|
Post by Johansson on Feb 27, 2024 1:40:18 GMT -5
We´ll see how fast this contraption can be spun before things go downhill, for the gen2 version I am tempted to try a single blade approach with hollow sheet metal blades locked by pins (similar to fir tree but with round holes in the disc instead).
With a series of 3D printed press jigs I think it would be possible to make them, a 1mm folded blade tig welded together along the inducer and exducer edges. Holes radially drilled in the disc to supply cooling air to the blade internals.
The compressor wheel shouldn´t have the same problems with the holes since it is in one piece, holding itself together and not pulling on the bolts like the segmented turbine blade arrangement.
|
|
|
Post by Johansson on Feb 26, 2024 15:43:55 GMT -5
True, but since the blades are laser cut to .1mm tolerances the balance in the un-profiled wheel is probably as good as it gets, profiling the blades and then statically balancing the wheel will most likely make it worse. I have to do any profiling by hand, flap disc style. Milling the blade profile would awaken vibrations from the outer regions of Hell... Shame you can't just get it red hot and squish it into shape with a few tonnes haha save the grinding nightmare. Sure you can rig up some setup to keep the grinder in the same position for each blade, like a diy surface grinder I don´t think it will be that much of a headache to profile the blades, if I weigh each segment before and after profiling I should be able to make them pretty identical. If this proves to work it is still a very easy way of making a turbine disc compared to turning a big chunk of inconel into a disc and manually grind out each blade from solid. Just need to improve on the lowest hanging fruit which seems to be the blade segment holes, until I can get a decent performance out of the design.
|
|
|
Post by Johansson on Feb 26, 2024 14:11:36 GMT -5
Remember turbos are usually balanced to a high level since even a tenth of a gram at high rpm is a lot of force so any weight you can shed that is further away from the axis of rotation will benefit you, plus profiling makes it better for flow so win win. You could probably profile before twisting, use an end mill or HSS radius cutter after you do the slot cut on each blade segment? True, but since the blades are laser cut to .1mm tolerances the balance in the un-profiled wheel is probably as good as it gets, profiling the blades and then statically balancing the wheel will most likely make it worse. I have to do any profiling by hand, flap disc style. Milling the blade profile would awaken vibrations from the outer regions of Hell...
|
|
|
Post by Johansson on Feb 26, 2024 14:10:56 GMT -5
Good find, it looks like 12.9 steel shoulder bolts are the way to go.
|
|
|
Post by Johansson on Feb 26, 2024 13:41:44 GMT -5
I was able to find a few moments today to do a few calcs. to get a better picture of when you will experience plastic deformation of your components. First, if I look at the stainless steel blade portion, due to the fall off of material capability of most stainless steels, it is limited. Not knowing which variant of SS you are using, I just made a guess of 304L. I got some properties from here: nickelinstitute.org/media/1699/high_temperaturecharacteristicsofstainlesssteel_9004_.pdfIf we look at Figure 3 on page 7, we can see that the 0.2% yield strength at your design point of 1050K, is below 30 ksi and the tensile strength is below 45ksi for almost all of the stainless steels. Setting the yield to be below 30ksi, the bolt holes will reach that level of stress at about 3500 rpm. They will exceed the tensile capability of the blade material by the time they hit 4200 rpm or sooner. I used a stress concentration factor, Kt ~ 1.2 from Peterson's "Stress Concentration Factors" Figure 147a, which is for a pinned or riveted joint. This is an approximation of course, but it should be roughly the order of magnitude expected and representational. If you haven't already, make sure that you use shoulder bolts in the bores of the through holes, because if you let the threads push on the inside of the holes, you will get larger peak stresses that could lead to local damage and potential failure. Also, making the bolt a tighter fit to the through hole will allow it to have better contact and distribute the stress somewhat better. If the bolt diameter is small relative to the through hole, there is less contact and it becomes more like a line contact with a sharp increase in local stress. As for the disk portion, the bore will likely hit yield first at around 7500 rpm and exceed UTS at 8450 rpms. A perfectly smooth through hole would approach yield at around 8500 rpm, but if they are tapped, then they will have even higher stress concentrations. I do not have a good estimate of how the stress concentration factor changes when the load is aligned 90 degrees to the thread centerline,so I can't estimate it. The smooth through hole radial stress would pass the UTS at around 8500 rpm. I would suggest that you take really good measurements of the bore, the o.d. and the bolt hole diameters, so that after you runs, you can the disk for any plastic deformation, indicating that you have exceeded the yield strength of the material. Remember, these are all "fuzzy" numbers with a lot of assumptions baked in to get you a rough order of magnitude idea of where the problem areas might be. Good luck, Chris Hi Chris, So the first thing to fail are the blade holes, good to know. I will use shoulder bolts and keep a close eye on the hole diameters. Just curious, if the bolts are a little smaller in diameter than the holes causing local stress and deformation wouldn´t the bolt "seat" itself and eventually get full 180 degree contact? If I can remove say 5% weight from the blades by profiling them when they have been given the right twist, how much would that affect the hole tensile capability rpms? 8000rpm is the usual max revs for a snowmobile clutch, so I think this will be my goal to eventually reach. Makes finding the right weights and springs easier when I run the clutch like it is supposed to. /Anders
|
|
|
Post by Johansson on Feb 25, 2024 15:33:53 GMT -5
According to Chris the bolts are not the first things to fail anyway, so it is perhaps overkill to spend a couple of hundred bucks on titanium bolts for the 5k rpm wheels.
|
|
|
Post by Johansson on Feb 25, 2024 15:30:50 GMT -5
I tried some Google-Fu on bolt single shear strength but didn´t find any site comparing different materials, and the rotational forces make weight a big factor for me as well.
|
|
|
Post by Johansson on Feb 25, 2024 15:01:35 GMT -5
What is the best suited M6 bolt material for fastening the blades?
12.9? A4? Titanium?
|
|
|
Post by Johansson on Feb 24, 2024 9:36:29 GMT -5
Today I made the blade twisting jig that I have been thinking about for some time, using a vice and a pair of pliers is not the way to go if I want to make 40 identical blades. A simple setup that should produce repeatable results, I heat the blade root with a propane torch to get the twist as close to the root as possible. I used the test blade section to find the right setting for a 30 degree twist. Some heat and gentle twisting later. This is the result. I am a bit undecided about any further profiling, I´d lose a little weight on the blades but it would be interesting to try them as they are first just to see how they perform. Cheers! /Anders
|
|
|
Post by Johansson on Feb 15, 2024 0:31:12 GMT -5
Chris, this is very helpful indeed. Thanks a lot for taking your time to make calcs and explain how an, in an uneducated mind, insignificant thing like holes around the disc outer edge can become a major issue at higher rotational speeds.
I´ll use plenty of bleed air behind the discs to keep them cool.
The segmented turbine blades are a mechanical challenge to fit to the disc properly, but they are very practical for production.
If I would keep them for mk2 and instead of a single disc use two steel/stainless discs back to back with the blades in between like a sandwich, how much would that improve things? If I would pick a better material for the discs (no inconel but the best "available" material) and silver solder the hub sections to the disc sides.
This means that I get a better working condition for the blade bolts, since they have twice the shearing points with DISC/BLADE/DISC instead of just DISC/BLADE.
How close to 10k would this take me before things start to stretch too much?
/Anders
|
|
|
Post by Johansson on Feb 13, 2024 0:16:33 GMT -5
A cluttered desk means a cluttered mind, an empty desk means...... Nah, I don´t have anything suitable for spinning it. It will have to wait until it pulls itself around.
|
|
|
Post by Johansson on Feb 12, 2024 16:25:33 GMT -5
No big update, but I finished welding the diffusor housing today. It would look nice if I could skip the comp cover and just run a sealing strip around the exducer rim. /Anders
|
|
|
Post by Johansson on Feb 12, 2024 0:09:34 GMT -5
I´ll check it out.
|
|
|
Post by Johansson on Feb 11, 2024 15:37:14 GMT -5
I´ve never heard of a car with belt drive other than the infamous DAF 66.
|
|
|
Post by Johansson on Feb 11, 2024 14:24:44 GMT -5
I got a fair bit of work done this weekend, yesterday I drilled and threaded the turbine disc with 40 M6 holes. A while later. This will be a big turbine wheel... I ended up milling a blade press jig from aluminum with a 3D printed insert. It works just fine, with 6 tons of pressure it profiles the blades just like intended. Despite its shortcomings in certain areas PLA can take a shitload of pressure like this. The printed layers are pressed into the 3mm stainless sheet so I can feel them with my fingers. Next up is to build a tool for twisting the blades, I´ll heat the root of the blades before twisting. Cheers! /Anders
|
|