTC Recommended Procedures and 2.1.2.1 Global Information Guidelines, 7.5-02-03-02.5: Experimental Scal- ing of a Wake to a Target Wake, which describes The main task is to provide a realistic simu- guidelines for experimental wake scaling and lated wake velocity pattern which will also give simulation. the proper speed of advance at the propeller disk location. Usual practice is to use the nominal 2.1.2.2 Dummy model wake distribution (either for the model or scaled to full scale) as the target wake for the experi- The main issue is to provide a simulated ment. wake flow to the propeller by using partial ship model or flow devices. There are scaling issues connected with try- ing to satisfy the three governing similarity pa- Alternative model schemes are used depend- rameters: Reynolds number, Froude number, ing on type of ship and facility size: and cavitation number. Cavitation tunnel pressure fluctuation testing does not satisfy Froude In small water tunnel test sections include or Reynolds number scaling, but is aimed at a (in case of a single screw ship): relatively high value of Reynolds number. Test- ing in a free surface cavitation facility offers the  Wire mesh screen placed perpendicular to possibility of matching Froude and cavitation the flow, in front of a flat plate simulating number, but at a rather low Reynolds number. the hull  Parallel plate wake generator, in front of a When Froude similarity is not satisfied, the flat plate ‘hull’ correct cavitation number is exactly satisfied only at the blade reference point for scaled sub- For small to medium size test sections, the mergence depth pressure. This can be compen- alternatives could include: sated for by simply adjusting the pressure head to the appropriate value for each of the blade po-  Inclined shaft with struts and bossing, sitions of interest. mounted below a flat plate or bump-like dummy model hull (in case of twin screw The low Reynolds number of model testing ship with propeller operating outside the presents more complicated problems. Because ship boundary layer) of Reynolds number scale effects, the boundary  Dummy model (‘after body model’) layer thickness relative to hull length along a ship model in the cavitation tunnel is relatively In large test sections: wide. The magnitude of this influence depends greatly on the type of hull form. Wake scaling  Shortened, but otherwise scaled ship mod- techniques such as the contraction methods offer els relatively simple methods for scaling a model wake to a target wake with full scale features. If  Half complete or shortened ship model at- Reynolds number wake scaling is to be applied, tached on a side wall (in case of twin the practical recommendation is to target the screw ship) three-dimensional full scale nominal wake. In  Complete scaled ship models ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 4 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

For all the model configuration types men- of vibration and noise generated by the motor, tioned, it is recommended to include as much of gearbox or any rotating device used in the trans- the stern appendages, such as the rudder, in the mission mechanism. An accelerometer shall be correct location behind the propeller. installed on the ship model to monitor the levels and the frequency distribution of the induced A difficult task is modifying a model config- hull vibrations. uration so that the measured wake reliably matches the target velocity profile by surface The alignment of the shaft and joints shall be mounted patches of screen, shortened model accurate in order to reduce propeller rotational hull length, wire mesh screens mounted perpen- speed variation during the revolution. Propeller dicular to the hull, slimmed after body dummy rotational velocity fluctuation shall be main- model shapes and flow liners in the corners of tained within 1%. In order to check the rota- the test section to prevent flow separation. All tional speed of the propeller the mounting of a these configurations strongly depend on the skill multi-pulse encoder (at least 1000 pulses per of the model basin and on existing correlation revolution) on the propeller shaft shall be pre- data between model and full scale. ferred.

When using dummy model, the simulated For the above reasons in case of twin screw wake shall be documented to verify that the de- ships two motors configuration (one for each viation from the target wake can be neglected. propeller) shall be preferred because this re- Measurement of the wake can be performed by duces the number of mechanical components re- using any suitable Velocimetry techniques. quired in the transmission and hence the induced When using pitot tube or laser Doppler Veloci- hull vibration. Furthermore a two motors config- metry, ITTC procedures can be adopted. uration allows the possibility to investigate the effect of different phase rotational angles be- 2.1.2.3 Simulation by a Ship Model tween the propellers for reducing the hull excitation induced by cavitation. The full ship model for cavitation induced pressure pulse test normally comes from previ- In case that for cavitation observation there ous resistance and self propulsion tests per- are windows required for cameras installation formed in a towing tank. Normally the specifi- on the model such windows should not affect the cation of the model construction shall comply propeller inflow. with the ITTC 7.5-01-01-01 Quality manual procedure regarding model construction. Fur- In case that electrolysis is used for cavitation thermore it shall comply with the low pressure stabilization, electrodes could be mounted on environment and the higher velocity encoun- the ship model at least 1 meter upstream the pro- tered in the cavitation test facility. peller plane.

The ship model shall be completed with all 2.1.3 Pressure Transducers the required appendages, fin stabilizer, -rudder and turbulence stimulators, that can influence Rugged miniature, high sensitivity, flush the propeller inflow. mounted pressure transducers of suitable range shall be used for pressure fluctuation measure- High quality mechanical components shall ments. Transducer range depends on the facility be adopted for the transmissions with respect to velocity and static pressure and could be abso- self propulsion tests in order to reduce the level ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 5 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

lute or differential. Band pass frequency re- 2.2.2 Corresponding Pressure Field sponse of the transducer shall be at least 1000 Hz and the measuring range shall not exceed 4- Set the facility pressure and flow velocity to 5 times the maximum measured pressure in or- obtain the correct full scale cavitation number  2 der to limit the measurement errors. Pressure = (p0 -pV)/(1/2Vref ); Where p0 = total static transducers shall be periodically calibrated with pressure consisting of atmospheric pressure plus respect to a standard calibration reference. If submergence depth pressure taken to a reference possible a minimum of 5 transducers shall be in- location on the propel1er blade, and with the ref- stalled on the ship model hull. If force calcula- erence velocity Vref taken as VA,. nD or nD. The tions shall be performed the number of pressure reference submergence depth used in the calcu- transducers shall be at least 20 located over an lation of the cavitation number is usually taken equi-spaced grid. Normally pressure pick ups at a point approximating the centre of the ex- can be located at 0.8D ahead of the propeller pected cavitation extent in the upper part of the disc to 0.6D behind the propeller disc (D = pro- disk, such as 0.8.R above the propeller centre- peller diameter). The maximum extension to line. port and to starboard amounts to about 0.6D to 0.8D, depending on the section shape. The dis- Inclusion of the effect of stern wave heights tances between the transducers are in the range can be determined based on discussions with of 0.15D and 0.35D. customers and/or experience of the model basin.

For Froude scaled cavitation testing in a fa- 2.2 Test Conditions cility with a free surface, such as a depressurized In a variable pressure water tunnel facility, towing tank or a free surface circulating water the model test conditions should satisfy the same channel, the standard results of a Froude scaled propeller working conditions as predicted for towing basin powering test may be used directly the full scale ships. to set the propeller RPM and speed for the vari- ous operating conditions of the experiment. It is The two basic parameters of propeller work- noted that the usual procedure for scaling model ing conditions are: powering results to full scale is based on satisfy- ing the thrust loading coefficient at full scale  Propeller loading Reynolds number, which is equivalent to a  Corresponding pressure field thrust identity approach. It is recommended to perform additional 2.2.1 Propeller Loading Condition tests at different drafts with off-design loading conditions. Satisfy the propeller loading through the kin- ematical condition for J = VA/(nD) in order to 2.3 Measurements and Instrumentation achieve the predicted full scale KT or KQ (thrust or torque identity). Here means VA = propeller The requirements for measurements and in- speed of advance, D = propeller diameter (m), n 2 4 strumentation for model pressure fluctuation = rotational speed (1/s), KT = T/(n D ), and KQ testing can be divided into two main groupings. 2 5 = Q/(n D ). Usual practice in water tunnel test- The following lists identify the quantities meas- ing is to satisfy the thrust identity, although there ured items and give special notes about the in- are circumstances where the torque identity ap- strumentation [in brackets]. proach is used. ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 6 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

(a) Basic Test Measurements  photographic and video records; strobo- scopic lighting; and time series and nar- Measurements ‘Absolutely required’ in- row band spectra of pressure signals clude: [spectrum analyzer machine].

 facility flow velocity; In the category of ‘worthwhile to have’ are  facility static pressure;  propeller thrust and torque;  measurements of vibration acceleration  propeller rotational speed; [accelerometers placed near the pressure  water temperature, transducers];  air content as % saturation or % oxygen  sound pressure level of noise [hydro- saturation [use Van Slyke Apparatus or phones]; Continuous Oxygen Analyzer],  time series and narrow band spectra for  wake velocity component(s) [most prefer- the accelerations and noise. ably use 5-hole pitot probes or Laser Ve- locimetry]. 2.4 Calibration

In the category of ‘worthwhile to have’ is For a successful hull-pressure measurement, the cavitation pattern on the propeller blade  the measurement of cavitation nuclei must be simulated properly. As part of the prep- number and size distributions [use a cavi- aration and set-up of the test, the following cali- tation susceptibility meter or Cavitation bration should be performed: Nuclei Counter device].  For the thrust and torque dynamometer, (b) Unsteady Pressure and Cavitation Observa- load response calibrations should be car- tion Measurements ried out with applied loads, and also long term stability of the calibrated data needs Measurements ‘Absolutely required’ in- to be confirmed. clude:  Thrust and torque of the bare hub are to be measured and corrected for each test con-  unsteady pressure signals [ use strain dition. Especially at the low revolution gauge diaphragm or piezoelectric type condition, the effects of internal friction transducers]; and gearing in the dynamometer are rec-  control pulses per shaft rotation for data ommended to be checked as the shaft sampling [shaft encoder device with min- RPM approaches zero. imum number of pulses per rotation =  The torsional or lateral vibrations of the 5*(highest BR harmonic)*(Z)]; model propeller shaft may have an influ-  data stream system [signal condition- ence on the steadiness of the cavitation on ers/amplifiers, A-to-D devices, link to the blades and the level of the pressure computer and data storage]; fluctuation. Attention should be paid to  viewing and photographic arrangements the vibration level of the shaft at each test [windows, viewing pods or ports]; condition.  Pressure gauges to measure static and differential pressure should be calibrated ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 7 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

within an established time period prior to physics of cavitation itself. This means that the the test. broad band level cannot be estimated in the  In order to enhance the reliability of the blade angular position but in the time domain. measurements, it is recommended that the calibration data of the pressure gages be 2.5.1 Global Data Collecting and Presenta- confirmed, simply by the change of the tion of Results static pressure inside tunnels before the measurement. The following types of data can be of interest to be presented in the report: 2.5 Data Collection and Analysis  Sequences of the unfiltered pressure sig- We must distinguish two types of analysis : nal p(t). A sufficient number of propeller revolutions have to be covered to yield a 1. harmonic analysis or blade angular posi- realistic impression of variations over tion domain analysis time, as intermittency etc. To get an over- 2. time domain analysis view of the characteristic of the pressure signal the frequency content of the signal The harmonic analysis is dedicated to pro- should be at least up to 10 blade frequency vide the pressure amplitude at blade rate or blade multiples but 20 or higher is preferable. rate frequency. This analysis has to deal with The amplitude scale has to be dimension- pressure as function of blade angular position less as Kp (definition see section 3.1) or in Pascal. The plot of p(t) can be very in- changing with time : p   p t. To get rid formative and should always be included of the potential fluctuation of the shaft revolu- in the report. As an important complement tion rate that might come to be a problem when at least one type of the spectra listed below focusing at the high harmonics, it is always bet- should be appended. ter to use an encoder on the shaft to sample the pressure as a function of the blade angular posi-  A table showing amplitudes and phase an- tion. It is helpful to use an encoder with a num- gles (mean values) for the first 5 harmonics. This is in principle a pure line spec- N  2 p  ber of pulses which a power of 2 , so trum (Kp or amplitude in Pa). that the Fast Fourier Transform might be used to  A plotted mean value spectrum, including speed up the calculation of the harmonic decom- the continuous part, for the first 5 harmon- position of the signal. ics (Kp or amplitude in Pa).  A plotted spectrum showing the mean val- If the shaft revolution rate is enough stable, ues of the X% highest blade passage the harmonic decomposition can be done in the pulses, including the continuous part, for time domain. the first 5 harmonics. X can be for example 50 or any value which according to full On the contrary the blade angular position scale correlation has been proven to corre- domain analysis is not anymore useful when spond to full scale data. The idea here is to looking at the broadband level of the pressure exclude the smallest pulses which in pulse signal. Because the broadband level is model testing may be a result of exces- generally related to the non-stationary cavitation, sively intermittent cavitation because of the fluctuations are not only related to the blade lack of nuclei or other scale effects. Other angular position of the blade but also to the ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 8 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

types of pulse statistics can give equiva- limited and no standard or praxis has been de- lent information. Sequences should be veloped. sampled during several minutes. Good complements are also video recordings In a standard test for pressure fluctuations showing variations over some time. the first two or three, possibly up to five harmon-  A plotted mean value spectrum, including ics are usually considered. The limit is set by the the continuous part, up to 15th or 20th har- number of harmonics that can be clearly distin- monics. Because of its relevance for guished above the continuous spectrum and be higher frequencies excitation and the predicted by some accuracy. Full scale correla- dominance of the continuous spectrum at tions usually indicate that the discrepancy be- these frequencies this spectrum can be tween model and full scale increases at higher presented in 1/3-octave bandwidths, harmonics. The first two are typically reasona- which simplifies the scaling (Kp or ampli- bly close to full scale while the values of the tude in Pa). higher harmonics are sometimes too low by  Additionally a plot showing the arrange- model tests. ment of the different pressure transducers should be given. The number of harmonics is strongly related to the shape of the pulses. Two main behaviours The presentation of continuous spectra in the can be distinguished. For sheet cavitation on the same graph as discrete spectral lines is problem- blade the pulse shape is related to the gradient of atic in the sense that these spectra have to be the wake - the higher the gradient of the wake scaled to full scale in slightly different ways. the more harmonics. For tip vortex cavitation Only for spectra of the constant percentage type, the higher harmonics are typically generated as the 1/3 octave, the same scaling can be ap- during pulsations of the cavity behind the blade, plied to lines as well as the continuous part. This due to mechanisms not fully known. problem can be managed in different ways: Two main reasons for too low values pre- Avoid plotting of mixed spectra in the same dicted for higher harmonics can be identified: graph, recalculate the scaled and modified spec- Spreading of energy from the harmonic spectral trum or write out the Pa or Kp scaled amplitudes lines by phase modulation or a less steep wake at the spectral lines only. Since the continuous in model scale. spectrum contains information about the cavitation as well as the energy exciting the ship it is Excessive modulation can be avoided by us- important to include it in the report. Above the th th ing a high Reynolds number, high nuclei content 5 to 10 harmonics the continuous part can be and by minimizing the unsteady wake by use of dominating. flow liners to avoid separation.

2.5.2 Higher Harmonic Components A fundamental lack of high harmonics due to wrong pulse shape can be cured by simulating The present meaning of higher harmonics the proper wake as much as possible. covers a frequency range from 5 up to 10 or possibly 20 multiples of the blade frequency. In When excessive modulation exists its effect practice these frequencies cause high frequency can be reduced by measuring/analyzing in wider vibrations and what on passenger ships would be bandwidths, or afterwards by integration of the called low or medium frequency noise. The ex- spectrum in some interval around the harmonics. perience of prediction at these frequencies is There is no standard for this type of analysis, and ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 9 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

therefore the analysis is strongly related to the sampling frequency and duration, pulse and correlation experience of the model basin. phase deformation due to filters, amplifiers etc. Errors from these sources are reduced simply by From the 5th harmonic it typically happens giving priority to a professional selection and that the energy of the continuous part of the operation of modern measuring systems. The spectrum dominates over the energy in the lines unavoidable errors from the measurement chain so obviously the continuous spectrum has to be have then to be added to the errors emanating considered. from the hydrodynamic approximations. Exam- ples of the latter are: Considering these facts a practical way is to analyze the first spectral lines, having levels 5-  Error in the specification of the loading 10 dB above the continuous spectrum, in the tra- condition (cavitation number and advance ditional way by a line spectrum. The entire spec- coefficient). The source of this error is the trum can however also be shown as a 1/3-octave propulsion test or an equivalent for the de- band plot (scaled according to the Kp -formula) termination of the loading condition. being particularly adequate at frequencies at  Error in the realization of the loading con- which the continuous spectrum dominates. This dition at the cavitation test. In this error are data presentation also corresponds to the mode included deviations in controlling the tun- characteristics of the hull at higher frequencies. nel setting (to properly fulfill KT-identity etc.). Also possible effects of use of ap- 2.5.3 Uncertainty Consideration proximate similarity conditions are included in this group of errors. Ignoring the Measurement of pressure pulses requires Froude number effect on the cavitation skill and insight into the physics of cavitation number (often considered to be small). but it is a straightforward process not resulting  Deviations in hull or dummy geometry in any extraordinary difficulties. Usually the  Deviations in propeller geometry main error sources appear because of hydrodynamic problems introduced by the approxima- The importance of these error sources can tions made in a model test. The hydrodynamic vary, not only between different facilities but problems result in lack of similarity between also between different projects. If the wake is model and full scale cavitation and pressure sig- extremely bad a small and local error in blade nals, a fact implying that analysis and interpre- geometry of a heavily loaded propeller will tation of model results become complex and can make a small difference while the same error on result in errors difficult to quantify. a high speed ship can be significant. It is very important therefore to analyze the error sources Adding up all errors in the most pessimistic individually, in every project. way in such a complex process may result in error estimates larger than the measured quantity. An engineering way to handle the hydrody- An estimate based on experience and full scale namically based errors which are often difficult correlation can be less conservative but usually to derive or estimate, is to consider key input more realistic. data, loading conditions etc., not as exact numbers but the nominal numbers +/- 5 or 10%, as a Obviously all sources of error have to be es- guess. Performing then the tests also at +/- 5 or timated and weighed in some way. Among the 10% variation of the cavitation number and/or standard errors related to instrumentation are advance coefficient will produce a plot from those emanating from transducers, selections of ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 10 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

which the sensitivity of the results for input er- established experience. To enhance the con- rors can be estimated. With later assumptions sistency of measurement results, it is recom- about the input error also the output error can be mended that the tensile strength of the water in estimated and the risk that a certain design fails the facility should be checked periodically. due to a realistic but unknown error can be eval- uated. 2.6.2 Influence of Wake Simulation

2.6 Other Essential Experimental Condi- One of the most critical problems when per- tions forming pressure fluctuation measurements is the correct simulation of the wake of the full This section deals with a lot of items that scale ship. When performing such type of meas- need to be taken into account when performing urements the following limitations regarding the pressure fluctuation measurements but cannot wake simulation shall be always kept in mind. be put into a direct rule neither specific values can be given. The points mentioned should be  Full ship models provide three-dimen- considered when discussing the results, depend- sional inflow at the propeller discs which ing on the experience with the specific test facil- is correctly simulating the full scale pro- ity and the correlation achieved there. peller inflow for twin-screw hulls pro- vided that the propeller discs are outside 2.6.1 Air Contents, Cavitation Nuclei and the ship boundary layer. In case of a sin- Stabilizing Model Cavitation gle-screw hull, where the propeller is working in the ship boundary layer, the It is generally accepted that testing at rela- large difference of the model and ship tively high air content, as a kind of nuclei, in a Reynolds numbers produces a wider wake water tunnel facility reduces the tensile strength and a different flow to the propeller in and improves the correlation of model and full model scale. scale results. When there are insufficient con-  Shortened full ship models proposed to re- centrations of nuclei, all forms of cavitation be- duce the differences between the bound- have intermittently and will therefore produce ary layer thickness between model and full non-periodic pressure readings at model scale. scale ship can be adopted but particular attention shall be applied when shortening When testing in a depressurized towing tank, the model in order to avoid possible sepa- the generation of the nuclei by the sand grain ration of the flow at the fore-body that will roughness on the leading edges of the model definitively deteriorate the propeller in- propeller blades or electrolysis in the boundary flow simulation. Prior to testing, oil paint layer flow past the hull stabilizes the cavitation or tuft visualization is suggested for veri- on the model propeller blade. fication. However, too high levels of air may deterio-  Wire screens are a suitable and practical rate the visibility inside the tunnel and introduce means to reproduce the full scale ship a damping effect on the measured unsteady pres- wake when the axial velocity distribution sure amplitudes. is the objective of the simulation. They are not effective in simulating the tangential Hence the optimum air content for a given and the radial velocity distribution. cavitation facility should be determined by long-  Dummy after-body model can be used to simulate the full scale propeller inflow, ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 11 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

but it requires a time consuming iterative high frequencies strong influence has been re- process of model modifications and wake ported from local plate vibrations. measurements to achieve the target propeller inflow. It is usually assumed that the influence of vibrations is small at model testing but can be im- 2.6.3 Facility Wall Effects portant on the ship having a more flexible struc- ture. If this occurs and no correction is made the The wall effect has been a subject of contin- comparison between the amplitudes at model uous discussion in the ITTC. The Round-Robin and full scale can be of limited relevance. tests in the 22nd ITTC report show that different levels of pressure fluctuations have been meas- Calibration procedures based on measure- ured in different test section areas in spite of ments of the vibrations at transducer positions similarity of wake field and cavity pattern. How- have been developed and tested. The experience ever, the effect of the blockage on the pressure from these problems is however limited and spe- measurement is not clarified quantitatively be- cific procedures are not engineering standard. cause the measured results include the effects of To avoid the problems the following simple different sources like tunnel vibration, Reynolds guidelines are recommended: numbers etc.

Hence, systematic studies on this effect will  Use a model/dummy being stiff enough to be needed, and it is recommended that each fa- reduce vibrations at relevant frequencies. cility gains experience by comparing at least the This is particularly important for at least results for two different sized propellers. one reference transducer in the centre line and close to the propeller.  If the model cannot be made stiff enough, 2.6.4 Effect of Induced Vibration avoid putting transducers far from the propeller or within or close to areas where vi- If a pressure transducer is mounted in a vibration can be expected. brating surface a pressure can be induced by the vibration. This pressure adds, with the compo-  Check vibration levels, at least for some nent of its phase angle, to the pressure that typical models, by mounting accelerome- would have been measured without the vibration. ters close to pressure transducers. Then The vibration can be excited by the propeller or apply an exciter to the model (without the by uncorrelated sources. They can be global, in- propeller in operation) so it performs vi- cluding the complete hull girder, or local plate brations similar to those generated by the vibrations. propeller. From such experiments a correction can be found. In principle the cor- Although the problem has been much dis- rection can alternatively be calculated by cussed there are only a few studies published methods being able to determine the added and according to a few full scale examples the mass of a body vibrating in water. How- influence of vibrations at lower frequencies ever, particularly for a model more or less seem for most cases and transducer positions to clamped to a cavitation tunnel this latter be modest. This, however, does not mean that method can be difficult. Take account of severe influence cannot occur, for example at phase relations and bandwidth effect in the transducer positions far from the propeller. At measuring system. ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 12 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

A different kind of influence occurs if the to be transformed by some procedure to include transducer is mounted at some distance from the the free surface effect. vibrating surface or part of the surface. An example of this case is vibration of the walls of a If not free surface and solid boundary effects cavitation tunnel which can disturb measure- for the correct shape of the hull are automati- ments significantly. The way around the prob- cally taken into account by use of a proper lem is to reduce the wall vibrations and avoid model/dummy tested in a free surface facility transducers far from the propeller and close to corrections have to be applied afterwards to find the walls. the relevant pressure amplitude. The correction factor can be empirical or computed as ex- Above resonant vibration of the tunnel walls plained in the 23rd ITTC Report on Cavitation there is also a contribution from acoustic reflec- Induced Pressure Fluctuations. tions in mainly stiff walls. At some distance from the propeller the free field will be disturbed For a correct solid boundary factor the by the reflections. The distance at which this oc- model/dummy has to be reasonably similar to curs depends on the size and stiffness of the test the ship and the transducer has to be in approxi- section and the amount of free gas bubbles in the mately the position in which the pressure ampli- water. This effect also limits the distance from tude is intended to be predicted. The reason is the propeller at which a transducer can record a that the solid boundary factor changes slightly relevant pressure signal. with position (due to hull shape) but also that the propeller has some directivity implying that a To find out such limitations an acoustical distance correction based on 1/r should be calibration of the tunnel is recommended. The avoided (This is particularly true for minor cav- calibration should, due to the influence of stand- itation, a condition at which the dipole contribu- ing waves, be made at realistic free gas content. tion from the blade loading cannot be neglected). In this way transfer functions can be determined by which for example the free field pressure can The correction for the absence of a free sur- be estimated. Particular for the higher harmonics face and possible shortcomings in the solid this procedure can be useful. boundary effect can be determined numerically. It is noted that empirical corrections based

2.6.5 Free Surface Effects on full scale correlation take into account also It is concluded that the influence of the free different scale effects on the cavitation, a reason surface on the pressure amplitudes has to be ac- to work with such correlations also if a numeri- counted for in most tests. Exceptions being am- cal free surface correction is applied. plitudes measured at transducers close to the propeller on fully loaded super tankers and sim- 3. PARAMETERS ilar.

The highest accuracy in simulation of the 3.1 Parameters to be Taken into Account boundary conditions, like the one at the free surface, can be reached in a very large cavitation Parameters that need to be considered during facility with a free surface. Data from all other pressure fluctuation measurements are basically facilities, i.e. from most cavitation tunnels, have the same as for cavitation tests (ITTC Procedure 7.5-02-03-03.1). For pressure fluctuation meas- ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 13 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

urements they can be categorized into "data ab- The following checklist shall help the test solutely necessary" and in "data worthwhile to engineers to settle all data: have" (section 2.4). If the latter are considered, the reliability and the quality of the measurements will considerably be improved.

Absolutely required Worthwhile to have

General infor-  Type of ship  Ship main particulars mation  Engine power and RPM  Propeller geometry data (Section, (Ship, propeller  Propeller main particulars Pitch Chord distribution etc) operating con-  Shaft immersion  Propeller design conditions ditions)  Tip clearance  Drawing of stern shape including arrangement of appendages Model propel-  Propeller model material  Detailed inspection of blade ge- ler operating  Flow velocity including wake distri- ometry conditions butions  Intrinsic unsteadiness of facility  Static pressure  Pressure drop through test section  Propeller thrust and torque  Level of turbulence upstream pro-  Propeller RPM peller Water quality  Water temperature  Tensile strength of the water  Air contents as % saturation or % ox-  Nuclei distribution number and ygen saturation size Instrumentation  Type and capacity of pressure trans-  Type and capacity of hydrophone ducer  Type and capacity of accelerome-  Type of amplifiers ter  Type of shaft encoder  Vibration characteristics of the measuring plate Measurement  Interval for signal acquisition  Acceleration measurement next to and analysis  Measuring time pressure transducers  Level of the pressure fluctuation at  Noise level near the propeller blade frequencies  Narrow band spectra of pressure, noise and accelerations  Recording all data stream

ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 14 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

BASIC MEASURED DATA DERIVED PARAMETERS

Representative static pres- PTunnel  PV  n,M,irtl,KE = . sure at propeller radius p Cavitation number  Tunnel 2 2 nM DM 0.8 2 Q Rotational velocity rps n Torque coefficient KQ =  n2 D5 T Tunnel speed m/s VTunnel Thrust coefficient KT =  n2 D4

Advance VA Propeller thrust N T J A = coefficient n D pV Propeller torque Nm Q Vapour pressure

Dimensionless PM(  ) = . Water temperature ° C t K P(  ) 2 2 pressure amplitude  Tunnel nM DM Air Content / Oxygen /S Phase angle  Content

ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 15 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

3.2 Recommendations of ITTC for Parameters

RECOMMENDED COMMENTS / PARAMETER VALUES RECOMMENDED IN Pressure adjustment to 0.7 ~ 0.9 R For wire screen, blockage is for propeller disk area. For Less than 20 % of test section Blockage dummy hull or full hull, size blockage is the fullest section of the hull. Number of revolutions of As high as possible in accord- ITTC 1996 Cav. Com. model propeller ance with tunnel speed Minimum value of 0.5 million Minimum Reynolds-number based on the blade chord length at 0.7 R Number of pressure 5 ~ 20 transducers As high as possible according to the facility ex- Air content / nuclei perience. ITTC 1984 Distribution Values of total air content ITTC 1996 Cav. Com. or Oxygen content should be mentioned Low values of the Noise ITTC 1990 Cav. Com. facilities At least two different rotation Reproducibility rates of the model propeller ITTC 1993 Cav. Com. should be tested Model propeller diameter > 200 mm

experiment. Customers should be informed of the uncertainty assessment methodology used 4. VALIDATION and which uncertainties can be expected for the tests (see also section 2.5.3). The uncertainty as- 4.1 Uncertainty Analysis sessment methodology should inform about

The 20th ITTC (1993) mentioned critical is- a) measurement systems sues concerning scale effects in cavitation test- b) error sources considered. ing. They were related to fluid effects (wake) c) all estimates for bias and precision limits and bubble dynamic effects. These must be and the methods used in their estimation taken into account when estimating errors of an ITTC – Recommended 7.5-02 -03-03.3 Procedures and Guidelines Page 16 of 16 Cavitation Induced Pressure Fluctua- Effective Date Revision tions Model Scale Experiments 2014 05

(e.g., manufacturers specifications, compar- Comparative Noise Measurement with Sydney isons against standards, experience, etc.). Express Propeller Model d) actual data uncertainty estimates. (17th 1984 pp.255-256)

The uncertainty analysis should be done in Comparison of Propeller-Induced Hull Pressure accordance with the following regulations/rec- Measurements for the "SYDNEY EX- ommendations: PRESS" Propeller Models ISO, 1992, “Measurement Uncertainty,” (18th 1987 pp. 209-210) ISO/TC 69/SC 6. ISO, 1993a, “Guide to the Expression of Uncer- Comparative Noise Measurements with "SYD- th tainty in Measurement,” ISO, First edition, NEY EXPRESS" Propeller Models (18 ISBN 92-67-10188-9. 1987 pp. 210-211) ISO, 1993b, “International Vocabulary of Basic Propeller-Induced Hull Pressures and General Terms in Metrology,” ISO, Second (19th 1990 pp.182-187) edition, ISBN 92-67-01075-1. ITTC, 1990, ”Report of the Panel of Validation Further Measurement of Pressure Fluctuation on th 'SYDNEY EXPRESS' Propeller Procedures”, 19 International Towing Tank th Conference, Madrid, Spain, Proc. Vol. 1, pp. (19 1990 pp.213-219) 577-603. Comparative Measurements on German Tanker "St. Michaelis" and the "Sydney Express" 4.2 Benchmark Tests (20th 1993 pp. 230-231)

ITTC. Standard Screw Cavitation Tunnel Tests Comparative Measurement of Pressure Fluctua- at Brodarski Institute (12th l969 pp.523-525) tion on the "St Michaelis" (20th 1993 pp. 236-240) Comparative Noise Measurements with the Sydney Express Propeller Model Measurements of Hull Pressure Fluctuation (21st (16th 1981 pp.447-453) 1996 pp. 65-69)

Comparison of Hull Pressure Amplitudes for Measurement of Hull Pressure Fluctuation, Sydney Express Propeller Round Robin Tests (22nd 1999, pp. 547-585) (17th 1984 pp.248-252)