Force to overcome friction of piston to move it

[deangardner]

I have an oil filled systems that is free from air, and at atmospheric pressure. As the temperature of the surrounding environment changes, the oil expands and contracts. I have designed an expansion cylinder into the system that allows for this change in volume which is a cylinder that caps off the oil volume, is open at one end with a standard piston inside which houses a NBR o-ring. This ensures that the pressure remains ambient for any given temperature.

In the past, as the temperature changes the static friction of the piston has been such that it imparts nearly 2 bar in the oil volume before it moves to its new position. The system has now been slim-lined such that the system is rated to 1 bar max, therefore the friction force (static and dynamic) has to be calaculated to quantify this value.

I understand that many factors are involved here: lubrication of o-ring; o-ring material; sliding face material; surface finish of contacting surfaces and rate of change in temperature.

I don't need help understanding the affects of the variable parameters in the system, I'm after a way of quanifying the force required to move the piston. Has anyone had to produce this data before and has knowledge of this affect? I would appreciate a formular if you have any idea..?

Thanks in advance

 

[israelkk]

I would not trust o-ring friction. In such delicate system it is better to use a rolling bellow piston or a diaphragm that can expand without friction.

 

[MintJulep]

Or a bladder type expansion tank or accumulator.

 

[desertfox]

hi deangardner

Friction as you say is very unpredictable,however if I understand your post correctly how about this:- calculate the axial force on the piston when the system is at 1bar pressure ie pressure * area, this would be the force to move the piston, now using a coefficient of friction divide this into the axial force and that would give you the maximum reaction force between cylinder wall and piston which needs to be overcome.
Difficult bit is the friction coeff, but you could start with what I've suggested and then maybe do some practical tests and measure the force.

desertfox

 

[JStephen]

It seems to me that testing would be the most reliable way to come up with any numbers.

Note that if other things remain the same, if you have a uniform friction force around the edge, then increasing the piston diameter will lower the pressure required to move it.

 

[israelkk]

JStephen

However, O-ring friction depends on O-ring diameter too, the larger the diameter the larger the friction. There is also a chance or O-ring rolling, etc. Do not forget stick (starting) friction especially after long resting time.

 

[tbuelna]

deangardner,

Simply quantifying the axial force required to move your piston is not so difficult (Fn x mu?). However, the problem gets very complicated and unpredictable when you must determine what friction coefficients your various boundary contacts can have. These friction coefficients can vary by a factor or two. And as israelkk points out, the difference in static and sliding friction coefficients can also be significant.

If you choose to perform testing to "quantify" your required forces, in order to achieve a statistically relevant sample, you will need to perform lots of tests, under a wide range of operating conditions, and with test articles that represent the range of fits/finish/clearances that you would expect in production.

Good luck.
Terry

 

[racookpe1978]

You've mentioned clearly your concern about "increasing" pressure that must move the piston to relieve the internal system pressure during operation. If the piston sticks, you've got a LOT of pressure (2 bar) for a system supposedly running at atmospheric.

But you haven't addressed the other side of your operation: when the system later cools, a sticking piston that can raise system pressure that much will have nothing but atmospheric (vacuum) pressure to move the piston back into the "neutral" point to begin the next startup cycle.

A 2 bar "push" to move the piston "out" cannot be equaled by a (max) 1 bar "pull" to close the piston. So you'd be left with an internal vacuum that is quite large.

 

[gadkinsj]

I've done testing on something very similar to this.

If you have the room to reduce your diameter and lengthen your cylinder then you will significantly reduce your static friction. Plus this actually optimizes weight of the unit since less material is required in a cylinder that is longer vs larger diameter.

But again there are alot of variables with this design so you won't know if you have reduced friction enough without building and testing.

The interesting thing that I found that is not easy to intuit, is that the static friction will increase dramatically if the o-ring is allowed to sit in one position for any extended time (days time frame) and that effect is exacerbated by temp. basically the O-ring flows into the surface variations of the material.

If it is very sensitive and you have some money to spend, I recommend a welded bellows system. You basically tune your internal pressure with the spring rate provided by the metal bellows, which are essentially a series of Belleville washers that are EB welded and a cap put on one end. Plus this has the added advantage of completely closed system, where the O-ring seal is not perfect and by nature you get small amounts of fluid loss.

 

[TeejT]

I've done testing on something very similar to this too.
If you reduce the diameter, you reduce the static friction from the O-rings as mentioned, but you also reduce the force exerted on the piston due to any given pressure. The force exerted on the piston due to the pressure difference grows faster than the static friction as you enlarge the diameter. Or put another way, the ratio of the surface area of the cylinder to the O-ring contact surface area increases as you increase the diameter. For example a 2" diameter piston moved at a fraction of the pressure that was required to move a 1" diameter piston, and that in turn took less pressure to move than a 1/2" diameter piston, all other variables being equal.
You should probably build some test apparata and determine things experimentally because there are lots of variables.