319 



Discussion 



ORVAR BJORHEDEN and TORE DALVAG 



We congratulate the authors of this very 

 interesting paper. For hull designers as well as 

 propeller manufactures the problem of predicting 

 the propeller induced vibration forces is a most 

 essential task indeed. In this context we wish to 

 inform you briefly about some recent developments 

 at the KMW* Marine Laboratory related to the model 

 testing technique applied in our cavitation tunnels. 



The first item concerns the method of hull 

 wake simulation. For some time the well-known 

 dummy technique, involving ship afterbody models 

 and transverse net screens, has been used in our 

 tunnels for the purpose of simulating model wake 

 pattern. This is a rather time consuming process 

 since the net screens have to be adjusted step by 

 step until the correlation with the wake pattern 

 obtained in the towing tank appears satisfactory. 

 Moreover, the method has some technical drawbacks 

 as regards the stability of the wake as well as 

 the interaction between propeller and hull and the 

 influence of the propeller on the wake pattern. 

 In connection with hydro-acoustic tests , cavitation 

 occurring on the nets may worsen the background 

 noise level. 



In order to eliminate the above drawbacks a 

 new technique involving longer afterbody hull 

 dummies has been introduced. The method aims at 

 simulating the full-scale ship wake pattern based 



upon the concept of equivalent relative boundary 

 layer thickness, i.e., the fractional boundary 

 layer thickness in relation to some characteristic 

 length, e.g., the propeller diameter should be the 

 same in the model and in full-scale. For ordinary 

 cavitation testing purposes utilizing propeller 

 model diameters around 250 mm and tunnels speeds 

 of 4 to 8 m/sec this criterion results in hull 

 dummy lengths of 2.5 to 3.5 m for most types of 

 vessels. In principal, the model stern contour as 

 well as the aftermost water-lines are made to scale, 

 whilst the maximum breadth of the dummy is chosen 

 on the basis of 2-dimensional potential flow cal- 

 culations comparing the ship water-lines in unre- 

 stricted water to the dummy lines within the bound- 

 aries of the cavitation tunnel test section and 

 aiming at similarity in the potential wake 

 dis tr ibut ion . 



Figure 1 shows a picture of a 3 m hull dummy 

 used for the testing of a 150 m, single screw, con- 

 tainer ship. In Figures 2 and 3 the model wake 

 distribution as obtained in the towing tank and 

 then corrected for scale effect according to the 

 so-called Sasajima method is given. In Figure 4, 

 finally, a comparison between the corrected model 

 wake and the wake distribution obtained in the cav- 

 itation tunnel is shown for a few radii close to 

 the propeller blade tip. As can be seen from the 

 diagrams , the agreement is quite good , particularly 

 as regards the wake peak in the 12 o'clock propel- 

 ler blade position. 



*Karlstads Mekaniska Werkstad 



Figure 1. Hull dummy for wake simulation m cavita- 

 tion tunnel. 



Apart from the advantage of a quicker and more 

 direct simulation of the full-scale wake, the 

 method with long afterbody dummies results in a far 

 more stable wake distribution which in turn implies 

 more consistent recordings of fluctuating propeller 

 forces, propeller induced pressure pulses against 

 the hull, etc. Probably, the interaction between 

 propeller and hull is also more realistic with this 

 method of wake simulation as compared to the method 

 utilizing transverse nets. 



The second item refers to the instrumentation 

 employed for recording of propeller forces and the 

 propeller induced pressure pulses on a ship's hull. 

 In both KMW tunnels a data collecting and evaluation 

 system consisting of an on-line connected desk com- 

 puter together with a printer and a plotter has 

 been used for several years. For the measurement 

 of propeller induced pressure pulses with the aid 

 of pressure pickups fitted into the hull, a pulse 

 sampling technique giving time averaged values from 

 a number of propeller revolutions at each blade 

 position has been the practice. With this method 

 the pressure signals are given in analogue form and 

 recordings can be obtained from only one pickup at 

 a time. Recently, a new data collecting unit was 

 put into service enabling simultaneous recording 

 on 6 channels and storing test results from every 



