Model Turbines

[Turbine GuyParticipant @turbineguy]

I ran my new turbine without the reversing chamber with my airbrush compressor and it obtained a speed of 21,500 rpm turning the EP2508 propeller. My estimation of the power required by this propeller to turn at that speed is 3.4 watts. The power available to the turbine from my airbrush compressor is approximately 18 watts so the efficiency is approximately 19%. This is an increase in power of about 1 watt over what was obtained by my last turbine for the same amount of energy. The increase in power is primarily due to the increase in rotor diameter, increase in number of pockets, and reduction of the pocket inlet angle. The increase in the rotor diameter was from 0.892 in. to 1.226 in. This increases the rotor tip speed and torque. The increase in rotor tip speed raises the power but also increases the rotational (windage) loss. The rotational loss is still low enough at 21,500 rpm that even with the extra set of pockets it did not appear to be excessive. The increase in number of pockets was from 48 to 60. With single nozzles, the number of pockets under the nozzle discharge opening becomes very important since energy is lost in filling the empty pockets. The pocket inlet angle decreased from 30 degrees to 25 degrees. The smaller the angle, the less energy is lost from the flow not being perpendicular to the direction of travel. The power achieved was greater than expected so my estimations of power from the propeller might not be conservative. However, the increase in performance with each of the changes using the same propeller and airbrush compressor verifies that Dr. Balje’s guidelines are useful in the design of model turbines.
The following photo shows my test setup. Not very pretty but it worked.

[Werner JeggliParticipant @wernerjeggli14222]

Hello Mike,

I do not think the loss in pressure between boiler and turbine entrance is only due to friction losses in the piping. The main difference must be the more or less static steam pressure in the boiler and the reduction of pressure in the turbine inlet pipe due to its reduced cross section and therefore high speed of the medium (steam). (airplane wing effect). Pressure loss will however be compensated by an increase in entropy (the molecules move in the same direction). How to take this mathematically into account is above my capability.

[Mike TilbyParticipant @miketilby23489]

Hi Werner

From my understanding, increase in entropy of the steam as it flows through the pipe will be due to friction and that will give a rise in thermal energy content of the steam which is basically random motion of the water molecules. In a perfectly efficient nozzle there is no increase in entropy as the steam accelerates and all the available thermal energy is converted to kinetic energy as the molecules tend to move more in the direction of bulk flow.

When you say “airplane wing effect” are you referring to loss of pressure in air as it accelerates as it passes over the upper surface of a wing? If steam pressure drops as it flows along the pipe then the steam will have a larger specific volume (i.e. volume occupied per kg). Since mass flow rate is constant that means the volume flow rate must increase and so the velocity must increase. That will require energy and so the temperature of the steam will decrease. A small acceleration of low velocity steam will require very little energy since kinetic energy is propotional to velocity squared.

If I understand you correctly, then you think the velocity increases quite a lot in your pipe. Does that mean the pipe is very narrow?

But whatever loss occurs in the pipe, is it not the steam condition at the nozzle entrance that is important? As Turbine-guy said, if you attach a pressure gauge just before the nozzle, then you will know the pressure and temperature at the point where it matters.

What is the internal diameter of your supply pipe? Is it very small? Knowing the temperature and pressure at the nozzle entrance would allow you to calculate the velocity in the pipe and hence the kinetic energy in the steam. However, it is usual for any pressure drop along the main supply pipe to be small because the pipes are generally much larger in cross-sectional area than the narrow point of the nozzle. By the time steam reaches the narrow point of the nozzle it has expanded to a lower pressure and so the volume flow rate will have increased a lot. This and the small size of the nozzle means a high steam velocity. In the supply pipe the pressure is high so the volume flow rate is small. This fact plus the large cross-section of the pipe means the velocity is low.

I'll be interested to hear if you agree with any of this.

Regards, Mike

[Turbine GuyParticipant @turbineguy]

I also ran my new turbine on saturated steam at a gage pressure of 50psi (3.4 bar) and an estimated water rate of 0.55 oz./min. The maximum speed it turned the EP2508 propeller was 28,000 rpm. The estimated power required for the EP2508 propeller at this speed is approximately 9 watts.

[Ian S CParticipant @iansc]

A generator run on a small turbine, or any small power source would be pushing it to get to 50% efficiency, so there will be quite a difference between the electrical output and the brake HP measured with a Prony Brake, an interesting little experiment on it's own.

Ian S C

[Werner JeggliParticipant @wernerjeggli14222]

Hello Mike,

I really do not feel at home in the pressure/enthalpy/entropy world. I'm trying to follow Norman Billingham's excellent series of articles in the SMEE Journal, but it's hard work and my grey cells are old and probably encrusted. If you tell me the measurements at the turbine entry can be taken at face value for further calculations – that's fine.

Yesterday, with the help of 2 friends (to also take simultaneous readings) we ran the test with 2 nozzles in action – and it was a total failure. The boiler couldn't keep up with the 2 nozzle steam demand. We will have to repeat it. With just one nozzle it should work.

To give you an idea of the hectic involved – have a look at the set-up!

and the expected data

[Turbine GuyParticipant @turbineguy]

Hi Werner,

Sorry your test didn't go well. Your test setup and spreadsheet look very professional. I look forward to seeing the results when you get a good run.

[Mike TilbyParticipant @miketilby23489]

Hello Werner

That certainly does look a good test set-up. I hope you have better luck next time.

I like your use of a condenser coil to measure steam consumption. Is it just a coil of copper pipe? Does it work well? What diameter and length do you use?

What type of thermocouple do you use? Are they encased or bare ended?

Regards, Mike

[Werner JeggliParticipant @wernerjeggli14222]

Mike,

The copper pipe is 10 x 8mm, 6 Windings, 350mm OD, height 140mm, installed in a can. Can filled with ice cubes and water prior to test.

The thermocouples are bare ended. They came with the handheld measuring instrument, *roline 307" , 2 measuring points. This was years ago.

regards Werner

[Werner JeggliParticipant @wernerjeggli14222]

Mike,

Sorry, the coil OD is 85mm!

Werner

[Turbine GuyParticipant @turbineguy]

My test setup for running on steam was quick and dirty but sufficient to get enough information to estimate the efficiency of my turbine running on a single row of pockets. The following chart summarizes the results of all the testing I have done so far. The Ns Ds efficiency shown is for the optimum axial turbine (the heavy solid lines) of the Ns Ds diagram I added to the 14/03/2019 post. The optimum axial turbine would require the blades machined similar to what is shown on the post of 18/04/2019. The details, descriptions of the changes, and tests of each turbine are described in previous posts on this thread and in the Testing Models thread. The steam was supplied by a Stuart Turner model 504 boiler and the assumed temperature was for saturated steam at the inlet pressure. I assumed the modest superheat of this boiler was lost to the cold turbine housing with air from the propeller blowing over it.

[Blue HeelerParticipant @blueheeler]

Great thread, thanks!

[Werner JeggliParticipant @wernerjeggli14222]

Gentlemen,

Yesterday, we repeated the test. It was halfway successful. In previous tests I got better output. In addition, there is a discrepancy in the turbine pressure/temperature values. What do you make out of these data?

[Mike TilbyParticipant @miketilby23489]

Thanks Werner, for posting the details of your condensor arrangement. I think I shall make something similar for my tests.

It is interesting to see the data from Turbine Guy and Werner.

Werner, why do you think there is a discrepancy in your temperature / pressure? As far as I can see, the saturation temperature for steam at 4.0 Bar gauge pressure is 152 deg. C. The temperature you measured (148 deg C) was only slightly below this. Do you not think it is within experimental error? Or is there another problem that I have over-looked?

Did you calibrate your thermocouple-based measurement against a known thermometer or are you relying on theoretical voltage outputs from the thermocouple and assuming an accurate amplification factor in your circuitry? The temperature of your exhaust steam is also slightly below the expected temperature of 100 deg C and that is why I ask the question. If both measurements were 4 degrees too low then would not the inlet temperature be as expected after applying a correction?

Mike

[Turbine GuyParticipant @turbineguy]

I hope that I’ve demonstrated that with careful attention to details, the open pocket tangential impulse turbine can be reasonably efficient and produce useful power. I still believe the axial impulse design as given in the post of 18/04/2019 would be the best if made as described. This would require the rotor cutter to move in a straight line at the entrance, run at a constant radius in the center, and then run in a straight line at the exit. It would also require a large number of blades (at least 60) and for small rotor diameters, very small cutter diameters. In addition, very small nozzle and blade angles would be needed and the blade height should be just slightly larger than the nozzle exit diameter. The intent is to keep the cross-sectional area of the flow passage through the blades constant so that the steam or air can’t expand and lose velocity. The post of 18/04/2019 described and illustrated the requirements given by Dr. Balje.
The open pocket tangential turbine’s biggest disadvantage is being open on the top. Even with close clearance to the housing, the flow is unrestrained in at least one direction. This allows the flow to expand and slow down as it moves through the pocket. The momentum of the steam or air holds the flow against the surface that’s changing the flow direction but does not prevent expansion. If you compare the turbine efficiency with the Ns Ds diagram efficiency in the chart of the 19/06/2019 post you can see what could be gained by using an optimized axial impulse design.

[Werner JeggliParticipant @wernerjeggli14222]

Hello Mike,

You were right. I calibrated T2 in the kitchen in a pot of boiling Water to 99°C. Now turbine steam exit is also 99°C. T1 I calibrated with steam going to atmosphere also to 99°C. The readings should therefore be correct in the future.

Schlieren is located 350m above sea level and the weather was nice – 99"C should therefore be good enough.

[Werner JeggliParticipant @wernerjeggli14222]

Hello Turbine Guy,

The different systems of units (imperial/metric) are really a pain in the ass !!

With a steam throughput of 0.555 oz/min (0.26 g/sec) at 50 psig (3.5 Bar) saturated steam you get a shaft output of 9.34 Watt.

My steam throughput is 0.287g/sec at 4 bar. Saturation estimated at 0.9 and shaft output of 8.7 Watt. So the results are more or less in the same ball park.

And then, how are the ball bearings doing? How much of the precious power are they eating away? My wheel side one went bust yesterday! I replaced it with a full ceramic 3x7x3mm. We will see how long it will hold.

Werner

[Turbine GuyParticipant @turbineguy]

Hi Werner,

You’re right about the converting data from different systems is a nuisance. I use the imperial system since all my books and most of my research material are in that system. Our results for the turbine performance are similar. You tackled designing a complete steam locomotive while I concentrated on optimizing the efficiency of a single type of turbine.

Your questions about the ball bearings are very good ones. They have been a problem for me when operating on steam. The steam/water always finds a way to get into the bearings and spoil the lubrication. I never found an effective seal that would run at high speeds. The last full size turbine I designed was for a boat and had a closed system with the turbine exhaust condensed and pumped back into the boiler. I got very frustrated trying to keep the steam/water out of the bearings and the oil out of the boiler. I finally resolved the problem by using graphite bushings lubricated with water pulled through the bushings by the vacuum of the condenser. I have not had any problem running with air using my airbrush compressor. The stainless steel ball bearings I use are shielded, contain lubricant, and are rated for speeds much higher than I am running. I have used the same ball bearings for both my turbines without any problems. I only made the one test with steam and hope that it will not spoil the ball bearings. When I turn off the air the propeller spins for several seconds before stopping. The ball bearing friction appears to be negligible. Avoiding bearing problems was one of the major reasons I wanted to use air for my testing.

[Werner JeggliParticipant @wernerjeggli14222]

Hello Turbine Guy,

Please have a look at your inbox.

Werner

[Mike TilbyParticipant @miketilby23489]

Hello Turbine Guy,

I was very interested to read about your experience with turbine bearings. For the tubine I am building I originally planned to use plain phosphor bronze bearings with forced lubrication. Then I was pursuaded that ball bearings would be more efficient, although I worry about power being lost in forcing water etc out of the way of the balls. I buy my bearings from a helpful UK company called SMB bearings. They sell stainless steel bearings rated at 60,000 rpm but they told me that is their absolute limit and if my design speed were 50,000 rpm they would not last very long. They suggested small bearings used for dental drills which are rated at 200,000 rpm. They are stainless with ceramic balls, are fairly cheap and can be run either dry or in water. I bought these but have not used them yet. More recently I have been thinking of using plain graphite bearings as such bearings were used in a turbine built by another model engineer in the past. I have obtained some rods of graphite of a grade made for making seals etc. I gather that some grades can be too abrasive for such an application. Did you use any particular type of graphite for the turbine that you mentioned?

Have you seen the articles in Model Engineer written by Prof. Chaddock in the 1950s? His turbine ran at over 80,000 rpm in plain bearings with forced lubrication. Also he determined the power output using a fan air brake. This consisted of plain rectangular wood bars of specified width and thickness mounted on a shaft mid-way along their length. That shaft was geared down 10-fold from the turbine shaft. For bars of various lengths he gave graphs which directly show the hp required to spin them at various rpm. Apparently the theory for predicting the power requirements for these bars is well established and the bars need no further calibration. Please let me know if you don't have his articles.

Regards, Mike

[Mark RandParticipant @markrand96270]

If the turbine is running into a condenser, or you can afford the space for a gland steam condenser, you could follow full size practice and use non-contact labyrinth seal glands vented to atmosphere at higher pressure and to the condenser at the LP end. Then the bearings, on the outside, don't see the wet.

[Turbine GuyParticipant @turbineguy]

Hi Mike and Mark,

I just got back in town and have a lot of catching up to do, so I will need to be very brief. The steam turbine with the water lubricated graphite bushings was made in the 70's so I am relying on memory. The turbine ran at a speed of approximately 20,000 rpm and produced about 8 horsepower. There was a single stage V-belt speed reducer that reduced the speed to what was required by the boat propeller. The V-belts required a very high belt tension to reduce the speed in a single stage. The graphite bushings supported the full load of the V-belts and ran with no measurable wear after several hours of testing. I assumed the lack of wear was due to the water supporting the shaft similar to oil supporting the crankshaft in automobile engines. I was thinking about using graphite bushings if I decide to run on steam again and found some information that I would be glad to send if you message me your Email address.

Since my my turbine uses a propeller for a load, it would be relatively easy to measure the actual running torque. I have not considered doing that because I can get enough information with the methods I have been using. I do not have access to the ME documents other than what friends have shared with me.

[Turbine GuyParticipant @turbineguy]

I checked my turbine that has not been run for several days after the last run I made on steam. When I tried to spin the propeller by flipping it with my finger, the propeller wouldn't turn even 1 full revolution. Up to this test using steam, if I flipped the propeller with my finger it would spin for several seconds. Apparently the steam/water got into the bearing and gummed up the lubrication as I was afraid might happen. I might try the bearings for dental drills Mike Tilby described in the 26/06/2019 post if I run on steam again. I plan on using the same type of stainless steel ball bearing I have been using for the testing with air since they worked fine until I ran on steam. Also, using the same bearings eliminates one variable when comparing the performance of each change.

I read the articles Professor Chaddock made in the 1950's that Mike Tilby was kind enough to send me and would highly recommend them to anyone with interest in model turbines. Professor Chaddock was never able to completely solve the problems he had with bushings. It was interesting that he tried carbon that he described as being very abrasive that worked somewhat successfully. Perhaps if he was not running with highly superheated steam from a flash boiler and had a material with more graphite in it, it might have been even more successful. I also found his use of spinning bars for a load similar to my spinning a propeller interesting.

[Turbine GuyParticipant @turbineguy]

I designed the rotor shown on the following drawing for Werner Jeggli. Werner had the rotor made with 3D printing by ECOPARTS. He did the necessary machining and ran a test with this rotor. The result was a disappointing. 14V, 100mA, 26'500rpm, at 3 Bar. He removed the shroud by turning it down and checked the blade profile. The picture in the next post shows the rotor with the shroud removed. It seemed to him that the blade shape was not exactly as designed, probably due to manufacturing restraints. It also seemed to him that some of the passage ways were obstructed. He put on a new shroud, put in new bearings and ran the turbine at 3 Bar with the 0.8mm nozzle. Result 18V, 120mA, 30'000rpm. Resulting shaft power of 3.2 Watt. This was much inferior to the results he got previously with his existing rotor. Apparently, the narrow passages and required close tolerances of this design was not appropriate for the 3D printing process. I’m sorry that this did not work out for Werner and hope that something can be learned from all his effort.

[Turbine GuyParticipant @turbineguy]

The following is the picture of the rotor with the shroud cut off I mentioned in the last post.

[Author]

Posts