250HP outboard Speed prediction for heavy boat

This is the one used here:

http://illustrations.marin.ntnu.no/hydrodynamics/resistance/planing/index.html

Are there better ones in the web? I have my own, which includes from Savitsky 2006 paper aero and whisker drag, both of which are not important in this case.

Joakim

 

Joakim
I do not like empirical models like the Savitsky predictor but I still point it out to people looking at planing hulls. It is about the best there is. I have only found one example out of about 20 or so where it was close to actual power. In practice it normally under predicts drag. The gap can be closed by considering windage and appendage drag. Some semi-planing hulls have very large keels.

I would always recommend having considerably more power than it predicts unless the design is meticulous in detail.

There is another predictor that uses analytical methods based on foil theory but it has not been widely used.

Michlet will also give good results for some hulls.

I suppose we both are suggesting that 2 X 250HP will be enough for 20kts.

Rick W

 

My Savitsky program has appendage, windage and whisker drag models included, but most importantly it takes roughness into account. I think the Cf allowance suggested by Savitsky is not enough unless the boat is very big and slow. To reach it requires very smooth bottom in "normal" boats.

I'm not a NA and I have no own data to compare with, but I have calculated some existing boats with rather good results.

There are too many unknown factors like the real propeller efficiency, thus I agree that it is better to have have some extra power, but 2x250 should be easily enough, if the LCG is somewhere near 5 m.

How big errors do you usually see with Savitsky? Do you measure LCG and VCG? Could you give me an example I could try with my program?

Joakim

 

Jaokim
If you go back through this thread you will get enough data to compare with Savitsky.
http://www.boatdesign.net/forums/projects-proposals/hull-modifications-23205-4.html
Also this one:
http://www.boatdesign.net/forums/inboards/prop-size-again-24565.html

If you are interested in doing the full analysis then have a look at JavaProp. It takes a bit to learn it but it will give good results for the props.

Rick W

 

I didn't find the data needed for Savitsky method from those and at least the latter boat was quite far from being a prismatic hull. I was hoping to get some data that has all the ACCURATE parameters (beam, deadrise, displacement, LCG, VCG, propeller shaft angle and location, spray rail configuration, area of all underwater structures, windage area). If you also had the detailed data of propellers and engines, you could really compare.

If you don't even know the displacement and hull shape, it is obvious that any metdod will be very inaccurate. Assuming a prismatic hull with ideal LCG will obviously give much lower resistance than the reality for a hull that is not designed for planning speeds.

Javaprop is some kind of lifting line method? How do you know the geometry of the real propellers? What more do you get from it compared to calculators based on the Wageningen B-series propellers?

Joakim

 

It is not all on a plate. You have to go through the posts and fill in the gaps with what else you can find on the web. Make some engineering estimates.

JavaProp is an analytical method that is much more flexible than an empirical series. I started using it for a quick check on the small props I design and now find it easy to use. What it does not do is determine the impact of inclined shaft. You are also stuck with a limited range of foils and Re#. I normally work on an EAR of 50% unless I have actual data.

Rick W

 

I have been trying JavaProp for a while and have some questions

1. I would like to compare it to a Wageningen B-series propeller calculator I have used a lot. How should I set JavaProp for this? Of course air->water, but what about profiles and angle of attack? At first look JavaProp seems to give clearly better efficiencies. Are they really achievable?

2. Quite often cavitation is the limiting factor in propeller selection/design. E.g. smaller EAR would give better efficiency, but is prone to cavitation. How to take that into account with JavaProp? It seems to design propellers with small chord at maximum R. This is not what most marine propellers are like. I would think the difference comes from cavitation and/or limited diameter.

Joakim

 

Joakim
I am designing props for low power applications and use my own program for this as I can design for any foil section I choose to use. I use JavaProp for quick results as it is has nice interface and is reliable enough. It also allows for velocity ratio which my model does not. I work at such low loadings that I do not need to bother with this. In these circumstances pressure differences are very low.

JavaProp is certainly accurate for the range of foils it has when lightly loaded. For the large powered boats the only section that comes close is the MH116 at Re# 500,000. So I use this. For the given conditions I adjust the AoA until I get an EAR around 50%. Usually this is 0 or 1 degree on the MH116 foil.

I did compare JavaProp once with Wageningen B-series data and got reasonable comparison. Irrespective there are a number of people using JavaProp and achieving very good results. I have measured power data within 1% that shows I get prop efficiencies up to 86% and this is what I predict with analysis almost same as JavaProp.

In real life there are a whole lot of factors affecting props that I aim to avoid. My props essentially operate in open water. I always run the shaft horizontal. I have no strut or use a very tiny strut. I use the optimum diameter consistent with blade stress.

For my props the constraint on chord is blade bending strength not pressure differential.

I am aware that JavaProp usually indicates higher efficiency for most power boat applications but in the past efficiency was not a high priority and you see a whole raft of terrible power sapping layouts. How do you cater for a shaft angle of 15 degrees when you apply the Wageningen B-series. I hate the thought of running props on inclined shafts after what I have done to shafts at very low power level.

So JavaProp is very good if you are seeking efficiency. Just be prepared for a shape that is not like a Wageningen B-series prop. I can usually get 1 to 2% better than JavaProp by optimising the foil to the application.

Attached link shows a prop milled to my design.
http://www.boatdesign.net/gallery/showphoto.php/photo/11202/ppuser/18624
It achieves 86% efficiency at design condition.

Rick W

 

My interest is not so much in designing propeller shapes, but analysing existing propellers and choosing a better one (just a hobby!).

I guess your designs are quite far from "normal" marine propeller due to very low drag and thus power?

How do you know EAR from JavaProp? Do you visually approximate it or calculate from the shape given?

I tried to compare the propeller of my sailing boat (actually it's a 2-bladed folding propeller which probably makes it far from optimal). It's a 14x8 and running at 1344 rpm 6.4 kn BSP, maybe 6 kn with wake factor. It's a strut drive thas no angle.

Using Wageningen B-series 50% EAR 3-bladed I get 800 N thrust and 54.5% efficiency. 800 N is very close to what my sailboat VPP-program predicts at that speed. The engine is 1GM10, so the real efficiency is much worse or there is something wrong with the engine, but that's another issue.

Using MH116 at 0.5 degrees and spinner dia of 0.112 m I get very similar results (about the same pitch using 1344 rpm and 800 N) with JavaProp, except that the efficiency is 61.2%.

So what is the explanation for this huge difference in efficiency? Which one would be more accurate for a normal good quality marine propeller I could buy? Are the Wageningen B-series really that far from optimum 14x8?

Joakim

 

Joakim
If the prop is of uncertain shape you can actually load the blade chord and pitch angle at various radii into JavaProp and do analysis from there. I produce numbers as a text file and import from there. If it is poorly made it could, at worst, have the performance of a flat plate. So you can use this to check the worst efficiency for the particular pitch. This is not quite a flat plate as it has a rounded nose and pointed tail. This should be as bad as it can be unless the pitch is extremely poor.

Your motor is rated at 4.8kW so the predicted power of 3.6kW is not a huge difference - what is the gearbox efficiency, seal drag and auxiliary load. I would have a lot of faith in the JavaProp result and then go about looking for reasons why there is a difference. Check the motor torque curve to see that it is not being torque limited by the prop. It could be overpitched. I would measure the prop and load the actual chord and pitch angle data. I normally load 20 points but I would only check the pitch angle at three points to confirm it is correct and then calculate the points in between. I also measure the blade chord at a few points and then extrapolate.

The more detail you do like this the more you will close the gap. If you have a gap it will not be with JavaProp. It will be your inability to define sufficient detail. Once you have got the performance of the existing prop nailed then you can use JavaProp to produce a better design. This is why I like analytical methods because you know they are correct providing the input variables are accurate.

The Re# of 500,000 is somewhat low but the foil performance does not alter a lot between this and the actual you have. This is one area that my model handles better than JavaProp but it is not a huge difference - typically less than 2% in efficiency prediction.

Rick W

 

I would not consider it a huge difference but the answer to your last question is YES. The Wageningen propellers are made for a typical range of applications where there are draft constraints and significant thrust for the size. They are nuts for a yacht application that does not have a draft constraint and are lightly loaded.

I expect you could design a prop for your boat that would achieve an efficiency of 80% but it would need different gearing. Something like a 24X16 2-bladed with very low EAR would be practical but it needs to spin at half current speed. You should be able to get 75% with the current gearing. The higher the blade aspect the more efficient.

Rick W

 

Yanmar 1GM10 is actually rated about 6.5 kW at propshaft: http://www.yanmarmarine.com/uploads/products/pdf/GM_YM/1GM10_TechData.pdf
http://www.yanmarmarine.com/uploads/products/pdf/Saildrive/Saildrive_20_TechData.pdf

The earlier version 1GM was probably that 4.8 kW. I have SD20, thus 1:2.64. The alternator is 35 A and I guess all else should be included in the rated power.

I know having much bigger diameter would give much better efficiency especially with lower shaft rpm, but it's a sailboat and primary interest is low drag while sailing. So I'm stuck to small size folding (or feathering) propellers, which obviously have high loading (close to cavitation limits) and I would assume Wageningen B-series would be rather good for that.

Also folding propellers needs heavy (very thick) blades in order to stay open while going backwards, thus the profile is far from optimal. I would guess flat plate is better than a very thick profile?

JavaProp doesn't "see" nor warn about cavitation. I guess that could cause significant errors, but in this test case it shouldn't be a problem.

Joakim

 

Would it really be possible to get 75% with current gearing? With JavaProp I can easily get that with 0.6 m diameter and 4 deg AoA. But the Cl would be close to one and tip speed over 40 m/s, which would mean that Cp (pressure coefficient, not power!) of less than -0.12 would lead to cavitation. With AoA of 4 Cp is bound to be far less than -1.

Even with 14" propeller and tip speed of ~24 m/s Cp is limited to -0.35 and thus Cl to very low unefficient values.

With very low AoA (Cl) the efficients with JavaProp drop dramatically (to more realistic values?).

Or have I understood something wrong? What kind of tip speeds have you used with your efficient propeller designs?

Joakim

 

Joakim
There are a lot of things to comment on.

The alternator would only draw significant power when it is charging battery. Typically just after a start so I would not get too concerned with this.

The drive leg could be sapping quite a lot of power. If the boat is taken out of the water you can check the seal drag. See what torque it takes to turn the prop when in neutral. You should be able to get around 98% on each set of gears. So you can do numbers on the drive leg.

We are making the assumption that the rpm is accurate and the motor is developing full power. These really need to be cross checked. Your fuel consumption can help with determining what power is being applied.

I would also do a drag analysis on the hull to see if the drag indicated by the prop loading is in the ball park. THis helps cross check rpm data and pitch data.

As far as the prop design goes if you get it more efficient then you can drop the engine rpm. Check what you can achieve with a prop doing 1000rpm at 3.3m/s having a diameter of 450mm. If you fiddle around with AoA at each of the 4 points available you should get very close to 75%. Now this is not the best foil available. It may not be a practical prop either because the blades will not be strong enough. I try to get down to 8% foils specifically designed for the required lift coefficient. I also optimise the chord and AoA at more than 4 points and include Re#. However my optimiser does not cater for velocity ratio. I cannot get a better shape for the MH foil then JavaProp is providing. However I have attached a text file that you can import to JavaProp geometry and try it. It will not be as efficient as JavaProp is getting but this is a shape that would have adequate strength for an 8% foil in your application. To use the file you set up all the prop parameters but do not do the design just load the file. You then use the analysis windows to get the data.

You need to end up with a propeller that is strong enough and it usually means a thicker section in the root.

Even your application has much higher velocity ratios than I normally operate at. Typically I am under 105%. I have not checked the minimum pressure in your application with the 450mm prop but the Cp is very low and dropping rpm reduces the loading.

I hope this gives you an idea of the options available. Working with JavaProp will give you greater insight then using series data.

Rick W

 

Dropping to 1000 rpm would make the motor run at 2640 rpm with almost WOT. This is probably not a good idea for a motor designed to work at 3400-3600 rpm WOT.

I have a very comprehensive VPP program, which calculates the drag at any speed up to Fn=0.6 for my sailboat, for which I know very accurately all the parameters needed (displacement, Cp, LCB, LCF etc.). The hull is very smooth, since it's used for racing. At 6.4 kn the drag is about 750 N.

The propeller turns easily while on neutral. The rpm has been measured from engine sound using FFT, which is a very accurate method. The speed has been measured with GPS and an accurately calibrated paddle wheel log and there are no currents around here (well up to 0.5 kn can sometimes happen).

I have doubts about the power output of the motor, which is one of the reasons I'm interested in this.

If I measure my propeller and import r, c and alfa to JavaProp, what would that actually help? I don't know the profile and it is certainly not any of the ones available . There is a huge difference in thrust and power at same rpm, if I change the profile. Of course I could start from the "known" thrust, but it would be nicer to use the pitch. How is the pitch defined in commercial propellers? Is it a geometrical value or is it derived from zero thrust when V equals the theoretical speed from pitch.

Joakim