most efficient propulsion?

Building a hull for drag testing is just a waste of time. You can get performance data within about 2% using Michlet. It also has the ability to optimise the shape for any given set of constraints so you will know you have the best possible hull.

I used to do things by trial and error until I started using Michlet for hull optimisation. My 11th pedal powered hull was a product of Michlet and currently holds the 24 hour distance record for human power, 245km. You can see the hull here:
http://www.adventuresofgreg.com/HPB/2008_09_15_archive.html

Michlet can optimise for any number of hulls so is suitable for catamarans. You can also get very good performance with little compromise from a single hard chine meaning the hulls can be formed from three timber planks or composite flat panels.

There is a thread on pedal powered craft that you might find interesting:
http://www.boatdesign.net/forums/boat-design/pedal-powered-boats-23345.html

Rick W

 

Uggliozzi

For your design, there are many aspects to consider. The main ones however, are the same as for any ‘real’ design, that is the hull and the power. (I don’t mean ‘real’ in a patronising sense, since your concept is ostensibly just a large “model”. My colleague and I have built ‘models’ for studies that are 5m in length and powered by small outboards.)

Taking power first. If you designed “a boat” suitable for this challenge of yours, would you get the same performance between say a 20year Olympic athlete or a 40 year old sumo wrestler or a 60 year old who keeps fit? Clearly there will be a major difference in the performance of the same boat given these 3 variables. Hence the design is to be “made” around the power plant, ie you.

As you peddle along, your fatigue will affect the power delivery, that is before even getting into peddling into a head wind, and/or waves or simply steering a straight course and many other factors. So what is it you are optimising for and is it cancelled out by a slight head wind, for example? Your post doesn’t mention about a record attempt, eg ideal conditions.

Hull lines. As I mentioned before, so long as you have your 15:1 or better and a low length displacement ratio, you’re onto a “winning formula”. You really need to keep a very close control on your weight.

Software’s such as Michlet, are only really useful for parametric studies. To use them for serious resistance and propulsion test means you don’t have the means or money for proper tank testing. Since, if Michlet is “so wonderful and accurate” all the tank test facilities all over the world would be obsolete over night and there would also be no need for further CFD devolvement! Try telling that to those in the Boundary layer Code development or those using Navier Stokes!

In a paper/report entitled ‘Hydrodynamic Drag of Small Sea Kayaks’ sums this up very well, one of their conclusion is thus:

“..The monetary rewards of sea kayak and canoe design have never been sufficient to justify tank testing or any other objective and methodical method of hull form development. Designers have, therefore, relied upon subjective evaluations of boats to determine performance values. This is neither reliable nor consistent. Test paddlers carry with them an extraordinary amount of baggage including personal and aesthetic bias, moods swings, the inability to duplicate test protocols and the more obvious inability to quantify or even sense performance variations in any reliable manner. This and the absence of formal design training for most designers results in a wide range of hull forms and little consensus on what is good or bad. Almost every shape has its proponents and detractors…”
And

“..Existing performance prediction methods based upon empirical data (Savitsky, Holtrop, Winters) have not proven satisfactory because there isn't enough data to span the range of hull forms and dimensions that constitute the genre. The fit between the predicted values and test values in this series is encouraging and computer programs such as Michlet (Lazauskas and Tuck, 1997) and GODZILLA (Lazauskas, 1997) may provide designers in this field with economical and reliable tools in developing hull forms..”

Full credit to Rick for what he has achieved. But this also needs to be put into a naval architectural context too in how that has been achieved and is it useful across all applications other than record attempts using peddle power. Since he has, as noted above, had at least 11 hull design iterations, this just adds to the ‘database’ of ‘knowing what to do for his design in the application required’. But this is true of any ‘database’. It is more about qualitative methodology than ‘design’ per se. The accuracy of many of these software’s like Michlet are ‘useful’ at certain speeds only. Yours and Rick’s design probably fall into this small window-long slender hulls of the program (so it’s not rocket science), but caution still needs to be exercised. Such as description of the free surface for starters…CFD is a major field in its own right and well beyond the scope of your post. If you stray outside the parameters of Michlet you’ll get poor correlations.

Software’s all have their limitations and to use them effectively one needs to understand the parameters of the database that runs the program. This is not to say they cannot be used. But to obtain any meaningful results one would need to add to the database, to get a real feel for any type of validation. The more data you have the better the predictions – qualitative research. Clearly Rick has this for his hull. But that is where the limitations of these programs are exposed. Slowly honing in on one type of hull to squeeze out every last drop, so to speak, means the program is good for that hull and its variations only. It does not mean it is accurate for all types and all speeds such a SWATH or a Class I offshore power Boat etc. Because the database that comes with the program is insufficient to accurately predict the resistance of these types.

See attached graph of accuracy. CFDs are getting better, but, still not as accurate as tank testing

Bottom line is, you need to establish what is important for you. What are the main variables and how to mitigate them, if possible. Then select a methodology that you feel comfortable with, regardless what we suggest here. It is your time and money.

But whatever you decided, it is all about weight, weight and er..oh yeah…weight, and then a high L/B ratio.

Get your weight prediction wrong and you’re on a losing battle before you even sit in the boat about to peddle…

 

I will run the Godzilla optimisation on these parameters and give you easy to build hull forms. I expect with 100W you should be able to get 6kph easily. Will let you know in a couple of days.

I am a slightly later vintage than you and design for 130W for about 8 hour duration but I wane 4 to 5 hours in unless I rest. In cruising you would be taking breaks to site see so should be able to hold around the 100W mark when actually motoring. The more you do it the better you will get or the easier it will become.

By the way Michlet/Godzilla is a product of University of Adelaide. The wave drag component is not based on parametric analysis but analytical physics known as Michell's Thin Ship theory. Michlet and its commercial variant, Flotilla, are in use by the US Navy and US Coast Guard amongst many others for hull optimisation. It is also used for olympic rowing shell design.

The main value of Michlet/Godzilla is that you will know you have the best possible hull for your nominated conditions. The optimisation does minor variations on the hull form each iteration and keeps selecting the result that gives the lowest overall drag. It can do about 1000 iterations per minute so in a few minutes you can arrive at a hull that would have taken a thousand years of trial and error to evolve. There is tremendous satisfaction in knowing you have the best possible hull shape. You can also see what certain constraints may cost in terms of performance.

There is a South Australian firm, Bolly Props, who make model aircraft propellers and they have one specifically for boat applications. I can see what performance you could expect with this prop in conjunction with the boat and compare it with the best possible prop. I am currently working on a folding prop in the hope that I can make it immune to logs and able to readily shed weed by letting it coast and fold. This might be something worthwhile for you as well.

Rick W

 

Rick and ugliozzi
It will be interesting how michlet compares to hullcalc4-9.xls.
I get 4.45kt with those above parameters and a square meter of CD 0.4 air drag (a miniscule drag contribution), and a 70% drive efficiency.
That is with a proa-ish hull form, very low prismatic and a beam from 0.24 to 0.27m. Going any longer did not increase speed. However a meter or two more length did not decrease speed either. So, I would consider those meters since they give a kinder motion and bigger load range.

Out of curiosity I optimised a single hull for the same parameters except length - it got a bit wider, from 0.31 up to .4, a bit longer, 6-7m, and 4.9knots at those same 100w.

Why do you want a cat? tent space on the tramp?

 

Thank you again guys. I have got a lot to think about. I could not get hold of the Michlet software (internet blindness I suppose) and have not given much thought to software modelling of hull shape. I was just going to use observation of fast fish and the profiles used in aero engineering.

I have finally worked out how to use JavaProp (read the manual and tell the software that the prop is operating in water, not air) and should get a good result once I have the hull parameters.

Rick,
I have had a look at your drives and am quite impressed, especially your self-locating flexible drive. Currently my plan is to use a 25H (quarter inch) twisted chain drive on a top-pivoted drive leg to try to make it resistant to snags and sand bars. As I will be on my own and I'm no spring chicken, I will keep to a strict regimen of cadence, pedal/rest cycles and rest days. Before I set out I will use my exercise bike to establish just what my pedal/rest parameters should be and then build in a safety margin.

Sigurd,
you have got it in one - the catamaran is so that I can put a tent on the back so that I don't have to make/break camp everyday. I will walk to caravan parks or service stations to shower when one is close enough. I will use a tent that can be folded down during the day to reduce windage. My plan is to use the recumbent pedalling position to reduce windage to the bare minimum.

 

oki - I think somebody said here and I think its true that as long as it is long and skinny it can be most any shape as long as it is fair and pointy. the hullcallc is a spreadsheet you can find here by searching I hope if you need it. Put big hatches there for storing stuffs!
Since I don't know what sort of catamaran configuration you have in mind, may I add, even if you are tenting, I'd consider a single outrigger, one loaded hull, the other skinny and short, but with enough reserve bouyancy (flare or tallness)?

 

Ugli

I did the Godzilla runs. The best catamaran for 200kg displacement and 100W power will get 2.4m/s (8.6kph) in calm conditions.

Best - requires 81W on the hulls and is 5.6m long.
Constrained length - limiting the length to 2.4m increases the power required at the hulls to 86W.
Easy to build length constrained - this has single hard chined hulls requiring four planks to make each hull increases power to 89W at the hulls.

I looked at a prop design and can get around 87% with a 400mm diameter prop running at 300rpm. The power at the pedals is 105W to get the 2.4m/s.

The attached image gives an idea of the boat. I just added one of my drive frames to give a bit more detail.

Rick W

 

Sigurd,

I will think on the proa design to see how I can use it. As construction has not yet started, I have time to be flexible. The big problem is my desire to use a "kick-up" drive leg. I must have a propellor assembly which can automatically get out of the way of obstacles because the river on which I will be travelling is notorious for hidden snags and sand-bars. (It is now very low and is unlikely to be any higher for years and so the snags and sand-bars will be a bigger menace than usual for some years to come. ) A "kick-up" drive leg is easy to do when suspended on a metal web between catamaran hulls - not so simple if I need to bring it through a hull. Thinking....thinking....maybe an offset drive with a bias to the main hull will work....thinking.


Rick,

Your results are very close to what I had in mind. The main change is that I was planning the widest part of the hulls to be 33% from the bow.

Thank you for the power calculations - they are very helpful and give me confidence. They show that my target of 6KPH (1.7M/S) is achievable. My natural cadence is 68 at the pedals which would give me 272 at the prop with the gearing I already have and that seems to fit the 300RPM of a 400mm prop.

 

Ugli, I think what Rick does on his pedal trimarans, is to have the propeller offset to one side of the main hull, which cancels out the turning moment of the rotating prop - I seem to recall he mentioned that. I too think through hull stuff should be avoided if possible.

Rick, cool that hullcalc was so close to michlet in this case - a bit slower but I had chosen a lower drive eff. than what you found achievable.

 

I've come across references to Kappel propellors which have surfaces at the tips of propellor blades which slope in the direction of travel. They are supposed to make a propellor more efficient by reducing tip turbulence.
One site: http://www.skk.mek.dtu.dk/English/Research/KAPPEL_Propeller.aspx

There is also a reference to bending propellor tips towards the rear of the propellor for improved efficiency :
http://www.zakpro.com/Tip_fins.html

I would imagine that shrouded propellors get some of their improved efficiency by reducing tip turbulence. To me, these tip fins seem to be propellor shrouding reduced to its minimal essence.

I see aircraft designers putting winglets on some aircraft wings. Some designers put winglets on the upper surface only and others cover both. Some birds such as eagles and buzzards have feathers which bend upwards at their wing tips.

With efficiency claims of 5 to 15% it would seem worthwhile taking a look at tip fins.
Does anyone know of tip fin experiments in the human powered boat realm?

 

Uggli
Yes. I have not left many stones unturned.

If you have the room for an efficient prop then design for efficiency in the first place and forget winglets.

If winglets were the be all and end all they would be used on gliders where efficiency is important. The main advantage with winglets on commercial aircraft is that you can improve lift to drag without increasing wing span. Extra wing span means all those airports gates would need to be separated more. Clashing of tips is already a problem.

Another example is a typical aircraft propeller blade. Again if winglets were really valuable you would see them here.

Winglets on sailing boats might offer a slight advantage by lowering the centre of pressure on the sail. This is also an advantage on a heavily loaded wing of course as the bending moment will be reduced a little by using winglets instead of increasing span.

Rick W

 

I have attached some pictures of the array of props I have played with. I have made between 30 and 40 props and have bought another 5. I have had one milled from my CAD file. It was the best to date. It performed slightly better than design because it was post-milling pollished and this reduced the blade thickness below design. It was still strong enough for the job.

It is surprising how well a couple of pieces of twisted flatbar will perform. But for best results you want high aspect blades to reduce induced drag and a blade section that has high lift to drag at the design condition. For lightly loaded asymmetric props the angle of attack is often negative.

The failed prop was an attempt at small winglets but the inner part of the blade was not up to the torque I could generate.

I also bent a few shafts due to having the shaft inclined when using high aspect blades. Best to run the prop in line with flow to avoid unbalanced loading.

Rick W

 

dam Rick Im going to have to try and enlist your assistance when it comes time to pick a prop

 

Rick,

You are indeed the master. To have made 30 to 40 props, your dedication must be exceptional, I am most humble.

I have plenty of room and will go for the high aspect prop as you recommend. Javaprop says I can have a prop with blades as narrow as 25mm, at the power and RPM I want to put in, at near 80% efficiency. The shrouded option in javaprop made me greedy for efficiency, hence my query. The 25mm width fits in nicely with readily available materials.

My plan is to use a 450mm diameter prop running straight in line with the flow with the prop tip coming no closer to the surface than 75mm. Is 75mm enough depth?

 

Uggli
I have not made a blade thinner the 3mm. I usually use 10% thickness or thinner so minimum chord is 30mm for me. (The milled one came in at 2.8mm after polishing)

You should seriously consider a Bolly prop. They are made in your state.
Bolly Props

Unit 8, 100 Hewittson Rd.
Elizabeth West, 5113
South Australia

Australia

Phone: +61 (08) 8255 9688 (please don't ring)

Fax: +61 (08) 8255 9666 (please don't ring)

Email is best : rc@bolly.com.au

(Emails are not likely to be answered)

Sales: Les, Lucy or Alex. (no more possible)

These are world class.

The attached test file is my profile data for a Bolly boat prop. This can be imported directly into JavaProp to check performance. You do this with the import function on the Geometry page - just copy and paste the data then import. You then use the Multi-analysis page to produce the data for the rpm set on the design page. DO NOT press the design button as this overwrites the imported data. The nearest blade section on the Geometry page is the E193 Re#100,000. Set all four points to this value. There is no need to set any particular value for the angle of attack.

Going on the JavaProp analysis you could expect 84% efficiency with the Bolly boat prop.

I usually try to set the prop 50 - 80mm below the surface. This can get air if I accelerate hard or get in some reasonable waves. Bear in mind that I do not have anything in front of the prop other than a tiny strut 2mm thick. With a drive leg you will get a low pressure trough on the back side that lowers the water surface. For this reason it is better if the prop pulls with clear water in front. There is also potential for a wave trough from the triangular bow wave that lowers the water level. It pays to have some ability to adjust height of the prop until these issues are assessed.

Problem with a forward facing prop is that it is prone to hitting things and suffering damage. I have gone to a lot of effort to get my shafts and props to be pliable so they bounce over things. The attached photo shows how my shaft will bend. I can pull the prop clear of the water if I want to inspect it. The CF Bolly prop is very strong though.

You will not be able to make any drive leg as efficient as a a curved shaft. I have seen drive legs that cost 30% of the total boat drag so they need careful consideration.

There is no benefit in shrouding a lightly loaded prop. The drag on the shroud will cost more than the reduction in induced drag.

Rick W