How to Reduce Hydraulic Pressure Variation? 3

[acspain]

Hello All,

What should I add to the hydraulic circuit to reduce the pressure variation?

We are running an endurance test of a component using hydraulic pressure ideally ramping from 0 psi to 1030 psi and back to 0 psi over the course of 30 seconds to complete a cycle.
The peak pressure is going as low as 1018 and as high as 1050 which is too much variation. The low pressure does not stoke the part enough and the high pressure causes too much stroke. I would like a ±5 psi tolerance.

The hydraulic circuit comprises of a variable piston pump driven by an electric motor generating the flow/pressure to a solenoid valve switching the part between the pressure line and the sump line. The pressure and sump lines have a metering valve set to ramp the pressure within the cycle time. The volume required between 0 and 1030 is fairly small at .25 cubic inch.

My suspicion is that the variation is coming from the on/off nature of the solenoid demanding flow/pressure or not from the pump and the electric motor adjusting to maintain rpm.

Thank you for your help.

 

[TugboatEng]

It sounds like you are using proportional controls. You should consider incorporating integral control if you need to maintain a specific setpoint.

 

[acspain]

The setup is not that sophisticated. The pump/motor is an independent setup that I don't believe we have access to or can change.

What I am looking for is a hydraulic component that I can add to the circuit in order to modify and/or steady the pressure supplied to the part.

 

[LittleInch]

Some sort of accumulator to soak up some of the volume of the variations would be my guess.

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[TugboatEng]

You're asking for ±0.5% accuracy which is going to be difficult to achieve using a proportional control scheme. With proportional control, as you approach your setpoints the control output proportions to zero so you can never quite maintain your setpoint. You need to usee a controller with an integral feature. If you don't have access to the pump, install a pressure regulating valve downstream of the component you are testing and use an off the shelf PID controller to regulate back pressure.

And yes, an accumulator will make this easier to achieve but it is not the answer on its own.

 

[PNachtwey]

What is the application? Chances are it has been done before. Chance are we have done it before. I do wonder how you get to 0 psi. The oil will not exit/flow unless there is a pressure difference so at low pressures, the oil will exit very slowly.

An accumulator will smooth out the pressure variations but it will also slow down the rate the pressure can changed from 0 to 1030 psi.

Controlling to with 1 psi at 1030 has been done before.

The question I have is are you really trying to control pressure or are you trying to control force?
Seriously, what are you really trying to do because chances are it has been done before or at least, something similar.



Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

http://forum.deltamotion.com/

IFPS Hall of Fame Member

 

[BrianPetersen]

To me, I don't see where proportional controls are involved. It sounds to me like there's a normal pressure compensated variable displacement pump whose output is being switched bang-bang between P and T, and a needle valve (and hopes and dreams) are crudely restricting rate of pressure rise and fall via flow restriction, with the flow in and out provided by the flexibility and compressibility of whatever it is that they are testing.

Is that it, or is more sophistication involved, and how so?

Schematic?

 

[LittleInch]

What's controlling the on off solenoid?

To ramp up 4 cc of fluid in 15 seconds is a pretty small flow. The accuracy of whatever you're using to measure / generate and on/off signal also needs to be considered in terms of accuracy and repeatability. Their accuracy can be 0.1% of span, same with repeatability and hysteresis.

If you want super accurate pressure why not use a dead weight tester? you'll get +/ 0.02%

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[3DDave]

Electrical users would use a zener diode to clamp the output. I'd say a relief valve would do the same function, but I do like the dead-weight tester as well.

 

[TugboatEng]

The problem with a relief valve is that it is a proportional control. A relief valve cracks open at it's set point but isn't fully open until some pressure above the set point.

The pump likely also has a proportional control. There is no feedback. Without feedback, unless the system is overdamped, there will be fluctuation above and below the setpoint. Small changes such as temperature will also impact the set point of the day.

 

[acspain]

Hello All,

Thank you all for your input it is appreciated.

LittleInch - I ruled out an accumulator as it would reduce the variation but probably not to the level I want.
Solenoid is controlled via a pair of timers and automated switch/cycle counter.

TugboatEng - Yes, I am quickly finding out that the approach we have chosen may not be the best and the accuracy I am looking for may be impossible to achieve with the current setup.

PNachtwey - Yes, we are not getting down to 0 psi which is fine. I am targeting pressure as that directly relates to deflection of the part. I am trying to stretch a bellows a fixed amount of distance for a number of cycles. The pressure variation is causing not enough stretch on the low end and too much stretch on the high end.

BrianPetersen - Yes exactly with no more sophistication. I have added a schematic which I should have done in the first post.

Unfortunately, this project is under time/resource constraints so I am stuck with the overall setup. I am hoping there is a way I can add a component to reduce the variation understanding that all the other pieces are adding variation.

For instance, if I add a relief valve set to 1030 psi, adjust the pump to run at say 1100 psi, then the pump would always have some flow going through the 1030 psi relief valve and back to sump. Then when the solenoid switches the small amount of flow directed to the part there would be little interruption to the pump/motor reducing the pressure variation.

Thank you for your help.

 

[LittleInch]

Solenoid is controlled via a pair of timers and automated switch/cycle counter.

Well there's your issue straight off.

You need to feed back pressure somewhere into this system of control. Using timers, valves and spring pressure reliefs you're not going to get close and it's a bit of miracle you're doing as well as you are IMHO.

does the pump run continuously but just dumps via the relief valve or how does that regulator work?

But if you put an accumulator between the filter and the metering valve your inlet pressure would stabilise and then rig a pressure switch into the solenoid circuit / timer and see what pressure you set it at to get your 1050 psi.

Also why the second piston set up for oil / water?

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[acspain]

LitteInch,

The pump runs continuously.
What I am calling a regulator adjusts the plate angle in the variable piston pump.
I am told the pressure relief valve is set to 1500 psi, which means it is staying closed and serving as a system safety valve only.

Requirement is to test the part with water, not oil, so the second piston is separating the two.

If I understand your statement correctly "You need to feed pack(back?) pressure somewhere ...". I need to added a pressure relief valve of some sort between the filter and first metering valve to give the fluid a circuit to run through at all times. Then the solenoid can siphon off the small amount required with little effect to this circuit.

Reading a fluid power book the other day runs through various types of Relief Valves, pilot-operated relief, unloading and pressure reducing valves all seem suitable. I don't know which to choose.

Thank you for your help!

 

[BrianPetersen]

The pressure "feedback" at the moment is the "pressure-compensated" bit of the variable-displacement hydraulic pump. It is a purely hydraulic/mechanical device in which, if the output pressure is sufficient to overcome a preload spring, an actuator inside the pump repositions the (probably) swash-plate to push it towards the zero-displacement position. If the output pressure drops, the preload spring pushes the swash-plate towards the maximum-displacement position. There may be some additional controls to stablise its output and fine-tune the relationship between flow and pressure, but that is basically the gist of it.

This is not intended to be a precision mechanism. It is a mechanical device with pivots and seals that all have friction, and parts (and thus, clearances) that expand or contract with temperature, etc.

You do not want to have such a pump pumping through a relief valve as part of its normal operation. The relief valve is there so that if there is a sudden pressure spike due to someone abruptly shutting a valve, or a cylinder abruptly running into the end of its stroke, and the pressure spike happens too quickly for the pressure-compensating device to react, the fluid has someplace to go (and the pressure spike is thus clipped of by letting it through the relief valve).

If the relief valve setting is below the pressure-compensating setting, the pump will adjust itself to max displacement and max flow through the relief valve in its futile attempt to reach the setting of the pressure-compensating mechanism ...

It sounds like what you need is a proportional valve (and no needle-valve flow control) instead of the bang-bang directional valve, and accurate feedback of the sensed output pressure back to the control system that operates the proportional valve. Then you could control the output pressure in any pattern you wish with accuracy determined mostly by how good the components are that you select.

To do this right, you need a proportional hydraulic valve (which is not inexpensive!) and a controller with analog inputs and outputs from the sensor and to the proportional actuator. This is a way more sophisticated system than what you have now.

For the accuracy that you wish, what you've got is not going to cut it, and I am also surprised that you have it as good as you have now with how crude it is; I would expect it to be worse. And it probably will be different tomorrow when the temperature in the shop is different, and worse after the system has been running for an hour to warm up, etc.

 

[LittleInch]

I think you need a back pressure regulating valve if this flow is continuous. Spring relief valves are not the best when it comes to this sort of service.

But pressure feedback (sorry for the typo) should be at the point where you need to measure it and then use it as a trigger to switch the solenoid. You should be able to tune the system to get the right pressure setting to get your 1050 psi with some sort of regularity and probably quite good accuracy.

Or listen to Brian!

Remember - More details = better answers
Also: If you get a response it's polite to respond to it.

 

[Stick.]

I am trying to stretch a bellows a fixed amount of distance for a number of cycles.

So it sounds like what you really want to control is the displacement of some portion of the bellows and you're just using pressure as a proxy for that. Is that correct? How tightly do you need to control the displacement (±10mm, ±1mm, etc)? Any chance you could use something as simple as a limit switch to indicate end of travel and use that signal to control the hydraulics? Adding a mechanical stop to the system might also be worth considering in combination with the limit switch (just make sure it's stout enough to handle the loads the system might impart).

 

[hydtools]

Use a hard stop to limit the bellows travel. Simplest.

Ted

 

[acspain]

Stick and hydtools - The end application for the bellows is a component in a larger hydraulic/pneumatic system hence the use of hydraulics for development testing. I was aiming for a .300 to .350 inch stroke. The component is in a pressure chamber so limit switches would need to be magnetic/proximity based which may bring in another type of challenge.

Yes, hard stops would solve the over stroke challenge but would result in an unbalanced pressure across the bellows. This particular bellows is not designed for an unbalanced pressure nor should the bellows ever see significant unbalanced pressure in the end application.

Thank you for your help!

 

[hydtools]

Pressure does not determine stroke. As long as force generated by pressure exceeds bellows resisting force, the bellows will continue to move as long as flow continues. Input volume of fluid does determine stroke. Seems you should be controlling input volume instead of pressure.

Ted

 

[Compositepro]

The bellows do have a spring constant. An external spring can increase the spring constant so that a higher range of pressure will work.