Efficiencies for Surface drives.

[sandhammaren05]

Hi Joe, I'm a bit surprised that there is such a discrepancy between your observations and the figures in those two reports. Could you give an example from your scene so that we can find out where the problem lies?

One side, where I would take a deep look is the dissolved air content in the various waters, since the bubble expansion (when microbubbles come out of solution) before the final expansion, is not linear with size or speed. This means that there is an "incubation period" that will vary with speed.

This is seen in cavitation in closed systems, where a velocity increase (say doubling the rpms of a pump) will not result in the expected loss of cavitation performance in terms of the NPSH required.

Hi Bodo, nice to hear from you. How about I simply email you the section from my book ms chapter that contains the example calculations? I think the problem is simple: all the available data have been taken at U≈10 m/s or less in a tank (including Hadler-Heckler but maybe they used a boat), limit of zero Froude nr.. Boats run more in the infinite Froude nr. limit. It's unrealistic to think that water flowing at a few m/s past a rotating prop simulates surface-piercing at high Froude nr. on a boat. If you want to read and comment/criticize, then I can send you the whole chapter on props. Best, Joe

 

[baeckmo]

Well Joe, I’m afraid you just shot yourself in your foot. You have missed something essential regarding the ventilated propeller; the working area is a variable. On top of that you have not considered the impact of the torque coefficient Kq, which is essential for the estimation of the effective working area.

The examples you refer to are, at best anecdotal; whenever I see a figure rounded off to double zeroes or to a fiver, I am sure that the number is at best a guesstimate. Your rpms are given as “8000 rpms”, speeds are given as 65 or 110 mph, there is nothing said about operating depth, propeller foil profile, blade profile, blade area ratio or the profile and depth of the steering fin in front of the propeller. The power figures are very uncertain, do you have dyno data to show?

Now, that said, let us check two of your examples, knowing that the basic info is vague. For the purists that may read this, I want to say that I am very well aware that I am stretching the extrapolations a bit beyond recommendations; it is for the good sake of demonstrating the principle. I have selected reference props from other sources than you used, in order to have a wider spectrum.

First the SST45 case. Using your data we have the following:
P: 40,5 kW @ 8000 rpm (133,3 rps)
V: 65 mph (29 m/s)
D: 8,4” (0,213 m)
P: 10” (0,254 m)
Advance coefficient (lambda) 1,024 (=V/(D*n);

This gives the following:

P/D: 1,19 and the torque constant Kq = 0,0062. With this P/D value, we check available propeller data.

A/ First I looked into the figures given by Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973 The report is highly recommended!). Here a 3-bladed Newton Rader propeller (without cup) with P/D = 1,33 is tested. The BAR is ~0,7. This means that you may use the Kt and Kq values for this prop if you check its graph at lambda 1,33/1,19 x 1,024; ie 1,144 for rough estimates.

At depth of 50 % the N-R propeller shows Kq=0,018 and Kt=0,07. As has been shown, f.i. by Ferando et al, the ventilated propeller is operating with reasonably constant ratio Kt/Kq at varying submergences. Hence we may (for a rough estimate) use a linear relation between the observed Kq (0,0062) and the tank value (0,018). From this we get a correction factor of 0,344, which we use to find a new operating submersion.

The relative area at 50 % submergence is 50 % (surprise surprise…!), and the needed relative area is then 17,2 %. Now if you calculate the effective disc area as a function of operating depth, this will correspond to a submergence of 23,4 %. The SST propeller blades are thus operating at a maximum depth of 50 mm. The thrust coefficient is then 0,024, giving T= 880 N and a power of 40,5 kW.

B/ Next we check the Rolla propeller tested by Rose and Kruppa (Published FAST 91). This series is also based on the Newton-Rader propeller, but with a cup. Here we use the diagram no 15 (30 % submergence, atmospheric ventilation, shaft angle 4 degr). From the lambda 1,02 (right side) you go to the line showing lambda asf of Kt/J^2 for the P/D propeller. You will find a Kt/J^2 value of 0,04. From this value, you go vertically to find the eta value ~0,61, which allows us to calculate the “chart power” (=Trust*speed/eta) to 72,4 kW, and Kq = 0,011.

With the same correction procedure as before, we get a correction factor =0,559. This time it's applied to a 30% submerged prop (relative area is 25,24 %), we arrive at an operating relative area of 14,1 %, which we get with a 20,3 % submersion, or 43 mm in this case. The difference to the original N-R is explained by the cup and a higher BAR (=0.80) on the Rolla prop. The thrust coefficient is 0,023 and the resulting thrust 859 N.

C/ Now we check the 4-bladed supercavitating propeller type 841-B (origin KaMeWa if I remember ok), used by Niclas Olofsson in his thesis (Chalmers uni 1996, Gothenburg). Foil profile is a modified cambered wedge (“Tulin” style) and blade shape is more like the classic cleaver, with an area ratio of 0,58 and P/D is 1,24. Shaft angle is zero degrees. At 30 % submergence it has Kt = 0,035 and Kq = 0,0091. Efficiency is ~67 %. Repeat the correction procedure and you end up with trust = 868 N and submergence 23,3 % or 50 mm.

D/ Check with data from Szantyr (FAST 97). Supercavitating propeller (modified wedge), gives nearly identical results as “C” above.

Ergo: Four different ventilated propellers studied for the same task, thrust results 850-880 N and operating depth mean value 48 mm. Differences to a great extent explained by differences in blade profile and blade area ratio.

When I go through the same exercise for the F1 Sports, using the Rose/Kruppa data and a 3-bladed supercavitation propeller (H. Ghassemi and M. Ghiasi, Amirkabir tech uni, Tehran), I get thrust 2258 and 2332 N and a submergence of ~17,6 %; in this case 45 mm.

In all cases studied the nominal submergence is slightly less than you will see in reality, since the nominal value is “hiding” the ventilated wake after the steering fin. I would guess that the gearcases in the two examples will have diameters of, say 60 and 80 mm respectively, thus clearing the waterline with a slight margin.

From the examples above, I must say that the various tank data are surprisingly consistent, considering the lack of precision in the observations from your examples. There is a lot more to comment upon regarding the influence of design variables, but i think I made my point. In real life there is also the influence from dissolved air in the water. It normally varies over the day, with a maximum between 14 and 15 o’clock and a minimum in the early morning. In the afternoon, the water is thus supersaturated with air, that will come out of solution very easily and will then cause a loss of performance of propellers and hydrofoils.

I will be back later with comments on the shaft inclination subject. I certainly do not agree with your previous statements on the quality of test data here; you really must know how to use the tools available.

 

[sandhammaren05]

Well Joe, I’m afraid you just shot yourself in your foot. You have missed something essential regarding the ventilated propeller; the working area is a variable. On top of that you have not considered the impact of the torque coefficient Kq, which is essential for the estimation of the effective working area.

The examples you refer to are, at best anecdotal; whenever I see a figure rounded off to double zeroes or to a fiver, I am sure that the number is at best a guesstimate. Your rpms are given as “8000 rpms”, speeds are given as 65 or 110 mph, there is nothing said about operating depth, propeller foil profile, blade profile, blade area ratio or the profile and depth of the steering fin in front of the propeller. The power figures are very uncertain, do you have dyno data to show?

Now, that said, let us check two of your examples, knowing that the basic info is vague. For the purists that may read this, I want to say that I am very well aware that I am stretching the extrapolations a bit beyond recommendations; it is for the good sake of demonstrating the principle. I have selected reference props from other sources than you used, in order to have a wider spectrum.

First the SST45 case. Using your data we have the following:
P: 40,5 kW @ 8000 rpm (133,3 rps)
V: 65 mph (29 m/s)
D: 8,4” (0,213 m)
P: 10” (0,254 m)
Advance coefficient (lambda) 1,024 (=V/(D*n);

This gives the following:

P/D: 1,19 and the torque constant Kq = 0,0062. With this P/D value, we check available propeller data.

A/ First I looked into the figures given by Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973 The report is highly recommended!). Here a 3-bladed Newton Rader propeller (without cup) with P/D = 1,33 is tested. The BAR is ~0,7. This means that you may use the Kt and Kq values for this prop if you check its graph at lambda 1,33/1,19 x 1,024; ie 1,144 for rough estimates.

At depth of 50 % the N-R propeller shows Kq=0,018 and Kt=0,07. As has been shown, f.i. by Ferando et al, the ventilated propeller is operating with reasonably constant ratio Kt/Kq at varying submergences. Hence we may (for a rough estimate) use a linear relation between the observed Kq (0,0062) and the tank value (0,018). From this we get a correction factor of 0,344, which we use to find a new operating submersion.

The relative area at 50 % submergence is 50 % (surprise surprise…!), and the needed relative area is then 17,2 %. Now if you calculate the effective disc area as a function of operating depth, this will correspond to a submergence of 23,4 %. The SST propeller blades are thus operating at a maximum depth of 50 mm. The thrust coefficient is then 0,024, giving T= 880 N and a power of 40,5 kW.

B/ Next we check the Rolla propeller tested by Rose and Kruppa (Published FAST 91). This series is also based on the Newton-Rader propeller, but with a cup. Here we use the diagram no 15 (30 % submergence, atmospheric ventilation, shaft angle 4 degr). From the lambda 1,02 (right side) you go to the line showing lambda asf of Kt/J^2 for the P/D propeller. You will find a Kt/J^2 value of 0,04. From this value, you go vertically to find the eta value ~0,61, which allows us to calculate the “chart power” (=Trust*speed/eta) to 72,4 kW, and Kq = 0,011.

With the same correction procedure as before, we get a correction factor =0,559. This time it's applied to a 30% submerged prop (relative area is 25,24 %), we arrive at an operating relative area of 14,1 %, which we get with a 20,3 % submersion, or 43 mm in this case. The difference to the original N-R is explained by the cup and a higher BAR (=0.80) on the Rolla prop. The thrust coefficient is 0,023 and the resulting thrust 859 N.

C/ Now we check the 4-bladed supercavitating propeller type 841-B (origin KaMeWa if I remember ok), used by Niclas Olofsson in his thesis (Chalmers uni 1996, Gothenburg). Foil profile is a modified cambered wedge (“Tulin” style) and blade shape is more like the classic cleaver, with an area ratio of 0,58 and P/D is 1,24. Shaft angle is zero degrees. At 30 % submergence it has Kt = 0,035 and Kq = 0,0091. Efficiency is ~67 %. Repeat the correction procedure and you end up with trust = 868 N and submergence 23,3 % or 50 mm.

D/ Check with data from Szantyr (FAST 97). Supercavitating propeller (modified wedge), gives nearly identical results as “C” above.

Ergo: Four different ventilated propellers studied for the same task, thrust results 850-880 N and operating depth mean value 48 mm. Differences to a great extent explained by differences in blade profile and blade area ratio.

When I go through the same exercise for the F1 Sports, using the Rose/Kruppa data and a 3-bladed supercavitation propeller (H. Ghassemi and M. Ghiasi, Amirkabir tech uni, Tehran), I get thrust 2258 and 2332 N and a submergence of ~17,6 %; in this case 45 mm.

In all cases studied the nominal submergence is slightly less than you will see in reality, since the nominal value is “hiding” the ventilated wake after the steering fin. I would guess that the gearcases in the two examples will have diameters of, say 60 and 80 mm respectively, thus clearing the waterline with a slight margin.

From the examples above, I must say that the various tank data are surprisingly consistent, considering the lack of precision in the observations from your examples. There is a lot more to comment upon regarding the influence of design variables, but i think I made my point. In real life there is also the influence from dissolved air in the water. It normally varies over the day, with a maximum between 14 and 15 o’clock and a minimum in the early morning. In the afternoon, the water is thus supersaturated with air, that will come out of solution very easily and will then cause a loss of performance of propellers and hydrofoils.

I will be back later with comments on the shaft inclination subject. I certainly do not agree with your previous statements on the quality of test data here; you really must know how to use the tools available.

Thanks, I'll see if I can unravel your assumptions above. I don't like fudge factors and there seem to be more than one in your reasoning. The RPMs are around 8000, as stated (they're not around 7000 nor are they around 9000), the speeds stated are the correct top speeds for a short course, and the shp is known from the manufacturers (actually, they do have dynos). Sorry, the Genoa data don't include the F1Sport case, p/D=1.6, J≈1.5. I've seen no way to extract a comparison from the Rolla data in your ref. of several days ago. High performance boats don't run with negative trim excepting sometimes with a tunnel into the wind. The trim angles of the propshaft relative to waterline are either zero or positive. These are not inboards with a driveshaft poking out at a negative trim angle. I guess that negative trim angles were studied for the usual reason in marine engineering research: naval contracts.

 

[sandhammaren05]

In order that we're in agreement what we're talking about we need to agree that the dimension L in the scaling relations (thrust, torque, advance coeff.) is always the measured prop diameter. Now, what are measured are thrust and torque. OMC did this well in the 1990s. They ran open water tests on a V-6 with a specially built gearcase so that the torque measurement device was inside the gearcase. The thrust was measured at the transom. The torque (N) and thrust (T) coefficients are obtained from T and N. With the scale fixed by L, those coefficients are then calculated as J is varied. The measured thrust and torque will depend on p/D, b.a.r., submerged fraction of the prop, F, and other control parameters that are of great interest for racing, and that will be reflected in the coefficients. So I don't see how, if T and N are properly measured, anything but simple direct calculation is required in order to estimate the shp required to do a given job, as DuCane originally saw and did. Having written this, I will certainly read your argument and respond later.

 

[sandhammaren05]

Bodo, here are a few preliminary comments/observations sandwiched in your text, if I may. Joe

Well Joe, I’m afraid you just shot yourself in your foot. You have missed something essential regarding the ventilated propeller; the working area is a variable. On top of that you have not considered the impact of the torque coefficient Kq, which is essential for the estimation of the effective working area.

Working area is reflected in the measured thrust and torque from which the coefficients are derived by arithmetic.

The examples you refer to are, at best anecdotal; whenever I see a figure rounded off to double zeroes or to a fiver, I am sure that the number is at best a guesstimate. Your rpms are given as “8000 rpms”, speeds are given as 65 or 110 mph, there is nothing said about operating depth, propeller foil profile, blade profile, blade area ratio or the profile and depth of the steering fin in front of the propeller. The power figures are very uncertain, do you have dyno data to show?

The RPMs are around 8000. They are not around 7000 or 9000. That is accurate enough for a scaling relation. The speeds of SST45 are 65 mph on a short clorse, 110 mph for F1Sport on a short course. The factors that you mention are reflected in the thrust and torque vs. J. They will be different for different prop and gearcase designs, but that can't be fudged into existing data!The power figures are from the motor manufactureres' dynos. Actually, the SST45 motor develops less than 60 hp (54 shp) at 8000 RPM, it's peak hp range is 5500-6000 RPM but that is not were we race it.

Now, that said, let us check two of your examples, knowing that the basic info is vague. For the purists that may read this, I want to say that I am very well aware that I am stretching the extrapolations a bit beyond recommendations; it is for the good sake of demonstrating the principle. I have selected reference props from other sources than you used, in order to have a wider spectrum.

Wider spectrum? Not at this point. You recommended and I used the Genoa data, so let's stick to that until you send me links to your other sources.

First the SST45 case. Using your data we have the following:
P: 40,5 kW @ 8000 rpm (133,3 rps)
V: 65 mph (29 m/s)
D: 8,4” (0,213 m)
P: 10” (0,254 m)
Advance coefficient (lambda) 1,024 (=V/(D*n);

This gives the following:

P/D: 1,19 and the torque constant Kq = 0,0062. With this P/D value, we check available propeller data.

But the smallest torque coefficient in the Genoa data is .015! Let's stick to that data, you recommended it.

A/ First I looked into the figures given by Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973 The report is highly recommended!). Here a 3-bladed Newton Rader propeller (without cup) with P/D = 1,33 is tested. The BAR is ~0,7. This means that you may use the Kt and Kq values for this prop if you check its graph at lambda 1,33/1,19 x 1,024; ie 1,144 for rough estimates.

You can't use p/D=1.33 for the case where it's 1.2. The Genoa data include the case 1.2. End of my comments for now.


At depth of 50 % the N-R propeller shows Kq=0,018 and Kt=0,07. As has been shown, f.i. by Ferando et al, the ventilated propeller is operating with reasonably constant ratio Kt/Kq at varying submergences. Hence we may (for a rough estimate) use a linear relation between the observed Kq (0,0062) and the tank value (0,018). From this we get a correction factor of 0,344, which we use to find a new operating submersion.

The relative area at 50 % submergence is 50 % (surprise surprise…!), and the needed relative area is then 17,2 %. Now if you calculate the effective disc area as a function of operating depth, this will correspond to a submergence of 23,4 %. The SST propeller blades are thus operating at a maximum depth of 50 mm. The thrust coefficient is then 0,024, giving T= 880 N and a power of 40,5 kW.

B/ Next we check the Rolla propeller tested by Rose and Kruppa (Published FAST 91). This series is also based on the Newton-Rader propeller, but with a cup. Here we use the diagram no 15 (30 % submergence, atmospheric ventilation, shaft angle 4 degr). From the lambda 1,02 (right side) you go to the line showing lambda asf of Kt/J^2 for the P/D propeller. You will find a Kt/J^2 value of 0,04. From this value, you go vertically to find the eta value ~0,61, which allows us to calculate the “chart power” (=Trust*speed/eta) to 72,4 kW, and Kq = 0,011.

With the same correction procedure as before, we get a correction factor =0,559. This time it's applied to a 30% submerged prop (relative area is 25,24 %), we arrive at an operating relative area of 14,1 %, which we get with a 20,3 % submersion, or 43 mm in this case. The difference to the original N-R is explained by the cup and a higher BAR (=0.80) on the Rolla prop. The thrust coefficient is 0,023 and the resulting thrust 859 N.

C/ Now we check the 4-bladed supercavitating propeller type 841-B (origin KaMeWa if I remember ok), used by Niclas Olofsson in his thesis (Chalmers uni 1996, Gothenburg). Foil profile is a modified cambered wedge (“Tulin” style) and blade shape is more like the classic cleaver, with an area ratio of 0,58 and P/D is 1,24. Shaft angle is zero degrees. At 30 % submergence it has Kt = 0,035 and Kq = 0,0091. Efficiency is ~67 %. Repeat the correction procedure and you end up with trust = 868 N and submergence 23,3 % or 50 mm.

D/ Check with data from Szantyr (FAST 97). Supercavitating propeller (modified wedge), gives nearly identical results as “C” above.

Ergo: Four different ventilated propellers studied for the same task, thrust results 850-880 N and operating depth mean value 48 mm. Differences to a great extent explained by differences in blade profile and blade area ratio.

When I go through the same exercise for the F1 Sports, using the Rose/Kruppa data and a 3-bladed supercavitation propeller (H. Ghassemi and M. Ghiasi, Amirkabir tech uni, Tehran), I get thrust 2258 and 2332 N and a submergence of ~17,6 %; in this case 45 mm.

In all cases studied the nominal submergence is slightly less than you will see in reality, since the nominal value is “hiding” the ventilated wake after the steering fin. I would guess that the gearcases in the two examples will have diameters of, say 60 and 80 mm respectively, thus clearing the waterline with a slight margin.

From the examples above, I must say that the various tank data are surprisingly consistent, considering the lack of precision in the observations from your examples. There is a lot more to comment upon regarding the influence of design variables, but i think I made my point. In real life there is also the influence from dissolved air in the water. It normally varies over the day, with a maximum between 14 and 15 o’clock and a minimum in the early morning. In the afternoon, the water is thus supersaturated with air, that will come out of solution very easily and will then cause a loss of performance of propellers and hydrofoils.

I will be back later with comments on the shaft inclination subject. I certainly do not agree with your previous statements on the quality of test data here; you really must know how to use the tools available.

 

[powerabout]

Sst 45 doing 8000 with stock pistons?
About 70 hp at crank at about 7000, ( the stock engine it came from was 55hp at the crank after all)
Thats data from computerised dyno that has seen dozens of 45's and I'm sure sandhammeron05 knows whose dyno that is (was, its at Seaway now)

 

[baeckmo]

Of course the Kt and Kq coefficients refer to the diameter as the relevant length dimension, but you must understand what they actually represent. The Kt is in fact a pressure constant comparing Thrust/Area (i.e. pressure) with the dynamic stagnation pressure meeting the rotating profile (=U^2*density/2; where U is peripheral speed of the blade). As often on the marine arena, dimensional factors have been left out. As long as we study fully submerged propellers, the pressure is referring to the full disc area, but when the prop is semisubmersed the active area will be different, but the test results are still referring to the D^2 that represents the full area.

In the case, where the propeller is working behind a transom, you have a fixed working submergence, as the bottom is “shaping” the inflow. Here you may decide on a submersion and then calculate the necessary propeller data. In the case with tunnel hulls or with a prop under the hull (“proprider”), the working position is flexible and you must seek the working submersion that matches the engine power curve. From there you find the possible thrust.

Now if we study Ferrando’s paper, it must be clear that the numbers are valid for one specific propeller series. I referred to it because of the principles for scaling of operational submergence. It has considerably more blade area than the racing cleavers, so we should expect it to produce similar performance with a slightly smaller submergence (reflecting the higher eta). Note that this is not to say that it would suit the conditions for a tunnel hull, although it performs very well in other applications!

Ferrando’s thrust coefficient Kt’ = [T/(rho*n^2*D^2*Ao)], where rho is fluid density, n is rps, D is diameter and Ao is submerged area. We can substitute Ao with [D^2*pi/4*(Ao/Adisc)], and we then get the relation between Kt’ and the “normal” Kt:

Kt/Kt’ = D^2*pi/4*(Ao/Adisc); with Ao = effective working area.

For instance with a 50 % submergence we get Kt’ = Kt*(1/(0.5*pi/4*D^2), where the 0.5 factor represents the percentage of the disc that is actively working. The same goes for the torque coefficient. To the example then:

Using the Kq’ linear regression for the SST case, we find Kq’ ~0,0515 for J=1,024 and P/D =1,19. The “normal” Kq is then 0,0515*pi/4*Ao. We know the Kq is 0,0062 (see previous post), we then get the relative area (=Ao/Adisc) is Kq/Kq’*4/pi; i.e. 0,153, or 15,3 %. This requires a submergence of 21.5 %, or 46 mm. At this submergence this specific propeller would require exactly the power available (40,5 kW@8000 rpm). At the selected operating point Kt’ is ~0.22, which gives Kt = 0,026 and the resulting thrust is 954 N.

Hadler and Hecker (“Performance of Partially Submerged Propellers”; Naval Ship Research and Development Center, Washington DC) use a more direct method where propeller test data are plotted as J (=advance ratio) and efficiency versus Ct’. This thrust factor is more “complete” than the Kt variant in the physical sense, as all relevant factors are included;

Ct’ = T/[(rho/2)*Ao*V^2]; where Ao is active area and V is boat speed.

Propeller number 3767 is a typical propeller designed for supercavitating operation. It bears some resemblance to the ventilated racing props in terms of blade profile and chord distribution, but its trailing edge is curved and the blade area is seemingly greater than you have on the SST cleavers. Anyway, check figure 21 on page 1471. There you find for J = 1.02 that Ct’ is ~0.35 and eta 0,61. With power 40.5 kW at 29 m/s you have a thrust of 852 N.

Insert the thrust value into the Ct’ equation and out comes the required active area Ao. In this case Ao = 0.0058 m2. With the 8.4” propeller this means 16.3 % of full disc area, which requires 22.5 % submergence, or 48 mm. This result is very close to the previous calculations.

For these calculations I personally prefer the Ct’ format, but the Kt and Kq versus J format is more common. The point is that thrust and torque may be regarded as linearly dependent on the active disc area for engineering purposes. The lower limit of application is unclear so far, but I would say that the question marks get very fat below a submergence ratio of ~18 to 20 percent.

Now it is up to you to do your homework and present the corresponding calculations for the F1 sports case!

 

[sandhammaren05]

Sst 45 doing 8000 with stock pistons?
About 70 hp at crank at about 7000, ( the stock engine it came from was 55hp at the crank after all)
Thats data from computerised dyno that has seen dozens of 45's and I'm sure sandhammeron05 knows whose dyno that is (was, its at Seaway now)

Typical SST45 motors do not develop 70 hp at the crank, 60 hp max. That's the hp of the stock Johnson 60. The stock Johnson 60 develops about 55 hp at the prop, 55 shp. 7500 RPM is pretty common in SST45.

 

[sandhammaren05]

Fudge factors are unnecessary and misleading. Avoid them like the plague. Water tunnel tests are misleading and should be avoided. The only reliable measurement of torque and thrust coefficients is via open water tests at higher than planing speeds. Special design of a gearcase allows the torque to be measured at the propshaft, specially designed swivel brackets with very low friction bearings allows the thrust to be measured on an outboard at the transom. That's how at least one major outboard manufacturer does it; OMC did that before 1990. The engineers agree that anything but an open water test is misleading.

Of course the Kt and Kq coefficients refer to the diameter as the relevant length dimension, but you must understand what they actually represent. The Kt is in fact a pressure constant comparing Thrust/Area (i.e. pressure) with the dynamic stagnation pressure meeting the rotating profile (=U^2*density/2; where U is peripheral speed of the blade). As often on the marine arena, dimensional factors have been left out. As long as we study fully submerged propellers, the pressure is referring to the full disc area, but when the prop is semisubmersed the active area will be different, but the test results are still referring to the D^2 that represents the full area.

In the case, where the propeller is working behind a transom, you have a fixed working submergence, as the bottom is “shaping” the inflow. Here you may decide on a submersion and then calculate the necessary propeller data. In the case with tunnel hulls or with a prop under the hull (“proprider”), the working position is flexible and you must seek the working submersion that matches the engine power curve. From there you find the possible thrust.

Now if we study Ferrando’s paper, it must be clear that the numbers are valid for one specific propeller series. I referred to it because of the principles for scaling of operational submergence. It has considerably more blade area than the racing cleavers, so we should expect it to produce similar performance with a slightly smaller submergence (reflecting the higher eta). Note that this is not to say that it would suit the conditions for a tunnel hull, although it performs very well in other applications!

Ferrando’s thrust coefficient Kt’ = [T/(rho*n^2*D^2*Ao)], where rho is fluid density, n is rps, D is diameter and Ao is submerged area. We can substitute Ao with [D^2*pi/4*(Ao/Adisc)], and we then get the relation between Kt’ and the “normal” Kt:

Kt/Kt’ = D^2*pi/4*(Ao/Adisc); with Ao = effective working area.

For instance with a 50 % submergence we get Kt’ = Kt*(1/(0.5*pi/4*D^2), where the 0.5 factor represents the percentage of the disc that is actively working. The same goes for the torque coefficient. To the example then:

Using the Kq’ linear regression for the SST case, we find Kq’ ~0,0515 for J=1,024 and P/D =1,19. The “normal” Kq is then 0,0515*pi/4*Ao. We know the Kq is 0,0062 (see previous post), we then get the relative area (=Ao/Adisc) is Kq/Kq’*4/pi; i.e. 0,153, or 15,3 %. This requires a submergence of 21.5 %, or 46 mm. At this submergence this specific propeller would require exactly the power available (40,5 kW@8000 rpm). At the selected operating point Kt’ is ~0.22, which gives Kt = 0,026 and the resulting thrust is 954 N.

Hadler and Hecker (“Performance of Partially Submerged Propellers”; Naval Ship Research and Development Center, Washington DC) use a more direct method where propeller test data are plotted as J (=advance ratio) and efficiency versus Ct’. This thrust factor is more “complete” than the Kt variant in the physical sense, as all relevant factors are included;

Ct’ = T/[(rho/2)*Ao*V^2]; where Ao is active area and V is boat speed.

Propeller number 3767 is a typical propeller designed for supercavitating operation. It bears some resemblance to the ventilated racing props in terms of blade profile and chord distribution, but its trailing edge is curved and the blade area is seemingly greater than you have on the SST cleavers. Anyway, check figure 21 on page 1471. There you find for J = 1.02 that Ct’ is ~0.35 and eta 0,61. With power 40.5 kW at 29 m/s you have a thrust of 852 N.

Insert the thrust value into the Ct’ equation and out comes the required active area Ao. In this case Ao = 0.0058 m2. With the 8.4” propeller this means 16.3 % of full disc area, which requires 22.5 % submergence, or 48 mm. This result is very close to the previous calculations.

For these calculations I personally prefer the Ct’ format, but the Kt and Kq versus J format is more common. The point is that thrust and torque may be regarded as linearly dependent on the active disc area for engineering purposes. The lower limit of application is unclear so far, but I would say that the question marks get very fat below a submergence ratio of ~18 to 20 percent.

Now it is up to you to do your homework and present the corresponding calculations for the F1 sports case!

 

[PetterM]

Very interesting reading.
I have also found the Rose and Kruppa/ Ferando et al papers quite useful.
I suspect the commercial style propellers tested by Rose and Kruppa/ Ferando et al have much thicker blade sections than what we use on fast offshore boats.
I am not familiar the with Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973, is it in German?
And I did not totally follow your calculation, it has been a while since i have been thinking about propellers, are you scaling the Newton-Rader Kt/Kq with immersed area?
I have been looking for testdata for surface piercing propellers with higher P/D for ages with no luck. Do you know of any?
Cheers,
Petter

Well Joe, I’m afraid you just shot yourself in your foot. You have missed something essential regarding the ventilated propeller; the working area is a variable. On top of that you have not considered the impact of the torque coefficient Kq, which is essential for the estimation of the effective working area.

The examples you refer to are, at best anecdotal; whenever I see a figure rounded off to double zeroes or to a fiver, I am sure that the number is at best a guesstimate. Your rpms are given as “8000 rpms”, speeds are given as 65 or 110 mph, there is nothing said about operating depth, propeller foil profile, blade profile, blade area ratio or the profile and depth of the steering fin in front of the propeller. The power figures are very uncertain, do you have dyno data to show?

Now, that said, let us check two of your examples, knowing that the basic info is vague. For the purists that may read this, I want to say that I am very well aware that I am stretching the extrapolations a bit beyond recommendations; it is for the good sake of demonstrating the principle. I have selected reference props from other sources than you used, in order to have a wider spectrum.

First the SST45 case. Using your data we have the following:
P: 40,5 kW @ 8000 rpm (133,3 rps)
V: 65 mph (29 m/s)
D: 8,4” (0,213 m)
P: 10” (0,254 m)
Advance coefficient (lambda) 1,024 (=V/(D*n);

This gives the following:

P/D: 1,19 and the torque constant Kq = 0,0062. With this P/D value, we check available propeller data.

A/ First I looked into the figures given by Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973 The report is highly recommended!). Here a 3-bladed Newton Rader propeller (without cup) with P/D = 1,33 is tested. The BAR is ~0,7. This means that you may use the Kt and Kq values for this prop if you check its graph at lambda 1,33/1,19 x 1,024; ie 1,144 for rough estimates.

At depth of 50 % the N-R propeller shows Kq=0,018 and Kt=0,07. As has been shown, f.i. by Ferando et al, the ventilated propeller is operating with reasonably constant ratio Kt/Kq at varying submergences. Hence we may (for a rough estimate) use a linear relation between the observed Kq (0,0062) and the tank value (0,018). From this we get a correction factor of 0,344, which we use to find a new operating submersion.

The relative area at 50 % submergence is 50 % (surprise surprise…!), and the needed relative area is then 17,2 %. Now if you calculate the effective disc area as a function of operating depth, this will correspond to a submergence of 23,4 %. The SST propeller blades are thus operating at a maximum depth of 50 mm. The thrust coefficient is then 0,024, giving T= 880 N and a power of 40,5 kW.

B/ Next we check the Rolla propeller tested by Rose and Kruppa (Published FAST 91). This series is also based on the Newton-Rader propeller, but with a cup. Here we use the diagram no 15 (30 % submergence, atmospheric ventilation, shaft angle 4 degr). From the lambda 1,02 (right side) you go to the line showing lambda asf of Kt/J^2 for the P/D propeller. You will find a Kt/J^2 value of 0,04. From this value, you go vertically to find the eta value ~0,61, which allows us to calculate the “chart power” (=Trust*speed/eta) to 72,4 kW, and Kq = 0,011.

With the same correction procedure as before, we get a correction factor =0,559. This time it's applied to a 30% submerged prop (relative area is 25,24 %), we arrive at an operating relative area of 14,1 %, which we get with a 20,3 % submersion, or 43 mm in this case. The difference to the original N-R is explained by the cup and a higher BAR (=0.80) on the Rolla prop. The thrust coefficient is 0,023 and the resulting thrust 859 N.

C/ Now we check the 4-bladed supercavitating propeller type 841-B (origin KaMeWa if I remember ok), used by Niclas Olofsson in his thesis (Chalmers uni 1996, Gothenburg). Foil profile is a modified cambered wedge (“Tulin” style) and blade shape is more like the classic cleaver, with an area ratio of 0,58 and P/D is 1,24. Shaft angle is zero degrees. At 30 % submergence it has Kt = 0,035 and Kq = 0,0091. Efficiency is ~67 %. Repeat the correction procedure and you end up with trust = 868 N and submergence 23,3 % or 50 mm.

D/ Check with data from Szantyr (FAST 97). Supercavitating propeller (modified wedge), gives nearly identical results as “C” above.

Ergo: Four different ventilated propellers studied for the same task, thrust results 850-880 N and operating depth mean value 48 mm. Differences to a great extent explained by differences in blade profile and blade area ratio.

When I go through the same exercise for the F1 Sports, using the Rose/Kruppa data and a 3-bladed supercavitation propeller (H. Ghassemi and M. Ghiasi, Amirkabir tech uni, Tehran), I get thrust 2258 and 2332 N and a submergence of ~17,6 %; in this case 45 mm.

In all cases studied the nominal submergence is slightly less than you will see in reality, since the nominal value is “hiding” the ventilated wake after the steering fin. I would guess that the gearcases in the two examples will have diameters of, say 60 and 80 mm respectively, thus clearing the waterline with a slight margin.

From the examples above, I must say that the various tank data are surprisingly consistent, considering the lack of precision in the observations from your examples. There is a lot more to comment upon regarding the influence of design variables, but i think I made my point. In real life there is also the influence from dissolved air in the water. It normally varies over the day, with a maximum between 14 and 15 o’clock and a minimum in the early morning. In the afternoon, the water is thus supersaturated with air, that will come out of solution very easily and will then cause a loss of performance of propellers and hydrofoils.

I will be back later with comments on the shaft inclination subject. I certainly do not agree with your previous statements on the quality of test data here; you really must know how to use the tools available.

 

[powerabout]

Typical SST45 motors do not develop 70 hp at the crank, 60 hp max. That's the hp of the stock Johnson 60. The stock Johnson 60 develops about 55 hp at the prop, 55 shp. 7500 RPM is pretty common in SST45.

Where did you get this information, whose dyno?
raised porting and raised rpm and it only makes same power as a stock engine?
Plenty of sst45's in socal make 70hp at 7000 whats wrong with yours?

 

[sandhammaren05]

Where did you get this information, whose dyno?
raised porting and raised rpm and it only makes same power as a stock engine?
Plenty of sst45's in socal make 70hp at 7000 whats wrong with yours?

Former OMC engineers, including racing division. The ports must be stock in SST45 Class. I'm not sure what you think you're talking about.

 

[sandhammaren05]

Hi Petter,

Brandt is in German. I simply use the experimental values for torque and thrust coefficients and efficiencies.

Best,
Joe

Very interesting reading.
I have also found the Rose and Kruppa/ Ferando et al papers quite useful.
I suspect the commercial style propellers tested by Rose and Kruppa/ Ferando et al have much thicker blade sections than what we use on fast offshore boats.
I am not familiar the with Brandt (Modellversuche mit Schiffspropellern an der Wasseroberfläche; Schiff und Hafen no 5 and 6, 1973, is it in German?
And I did not totally follow your calculation, it has been a while since i have been thinking about propellers, are you scaling the Newton-Rader Kt/Kq with immersed area?
I have been looking for testdata for surface piercing propellers with higher P/D for ages with no luck. Do you know of any?
Cheers,
Petter

 

[ChrisN67]

I have understood one thing is for sure, that unless there is a synergy of design of hull and propulsion system one can rarely achieve optimal results by accident. But on the other hand, it is the only "science" where multiple wrongs can create a right.

There is no replacement for sea trials and sea trials. Being close to the water is better than 1000 computations.

I stretched my intrepid and get 48 knots with a pair of Yanmar Diesels and SBM surface drives when ambient temps are over 50C. On a cold day, it can get over 50 knts. That's with 500 liters fuel, 200 liters water and 4 persons on board.

https://www.youtube.com/watch?v=nT3JSDxIRYg

The boat had been stretched when I got it and was coming apart at the seams from poor workmanship. I had to take it down to the skin, straighten it out and build it back up.

I was very happy with the performance ; but I think I was lucky.

On a separate subject; can someone advise if there is a thread dealing with trailers? I could not find one. The boat weighs 18000 on the ramp and I keep breaking the Fulton (expensive) 2 stage winch. I spent the money for an electric IP67 winch with synthetic rope but not sure what the best mounting configuration would be... I was thinking about putting it near the trailer tongue and setting up a a pillow block roller to the boat eye.

Once secure, I would detach the winch line and connect a Spring Load Binder to lock the boat to the trailer.

https://www.dropbox.com/sh/z3btl56ontrqtag/AAAXBVX0cgJOzhys_d86llOCa?dl=0

Thanks guys

 

[ChrisN67]

Sometimes you get lucky

I have understood one thing is for sure, that unless there is a synergy of design of hull and propulsion system one can rarely achieve optimal results by accident. But on the other hand, it is the only "science" where multiple wrongs can create a right.

There is no replacement for sea trials and sea trials. Being close to the water is better than 1000 computations.

I stretched my intrepid and get 48 knots with a pair of Yanmar Diesels and SBM surface drives when ambient temps are over 50C. On a cold day, it can get over 50 knts. That's with 500 liters fuel, 200 liters water and 4 persons on board.

https://www.youtube.com/watch?v=nT3JSDxIRYg

The boat had been stretched when I got it and was coming apart at the seams from poor workmanship. I had to take it down to the skin, straighten it out and build it back up.

I was very happy with the performance ; but I think I was lucky.

On a separate subject; can someone advise if there is a thread dealing with trailers? I could not find one. The boat weighs 18000 on the ramp and I keep breaking the Fulton (expensive) 2 stage winch. I spent the money for an electric IP67 winch with synthetic rope but not sure what the best mounting configuration would be... I was thinking about putting it near the trailer tongue and setting up a a pillow block roller to the boat eye.

Once secure, I would detach the winch line and connect a Spring Load Binder to lock the boat to the trailer.

https://www.dropbox.com/sh/z3btl56ontrqtag/AAAXBVX0cgJOzhys_d86llOCa?dl=0

Thanks guys