Modelling Hydraulic Systems

On Mon, 01 Oct 2007 19:03:31 -0700, pnachtwey wrote:
-- snip --


Unless I'm sadly mistaken a Kalman or H-infinity filter, when used in a control system, is nothing more than a formally constructed time-varying observer. If you don't like (or don't need) the time-varying part, then a steady-state Kalman or H-infinity filter (if done right) pretty much meets the criterion for "well constructed observer".

AFAIK, yes.

Probably not. But the folks working on those models and simulations put a whole lot more effort into them than usually goes into a model for an industrial system. Certainly, if you are _really_ careful with your math then you can squeeze some valid intuition out of it -- but it doesn't just happen automatically.
--
Tim Wescott
Control systems and communications consulting
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Jul 20, 6:19 am, Paul M <PaulMatWiredogdotcom> wrote:

The link is to a Mathcad worksheet where I used data from our hydraulic test system and compute the model. The columns are time, target position, actual position,?,?, control output. The target and actual positions are the same because I used a an open loop output. I only used the control output (5) and the actual position (2) to generate this data. ftp://ftp.deltamotion.com/public/NG/Mathcad%20-%20Sysid2A2BV70%20T02.pdf
One can see there is some oscillation when the control output changes rapidly. First order lag systems would not overshoot the steady state velocities. One can also see my model generated an estimated velocity which closely matched the actual velocity. This is more than good enough to use for tuning. I would use a PIDD with velocity, acceleration and jerk feed forwards to tune this system. All those following my posts know I have formulas, where I plug the actuator gain, damping factor, frequency and the desired closed loop pole locations, to generate the controller gains.
One must design the hydraulic system properly to identify a hydraulic system this well. Hose, long tubing, non-linear valves and not enough accumulator capacity will make identification harder or less accurate. A well designed system tunes easily and is easy to keep tuned. A poorly designed system will be a nightmare until it is ripped out.
Peter Nachtwey
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Tim Wescott wrote:

Spool valve: see fig.5-13 at http://tinyurl.com/34hf4f A steam engine I built uses a spool valve instead of the traditional slide valve. The ports open to annular grooves in the valve body, so it opens as a complete circle. The valve is only 3/8" dia., but the effective width is about 1.2" Unlike slide and poppet valves, fluid pressure exerts no net force on the spool.
Springiness arises from compressibility of the oil and from compliance of the hoses and tubing. (Do you know that the lines from fuel pump to injector on a Diesel engine are cut to the same length regardless of the actual run? The amount of delivered fuel is less than the pumped volume because of the compliance, and equalizing the lines delivers the same amount of fuel to each cylinder.)
When building precision machinery, it is well to design and think as if all structural elements are made of rubber. Homework problem: A steel cylinder, full of hydraulic oil at zero (gauge) pressure is 3.000 ID with .250 wall. When the pressure is increased to 4,000 PSI, what is the new diameter? If the fluid were truly incompressible, how much would need to be added? (Which is more compressible; hydraulic oil of steel?)
Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
: : Springiness arises from compressibility of the oil and from compliance : of the hoses and tubing. (Do you know that the lines from fuel pump to : injector on a Diesel engine are cut to the same length regardless of the : actual run? The amount of delivered fuel is less than the pumped volume : because of the compliance,
Where does the oil go?
and equalizing the lines delivers the same : amount of fuel to each cylinder.) : : When building precision machinery, it is well to design and think as if : all structural elements are made of rubber.
http://www.deltamotion.com/pdf/hydr2.pdf I have written a series of articles for Hydraulics and Pneumatics. These articles are not very 'deep' but provide a basic understand of many topics relating to hydraulic motion control using servos and motion controllers. Most hydraulic control is still bang-bang or manual.
: Homework problem: A steel : cylinder, full of hydraulic oil at zero (gauge) pressure is 3.000 ID : with .250 wall. When the pressure is increased to 4,000 PSI, what is the : new diameter?
The diamater would increase by about 0.002. I have been asked about the hydraulic capacitance of the cylinder. It is too small to worry about. See http://www.danzcoinc.com/html/step_lock_piston_seal.html This is why one needs seals.
: If the fluid were truly incompressible, how much would : need to be added? (Which is more compressible; hydraulic oil of steel?)
If the fluid is incompressible then anything over of the volume required to fill the cylinder would cause the pressure to increase infinitely. The formula for calculating the change in pressure is
DeltaPressue = BulkModulusOfOil*deltaVolume/Volume
Oil is much more compressible.
Peter Nachtwey
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Peter Nachtwey wrote:

The injector creates a large back pressure. The steel line expands (increasing its volume) and the fuel compresses. Think of a constant-displacement pump intermittently pushing air into a gum-rubber hose that has a relief valve at the far end.

Most of the reason for making the pump-to-injector lines the same length is so that they hold the same amount of fuel. Expansion of the lines is secondary. but measurable.
Jerry
--
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I like to browse here but most of the topics are well beyond the stuff that I used to get involved in. Having said that, I did get to play around with hydraulic control systems. We kept things pretty basic--defining the system dynamics using the gain of the spool valve (Flow as a f(inlet stroke), the characteristics of the various components, actuator, the various flows in the circuit and the feedback mechanics. Obviously, one of the biggest factors was "air" and the manner in which it affected the stability and the transient response of the system. Calculations are nice (if you can predict everything) but nothing did the job better than a good set of instructions for bleeding the air out of the system and then a means of verifying it with the system running. As noted in another post, at times things move pretty fast and I've seen a 30-50 millisec delay (no actuator movement due to compressible flow) cause a control parameter tracking error which eventually resulted in compressor blade failures (jet engine). In this case, putting a .010 dia hole at the top of the actuator pistons was enough to allow the trapped air a way of getting out of the actuator lines. Cheap solution for such an expensive, catastrophic failure. MLD
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
MLD wrote:

Bleeding air out of the system is basic hydraulics 101. If designed right the air will flow out of the cylinder on its own. See below. What gets tricky is keeping entrained air out of the oil. A major culprit is the return line to the tank. The oil must be returned without any splashing. Usually the return oil is released below the level in the tank and baffles slow down the motion of the oil.
As noted in another post, at times things move pretty

You didn't describe your system but I recommend putting the hydraulic servo valve directly on top of the cylinder for two reasons. One is the trap volume of oil between the valve and the piston is minimized and the other is that trapped air will naturally flow out the valve because the lines are at the top of the cylinder.
I see too many take the easy way out and put the valve on a manifold that is often below the cylinder and then hose is run to the cylinder. This is bad for three reasons. Air can't escape easily, the trapped volume of oil is larger than it needs to be, and the hose adds capacitance. All three factors lower the natural frequency of the actuator.
Peter Nachtwey
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

instructions
with
eventually
putting a

for
actuator.
Peter: I didn't get into too much detail but what you're referring to applies well to something that is usually large and stationary. Jet engine control systems are fairly complex and today's digital controls are complimented with many hydromechanical components. Most (almost all) gear driven components are at the bottom of the engine and the actuator locations are dictated by their function (moving inlet guide vanes, jet nozzle etc) as well as the envelope constraints and are usually not very close to the high pressure source. I agree that getting air out of a system should be "basic hydraulics 101" as you put it. Again, not so basic or simple when the system is all assembled and then is filled with oil or fuel. No tank but a complete closed system; return flow just goes to the low pressure side of the system (pump inlet); maybe there is an accumulator. Arbitrarily cracking fittings or lines to bleed them is not acceptable and is not allowed. It usually takes a systematic series of steps to ensure air is removed--in some systems, there are means to extract the air via push button type of relief valves as it is being filled. A well defined specification has to be in place since this is done world-wide using skill levels from experienced to basic rookie, and can't be left to everyone's own idea of how to do it. MLD
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.