Model Turbines

The following table is the test results for turbine 3 running on steam. I added the mass flows, input energy, torque, and efficiency for comparison with the test results running on air shown in the last post. I added the torque to the chart because it is much more relevant when I use a larger propeller. The propeller speeds were getting higher than the speeds I could verify the required power of my GWS EP 2508 propeller, so I decided to use an APC 4x3.3 EP propeller for tests with steam. Because the larger propeller requires more torque to turn, the resulting maximum speeds will be much lower and consequently the power and efficiency. With the propellers, the torque is the best measure of performance. For a given energy level, any improvement will increase the stall torque regardless of the size of the propeller. Once the stall torque is determined from testing, the power can be estimated for higher running speeds using the formulas shown on the stall torque test described in the post of 01/02/2020. I tried Krytox GPL 105 oil that Werner Jeggli recommended, and it appears to be working well with my shielded ball bearing. The oil is very viscous when cold but thins out with the steam temperature. I put a small quantity of oil on the outside face of each ball bearing before each test with steam. The oil appears to be getting into the ball bearing through the shields and each run required a short time for the oil to thin out and reach the optimum amount in the bearing. Once the optimum conditions were met, the turbine ran almost constant speed for several minutes. I ran three tests with the 0.031 in. nozzle and the difference in turbines speeds was less than a 100 rpm and the total run time for the three tests was about a half hour. Now that I am getting consistent test results for the 0.031 in nozzle, I will try to come up with a method to measure the mass flow.

The following table is a test of the ball bearing oils used in my turbine 3. The ball bearings are supplied with AeroShell Fluid 12 oil. I compared the performance of new bearings with this oil to the performance of bearings with this oil that had been run for quite a while on air and never exposed to steam. I also compared these with ball bearings that had Krytox GPL 105 oil added and had been run on steam for quite a while. Because my turbine 3 is very sensitive to the position of the rotor relative to the nozzle, I ran all these tests with 3 shims and a clearance between the set screw collar and the bearing of 0.002 in. This ensured that the position of the rotor relative to the nozzle was within 0.002 in. for each of the tests. This is not the optimum position for both propellers, so the powers achieved in this test are not as high as I have obtained running on air. The loss in power with the Krytox GPL 105 oil is due to very high viscosity at room temperature that is approximately 14 times as much as the AeroShell Fluid 12 oil at room temperature.

The low power for the tests with Krytox GPL 105 oil in the ball bearings shown in the last post seemed lower than expected. I ran several more tests with the Krytox GPL 105 oil and each of my propellers. I found that if the turbine was run for several minutes the speed would go up in small jumps as the thick oil thinned out and got pushed into the optimum position. The first test with each propeller took the longest to obtain the maximum speed. Each later test took progressively shorter time to reach maximum speed. Once the maximum speed was reached, the speed would stay approximately the same for the rest of the time of the test. I repeated the tests several times and each time, the maximum speed stayed approximately the same. The following chart is updated to show the results of these tests and is more like what I expected.

I tried to find a source where I could purchase AeroShell Fluid 12 oil. The only sources I found sold the oil in gallons and consequently was VERY expensive and much more than I would ever need. I looked at similar oils and the most promising oil available in small quantities was Krytox GLC 102 which could be purchased from Amazon at a cost of about $35 for 1 oz. This is the same as what 1 oz. of Krytox GLC 105 costs. The two oils are similar, and the biggest differences are low temperature viscosity and temperature limit. The properties are shown in the following table provided by Chemours. I will replace the factory oil in my older ball bearings with the Krytox GLC 102 and see how it compares with the factory oil. If it works as well as AeroShell Fluid 12, it will allow me to periodically add oil as needed. I still plan to use the Krytox GLC 105 when I run with steam since it appears to be working well. Both oils are very expensive, so hopefully they will last a long time. If I only ran my turbine a few minutes at a time periodically on air, I would probably never need to add oil.

I decided to test the three different oils discussed in the last post for a running time long enough to see how their performance holds up. I decided that a total run time of over 20 minutes would give an indication of their live expectency and consistency. I also decided to run all the tests with the APC 4x3.3 EP propeller since it is what I plan to use for future testing. I made one short run with the GWS EP2508 propeller with the AeroShell Fluid 12 oil at the end of the tests with this oil. The maximum speed reached was 24,500 rpm, the highest speed I have been able to obtain with this propeller running on air. This showed the maximum power was still the same after running quite a while with this oil. The maximum speeds for the APC 4x3.3 EP propeller for all the oils were at least as high as when the oil was new. The following table gives the results of these tests. The Krytox GLC 102 oil did not perform as well as I hoped. I don’t know if this is because of the way I added it or if the oil is not as suitable for my application. I can’t remove the shields from the bearings I use, so I tried removing the existing oil by soaking the bearings in Acetone and then adding the new oil by letting it seep into the bearings through the small gap in the shields. I don’t know how effective that was in removing the old oil or adding the new oil. When I added the Krytox GLC 105 oil I didn’t do anything to remove the existing oil and just relied on the steam to flush it out or the existing oil to be compatible. I applied a small amount of Krytox GLC 105 to the faces of the ball bearings before each run with steam which appeared to work for adding this oil. Even with the much higher viscosity at low temperatures, the Krytox GLC 105, worked better than the Krytox GLC 102 running with air. Based on these tests, I believe the ball bearing as supplied with the AeroShell Fluid 12 will work fine and have adequate life when running on air. The Krytox GLC 105 appears to work well with steam and results in only a small loss in power running on air. I will use the ball bearings with the Krytox GLC 105 for all my tests that don’t require the maximum power running on air.

Since my turbine 3 has multiple openings required for the different options, I can’t channel all the steam into an exhaust tube and measure the mass flow by cooling and collecting the exhaust. I thought I would try making a second nozzle as shown in the following drawing. This would increase the flow and drop the pressure in my boiler below the pressure the relief valve opened. All the mass flow out of the boiler would then go through the turbine nozzles. With the wick type of burner I use, I can run the boiler until it is empty before pulling out the burners. I can then fill the boiler with a measured amount of water and make my test run the amount of time to boil out all the water. This was the plan, unfortunately even with my very careful setup the drill drifted out of position and the second nozzle could not be used. Fortunately, I am still able to use the existing nozzle. I ran a test with steam to check if there was any damage to the existing nozzle and the maximum speed stayed at almost a constant 6,150 rpm for the approximately 7 minutes of the test. This is slightly higher than the maximum speed of 5,900 rpm I was able to obtain in prior tests with steam at approximately 45 psig, with the APC 4x3.3 EP propeller, and using Krytox GLC 105 oil. More of this oil was able to get inside the shielded ball bearings since the speed running on air after the test with steam was lower than before the test with steam. I am expecting the speed to gradually increase when running on air after the oil wears and thins out like it has after prior tests with steam.

The safety valve on my Stuart Turner 504 boiler that was purchased in the 1970’s was releasing steam at a pressure of around 45 psig (3 bar). The maximum operating pressure for this boiler is given as 60 psig (4 bar). I don’t have a way to measure the size of the threads on the safety valve, so I searched the internet. In this search, I stumbled on the video shown in this this link https://www.youtube.com/watch?v=O05uMDICFWk. The video gave the thread size but also gave suggestions for repairing the safety valve. I tried adding a washer and sealant as suggested in the video and checked the release point with these changes. The release point was slightly over 60 psig and almost all the leakage from the valve was stopped. I ran the boiler dry and then added a carefully measured ¾ cup of water. I then ran the boiler with my turbine 3, APC 4x3.3EP propeller, and Krytox GLC 102 oil. After a short time, a turbine speed of 6,800 rpm was reached and held approximately constant for the rest of the run. The total run time was 7.5 minutes. I wanted to learn two things from this test, the mass flow out of the boiler into the turbine and if the Krytox GLC 102 oil worked better after running on steam. Since there was no release or leakage from the safety valve, all the flow went to the turbine and the pressure reached 50 psig (3.5 bar) with the 0.031 in. nozzle. The mass flow was approximately 3.1 lb/hr (23 g/min). After running on steam, the maximum speed running on air was 4,200 rpm so the performance of the Krytox GLC 102 did improve as shown in the following table. Apparently the high temperature of the steam thins the oil enough that it can expel the excess and obtain the optimum amount.

Edited By Turbine Guy on 25/03/2020 16:43:47

I updated the turbine test results for running on steam as shown in the table below. As described in the last post, fixing the safety valve on my boiler stopped the leakage and allowed the pressure to get high enough that all the flow went to the turbine. This allowed me to measure the mass flow and estimate the actual percent moisture. The estimated 15% moisture found in this test was higher than the 10% I had been assuming.

Hello Byron,

I do not understand why the turbine is running at such a low speed. With steam and at this pressure, this should be in the vicinity of 30'000 rpm. The torque will stay approx. the same, but the power output will rise proportionally.

Hi Werner,

The output torque drops as the speed increases due to the rapidly rising rotational losses and smaller difference in velocity between the rotor and the nozzle. The table shown in the post of 01/02/2020 shows an example of this. Earlier posts in this thread explain what was involved in estimating the values shown in that table. I have developed spreadsheets to help estimate the performance with different energy levels, different size propellers, different rotors, and different nozzles. I compare what is estimated in these spreadsheets with any test results to check their accuracy. The post of 26/03/2020 is an example of this. With the spreadsheet used for that test, the estimated power at a speed of 30,000 rpm would be 7.5 watts and the torque would be 0.338 in-oz. If I had the ability to adjust the load to a power my turbine was capable of producing at 30,000 rpm, I could check these values. The down side to using propellers, is that the torque required to turn the propeller determines the speed the turbine can reach. The torque I obtained in this test would spin my GWS EP 2508 propeller at above its recommended operating speed and that is why I am using the larger propeller.

Hope this helps,

Byron

I made the following table using the equations and values shown for estimating the performance of my turbine 3 running on steam at a pressure of 50 psig, 15% moisture, 3.12 lbm/hr, and with the 0.031 in. nozzle. The table gives the estimated performance at given speeds, and the equations can be used to find the power and torque at other speeds. The table and equations assume that the torque required by the load does not exceed the torque of the turbine at that speed. Earlier posts detail how I determined the nozzle and rotor velocity coefficients.

The following drawing shows a concept I am going to try. This is an attempt to allow the flow exiting from the nozzle to reach its maximum supersonic velocity before contacting the rotor. Detail B is an enlarged view of the section through the nozzle. The 0.032 in. diameter is the existing diameter of the nozzle and is the largest diameter I can fit into the turbine 3 housing. It is also larger than the maximum nozzle size recommended for the rotor pocket size made by the 0.094 in. diameter end mill. The optimum size nozzle for the 0.094 width pocket is 0.029 in. When I increased the size of the nozzle to 0.032 in., it is the largest I felt I could use. My plan is to make an insert with a diameter small enough that when the flow exits the insert it can expand to the diameter required for the maximum supersonic velocity without contacting any surface. The discussion of flow exiting the converging only nozzle into a pressure below the critical is given in this link https://www.model-engineer.co.uk/forums/postings.asp?th=140195&p=4. From this discussion, it appears that the diameter that the flow from the convergent only nozzle exits into must be greater than the diameter required for the supersonic flow. The distance from the exit of the convergent only nozzle to the first contact with the rotor must also be long enough for the flow to stabilize. The dimensions and placement of the insert shown in detail B of the following drawing are my estimates of what might work for steam at 50 psig.

I tried making the insert shown in the last post out of aluminum. Even with a sharp cutting tool and very small cantilevered length, I could not turn the OD down to 0.032 in. The cantilevered portion would bend before I could get down to that small of a diameter. Werner Jeggli has used injection needle tubing to make similar inserts, so I thought I would try that. I was able to order a short length of 304 stainless steel hypodermic tubing with an OD of 0.032 in. and an ID of 0.022 in. I also ordered a center drill with a 0.020 drill size and 60 degree included angle to make the tapered entrance into the insert. I will try making the insert after I receive the tubing and center drill.

I received the hypodermic tubing and center drill mentioned in the last post and made the insert. The following picture shows the center drill in one drill chuck and the insert extending out of another chuck. I use the center drill to make the 60 degree taper on the entrance of the insert. The next few posts will show some more pictures I made while working on and installing the insert.

The following picture is of the insert on the 0.022 diameter drill. This picture shows how tiny the insert is.

I didn’t realise that you can buy such a small centre drill “off the shelf”. Keep up the good work!

The following picture shows the insert starting into the existing nozzle hole of my Turbine housing. I apologize for the rough appearance of the housing. It had so many changes to try different ideas and has been assembled and disassembled so many times it shows all the scratches and wear. I will add an updated drawing showing the actual dimensions of the insert and its position in the existing nozzle in the next post.

I bonded the insert to the existing nozzle bore with Loctite 290 and allowed it to cure overnight. The following drawing shows the finished dimensions of the insert and the position in the existing nozzle bore. I started the cleanup of the Loctite and found it has got into a lot places I did not want it to get into. It will take me a while to finish the cleanup, but if I can't get the Loctite out of all of the flow path, the insert won't work correctly.

I cleaned out the bore of the nozzle insert with a 0.022 in. drill and then cleaned the 0.032 in. diameter at the discharge end of the nozzle. I tried to clean the tapered entrance of the insert. Because of the metal filled epoxy I used to fill an opening where the drill broke through, I could not get a tapered pin down to the entrance of the insert. With the entrance taper still blocked with Loctite, the flow coefficient of the insert is reduced considerably. My original nozzle has a 60 degree included angle entrance and an estimated flow coefficient of approximately 0.93. With the tapered entrance of the insert partially blocked with Loctite, the flow coefficient is reduced considerably. I ran Turbine 3 with the APC 4x3.3EP propeller, Krytox GLC 105 oil, and 50 psig steam. The maximum speed was 4,150 rpm and the speed stayed approximately the same for 13 minutes until the boiler ran dry. The boiler had ¾ cup of water at the start of the test, so the mass flow rate was approximately 1.8 lbm/hr (19 g/min). The performance indicated that the larger diameter at the discharge end of the insert did not allow the steam to reach supersonic speed. Apparently, this was not one of my better ideas.