High Temperature Seal Material 1

[goldenfab]

I am working on a design that uses a ball valve type sealing device to seal hot gasses resulting from hydrocarbon combustion (diesel/gasoline w/air). The first prototype I used bronze for the seal and 6061 for the ball. This worked for a proof of concept for the overall design but as expected much galling occurred in a very short period of time. Here are the loose design requirements:

-Gas temperature: < 1,800 F
-Expected maximum seal temperature: < 1,000 F
-Expected maximum ball temperature: < 400 F
-Maximum gas pressure < 2 kpsi
-No lubrication
-4 in diameter ball rotating at 4,000 rpm -> surface speed = 70 ft/s

My first thought was to nickel plate the ball and use a carbon graphite seal. I would think the nickel would do fine against graphite but I'm not sure if the graphite would have the strength to hold together. I'm wondering if instead of plain graphite I could sinter a ceramic with some type of lubricant like graphite? Then, instead of the nickel plating would a super hard plating like TiN be better. Or maybe something with Boron for lubricity? I'm guessing it would not be a good idea to use the same material for both moving surfaces? Materials suggestions as well as sources of information to read up on welcomed.

 

[btrueblood]

Does it have to be a ball? Think about how the majority of devices doing this achieve a seal (i.e. reciprocating engine exhaust valves, operating with a linear poppet-seat).

If you've got to use a ball, then yes, a very hard ball and a slippery and (relative to the ball) soft seat material (boron nitride might work), with the seat spring-loaded to ensure good contact with the ball. Make the seal easily serviced, as it will be the wear item. 4000 rpm...I would not predict a very long service interval...

 

[EdStainless]

I would go with Stellite for the seats. These are hard and will need to be lapped in with the ball. They will take the temperature and resist corrosion and wear.
The ball should be hard and slippery, and a different material. Look at coating with TiN, TiC, or a bunch or others.
And you need a substrate that is strong enough to support the coating, maybe H13 tool steel. Actually H13 might work without coating it.
I would dry film lube with MoS or HBN.


= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[goldenfab]

I am trying to avoid a poppet valve due to its complexity and friction losses.

I'm checking into Stellite now. I have not heard of it. Thanks for the recommendation. I would like to use aluminum for the ball if possible (due to machinability and weight) although its not a hard requirement.

 

[EdStainless]

No, don't use Al. It is very soft (when hot) and has very high thermal expansion which will complicate issues.
The other option is to look at using SiC for the ball (though a plug valve might be a better option).

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[goldenfab]

I have been sidetracked working out other aspects of this design but I'll be getting back to the sealing problem soon. I'll go with something other than Al if I absolutely have to. The design of the port shape is critical and would be difficult to machine. I want to stay with Al because I can cast and machine it in house and because it would presumably have good heat transfer.

Ideally I would want the ball to not wear out and the seal to be the replaceable wear item. Does anyone have a clue if I used something like Satellite for the seal and one of the thin film coatings mentioned which would wear out faster and about how much. I'm wondering if I would want to go with a softer seal material so the ball does not wear out since it would be a costly replacement item vs the seal. The last idea I had was a AlTiN/TiAlN coating for the ball and free machining alloy for the seal. Thoughts as to if this is not a good idea and why welcomed. I have read about how AlTiN/TiAlN forms an oxide layer in hot conditions but what I have not been able to find out is if you don't subject it to a hot enough temperature are you going to damage the coating?

I've been searching the web for literature on wearing of dry sliding materials and am getting lost in the weeds. Does anyone have any good resources to get an idea of what the wear rates would be for something like I am doing?

While I may be better off just sending parts out to be coated I have acquired most of the parts to build a DC/RF sputtering machine. If anyone has any good practical resources for the design of sputtering guns and the process for sputtering different materials please let me know.

 

[EdStainless]

You can buy SiC ball and seal trim for commercial ball valves.
Or you could go with SiC ball and Stellite seats, that might give a good combination (I have used that combo for bearings in an abrasive environment).
I still feel that Al for the body is a mistake.....

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[goldenfab]

Thanks again for your input Ed. I may have to outsource prototyping the ball. Unfortunately there would be nothing on ball valve market that would be close to what I need. Did the SiC / Stellite combination run without lubrication? Can you offer any direction on where to get data on wear rates for this and other material combinations like it?

I was looking into Hexagonal Boron Nitride ceramics. I have yet to see it used as a seal or bushing like graphite is sometimes used. Does anyone know why this is? Is it the material properties or just cost? The prices I have seen for it are a bit high.

 

[Rustam1]

Dear adamflyer,
pleas specify:
1. Maximum gas pressure < 2 kpsi - this is 2 ksi (140 bar) or 2 psi (0,14 bar)
2. Maximum seal temperature: < 1,000 F , maximum ball temperature < 400 F.
You want to use the aluminum or aluminum alloy for a ball which is in contact with the saddle.
You know - melting temperature aluminium (for the alloy slightly below the) 1,000 F.
The ball is in contact with the saddle and the ball's surface temperature can reach 1000 F.
For me, is not very clear proposal to use aluminum for the ball.
3. What level (e.g., 1000 F - 100 F) and the frequency thermal shocks.
4. What degree of leakage is permitted between the ball and the seat? As a percentage of the passed volume of gas. As far as this is critical?

 

[EdStainless]

HBN is soft, slippery, and very chemically inert.

We used to buy SiC balls without a special order. We selected a commercial ball valve based on availability of ceramic trim. It is all aftermarket, the valve makers don't want to sell something that durable. Try Coorstech.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[metengr]

Use a Cr-Mo (2.25% Cr-1%Mo) low steel ball with a carbide wear coating against Stellite seats. The Cr-Mo valve ball material will be stable and perform satisfactorily at 1000 deg F, in terms of elevated temperature strength and resistance to creep deformation. This combination is used in many large steam turbine throttle and stop valve applications in power generation.

 

[tbuelna]

Mazda spent considerable amounts of money and engineering effort trying to solve these same design problems with their rotary engine apex seals. However, the max seal temp and max gas pressure requirements were less demanding. To get acceptable service life from your sliding gas pressure seals you will need to provide for a very carefully controlled oil film at the sliding gas pressure seal interface. This no different than the situation with piston rings on ported 2-stroke engines. The trick is keeping the oil film extremely thin and evenly distributed, making sure the surface the oil film is applied to is kept cool enough to prevent flashing of the oil, and making sure that only extremely small amounts of this oil ends up in the intake/exhaust flows.

 

[goldenfab]

Rustam1,

1. Maximum gas pressure < 2 kpsi - this is 2 ksi (140 bar) or 2 psi (0,14 bar) Yes 2ksi (140 bar)
2. Maximum seal temperature: < 1,000 F , maximum ball temperature < 400 F.
You want to use the aluminum or aluminum alloy for a ball which is in contact with the saddle.
You know - melting temperature aluminium (for the alloy slightly below the) 1,000 F.
The ball is in contact with the saddle and the ball's surface temperature can reach 1000 F.
For me, is not very clear proposal to use aluminum for the ball. My requirements are very vague I know. Basically the valve would have a similar environment as a piston in a gasoline engine. This is why I think I can get away with aluminum.
3. What level (e.g., 1000 F - 100 F) and the frequency thermal shocks. The valve would be subject to ambient air temperatures to very high flash temperatures several times a second aso thermal shock could be a real issue.
4. What degree of leakage is permitted between the ball and the seat? As a percentage of the passed volume of gas. As far as this is critical? To be honest I have no idea. My plan is to determine this experimentally.Does anyone know of any resources that detail leakage of poppet valves in gasoline engines?

All,

Thanks again for the suggestions. I have cast some prototypes in aluminum alloy 356 (see picture). My plan is to hard coat anodize them. I think some of the other materials suggested would be a better choice but right now I'm going to see how the aluminum works out. I'll post back the results for those that are interested.

To those of you who have not guessed already I am developing a rotary valve train for a regular gasoline engine (nothing to do with the wankel). I know it has been done many times and never succeeded for many reasons, mainly the wear/sealing issues. I don't see this technology going anywhere for most engine applications but I have a market in mind in which I think it just might take. I'm a bit behind schedule trying to remedy casting defects of the cylinder head casting so work on the valves has been halted for a while. If anyone has experience in sand mold/core design -specifically solving core blow issues any help would be much appreciated. I'll post more in about this in a new topic.

 

[1gibson]

Well, this one is obviously not a homework question. Pretty impressive so far.