Impulse Pressure Waveform 1

[djmccann]

I have a requirement to test a hydraulic filter assembly 280,000 pressure cycles with pressure from zero (0) to 300 psi and back to zero (0) at a rate of 10 to 25 Hz with a superimpose impulse cycle of zero (0) to 90 psi and back to zero (0) at a rate of 210 Hz. A representative image would look like

. Does anyone know of a way to generate this waveform. I need to be able to input a vibration controller and run it "open-loop" to drive the pressure into a bellows assembly to generate this waveform, but I need a function...??? PLEASE HELP!

 

[Compositepro]

Perhaps I am missing something but that is just two sine waves of two frequencies added together. Add a constant to the lower frequency wave so the lowest point is zero.

 

[gruntguru]

Because the frequency response of your actuator will not be flat, you will need the ability to control the amplitude of the two sine waves independently. You can then adjust each amplitude until the pressure signal is correct.

Not sure if you can do this with an off-the-shelf signal generator at a reasonable price, but adding two sine waves electronically is not difficult. Alternatively something like Labview would probably do it for you.

je suis charlie

 

[PNachtwey]

Whoa guys!!!! This is not as easy as it looks. How does the pressure drop to 0 so quickly? It isn't as easy as just letting oil flow out of the cylinder. The flow through the valve will approach 0 as the pressure drops to 0. There must be a pressure drop to get the desired flow.

To do this there will need to be a cascaded loop system. The inner loop is a position control loop that moves another piston that creates the pressure. This way the motor draw back the piston and even create a vacuum if necessary. The inner position control loop can be either hydraulic or a servo motor.

Generating the wave form is no problem. Designing the test system is.
This will do the trick if you can find a valve that is fast enough and solve the pressure drop problem.

http://deltamotion.com/applications/appnotes/documents/hydrostatictesting.pdf

Simply adding 2 sinewaves together will not result in the actual pressure following a profile as described. There must be closed loop control.
The RMC75E interfaces easily with LabView so it is easy to make your own test system.
I just have doubts about find a valve that is fast enough to make the 210Hz sine wave.
Even 25Hz can be tricky. A lot depends on the volume of oil that must change pressure quickly.
A fast pressure sensor is also required.



Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[djmccann]

Ordinarily we would just use a servo valve and LabView to generate the waveform. The problem is the frequency of the superimposed waveform. I can't find a servo that will act that fast. The test doesn't require flow and essentially all the lines would be "fluid-solid" and therefore the pulse would just be a 'slug' of fluid moving with the displacement of the bellows. This is the only way I can imagine getting that cycle rate and being able to generate the waveform as the shaker system can be run "open-loop" so that I can can control the zero-offset as well as the input waveform.

I still think this is like a FFT type waveform where one sinusoid is used to make the other, but I need some way of generating that function. This isn't clear to me.

 

[PNachtwey]

http://www.jsme.or.jp/English/awards/eyushu10-5.pdf

Yuken makes fast valves. Here they mention they have a 400Hz valve.
There is another company in Seattle WA I think that makes very fast valves but I don't remember the name. I don't know if they sell to other companies.
Finding a 400Hz valve isn't that hard if you have money.
I forgot about this thread during work. I will show how to generate the wave form. It isn't that hard.
There is a formula that you should be aware of
dP/dt=B*Q(t)/V
B is the bulk modulus of oil.
Q(t) is the flow into or out of the volume of oil under compression
V is the volume of oil under compression. You need to keep this volume small so you can use smaller faster valves and still achieve the pressure rate.
I haven't done the math yet on what the maximum pressure rate needs to be.

I get involved in a lot of servo testing applications.


Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[PNachtwey]

oops, it looks like you have a solenoid pushing the bellows. That should be able move at a high enough frequency but can it provide the enough stroke.
Then the formula simplifies since there is no flow and only the volume is changing. ΔP=β*ΔV/V





Peter Nachtwey
Delta Computer Systems

http://www.deltamotion.com

 

[gruntguru]

djmccann.
The filter assy should be direct coupled to the bellows (large passage, minimum volume). The bleed/top-up system can be small diameter pipes since these are only controlling the "DC offset". This loop will have a long time-constant so I can't see any need for fast valves.

je suis charlie

 

[Panther140]

Can you keep a set amount of oil in the system?

Think along the lines of a beefed-ed up subwoofer pushing/pulling hydraulic fluid inside of a contained hydraulic system. Some of the ideas in this thread are taking the LOONNNNGGGGGG way around problems. 210hz is not hard to do, even if you're on a DIY budget. That's not even fast by the standards of our slowest modern DAC/ADC systems, and the math to get that wave made is mid-level algebra. Designing your system to facilitate precisely controlled waves, without interference waves, is the hard part IMO.

 

[3DDave]

I'm just curious as to what system is likely to do this in real life. If it's an accelerated test, then what is the gain in power dissipation going into the test item?

By way of background we had a 3rd party (antagonist hired by customer) do a life test on a large bearing and reported the bearing would overheat. The bearing in test was ordinarily operated at less than 1 RPM for a 90 degree cycle, at a minimum interval of 20 minutes. The interval was usually hours or days. Even the highest use rate allowed re-establishing an oil film from the grease gallery. The highest load was applied over a 10 degree rotation at the initial operation of the linkage.

The 3rd party put the maximum allowable load on it and ran it at a continuous 60 RPM for hours on end. When we mentioned the grease and oil film, they said they followed the suggested lubrication interval - not even an accelerated interval to match the accelerated test. I guess they could have run the bearing at 20k RPM and reported that it self-destructed from inertial forces. They never even approached the fact that the oil film was being forced out of the bearing by the high continuous loading and the high speed.

 

[gruntguru]

"I still think this is like a FFT type waveform where one sinusoid is used to make the other, but I need some way of generating that function. This isn't clear to me. "
The two sinusoids don't need to be related, in fact as I said earlier the amplitude of each needs to be independently adjusted. An electronics guy would find it trivial to generate this signal. In simple terms you could think of it this way:

Use two signal generators - one to generate each sinusoid. Feed the two signals into a summing amplifier. Job done. As you say, this part could be done open loop - simply adjust the amplitude of each sinusoid until the pressure waveform is correct.

Maintaining the correct "DC" pressure level probably needs to be closed loop for a 3 hr test. Alternatively since there are no potential leakage sites and DC drift will probably be mostly thermal, it may be sufficient to check the apparatus every 10 minutes and manually pump the pressure back up.

Your first diagram is not clear but it looks like the peak-peak limits of your pressure envelope are 0 - 300 psi? From the wording of the problem, I would have assumed a 0 - 300 psi waveform with 0 - 90 psi waveform superimposed for an overall envelope of 0 - 390 psi p-p.

je suis charlie

 

[someguy79]

Panther140 has the right idea. I also think this will need an electronic drive system of some sort.

Perhaps an "electrodynamic exciter" that acts on a low density incompressible fluid could do the trick. Be careful though, lots of those low density, low viscosity liquids are also quite flammable.

One thing that seems to be missing in the above posts is that there isn't a mention of WHERE in the system these pressure values are needed. You will likely have to tune the system to get the waveform you're looking for. Like it's been said, there will likely be interference/noise in the pressure waves.

 

[gruntguru]

The OP has posted a diagram above using a shaker and bellows to provide the input. Perfectly sensible. No valves are needed except for initial pressurisation and correcting drift in the static (DC) pressure level.

A pressure transducer plumbed close to the filter should satisfy the client that his tst specification is being met.

je suis charlie