DIY Tesla Impulse Turbine

[The Ignoble Troll]

I'm building a giant leaf blower

The turbine output shaft will be coupled directly to a centrifugal air compressor, either a standard bladed design, or a Tesla pump, if it proves workable.

I remember reading you wanted it to pump air, I did not realize it was the pure goal. Awesome.


Can you use water cooling on an intercooler to preheat your water, or are you going for a single stage blower?


What kind of pressure are you looking to build?

 

[The Ignoble Troll]

https://www.google.com/url?sa=t&sou...ChAWegQICBAB&usg=AOvVaw08H6Q2BOi4nq0fBHlZGE_M


This paper examines chocked flow conditions.

 

[Toymaker]

I remember reading you wanted it to pump air, I did not realize it was the pure goal. Awesome.


Can you use water cooling on an intercooler to preheat your water, or are you going for a single stage blower?

My final design calls for a closed system, so I will be using condensers to recover & re-circulate the water. Using circulation pumps, temperature sensors, and another Arduino computer (digital microcontroller), I should be able to keep the condensed water at close to 100 C in the hot well tank, (which is another factor driving my decision on which feed pump to use). The pumped feed water is already pre-heated

A single stage is actually best for my needs.

What kind of pressure are you looking to build?

Typical bladed centrifugal compressors using aluminum impellers can develop 40 to 60 psi, which should work well in my application. Using stainless or titanium impellors can provide 90 to 110 psi, which I hopefully will not need.

I've never been able to find max air pressure output for Tesla compressors, so I'll be experimenting again

 

[The Ignoble Troll]

https://www.researchgate.net/figure...-SPACINGS-2500-RPM-MOTOR-SPEED_fig3_280385113

 

[The Ignoble Troll]

https://www.researchgate.net/figure...-SPACINGS-2500-RPM-MOTOR-SPEED_fig3_280385113

I wonder if it would follow water trends.

 

[Toymaker]

I wonder if it would follow water trends.

I have doubts. The difference in mass/volume between water and air, and the difference between compressible vs incompressible flow are just too great.

 

[The Ignoble Troll]

I think the air would be incompresable until you hit supersonic ejection from the blades, so there might a neat tick in the graphed output.

Will you use your arduino to run a mass flow sensor, anometer and thermometer and graph the data?

Edit: the pump is dynamic displacement which is why I have the hypothesis that disc tip speed will need to exceed the speed of sound to get good pressure.


Edit again for spelling

 

[Toymaker]

I think the air would be incomprehensible until you hit supersonic ejection from the blades, so there might a neat tick in the graphed output.

Will you use your arduino to run a mass flow sensor, anometer and thermometer and graph the data?

Edit: the pump is dynamic displacement which is why I have the hypothesis that disc tip speed will need to exceed the speed of sound to get good pressure.

At the moment, I have no plans to gather that much data,...but that could change

 

[Nerd1000]

My final design calls for a closed system, so I will be using condensers to recover & re-circulate the water. Using circulation pumps, temperature sensors, and another Arduino computer (digital microcontroller), I should be able to keep the condensed water at close to 100 C in the hot well tank, (which is another factor driving my decision on which feed pump to use). The pumped feed water is already pre-heated

A single stage is actually best for my needs.



Typical bladed centrifugal compressors using aluminum impellers can develop 40 to 60 psi, which should work well in my application. Using stainless or titanium impellors can provide 90 to 110 psi, which I hopefully will not need.

I've never been able to find max air pressure output for Tesla compressors, so I'll be experimenting again

I would assume the potential pressure ratio of a tesla compressor is similar to that of a normal centrifugal one. So practically speaking perhaps 3:1 or 4:1 per stage.

What will be interesting is to see if it has a surge line like a bladed centrifugal compressor.

 

[HMEL]

Indeed it is viscous drag. Take note of the differences between turbulent and laminar boundary layers too, they will have an impact on the turbine's behaviour. However this doesn't explain how it can be operating as a reaction turbine.
One potential hypothesis that comes to mind is that the fluid flow within the rotor actually acts as a nozzle of sorts, causing the fluid to accelerate tangentially as it spirals inwards and giving extra momentum that can be transferred to the rotor.

The classification of this being a reaction turbine does not fit the definitions of classical turbine design because of how the pressure drop across reaction blades is defined. It appears that the maximum disk velocity should be no greater then the fluid velocity leaving the nozzles. As the disk velocities increase the centripetal forces will force the fluid to move outward in a curved path also increasing the area of contact with the fluid and disk. This is a matter of conservation of energy. There is a pressure drop from the inlet to the outlet or the fluid would not flow. And if energy is transferred to the disk to do work there has to be a drop in enthalpy if one believes in the first law of thermodynamics.

There is probably a way to calculate these parameters but it would be interesting to read Teslas original work on the subject. However it is obvious that unless one reaches the high rpm the turbine will never achieve a 90% efficiency. Teslas original turbine was only able to reach 10,000 rpm before mechanical failures occurred. His first turbines exceeded 30000 rpm with resultant disk failures. I see two problems to solve. One a bearing capable of handling this rpm and two a disk to handle the stresses.

 

[Steamchick]

I know they work, from models I have seen spinning Tesla turbines with compressed air. But I am really curious as to how they work with condensing steam.... The steam as a gas has gas characteristics, the water as a liquid is completely different, and there is a mis-match in my "spinning" head as to the gas flow versus liquid flow that will occur. The gaseous steam will work - as planned - but then the liquid condensate - presumably at the surface of the discs - will be shed radially outwards, carrying with it some rotational momentum/energy that will be lost to the outside casing... and must be drained to avoid liquid drag on the edges of discs..
I don't know how Tesla managed this, or anyone else?
But I may be completely wrong, if the condensing steam acts like an aerosol and the water droplets are simply fine particles carried through with the gaseous steam? - In which case, I figure their momentum will simply continue to be able to be transferred to the discs if the boundary layer does not deposit water onto the discs?
Any ideas? - or have I missed the understanding of steam condensing inside the disc slots?
K2

 

[NapierDeltic]

I have the same view. That's why I recommend trials avoiding condensation as much as possible. I'd say there's no big deal in using different setup - just for testing.
I would say for a reasonable fraction of condensate, it remains in suspension.
On another point of view, the model considering entire pressure drop in the nozzle is not realistic (not entirely). Any resistance to gas flow - section change, direction change ... translates to static pressure drop. This will confirm anybody who deals with fluid dynamics. And I see a lot of reasons for this happening in Tesla turbine's design.

 

[Nerd1000]

I have the same view. That's why I recommend trials avoiding condensation as much as possible. I'd say there's no big deal in using different setup - just for testing.
I would say for a reasonable fraction of condensate, it remains in suspension.
On another point of view, the model considering entire pressure drop in the nozzle is not realistic (not entirely). Any resistance to gas flow - section change, direction change ... translates to static pressure drop. This will confirm anybody who deals with fluid dynamics. And I see a lot of reasons for this happening in Tesla turbine's design.

It would be more accurate to say that (almost) the whole enthalpy drop happens in the nozzles. Other pressure drops from 'non ideal' characteristics of the machine represent a rise in entropy, which doesn't do work.

A nice side effect in pressure compounded turbines is that some of the losses in each stage (due to friction of the steam on the blades and so forth) reappear as heat added back to the steam. That increases the enthalpy available to subsequent stages, and thus the overall turbine efficiency. Or you could look at it as expanding in smaller steps being more reversible and thus creating less entropy.

 

[NapierDeltic]

Fair enough!
Speaking about the real Napier Deltic, that was a nightmare for its designer!
That-s why I can't understand why a relatively simple mechanism and thermodynamic device like Tesla Turbine is not scientifically described with a reasonable degree of precision; either as a pure theoretical study or an application study. If it is a challenging device, that would be exactly the reason to be approached by those highly qualified scientists who have all means in their hands.

 

[NapierDeltic]

I know FEA is traditionally linked with solid materials and their behaviour correlated to physical shape and dimensions; but as a principle I am sure it translates to programs dedicated to fluids and their properties, considering interactions and modifications of elemental volumes of fluid inside a finite, geometric well described volume (like Tesla Turbine is) .

 

[NapierDeltic]

Deleted (just edit error).

 

[Nerd1000]

Fair enough!
Speaking about the real Napier Deltic, that was a nightmare for its designer!
That-s why I can't understand why a relatively simple mechanism and thermodynamic device like Tesla Turbine is not scientifically described with a reasonable degree of precision; either as a pure theoretical study or an application study. If it is a challenging device, that would be exactly the reason to be approached by those highly qualified scientists who have all means in their hands.

I think the reason is that practical Tesla turbines usually have rather poor performance compared to bladed turbines, so they are just not often used.

As I understand it Tesla turbines have received a bit more development as pumps, they have certain advantages when handling fluids with entrained particles and so forth. And they work better with viscous fluids than they do with gases.

 

[HMEL]

I know they work, from models I have seen spinning Tesla turbines with compressed air. But I am really curious as to how they work with condensing steam.... The steam as a gas has gas characteristics, the water as a liquid is completely different, and there is a mis-match in my "spinning" head as to the gas flow versus liquid flow that will occur. The gaseous steam will work - as planned - but then the liquid condensate - presumably at the surface of the discs - will be shed radially outwards, carrying with it some rotational momentum/energy that will be lost to the outside casing... and must be drained to avoid liquid drag on the edges of discs..
I don't know how Tesla managed this, or anyone else?
But I may be completely wrong, if the condensing steam acts like an aerosol and the water droplets are simply fine particles carried through with the gaseous steam? - In which case, I figure their momentum will simply continue to be able to be transferred to the discs if the boundary layer does not deposit water onto the discs?
Any ideas? - or have I missed the understanding of steam condensing inside the disc slots?
K2

I like to think of it this way. First you really do not want moisture in the steam or air. But if it is a mixture there will be a mass difference and the liquid particles will likely move to the outside like the old dairy cream separators. But I do not see the steam condensing because that requires either compression or a heat loss. But you have proposed a good physics experiment and asked a dam good question.

 

[HMEL]

I know FEA is traditionally linked with solid materials and their behaviour correlated to physical shape and dimensions; but as a principle I am sure it translates to programs dedicated to fluids and their properties, considering interactions and modifications of elemental volumes of fluid inside a finite, geometric well described volume (like Tesla Turbine is) .

The technique used is computational fluid dynamics. The trick is putting the paramaters of the equations together , A 3d compliation on a computer costs a minimum of 70 dollars / hour not including the time to do the coding and check the results. Most of this work is proprietary and its difficult to obtain good information. There are graduate engineering classes on the subject but software is difficult to come by. But I have a couple of texts on it and you have made me curious if its possible to do something with it at a minimal level.

 

[Steamchick]

HMEL. I did the test today. Having added a 2mm drain hole at the perimeter of the end cover, the turbine now drips a lot of condensate when warming-up, then gives a continuous drip when running hot. A lot of condensate is expelled as wet steam from the exhaust.
My conclusion is that while the condensate is probably removing energy/ momentum from the rotor, it is being expelled to the casing to drain out the small drain hole. Aerosol water vapour travels through the turbine releasing most of its entropy to the discs.
I'll attach a video later.
K2