Pressure regulator model: how to study its stability

[Q_Goest]

Hi serbring. Something just occurred to me that might help you. Whenever my company has had to buy more than a few dozen valves, I've often talked to someone at the supplier and asked them to send me one out of courtesy so I could evaluate it. I'd take the valves apart, do stress analysis, fluid flow analysis, spring and dynamic analysis of poppets, thermal analysis, seal analysis, pretty much any analysis needed to understand how good the valve was. I actually found many valves that were poorly designed and could potentially fail in service. It wasn't unusual to find substandard designs even from reputable manufacturers. I'd also perform testing as needed, though it was amazing what you could find out just by doing analysis.

It's actually not unusual for companies to do that. If you know your company will be purchasing more than a dozen or so per year of this valve, I'd suggest you contact them, explain what you're doing, and ask them to send you a sample valve for evaluation. Do that for any valve you expect to use for this application. I'd suggest getting at least 3 different valves in for test before you decide on one. If you only want to test valves (not take them apart and do analysis on them) you should be doing, at the very least, performance testing and life testing. You want to know how well the valve is going to work and for how long. Valves can wear out quickly or they can last for many years without problems.

I'd be interested in hearing how to works out for you. Feel free to post your progress.

 

[serbring]

Hi serbring. Something just occurred to me that might help you. Whenever my company has had to buy more than a few dozen valves, I've often talked to someone at the supplier and asked them to send me one out of courtesy so I could evaluate it. I'd take the valves apart, do stress analysis, fluid flow analysis, spring and dynamic analysis of poppets, thermal analysis, seal analysis, pretty much any analysis needed to understand how good the valve was. I actually found many valves that were poorly designed and could potentially fail in service. It wasn't unusual to find substandard designs even from reputable manufacturers. I'd also perform testing as needed, though it was amazing what you could find out just by doing analysis.

It's actually not unusual for companies to do that. If you know your company will be purchasing more than a dozen or so per year of this valve, I'd suggest you contact them, explain what you're doing, and ask them to send you a sample valve for evaluation. Do that for any valve you expect to use for this application. I'd suggest getting at least 3 different valves in for test before you decide on one. If you only want to test valves (not take them apart and do analysis on them) you should be doing, at the very least, performance testing and life testing. You want to know how well the valve is going to work and for how long. Valves can wear out quickly or they can last for many years without problems.

Hi Q_Goest,

Thanks for your interesting suggestion, unfortunately we have to build just one sample of the rig, therefore hopefully we will need only one valve, otherwise it would mean I have made any mistake.

I'd be interested in hearing how to works out for you. Feel free to post your progress.

[/quote]

I surely keep you updated about this project. Hopefully in a month a I will be able to show you some interesting results. Actually I'm trying to model the festo mpye valve. There is something I don't understand about the valve behaviour and probably you are able to help me. In the catalogue you can find the voltage-spool position behaviour:

So with 5V the spool is positioned in the middle and with 0V the spool is positioned in one of two ends. But 0V means 0N, right? Moreover in the catalogue it is written:

Safety setting Active: mid-position in the event of setpoint value cable break

So if the setpoint cable break, the solenoid voltage is 0V, but the spool is positioned in the middle? This is in disagreement with the previous statement. I try to explain it better. For modeling the valve I need to set the spring stiff for the mechanical model of the spool valve. Here you can find the dynamic equation of the spool that is:

$m*\ddot{x}+c*\dot{x}+2*K*x+F_f=F_c$

Equation 37 of this paper. From this equation, I would believe that with 0V the spool is placed in the mid position and not with 5V.

What's wrong in my thinking?

thanks

Cheers

 

[Q_Goest]

So with 5V the spool is positioned in the middle and with 0V the spool is positioned in one of two ends. But 0V means 0N, right?

Yes... it means that port 1 (supply pressure) is connected to port 2 so port 2 is pressurized. Looking at the schematic diagram at the top of page 5, you can also see that port 4 is connected to port 5 (vent) and port 3 is sealed off.

Moreover in the catalogue it is written:

Safety setting Active: mid-position in the event of setpoint value cable break

So if the setpoint cable break, the solenoid voltage is 0V, but the spool is positioned in the middle? This is in disagreement with the previous statement.

I agree, that's a bit confusing. You need to supply power to the valve which is 17 to 30 V (see the top of page 6). I'm assuming that means 24 VDC nominal power supply which is very common in the industry.

You then send a control signal depending on if you purchase the 4 - 20 ma current version or the 0 - 10 VDC version of the valve. So I suspect the valve is spring loaded to be in the middle position when disconnected and no power (no 24 VDC) sent to it. I believe they're referring to the power not the signal. With no power, the valve is in the same position as it would be if it had power and you sent a 5 VDC signal to it. With power, if you send it a 0 VDC signal, the the valve shifts to one of the two ends as described above. Also, if it has power (24 VDC) and you send it 5 VDC signal, the valve position is in the middle with no flow. You should verify that with the sales rep at Festo as I'm not absolutely positive.

 

[serbring]

Hi Q_Goest,

following your suggestion I have a good model of the spool valve (by luck I found a paper with the flow rate data of the specific spool valve I need) and I can well actuate the brake. I'm really greatful to you.
There is still two unresolved points:

• Pipelines design: how can I design the pipeline diameter and length? Regarding the pipeline length, in the spool valve datasheet it is written that the length between the valve and the actuator should be no longer than 2 meters, so can I freely choose any value between 0 and 2 meters?
• Cylinder stability: Should I perform a motion stability analysis of the pneumatic cylinder or it is not usually necessary?

 

[Q_Goest]

• Pipelines design: how can I design the pipeline diameter and length? Regarding the pipeline length, in the spool valve datasheet it is written that the length between the valve and the actuator should be no longer than 2 meters, so can I freely choose any value between 0 and 2 meters?
• Cylinder stability: Should I perform a motion stability analysis of the pneumatic cylinder or it is not usually necessary?

Hi Serbring. I'd suggest staying with the recommended 2 meters or less. That minimizes volume and pressure drop through the pipe. Regarding diameter, as long as the pipe/tube is the same as your port size, you should be fine. As for motion stability analysis, I would suggest you simply allow yourself some time to tune the system once you put it together. You should count on some time for testing including upset conditions and just monitoring overall performance over some period of time. You might consider having automated data collection so you can review data every so often and verify things are functioning as anticipated.

 

[serbring]

Hi Serbring. I'd suggest staying with the recommended 2 meters or less. That minimizes volume and pressure drop through the pipe. Regarding diameter, as long as the pipe/tube is the same as your port size, you should be fine. As for motion stability analysis, I would suggest you simply allow yourself some time to tune the system once you put it together. You should count on some time for testing including upset conditions and just monitoring overall performance over some period of time. You might consider having automated data collection so you can review data every so often and verify things are functioning as anticipated.

Hi Q_Goest,

I supposed for pneumatic pipes there are similar rule of thumbs as for hydraulic pipes. As example this:

http://www.google.it/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&cad=rja&uact=8&ved=0CDEQFjAA&url=http%3A%2F%2Fwww.ihservice.com%2FPDF%27s%2FTube%2520Selection%2520Chart.pdf&ei=Gy-MU6e-PJD70gXrnIG4Dg&usg=AFQjCNEmCwyaFTVE7rr5xcGoKAAhT6xWPQ&sig2=UpFhBxa-7hyRP9nnQHBb8A&bvm=bv.67720277,d.d2k

The only parameter I can do for stabilizing the system is the pipe length, right?

 

[Q_Goest]

Hi serbring. Tables and graphs aren't generally used for compressible flow. In fact, they're generally not used for hydraulic flow either. The most common method of determining flow rate and pressure drop for a fluid through a pipe or tube is to use the Darcey Weisbach equation coupled to the Benoulli equation. Unfortunately, it's a lot more complex than looking at some tables such as the ones you linked to.

One of the best references for this is the Crane paper, TP 410. It's the most widely recognized text for determining pressure drop and flow rate. The paper (thin textbook really) is rewritten regularly.

If you don't want to pay for it, and I'd suggest if you aren't going to spend weeks or months studying it to understand how to perform a flow analysis, you might try an online calculator such as the one here: http://www.engineeringtoolbox.com/darcy-weisbach-equation-d_646.html

To be honest though, if you're keeping the pipe length short (2 meters), there's no need to even look at pressure drop through your pipe. It will be very small and insignificant. That's one of the reasons they suggest keeping your pipe length so short. Making it shorter than that won't change anything. The longer you make it, the more impact it will have on your system so I'd suggest not making it longer unless you really need to and then step up a size in diameter if it gets significantly longer.

 

[serbring]

Hi serbring. Tables and graphs aren't generally used for compressible flow. In fact, they're generally not used for hydraulic flow either. The most common method of determining flow rate and pressure drop for a fluid through a pipe or tube is to use the Darcey Weisbach equation coupled to the Benoulli equation. Unfortunately, it's a lot more complex than looking at some tables such as the ones you linked to.

One of the best references for this is the Crane paper, TP 410. It's the most widely recognized text for determining pressure drop and flow rate. The paper (thin textbook really) is rewritten regularly.

If you don't want to pay for it, and I'd suggest if you aren't going to spend weeks or months studying it to understand how to perform a flow analysis, you might try an online calculator such as the one here: http://www.engineeringtoolbox.com/darcy-weisbach-equation-d_646.html

To be honest though, if you're keeping the pipe length short (2 meters), there's no need to even look at pressure drop through your pipe. It will be very small and insignificant. That's one of the reasons they suggest keeping your pipe length so short. Making it shorter than that won't change anything. The longer you make it, the more impact it will have on your system so I'd suggest not making it longer unless you really need to and then step up a size in diameter if it gets significantly longer.

Thanks for your reply. I have asked you for it, because I'm always curious to understand things.

I found the an ald edition of the book you meantioned, I'll take a look on it. What about the stability? Shorter the pipes are and less stable is the system, right?

 

[Q_Goest]

Shorter pipes should be more stable because you have less volume and flow restriction. But with a pipe as short as 2 m, having something shorter won't make a noticable difference on stability.

 

[serbring]

Shorter pipes should be more stable because you have less volume and flow restriction. But with a pipe as short as 2 m, having something shorter won't make a noticable difference on stability.

Ok thanks again. I supposed: longer pipes are equivalent to damping for a mechanical system, so longer is the pipe the motion magnitude at the resonance frequency will be lower.

I'm updating the pneumatic circuit, is there any standard way to point out the line length? As example indicate that a specific line length must be shorter than 2m?

thanks

 

[Q_Goest]

Are you creating a drawing for the installation? Something to show where the valve is to be mounted and where the brake is, what tubing to run, etc... ?

I'd show something like line length on a drawing and maybe even cover it with a note which indicates not to exceed 2 m in length.

 

[serbring]

Are you creating a drawing for the installation? Something to show where the valve is to be mounted and where the brake is, what tubing to run, etc... ?

I'd show something like line length on a drawing and maybe even cover it with a note which indicates not to exceed 2 m in length.

Thanks Q_Goest.

I have another question. It's necessary that the pressure in the brake cylinder rises of 3*bar/s, that is 0.33 s/bar with a volume of 2l in the worst case (that is when the pad are woren out). This would allow me to verify the valve flowrate. So using the Ideal gas law and supposing the pad is in contact with the disc (when the brake cylinder is completely extended), I can compute the air mass necessary to increase the brake pressure of 1 bar (from 4bar to 5bar) at 300 K is:

$\Delta M=$$\frac{\Delta P V}{R T}=\frac{1e5*2e-3}{287.058 *300 K}=0.0023 kg/s$

I have the feeling the mass is very low and I believe this is because this formula doesn't take into account any comprimibility effect and therefore the mass can be understimated. Checking in the following the compressibility factor is close to the unity so in this condition it is plausible to assume air as ideal. Is it correct?

http://www.enggcyclopedia.com/2011/09/air-compressibility-factor-table/

 

[Q_Goest]

Correct - air under ambient conditions can be approximated as being an ideal gas. Your interpretation of the chart is correct.

I checked your work and came up with the same answer (0.0023 kg/s) for increasing a 2 liter volume by 1 bar. So you can use this to determine how large a valve you require or how far open the valve needs to be to get that flow rate.

 

[serbring]

Correct - air under ambient conditions can be approximated as being an ideal gas. Your interpretation of the chart is correct.

I checked your work and came up with the same answer (0.0023 kg/s) for increasing a 2 liter volume by 1 bar. So you can use this to determine how large a valve you require or how far open the valve needs to be to get that flow rate.

Thanks,

under this assumption, a valve the valve with a similar pneumatic connection to the one of the brake (G3/8) has a too high flow rate. So I need a pneumatic adaptor, is there any issue in using adaptors that I should consider?

 

[Q_Goest]

If a valve with a smaller port size on it will give you the correct flow rate and that port size is smaller than what is on the brake, go ahead and use a reducer somewhere. The only other option is to increase the volume of your brake, and that's not a particularly good answer.

 

[serbring]

If a valve with a smaller port size on it will give you the correct flow rate and that port size is smaller than what is on the brake, go ahead and use a reducer somewhere. The only other option is to increase the volume of your brake, and that's not a particularly good answer.

I'm sorry my question was specific enough. The brake has G3/8 and the valve G1/4, so there is no problem, right?

 

[Q_Goest]

I'm sorry my question was specific enough. The brake has G3/8 and the valve G1/4, so there is no problem, right?

That's correct.

 

[serbring]

That's correct.

great. Q_Goest, thanks a lot for your precious help! :)

 

[vishi6486]

I am new to simscape and trying to setup a similar pressure regulator model. It seemed like I connected all the elements correctly but I am not getting the expected results.

I have attached a pic of the regulator and of the simscape model as well below.

Can someone please let me know if the elements are correctly connected ? Any advise, help , suggestion is appreciated.

Regulator Model

Simscape Model