Model Turbines

[duncan webster 1Participant @duncanwebster1]

I seem to think that interstage seals in Rateau turbines are non contacting labyrinth, so no seal friction.  If the bearing at the HP end is outside the casing then it needs a seal, but the LP end will be atmospheric

I’m not suggesting this is easy. A reciprocating engine/turbine compound had a lot going for it, used in big steam shops before they really hot turbines sussed

[Mike TilbyParticipant @miketilby23489]

In response to your question Byron and to Duncan’s comments, here is a summary of the design.

My plan is to initially get a single stage turbine running in the 4-stage housing / cylinder. Then add further stages into that housing, one at a time. The current cylinder is not split but I plan to make a new one that will be split before adding more stages. Both types of cylinder will use the same end covers which are not split and these have bores which will house inter-changeable bearings (see drawing). The first stage nozzles (see photo) are a bit rough being the first things cut with my micro-milling contraption and using a poor type of cutter. In Rateau turbines the stages are separated by so-called diaphragms. These house the nozzles. Unlike the stationary blading of typical reaction turbines, the diaphragms have small diameter bores which minimise the steam leakage area. As Duncan says, these bores are often sealed with labyrinths but quite a few turbines (e.g. those made by John Browns) had graphite ring seals. These were complicated components with over-lapping segments held together and held against the shaft by garter springs. I can post some pics if anyone is interested. Labyrinth seals are not suitable for tiny models as the necessary clearances would be unworkable. Multisegmented graphite seals are also not practical so I’ve adopted so-called floating ring seals. In these the graphite ring is a close running fit on the shaft and is held against a flat face of the stationary diaphragm. It is free to slide across the flat face and so can move with any vibration in the shaft.  Full size rateau diaphragms were split in order to allow them to be assembled in between the rotor discs. The central discs of my inter-stage diaphragms will also be split (see photo). As with the first stage nozzles, the edge of the disc is tapered to fit closely in a single piece ring with a matching internal taper.

The design of the hp end bearing keeps getting changed and the drawing shows my latest idea. The sleeve that holds the graphite liner will be shrunk onto the graphite. i.e. it will be heated, the graphite inserted and then allowed to cool. (This method is recommened by one of the graphite manufacturers). The sleeve will then be centred in lathe and the graphite bored.

Published designs for graphite bearings show delivery of lubricant at the mid-point and though a hole leading into an axial groove along the bore of the bearing. Arranging for water to be delivered via a radial hole at the mid-point will be difficult so I plan to deliver it to one end – if it is needed at all. (As mentioned in a previous post, the PV value may be less than the 15,000 limit for graphite). I think end delivery of water is worth a try and I’ll be able to assess the effect of lubricant delivery on rotor speed etc. I think enough water will make its through the running clearance of the bearing but, if not, then one possibility could be to make a very fine axial groove in the graphite to allow water to penetrate further along the journal.

That’s more than enough from me. Sorry to have taken up so much space on your thread Byron.

Mike

 

[Turbine GuyParticipant @turbineguy]

Hi Mike,

Great job showing what you plan to do. I tried to find examples of Rateau turbines but couldn’t find any that showed a better way. The way you show adding water to the graphite bearing would also work for cooling ball bearings like the dental ball bearings I am using that are compatible with water. Controlling the water with a valve will allow you to only add what is necessary.

You didn’t mention what the last picture showed. I assume they are the diaphragms from a large Rateau turbine. If so, they are not split and would have to be slid onto to the Rotor shaft and then the next Rotor added.  This complicates assembling the parts but might be worthwhile since it could simplify the housing enough to make this worthwhile.

Thanks for showing all this information,

Byron

[Mike TilbyParticipant @miketilby23489]

Hi Byron

I am encouraged by your comments about the bearing design – many thanks for your thoughts.

I know what you mean about those diaphragms looking like they cannot be split. My original design for my turbine was to have solid diaphragms because i thought making them in halves and steam tight would not be possible in such a small scale. But I later realised that it would be extremely difficult to get a balanced rotor by sliding onto the shaft rotor discs alternating with diaphragms. Once all assembled it would be  impossible to balance and if I balanced the rotor on its own and then dismantled and reassembled it with the diaphragms, the balance might well have been upset. Balancing a long rotor will be challenging even under the most favourable circumstances. I shall balance each disc separately but when they are all assembled things might change a bit and there s the problem of static balancing not revealing imbalances due to couples at different points along the axis. I  have been looking into making a dynamic balancing machine. With split diaphragms I hope to be able to assemble and balance the rotor and then not disturb it again.

I wonder if the poor quality of that old photo of the large diaphragms masks a fine slit line across them otherwise, maybe balancing was not done so carefully in those days. Here is a photo of a more modern turbine showing split diaphragms and also a photo of a Rateau rotor which clearly shows the deep narrow gaps between the rotor discs.

Mike

 

[ChrisLHParticipant @chrislh]

Just in case you are not aware, the model gas turbine people have made and used what looks like a rather sophisticated rotor balancing machine.

[Turbine GuyParticipant @turbineguy]

Hi Mike,

The last pictures you posted are what I was trying to find. They really show how the parts fit together on one example of this type of turbine. They also show the spherical journal bearing like Mr. Gordon used, and the labyrinth seals like Duncan mentioned. I like your idea of using the close fitting floating graphite rings to control leakage. I think floating the ring will keep the friction low enough to keep the graphite from over heating.

I agree the balancing of the rotor needs to be very good for the speeds you intend to run. I believe you planned to operate your turbine at speeds around 50,000 rpm.

Keep us up to date on your progress.

Byron

[Turbine GuyParticipant @turbineguy]

The Reentry Rateau Link is a report on a pressure staged reentry turbine. It collected the exhaust from the first stage and routed the flow in a duct with gradually reduced area that acted like a nozzle for the second stage. The following drawing copied from the report shows the concept. This looks very straight forward but the following photo shows that adding the ducts gets very complicated. Anyway, this was a test of the Rateau type of turbine that might have some useful information.

[Mike TilbyParticipant @miketilby23489]

Regarding Duncan’s earlier post where he mentioned the use of turbines operating on exhausts of reciprocators (the Bauer-Wach system) there is an example of this in the Model Engineer. Mr. Jan Hook described using his turbine with a ST D-10. e.g. the 6th and last part of his series is in M.E. issue of 3rd June 1988). This is a picture from his article – let me know if you want copies.

Hi Chris,

thanks for reminding me about the GTBA balancing info. I joined that association a couple of years ago to find out more about this and their other impressive achievements but I’ve not looked at the balancing stuff for some time. One off-putting aspect of this is that building a balancer will be another significant project. That is another reason why I hope to get a single stage turbine running initially since I hope that static balancing of that will give acceptable results.

Mike

[Mike TilbyParticipant @miketilby23489]

Hi Byron,

that paper you posted is fascinating – thanks for making it available. For me, the paper is also rather frustrating since the aerodynamiscists/gas turbine engineers who seem to have written it use different approaches to the analysis compared to what is used in my text books about steam turbines. It took me a while to understand the pressure compounded basis of the turbine which contrasts with the velocity compounding that had been the used for many years with re-entry steam turbines – as also in some of your own model turbines I recall.

It seems to me that a really big advantage of a multi-stage re-entry turbine (which they did not list on page 3) would be the avoidance of the fannage loss and disc friction loss that occurs in a multi-disc axial turbine with only a few nozzles per stage. I’m sure that such losses will be significant in my turbine especially as, unlike model gas turbines, steam entry has to be restricted to a tiny percentage of the rotor circumference. If there were nozzles all around the circumference, as in model gas turbines, I’m sure the turbine efficiency would greatly improve but one would need a huge boiler to supply sufficient steam.

I was surprised that pressure compounding would be attempted in a re-entry design because of the impossibility of preventing escape of air (or steam) through various clearances. Their blade clearances of 15 and 20 thou seem quite large to me. Have you heard of any other pressure compounded re-entry turbines?

Mike

[Turbine GuyParticipant @turbineguy]

I also didn’t like the way the report presented the information.  The information I found in the report was an inlet pressure of 54 psia, an inlet temperature of 280 degrees F, an overall enthalpy drop of 60 btu/lbm (20 btu/lbm per stage), an design blade speed of 382 ft/sec, and an efficiency of 62.2%.  For a 20 btu/lbm enthalpy drop the spouting velocity is approximately 1000 ft/sec and the blade speed/spouting velocity (U/Co) ratio would be approximately 0.38.  This speed ratio is where the impulse turbine should be at its peak efficiency which can be around 80% as shown on the following chart. The chart assumes maximum nozzle and rotor velocity coefficients and full admission. Each stage had partial admission so that reduced the efficiency but should not have been a very large reduction with each stage having several rotor blades overlapped and the three stages overlapping almost all the blades. As Mike said,this is a major advantage of this concept. The report stated that the pressure drop in the ducts reduced the efficiency about 7% so assuming their blade designs were good, the major part of the rest of the loss would be due to leakage around the blades. I agree with Mike, the clearances seemed very large.  The very low enthalpy drop and very large mass flow would make this turbine suitable for only special applications.

Dr Balje in his 1958 report “A Study of High Energy Level, Low Power Output Turbines” looked at this concept using his specific speed Ns and specific diameter Ds  explained in the 7 July 2021 post on page 13.  His report estimated the efficiency of up to 16 stages for the following conditions:
Stator angle 15 degrees
Number of blades 100
Minimum aspect ratio 0.75
Minimum h/D ratio 0.04 (h= blade height, D= rotor diameter)
Axial clearance ratio c*/D 0.001 (c*=clearance between the rotor blades
and stator blades)
Pressure ratio 16:1

If these conditions appear to be near enough to your conditions Mike, give us the information needed to find the Ns Ds values and I will see what Dr. Balje estimates for the reentry turbine with four stages. You can see in the above clearance limit his allowance is quite generous for larger turbines but is pretty hard to meet with tiny models.

 

 

[Turbine GuyParticipant @turbineguy]

The following spreadsheet I used to check the performance of Axial Turbine 4A (that uses the rotor Mike made for me) shows the data that is actually needed and converts the units to those used by Dr. Balje in his Ns Ds diagrams. For the 4 stage turbine, the total enthalpy drop would be used and what I show as specific volume, V3, would be the specific volume of the steam exiting the nozzle into the last chamber. I hope this makes clear what is actually needed to use the Ns Ds charts and I would be glad to send a copy of this Excel file.

I will see if I can find enough information in the link I gave in the 7 August 2024 post to fill out this spreadsheet and see what Dr. Balje estimates the performance would be for that 3 stage Rateau turbine.

[Turbine Guy]

On Turbine Guy Said:

I mentioned in the 29 July 2024 post that the steam stayed superheated all the way through the turbine in the test described in that post. I found the mass flow out of the boiler to be 1.4 lb/hr by the 11 minute time required to boil out ½ cup of water. I estimated the temperature of the steam entering the turbine by finding what the temperature and pressure at the throat of the nozzle had to be for this mass flow. This was a process of iteration guessing the temperature and pressure and using the corresponding enthalpy drop to find the throat velocity and specific volume of the steam.  The mass flow for dry steam can be determined with the throat velocity and specific volume. It took several iterations to find the estimated pressure and temperature at the inlet of the turbine that resulted in the mass flow found in the test. The following spreadsheet shows the estimated inlet temperature, pressure, and enthalpy drop for the test described in that post and the performance of the turbine for that test. On the second test I measured the housing temperature and it stayed at 211 degrees F for the entire length of the run after the turbine warmed up to its operating temperature. I also added below the spreadsheet for the test with air for comparison with the test with steam like I did in the 7 April 2023 post on page 20 for Tangential Turbine 5C. This post and the 7 April 2023 post show the performance on air and steam of the two best turbines I have tested.  Axial Turbine 4A discussed in this post is a traditional De Laval type of Axial Turbine and Tangential Turbine 5C discussed in the 7 April 2023 post is similar to the open pocket Stumpf type of Tangential Turbine but with overlapping pockets.

When I tried running Axial Turbine 4A on air after the runs on steam discussed in the quoted post, it would not run correctly. I disassembled the turbine and found the inner ball bearing and the spacer washers that contacted the inner ball bearing were covered with a brown film. This has never happened before running on steam but these last runs had estimated steam temperatures quite a bit higher than previous runs (380 F vs 290 F). The dental ball bearings I have been using are used in the Star 430 handpiece and the manual for that handpiece states the maximum temperature for steam used for cleaning is 135 C (275 F). Apparently the 380 F estimated steam temperature of these last tests is enough to damage the bearings. Running at close to the 290 F of the saturated steam at 40 psig of previous tests is probably close enough to the 275 F maximum recommended by Star. I have not had any problems running on steam when the superheat is low enough that the steam is visible exiting the exhaust.

 

[Turbine GuyParticipant @turbineguy]

I got new bearings for Axial Turbine 4A and ran it with air.  The performance was the same as before running on steam. Both Axial Turbine 4A and Tangential Turbine 5B have run better since replacing the set screw collars with springs and optimizing the position of the rotor. I updated the following table to show the best performance of the tests I performed today that were repeated several times to ensure the results were valid.  

[Turbine Guy]

On Turbine Guy Said:

I just read Mike Tilby’s very informative Steam Turbines Large and Miniature Article 23. He generously included a reference to this thread. This article is on Stumpf-type (Terry-type) turbines. He correctly stated that I believe the open pockets can be as efficient as pockets with an outer wall, especially when the spacing between the rotor and the casing is very small. To back this up, I will compare the performance of the last test of my Turbine 3 running on air with the performance estimated from the following chart. This chart, also shown in the post of 14/03/2019, is copied from ‘A Study Of High Energy Level, Low Output Turbines’ prepared by Dr. O. E. Balje for the Department of the Navy in December 1957. This chart illustrates the maximum performance he estimated for various types of turbines when optimized. The heavy solid lines in the chart are for axial turbines. The dashed lines are for Terry turbines. The thin solid lines are for Drag turbines. The drag turbines are like turbine pumps. The blades circulate the flow in a way that increases the drag force on the rotor. The units for the volume flow are ft*3/sec and lb/ft^3 for the gas density. The values of Ns and Ds for Turbine 3 for the test of 3/4/2020 are 1.0 and 14.9 respectively. The efficiency of an optimized Terry turbine with these values of Ns and Ds from the chart is approximately 16% and the efficiency of Turbine 3 was 14.3% based on my most conservative values of the power required by the propeller. I’ll get into the Reynolds number effects and how they lower the efficiency of the optimum Terry turbine below that of my turbine for this particular case in the next post.

I go on in the quoted post to show how my open pocket Stumpf (Terry) type of turbine matched the efficiency of the bladed Terry Turbine shown on the chart. After several improvements the actual efficiency of my Tangential Turbine 5B exceeds the efficiency of Axial Turbine 4A in the 10/30/2024 tests shown in the table of the last post. These two turbines are almost identical except for the rotor and changes required for direction of flow as shown by the following drawings, so this is a good comparison. In Mike Tilby’s article 23 he mentions that Jim Bamford found in tests with Stumpf type and De Laval type rotors, the Stumpf type was about 10% more efficient. In the tests shown in the table of the last post my Stumpf type Tangential Turbine 5B had an efficiency of 30.7% and the De Laval type Axial Turbine 4A had an efficiency of 27.3%. This is about the same difference in efficiency Jim Bamford found. In my opinion, the reason for the better efficiency of the Stumpf type of rotor for turbines with single nozzles is the much larger overlap of the nozzle shown as dimension “a” in the drawings. This more than offsets the better rotor efficiency of the De Laval type of rotors.

[Turbine GuyParticipant @turbineguy]

I started to have rubbing when running Axial Turbine 4A regardless of placement of the rotor. It appeared to be due to excess tilting of the rotor. Axial Turbine 4 has the option of adding a ball bearing in the cover as explained in the 15 October 2022 post on page 17. The combination with three ball bearings is called Axial Turbine 4C as shown on the following drawing. I thought this option would significantly reduce the rotor deflection and allow the spring to keep the rotor pressed against the inner bearing but might reduce the performance due to the added friction of an extra ball bearing. When I ran a test of Axial Turbine 4C the results were the same as I was getting with Axial Turbine 4A before the tilting became a problem. Apparently the extra rigidity allowed the rotor to get closer to the cover increasing the performance enough to offset the added friction. I will run a few tests to break in the new ball bearings and record the performance in the next post.

[duncan webster 1Participant @duncanwebster1]

Here’s a challenge for someone. Parsons Epicyclic Engine. Four single action cylinders but no crossheads/connecting rods. Ii reckon you could do a 2 cylinder version

http://www.douglas-self.com/MUSEUM/POWER/parsep/parsep.htm

[Mike TilbyParticipant @miketilby23489]

Hi Byron

Why do you think your started to have rubbing problems that did not exist before? Did you try to reduce the clearances or has something changed such as the bearings are wearing?

Mike

[Mike TilbyParticipant @miketilby23489]

Hi Duncan

I found the link you posted very interesting. I’d read about the epicyclic engine designed by Parsons before but had not seen any detail.  I read somewhere that Parsons made a carboard model to illustrate the principle of the epicyclic engine while he was a student of mathematics at Cambridge. Apparently he kept the model in his college room to show visitors.

Gerd Niephaus’s model as shown in the links from the Douglas Self webpage certainly does look a fascinating project.

Mike

 

[Turbine GuyParticipant @turbineguy]

Hi Duncan,

That is a very unique design showing the creativity of Parsons. Thanks for sharing the links.

Hi Mike,

Because of the wear in my very old lathe I was not able to get the ball bearing bores quite as close to the diameter as I needed, so Axial Turbine 4A always had more clearance on the ball bearings OD than I would like. I made the clearance between the rotor OD and the housing bore as close as I could with the allowable movement of the ball bearings. Apparently sliding the ball bearings into and out of the housing and onto and off of the rotor shaft loosened the fit of bearings enough to allow the deflection to make slight contact. I have made many tests that required me to do this in order to try various options. I took the simplest method to solve the contact by adding the extra ball bearing. I felt that since this solved the problem without any loss in efficiency it showed how important having a ball bearing each side of the rotor is in reducing the tilting of the rotor shaft. I got no change in performance when I added new ball bearings so wear of the bearings appeared not to be the problem.

Thanks both of you for adding a little life to the thread,

Byron

[Martin Johnson 1Participant @martinjohnson1]

Some years ago I thought about building the “Turbomotive” in 5″ gauge.  I eventually put the project in the “too difficult” box.

If this thread tells me anything, it is that I made the right decision back then.  Hats off to those that continue to try!

Martin

[duncan webster 1Participant @duncanwebster1]

One of my pipe dreams is this, a sentinel geared loco, but 2 engines, each gear coupled to a driving axle with a loco boiler. The wheel arrangement seems to be 1 A A 1 in modern terms, the carrying wheels are in the main frame, not in trucks

It will never got done, too many other half built projects, but it looks a lot easier than a turbine.

 

http://www.douglas-self.com/MUSEUM/LOCOLOCO/egypt/egypt.htm

[Turbine GuyParticipant @turbineguy]

Hi Martin and Duncan,

I believe my Tangential Turbine 5B shown in the 1 November 2024 post is easier to machine than even simple oscillating cylinder steam engines if the housing and cover is printed. It could even be done with just a small lathe if the rotor was also printed. Since I have a milling attachment and an indexer that can be set to 60 increments per revolution making the rotor only takes a couple hours.  I envy anyone who makes the engines you describe.  They are WAY beyond my capabilities. Thanks for sharing the links.

Hope you are both doing well,

Byron

[Author]

Posts