Valve Seal Failure

I don't think your going to find anything that will stand up to that kind of beating.

Your using this for some type of temperature control at a guess? If so, you would be better off with a couple of modulating valves, 1 for cool water and 1 for hot water. With a properly tuned P.I.D. loop, you will get much tighter control than on/off valves.

The test rig supplies water at 10C for 9 secs and then 90C for 9 secs to the items under test so a PID control system is of no use in this situation.

Two months of turning a ball valve on and off every 9 seconds sound like you got your monies worth. Don't know of many that would preform that well. I guess that why there is solenoid valves. You can find them with service life cycles to 10 mil. At 9 Sec a cycle that's close to three years.

The Teflon seals are the problem. Yes, Teflon lowers the coefficient of friction but it does it at it's own expense as the number of operations increases and the temperature rises. 600,000 operations and at 95C is more than what can be expected of it.

Also if the edges of the hole in the ball are not perfect each time it opens or closes those edges of the ball abrade a micro amount of Teflon. A micro (tiny piece) times a lot of operations at either a lower temp or raised temp will cause this. Also Teflon will "flow" as the temperature rises. The Teflon looses it's strength and flows to another more relaxed configuration.

Use a more suitable packing for the conditions you have. You are certainly not the first to encounter this problem.

Good Luck, Old Salt

More than likely, there is corrosion taking place somewhere in the system due to oxygen transfer during the switch from cold to hot. Oxygen dissolves well in cold water up to approximately 12 mg/L under 1 atm pressure air at your stated temperature. Once the system switches (under flow) to the high temperature water, there will be oxygen (and nitrogen, and tiny amounts of carbon dioxide) introduced from the cold side (assumption is the cold water reservoir is open to the air). Any metals such as carbon steel, galvanized steel, etc. will corrode, and the corrosion products will abrade the valve seats over time. Also of import is the open/close time. Faster switching of the valves will probably cause enormous stresses at the interface betwee the ball opening and the valve seat.

There is much more elegant solution to your problem: "valveless valve" that switches based on fluidics. If you can't find one, make one. After all, you should be able to use the fluidic analogy to electrical current, and arrange to produce a stream oscillator. Other than that, use solenoid valves as suggested by others, but you might want to invest in the highest grade ones you can find.

Some diaphragm valves are designed to be cycled several million times without a failure. I don't know how they would perform in your application, but it might be worth a phone call.

At present the only type of valve being offered by several manufacturers that will do 2 million ops at 90C are angle seat valves. Sceptical now as it was one of the original manufacturers who promised that the ball valves would be ok, but would not put in writing. Anyone with experience of angle port valves?

Another consideration is that you used cheaper ball valves than what would be required in such rigorous testing, although I liked the idea of an angle seat valve (presumably with electric actuation. I noticed one company offers valves of this type that are three-way valves. Although I did not find instances of anyone using fluidic switching for repetitious control of fluid streams, I do not really doubt that it could be done, just no one seems to have the "want to".

Swagelok® ALD Series Valve Testing Shows Increase In Cycle Life, Up to 100 Million Cycles

Man. That is outstanding! I bet those valves are priced somewhere around the gold standard.

If you are still interested in an answer I would have a look but I need some more input:

is is a spherical valve ? which is the seal which fails : the one sliding on the sphere or the one on the shaft ?

do you notice any surface damage on the sphere when the seal failed ?

dimension of the sphere ?

does the seal present any local wear of is its wear uniform on the circumference ?

I am sorry but not being so clever to give an answer without data I can only do a step if I see what happened and interpretate it.