Hydraulic fixed deplacement pump and VFD.

Hello,

Does anyone have any experiences using a variable frequency drive and a hydrulic pump with fixed deplacement insteade of using a pump with variable deplacement.

I was considering to use a pressure transmitter connected to the built in PID controller in the VFD. I will probably need an accumulator to cake care of fast variations. Is this possible? What do I need to consider?

Yes, I have experience with this. It doesn't work very well.

Biggest problem is shedding excess pressure without generating large amounts of heat.

It is a considerably more expensive solution than a variable displacement pump.

Take a look at this site:

http://www.kadantunigy.com/

From what I understand of this way of controlling flow tere is less wasted energy that using Pressure Compensated pumps unless you can use Load Sensing.

Alaric said:

It doesn't work very well.

Click to expand...

Did you have an accumulator with a servo valve or were you controlling the load directly and blowing the excess pressure through a relief valve?

It is a considerably more expensive solution than a variable displacement pump.

Click to expand...

I agree but how much more is considerably more? The variable displacement pumps are so slow and the controls are so poor that I have often wondered if a VFD fixed displacement pump would be better for high performance servo applications. This way the pump VFD can get a feed forward signal from the hydraulic controller so it doesn't have to wait for the pressure to change before it responds. The VFD would also know when the actuator isn't moving so the VFD can stop too. This would reduce wasted energy.

MR Q said:

I was considering to use a pressure transmitter connected to the built in PID controller in the VFD. I will probably need an accumulator to cake care of fast variations. Is this possible? What do I need to consider?

Click to expand...

Are you doing servo control using valves and the VFD driven pump is just used to supply oil? If controlling actuators with servo valves, how many?

As Alaric said, this can be overkill unless you have a very demanding application. The only advantage I see the the VFD driven pump is that is would respond much faster. This is key on higher performance applications where the motion is half over by the time the variable displacement pump comes on stroke.

Ideally the VFD and the servo valve control signals would be proportional to each other. I would use the pressure feedback as a trim or fine adjust to the main control signal from the hydraulic controller. Better yet, the hydraulic motion control controls the servo valve and the VFD.

Now things get tricky when mulitple actuators are controlled. The VFD would then need a control signal proportional to the sum of the magnitude of the control signals to the valves.

The problem with positive displacement pumps is the pulsations in the output. Most systems use a dampner/accumulator to cushion the pulsations.

When you use a VFD you need to make sure it runs in Constant Torque mode. A positive/fixed displacement is not a variable torque load.

I have seen the motor and drive over sized so they do not run at full capacity. This allows the drive to handle the pulsating current better.

I haven't done it with pumps, but I have with blowers.

It seems to me that it would make a huge difference whether or not you are using open center hydraulics or closed center hydraulics. (It has been 30 years since I did much hydraulics, but I assume these terms are still in play!) With open center I would expect the efficiency to be better, not worse.

There is an article in the August issue of Hydraulics & Pneumatics magazine about this that I just read a few minutes ago - apparently some people are doing OK with it.

In our particular application, I think that motor inertia is such a huge factor that we simply can not get fine enough control for it with just one large 100 hp motor to be viable over just using a servo valve and pressure compensated pump. Maybe with a large motor and a small motor/inverter it could work, but that will really drive up the costs.

It could certianly be viable for some systems so my earlier statement may not be true for you, its only reflective of my experience with it.

Alaric said:

There is an article in the August issue of Hydraulics & Pneumatics magazine about this that I just read a few minutes ago - apparently some people are doing OK with it.

In our particular application, control of an intensifier, I think that motor inertia is such a huge factor that we simply can not get fine enough control for it with just one large 100 hp motor to be viable over just using a servo valve and pressure compensated pump. Maybe with a large motor and a small motor/inverter it could work, but that will really drive up the costs.

It could certianly be viable for some systems so my earlier statement may not be true for you, its only reflective of my experience with it.

Click to expand...

With todays vector capable drives, you can have 0.1% speed control on the motor. If you have an inverter duty motor that is designed to operate at low speeds at high torque, you actually have smoother shaft speed. At low speeds a VFD has 0.80 to 0.85 power factor. Above 50% speed it climbs over 0.9 to 0.95% power factor. A well tuned DC drive is typically 0.3 to 0.5 power factor and 0.7 at the best. This translates into the power to drive the pump is quite a bit less with AC.

IMO today's VFD's are leaps and bounds better than DC drives.

MrQ

Why do you need to control the speed of the pump!

i have had some experince with positive displacement pumps over the years and had limited successes with VSD control, this is partly due to the pulsation you will need a fairly large pulstation damper/accumulator to compensate this also you will find that if you use a relief valve the exses heat generated will distroy the oil very quickly. we currently run to hydraulic pump systems and use variable flow piston pumps with very little heat genrated, simple control and maintenance procedures.

hope this helps scairn

Right Idea Wrong Application

Where I would try to use something like this (fixed pump w/ VFD on motor) is where I had several actuations but only one working at a time.

Say traditionally, the actuations would each have a proportional valve and ramping card (open loop). Since they don't overlap in time, I could replace all the proportional valves and cards with bang/bang valves and do the ramping with the VFD.

I believe a Japanese outfit had a setup like this for a plastics machine but I can't remember the name....

scairn said:

MrQ
Why do you need to control the speed of the pump!

Click to expand...

I most of you know I get into a lot of hydraulic motion control applications. Some of these require very high flow and response. For instance a customer wanted to extend 60 inches in .8 seconds, dwell for .2 second and wanted to retract 60 inches in .8 second and dwell for .2 seconds. That is really moving.

This means that a variable displacements pumps swash plate will be moving back and forth quite a bit.

Two of the key things in hydraulic motion control is the cylinder diameter and keeping the system pressure constant. The larger cylinder diameters make the system stiffer and increase the natural frequency. Keeping the system pressure constant keeps the motion controller gains constant. If the system pressure drops then the system gain drops. This means the motion controller gain must increase to keep the overall gain equal to 1. The problem is that the motion controller can't output more than +/- 10 volts to make up for falling system gains. The problem with pressure compensated variable displacement pumps is that they respond to pressure changes and can't supply full flow at the system pressure. Variable displacement pumps have a swash plate that causes the flow to vary. When the error between the desired system pressure and actual pressure small set point then the flow is very little if any. This keeps from wasting energy. As the actual pressure drops 200 to 300 psi the swash plate will move to full stroke, but this means the pressure compensated pumps can not output full flow at the system pressure which is bad. The problem is not only must the pressure drop but then we must wait for the swash plate to move so the pump will respond. Typically the swash plate control is a very crude spring adjustment that provides some sort of swash plate motion in proportion to the difference between the actual pressure and the desired system pressure. This just just control using a proportional band. This type of control is slow because one can't make the gain larger or the proportional band smaller with out risking instability. Pump controls lack the tricks I like to use like feed forward gains and derivative gains for added stabilty. Stable proportional only control is over damped control so it is slow and unacceptable. In my example application above the pump response will lag the application demands by many tenths of a second. Pressure compensate pumps with only proportional control can't see the rate of pressure change and use this derivative action to react faster. Finally the pump is always reacting to pressure changes and not the actual demand. Large accumulators will help reduced the drop in pressure but for some reason I find it hard to convince people that 50 gal accumulators are necessary to keep the pressure relatively constant. I get a little excited when a hydraulic guy looks at the pressure gage and tells me the pressure is only dropping 200 psi when in fact it is dropping three times that. Most hydraulic people don't think in terms of milliseconds yet.
Adding extra accumulators slows down the pressure changes so that the variable displacement pump has time to react, but at first all the extra accumulators do is slow down the pumps ability to react to the changes in demand because the pressure changes more slowly with changes in demand with more accumulators.

Eventually adding larger accumulators just means that the pressure is relatively constant but at the pressure where the pump is on nearly full stroke. This is why hydraulic axes need to be tuned under average load and with all the other axes moving at their normal rates.

If I can ramp up a VFD driven pump at the same time I am opening the valve I can get whatever flow I need without waiting for the system pressure to drop. This will keep the system gains constant and the control easier. The accumulator would be there only to cover for small mismatches between the valve flow and the VFD driven pump flow. The problem here is that the controller must be able to do this. I have had a few customers control VFD driven pumps they didn't uses them the way I would. The customers just wanted something to work and didn't want to make a research project out of it.

BTW, what difference does it make if the pump is a fixed or variable displacement pump in regards to pressure pulses? The only problem with a fixed displacement pump is that it is always at full stroke but the strokes would be slower when not at full speed so I can't see what difference the pump makes in regards to pressure pulses. I can see pressure pulses on variable displacement pumps too.

Why does the VFD driven pump heat the oil? You have control over the speed don't you?

Now the question is what is MR Q doing that would justify all this because it is cheaper to add more accumulators to keep the pressure constant.

Hello all,


It was interesting to read all the replies.

My application is for on/off controll of 4 cylinder functions. They have small volume but need to run really fast. They run cyclic and one cycle runs in about 1.4 secconds.

We use a variable pump with max flow 30 litres/min and a 10 litres accumulator. The thing is that there is really not space in the machine for this large accumulator. And the customer are also looking for ways to cut energy costs since this runs 24/7.

MrQ said:

My application is for on/off controll of 4 cylinder functions.

Click to expand...

Then why bother with a VFD? This does sound like gross over kill to save on the size of the accumulator. Since we don't have the specifications we can't figure out the minimum accumulator size you need but I bet you were going to try to get by with out an accumulator.

My customers have only used VFD pumps on servo applications and they worked well. I am pretty sure that one doesn't even use an accumulator but it is not high speed. In addition they monitored the pressure and position. I think you could have trouble with pressure spikes and dips if you try to get by without an accumulator no matter what pump you use. I can't see why there would be any difference between a variable displacement pump and the VFD pump except the VFD pump may be able to respond much faster. It is a shame you will only have pressure feed back to control the pump speed.

I work in Injection Molding and it has been my experence that if you are just needing exra volume then Accumlators are the way to go. They store oil during the low demand periods then give it back when you need it. We are doing think wall injection and we run 850 ton machines at 4.7 and less cycle times. This is on 4 level molds with huge injection shots. Everytime we needed speed (which is a function of volume) we added an accumlator. This method also reduces the amount of "recycling" the oil you would normally get.