My Jungle hydro project- 5kw pelton to power future eco-hostel

[junglehydro]

Hi All, after some reccomendations from other users I decided to start a new topic to document the entire construction of my pelton turbine, and to share as many resources as possible, that they might be of use to someone else in the fututure.

A little info:

I live with my girlfriend and our 3 dogs and 3 horses in the Sierra Nevada mountains/rainforest of Colombia, where I am building my dream hostel. A truly sustainable off-grid adventurous neverland for grownups. Think treehouses along a jungle river, waking up with the sound of monkeys and toucans all around and gorgeous waterfalls within short walking distances in all directions.

Our site is on our neigbors property and I have been scouting for the perfect location to place the turbine and I believe I have found it. It will take a few more measurements that I hope to complete in the next few days and then the site preparation can begin.


the creek

Me (on the left) and a volunteer measuring the height of the hill next to the creek.
We use a very crude system of a hose with water on a Y shaped stick as a level, and a stick marked every meter to measure the height. It is crude but works quite well.

The system:

I am planning to build a 4" penstock with about 50 meter head, to power a 250mm pelton wheel connected to a 5000 watt dynamo. Since the power this system will produce will hopefully be near 5000 watts, this will be our peak usage and we are opting to go without any batteries or inverters and use our power directly from the turbine. This means that we will need to create a diversion load controller that will wisk away any power not being used to a water heating element, thus producing warm water for the hostel as a waste product from the turbine.

I have found that reading the book "the micro-hydro pelton turbine manual" by Jeremy Thake has helped me a great deal in understanding the task that lies ahead. And even though the books goes way too deep into the material for my understanding at times, it is very complete and has been a great starting point for me. I also recommend anyone to look at the materials, videos and calculators on the powerspout website.

I found a great 3d model that was made according to the specifications of Jeremy Thake and that was free to download. I 3d printed it ad a workshop and now I am having it cast in aluminum 20 times, and then sandwiched between two stainless steel plates to make the wheel.

And have designed an enclosure for the wheel with two jets. the piping will have threading so we can remove nozzles and experiment for optimization, and both inlets will have a ball valve to seal them of separately.


I am planning to expand this first post in the future to make the info more complete as the project moves along.

[junglehydro]

Currently I have no idea on the sizing of the belt-drive wheels, since I don't know how many revolutions I will get with my system.
My dynamo needs 1800 rpm for 5000 watts, so say my pelton wheel does 600 I need a 3:1 ratio.

Anyone know of the formula to calculate my expected RPMs so I know what size wheels to use?

[SparWeb]

You need the exit velocity of the water from the nozzle.  That depends on nozzle geometry and the head.
Do I recall correctly that the charts I gave you on the previous thread give you some idea of that?

I don't have Thake's book, so I need to ask:  Does Thake get into the definition of power?  You need both flow and pressure.  The nozzle you have diagrammed may have a 1/2" port.  Is that enough flow to make the power you want to get?  Do you only have 1 nozzle?

[SparWeb]

Here's a photo of a pelton with 2 nozzles pointed at it.
I have seen an arrangement with 4 before, perhaps in Home Power magazine.

[JW]

https://www.fieldlines.com/index.php?topic=148325.25



This is from FL

[SparWeb]

That's a poncelet wheel.  It's for trapping flow without having pressure, or a conduit.  More like a water wheel in a river.

[clockmanFRA]

This is a low cost manufacturer in New Zealand.

https://www.powerspout.com/

They have been making them for many many years, and have perfected small hydro systems and have a good track record.

[junglehydro]

You need the exit velocity of the water from the nozzle.  That depends on nozzle geometry and the head.
Do I recall correctly that the charts I gave you on the previous thread give you some idea of that?

Alright so from what I can figure out I would get a velocity of discharge at around 102 feet per second at my site(50m/165 feet of head), giving me a discharge of 250 GPM (per nozzle, so total would be double that) However from the powerspout advanced calculator I get a total possible water flow of about 370 GPM for my give penstock size (4"/ 10cm)
The size calculations for my nozzles (a total of around 850mm²) do match up between the powerspout calculator and the table that you posted in the other thread. But looking at the info in the goulds sheet it looks like 2  25mm nozzles would drain my penstock...

So how would I convert those 102feet per second into RPM's for my wheel?

I am planning to put ball valves on both pipes that lead to the nozzles, should I regulate my flow with them to prevent my penstock from draining? Or will that give me an efficiency loss that is too great and should I opt for slightly smaller nozzles?
Right now I am leaning to make the nozzles tapered cones with an opening of about 22mm, and that I can just shave of a bit of the tip to make the opening wider to experiment with the outcome.

I don't have Thake's book, so I need to ask:  Does Thake get into the definition of power?  You need both flow and pressure.  The nozzle you have diagrammed may have a 1/2" port.  Is that enough flow to make the power you want to get?  Do you only have 1 nozzle?

From the table i concluded that with my head and flow, and the fact that I am using 2 nozzles the nozzles are going to be 25 mm in diameter. They re going to be removable however, so I will be able to experiment with different sized nozzles and see what works best. But your chart definitely helps taking most of the guessing out of it.

The book goes into power as in it gives the formula :

So let me see if I can work with that and figure it out :-)

This is a low cost manufacturer in New Zealand.

Powerspout

Hi Clockman, yes as you can read in my start post I know of powerspout. I have chosen to build something with local parts so that I figured it would fit our usecase better and the maintenance in the future might be easier if it is made with local parts. Plus this process allows me to learn more about these systems than just installing something off the shelf, so I am better equipped to do any future work on the system.

Here's a photo of a pelton with 2 nozzles pointed at it.
I have seen an arrangement with 4 before, perhaps in Home Power magazine.

Isn't that losing a lot of efficiency by the nozzles being so far away from the wheel? I am trying to get them as close to the wheel as I can get them without them actually touching, but from looking at this design I'd think that having 10cm between my wheel and Nozzle doesn't present such a problem.

[SparWeb]

Hi JH,
One thing to check is the volume of the flow in the penstock (Area * velocity) and compare with the nozzle flow volume (Area * velocity).  Put that way, kinda obvious that they'd be equal, but the reason to point it out is to make sure it's considered in your math.  It can give you a solved variable that otherwise wasn't certain.  A few calculations to size up the nozzles, penstocks, and wheel in general, followed by experimenting with nozzle sizes, should allow you to tune in the generator to the point you're happy with it.

I haven't experimented with these personally, so I can't say anything useful about the distance of the nozzles from the wheel.  I guess it would depend on how well... "collimated" the "beam" is... I don't think I'm using the right terms; I hope you get what I mean.  There's another factor that getting the water OUT of the compartment may also be an issue.

[bigrockcandymountain]

So to calculate rpm from velocity, you can just use your basic geometry.  Circumference is diameter x 3.14
Say a wheel with diameter 2 feet x 3.14 is 6.28' circumference. 

102fps discharge divided by 6.28 is 16.24 rps or 975 rpm.  That is unloaded rpm.  Under load, your wheel will run slower than that.  How much slower is optimum? I'm not sure but i bet the knowledge is out there.  I would estimate 25% to 50% slower might be close. 

Another question would be is it better to have a large low rpm pelton wheel or a small high rpm.  I would think you are close thinking a 3:1 ratio with your pelton running 600rpm and your generator running 1800 rpm but that would be a good question to ask. 

I really know nothing about hydro, so I'm mostly just thinking about questions that need to be asked. 

I would suggest the inside of the nozzle is pretty important to make the right smooth shape. You want laminar flow exiting the nozzle, so any valve pinching should be done a distance back from the nozzle and in the larger diameter pipe.  That is just a guess, but I could see messy turbulent flow near the nozzle really hurting efficiency.  As spar would say, you want a nice "collimated beam" of water.    I don't know any better term, so i better not laugh.

[mab]

I think when i was looking into my pelton system it was stated somewhere that the optimum spoon velocity was approx 1/2 the jet velocity for max power xfer to the wheel; in practice it will vary slightly depending on the exact spoon shape and where the jet is actually hitting the spoon - if the water hits the spoon and comes off still with forward velocity the wheel's spinning too fast; if the waters coming off back towards the nozzle the wheel's too slow; if the water comes off sideways it's about right. In practice however, with my small wheel the water's going in all directions so fast I can't really tell TBH.

At the risk of being a pedant: the generator Sparweb showed looks like a Turgo wheel rather than a pelton (judging by the angle of the nozzles to the wheel); possibly H-hydro blue spoons? It may be worth considering a turgo - I found that it was possible to get a smaller diameter turgo than pelton which accepts the same max jet/nozzle diameter, which in turn allowed me to get higher shaft RPM (i only have 20m head). I say this as direct drive of the generator will always be more efficient than a transmission, and with one less thing to wear out. The efficiency of a turgo is comparable to that of a pelton i believe.

[junglehydro]

Hi JH,
One thing to check is the volume of the flow in the penstock (Area * velocity) and compare with the nozzle flow volume (Area * velocity).  Put that way, kinda obvious that they'd be equal, but the reason to point it out is to make sure it's considered in your math.  It can give you a solved variable that otherwise wasn't certain.  A few calculations to size up the nozzles, penstocks, and wheel in general, followed by experimenting with nozzle sizes, should allow you to tune in the generator to the point you're happy with it.

Yeah so I am pretty sure that we'll start with less water and a less than optimal power production, since we don't have enough length of pipe to feed the penstock with enough water for 5kw. But we don't need that much know, and it's something we can easily scale up in the future by adding another pipe to the water bay that feeds the penstock. and we'll have to adjust the nozzles probably a few times as we go up in flow. Right now I would settle with a constant 500-1000 watts so we can run everything we need to build our place.

I think when i was looking into my pelton system it was stated somewhere that the optimum spoon velocity was approx 1/2 the jet velocity for max power xfer to the wheel

Yeah I remember reading that as well, that usually runaway(unloaded) speed is about double of the optimal speed. ( I think in powerspout documentation, so not sure if thats a general rule or applies to powerspouts only.) If I use the calculation from bigrock I come to about 2400 rpm unloaded, so then loaded would be 1200. so a 2:3 wheel ratio should do the trick in that scenario. I guess I should see if I can find a Tachometer to measure rpms somewhere once I have everything setup is still the most reliable way to figure out my true RPM.

[bigrockcandymountain]

If you have the generator hooked up, you can just measure the frequency in hz for rpm indication too.  A tach is a good idea though. 

[junglehydro]

Hi guys and girls a quick update here, with some new questions!

So I've decided to start a bit smaller with a creek that is quite a bit closer, but will dry up a bit in de dry season, so somewhat less production then. The big turbine is too much of a hassle with transport and infrastructure being such a problem at the moment, and that capacity will not really be needed for a few years anyway..

I am going to install a powerspout in a closer creek with 85m of head and am thinking about hooking it up to our current solar system.
Am I correct in thinking that basically all it needs extra is a diversion load so my turbine won't overspeed?
I am thinking about hooking up a Xantrex c40 in diversion mode.
I now have a all in one inverter chargecontroller, a powest 3kva, hooked up to two 100ah 12v lithium batteries

The turbine will generate about a 1000 watt continuous for about 9 months a year and then drop down to 400 watts at the minimum for the rest.

[bigrockcandymountain]

1000w continuous is awesome anyway. 

A c40 is only able to do 40a.  I can't remember if you are 12v or 24v.  At 12v you'll need to dump 83a and at 24v 43a.  Either way a c40 is a bit small.  Float voltage of a 24v system is 26v or so, so i guess the c40 might be ok.  I like to oversize a bit but that's up to you. 

If you are 12v, 2 c40s or something like that will do the trick but will be close to max just like 1 at 24v.

A ts45 tristar would give more margin for error and they have a better reputation.

Are you heating water with the dump load?

Good to hear from you again.  I have been wondering how your project was working.

[junglehydro]

Yeah my system is 24v, thanks for the tip on the tristar, it is good to have a little bit more headroom I guess, so that does look like a better bet than the xantrex. And thanks, I didn't know about having a better reputation, so that is definitely good. Even if this system fails and flips a breaker, it should be within safe overspeed velocities for the turbine according to the calculator.
But from what I think I've read powerspouts have their own diversion load built in as well right? That functions like a last resort in case there is no load measured at the turbine to prevent overspeed I suppose?

Yeah my plan would be to heat water with the diversion load. Will be a nice side project to build a nice hot water system that doesn't overheat and provides us with nice hot water for our showers, but that is for later. For now I' ll just stick the heating elements under flowing water.

Yeah glad to be able to post something again, hoping to make some faster progress with some updates this time around!

[camillitech]

But from what I think I've read powerspouts have their own diversion load built in as well right?

Hi JH,

yes the Powerspout has a perfectly adequate dumpload that is cooled by the exiting water. I have run one for years, it is a very robust and simple unit.

Good luck with your project, Paul

[camillitech]

But from what I think I've read powerspouts have their own diversion load built in as well right?

Hi JH,

yes the Powerspout has a perfectly adequate dumpload that is cooled by the exiting water. I have run one for years, it is a very robust and simple unit.

Good luck with your project, Paul

Just bought another Powerspout cos my Stream Engine failed after 12 years service and it was uneconomical to repair. Figured it made more sense to run two turbines from the same manufacturer then I could utilise the spare parts I already have. The turbine that arrived doesn't have a built in dump load like my original, not sure if this is because my ten year old one is a grid tied version though. I run a grid tied version at this site cos of the lower transmission losses at the higher voltage. I then run it through a GTI into my own mini grid



This one charges my 48V bank via MPPT controller, well it will once I finish wiring it up :-)

[Scruff]

This one charges my 48V bank via MPPT controller, well it will once I finish wiring it up :-)

^ ? MPPT?

[camillitech]

Aye Scruff,

I've got the Victron one, my original turbine was three phase up to the shed then rectified and transformed to charge my batteries. This one is rectified at the turbine and I'm gonna put my Powerclamp and dump loads in the shed.

Cheers, Paul

[Scruff]

Cheers Paul.

Apologies about my last response...I've regained my faculties now.

I've quite curious how to MPPT a turbine because I expect it's a solution MS don't cater for very well. The TS MPPT has a forced power curve profile that you can plot...I expect from trial and error. I hear it's prohibitively complex to implement with seems to make it exclusive to average users who often prefer a MidNite instead for simplicity.

I haven't tried either so I can't say. Just curious...
If I ever get a proof of concept together I'll let you know.
I've bookmarked your home site so I can copy your answers later.

Peace Out.

[mab]

If i might answer your question to Paul:
It's actually quite simple to use mppt on a hydro generator as the output has a positive 'knee' I-V curve same as a solar panel (albeit not nearly so sharply curved) so a standard solar mppt controller will find the MPP.

I sometimes use a sunny boy 1200 on mine. The only 'issue' is that the 'E-today' figure on the display keeps going up and up until about 128kWh when it clocks round to zero, as it normally resets when it goes 'dark'  .

Oh, and you still need a dump/diversion load in case the mppt cuts out/unloads the generator  - not as critical as a wind turbine  as the hydro generator will cope with free running, but the unloaded voltage (double or more) may well exceed the max input of your mppt controller.

[clockmanFRA]

 

[camillitech]

If i might answer your question to Paul:
It's actually quite simple to use mppt on a hydro generator as the output has a positive 'knee' I-V curve same as a solar panel (albeit not nearly so sharply curved) so a standard solar mppt controller will find the MPP.

I sometimes use a sunny boy 1200 on mine. The only 'issue' is that the 'E-today' figure on the display keeps going up and up until about 128kWh when it clocks round to zero, as it normally resets when it goes 'dark'  .

Oh, and you still need a dump/diversion load in case the mppt cuts out/unloads the generator  - not as critical as a wind turbine  as the hydro generator will cope with free running, but the unloaded voltage (double or more) may well exceed the max input of your mppt controller.

Hi Mab,

I ran an SB1200 for a year or two and found the same, it was a perfectly adequate inverter but then Hugh Piggott asked me to run a comparison with a Gin Long inverter for Michael Lawley of Powerspout. This I gladly did cos he supplied the Gin Long (now Solis) and I was doing a spreadsheet of readings twice daily anyway. So I ran the SB1200, Solis 1.5kW and an SB3800 for a month each and submitted the data to PS which they published at the time. The outcome being that whilst there wasn't much overall difference in performance The SB3800 being ever so slightly more efficient than the Solis which was better than the SB1200. However the Solis was by far the better inverter in features, menu, operation and price.
I'm only going down the DC coupling route cos I'm maxed out on AC coupled generation.

Cheers, Paul