Hydraulic ram position / pressure control

[rak_plc]

Hi All,

I could do with some advice on a hydraulic control system. It is necessary for me to accurately position a vertical hydraulic ram with an upper tool part. This will be subjected to an upwards force from below due to an hydroforming process. I've not been involved in the specification of the hydraulics but I have the following equipment to control:

String pot for the ram position
Pressure transducer for ram
Analogue valve for ram pressure
Analogue valve for ram flow
5cc digital valve back to tank

So the process will be - move the ram to a position - this is OK and I can stop to 1/10mm (very low speeds). Then during hydroforming there will be a gradual upwards force reaching the order of 55-100 tonnes which I need to then hold the ram in position by increasing the downwards force.

The only way I can think of tackling this is using a PID instruction in the Siemens S7 1500 plc which is what I will be attempting.

As soon as I saw it though I thought this is an hydraulic servo application - but I don't have the knowledge to offer an alternative to what they have got.

Any suggestions would be much appreciated.

 

[504bloke]

You should be able to hold the ram position with the ram position feedback (string pot?), You can use the Siemens PIDs or roll your own logic if its all relativley slow, presume the ram flow valve is for speed ? hard to really advise without the hydraulic circuit with controls etc

 

[Tom Jenkins]

You can't control the pressure. The load dictates the pressure required to move or hold a position.

 

[rak_plc]

As the hydroform pressure below increases, I need to increase the pressure in the upper ram to maintain the position of the upper part. So with 55 tonnes acting upwards I will need to apply 55 tonnes downwards to keep it in position.
The theory at the moment is I will move the upper ram into position, then when hydroforming starts I will need to look at the ram position and take action when the ram starts to push back - using a PID instruction maybe to drive the analogue valve to give me the pressure in the ram to keep it where it is.

Its a strange one because the ram has no load on it and is not subjected to an upwards force until hydroforming is taking place - the hydroforming of the titanium part is complicated in itself has the metal yields and flows causing fluctations in the upwards force.

 

[Steve Bailey]

How does the hydroforming cause any upward movement of the ram? Once you have positioned the ram, why can't you simply close off any flow path to it?

 

[rak_plc]

How does the hydroforming cause any upward movement of the ram? Once you have positioned the ram, why can't you simply close off any flow path to it?

good point, I need to ask the question, too busy down this blind alley!

 

[lfe]

I think you have to do sequential programming, I imagine these steps:

- Raise the piston until the pressure begins to increase. I don't know if you can go at full speed or if you have to regulate the position during the ascend, in this case you will have to use PID and generate a setpoint ramp.

- Once there is initial thrust pressure, another PID begins to control the pressure according to a pressure setpoint ramp.

- Once the maximum pressure is reached, the pressure continues to be regulated but the setpoint no longer increases. This for a while.

- Finally the piston descends to the initial position, normally in this type of press this is done at full speed

 

[ndzied1]

Lot's of questions here. I think I have the general idea of what you are trying to achieve but there are a lot of gaps in the details.

One big questions is when you say "hold the ram in position" what do you really mean?

Do you need to hold to 1/4", 1/100", 1/1000" ? these require different equipment and control methods.

As Steve mentions you can close off the flow path but you will get some motion doing this. This is because oil is compressible. It's not much but does make a difference. Back of the napkin formula for this is that the volume of oil decreases 0.5% per 1000 psi of compression. Your cylinder size and the opposing force will determine the volume of oil under compression and the pressure increase in the cylinder. If you can live with that motion, this is probably the easiest way to go.

If you are going the closed loop control route and you need to keep the ram where it is you really are doing position control. The fact that the load is varying doesn't change the fact that you are trying to maintain a position. It does mean that your cylinder needs the capability to produce more force than is opposing it but you will still always be in position control.

Depending on the speed and accuracies required a PID in a PLC may not be able to do what you need to do. For sure, your job will be made infinitely easier and more robust by using a closed loop controller designed specifically for hydraulics such as a Delta RMC75.

You mention that the hydraulic components were selected by someone else. It could be that the components that they have picked will make your job very difficult or even impossible. Besides the valves, this also includes the cylinder type and size and even the position feedback type and resolution.

Again, it will all come down to how accurately you need to hold the position and how fast the opposing force is changing. You should start from those items to come up with requirements and then be designing the hydraulic system. Remember to take into account all worst case scenarios as well as things like emergency stop / breakage situations.

It sounds like a fun application. Please share more details if possible

 

[rak_plc]

Lot's of questions here. I think I have the general idea of what you are trying to achieve but there are a lot of gaps in the details.

One big questions is when you say "hold the ram in position" what do you really mean?

Do you need to hold to 1/4", 1/100", 1/1000" ? these require different equipment and control methods.

As Steve mentions you can close off the flow path but you will get some motion doing this. This is because oil is compressible. It's not much but does make a difference. Back of the napkin formula for this is that the volume of oil decreases 0.5% per 1000 psi of compression. Your cylinder size and the opposing force will determine the volume of oil under compression and the pressure increase in the cylinder. If you can live with that motion, this is probably the easiest way to go.

If you are going the closed loop control route and you need to keep the ram where it is you really are doing position control. The fact that the load is varying doesn't change the fact that you are trying to maintain a position. It does mean that your cylinder needs the capability to produce more force than is opposing it but you will still always be in position control.

Depending on the speed and accuracies required a PID in a PLC may not be able to do what you need to do. For sure, your job will be made infinitely easier and more robust by using a closed loop controller designed specifically for hydraulics such as a Delta RMC75.

You mention that the hydraulic components were selected by someone else. It could be that the components that they have picked will make your job very difficult or even impossible. Besides the valves, this also includes the cylinder type and size and even the position feedback type and resolution.

Again, it will all come down to how accurately you need to hold the position and how fast the opposing force is changing. You should start from those items to come up with requirements and then be designing the hydraulic system. Remember to take into account all worst case scenarios as well as things like emergency stop / breakage situations.

It sounds like a fun application. Please share more details if possible

The specification is to not allow the ram to move more than +/- 1mm. The press is 300 tonne, ram is 382mm diameter. The hydroform part can vary in size but the 1st part is 114mm diameter, with 530 bar in it this will produce 55 tonnes upwards force. Initial testing has revealed that the main ram is not responding quickly enough and is getting pushed back. Hence I will be introducing the PID function to try and improve the situation.

I'll see what happens next week!

 

[ndzied1]

When you say main ram. Is it truly a ram in that it can only push down. Or is it a cylinder that has ports for extending and retracting?

If just a ram, how does it go back up? Maybe separate jacking cylinders.

Is tonne 1000kgf?

Do you have a hydraulic schematic you can share?

 

[drbitboy]

What is the timescale of this process? E.g. assuming perfect position control by the hydraulic ram, what does the hydroforming-side pressure trend look like over time? How long should each piece take to complete?

Have there been any FEA or FEM simulations performed of any part ofhe process? What sort of natural frequencies are anticipated?

What is the compressibility of the hydroforming fluid (kPa**-1)?

What are the volumes of the hydroforming-side and hydraulic-side spaces?

We know the position-normal areas on each side, so if the hydroforming fluid side pressure can be measured, then the steady-state (balancing) hydraulic pressure can be calculated. Would it be possible to simply scale the position error to a hydraulic trim offset pressure, and add that to the calculated balancing pressure as a feed forward signal to the "Analogue valve for ram pressure?"

What kinds of pumps move the two opposing fluids? Specifically, what are the temporal and noise characteristics of the pressures on either side?

 

[drbitboy]

Initial testing has revealed that the main ram is not responding quickly enough and is getting pushed back.

How far back? What type of control is in place on the hydraulic side? E.g. is it some form of position control feeding forward to one of the control elements (valve for ram pressure or flow)? What is allowing the ram be pushed back e.g. compressibility of hydraulic fluid, and/or backflow through a valve, and/or leakage, and/or deformation of press, etc.? Over what timeframe is it getting pushed back?

 

[drbitboy]

This article, by forum member @Peter Nachtwey, suggests that hydroforming processes require special, fast controllers, as suggested by @ndzied1 in an earlier post.

 

[rak_plc]

What is the timescale of this process? E.g. assuming perfect position control by the hydraulic ram, what does the hydroforming-side pressure trend look like over time? How long should each piece take to complete?

Have there been any FEA or FEM simulations performed of any part ofhe process? What sort of natural frequencies are anticipated?

What is the compressibility of the hydroforming fluid (kPa**-1)?

What are the volumes of the hydroforming-side and hydraulic-side spaces?

We know the position-normal areas on each side, so if the hydroforming fluid side pressure can be measured, then the steady-state (balancing) hydraulic pressure can be calculated. Would it be possible to simply scale the position error to a hydraulic trim offset pressure, and add that to the calculated balancing pressure as a feed forward signal to the "Analogue valve for ram pressure?"

What kinds of pumps move the two opposing fluids? Specifically, what are the temporal and noise characteristics of the pressures on either side?

"We know the position-normal areas on each side, so if the hydroforming fluid side pressure can be measured, then the steady-state (balancing) hydraulic pressure can be calculated. Would it be possible to simply scale the position error to a hydraulic trim offset pressure, and add that to the calculated balancing pressure as a feed forward signal to the "Analogue valve for ram pressure?" "

I have done an initial study of the ram force graph for a given % analogue output from the PLC, my idea being to use this data as the starting point for controlling the upwards force. The hydroform pressure is measured and I can then calculate the force upwards and then using the equation of my graph use this as the starting point for ram control. It is this method that failed to counteract the upward force. As you state I now need to apply some feed forward signal based on the error to modify this value.
The process takes around 60 seconds. The hydroforming pressure trend is quite erratic at the maximum pressure.

The hydroform volume is probably 3 litres with its pipework, the press ram is probably 10x that

 

[rak_plc]

When you say main ram. Is it truly a ram in that it can only push down. Or is it a cylinder that has ports for extending and retracting?

If just a ram, how does it go back up? Maybe separate jacking cylinders.

Is tonne 1000kgf?

Do you have a hydraulic schematic you can share?

Its an hydraulic cylinder, there is a separate circuit for retracting it and advancing it which is closed when hydroforming. An additional analogue valve then takes over to change the pressure in the ram to keep it in position to counteract the upwards force of the hydroform. If it moves down too far it can't move back on its own this would only be caused by the upward hydroform pressure. I've been asking for a schematic - its getting to the point where I will have to draw it myself. I've replied to drbitboy with some detail of the % analogue - force graph I have produced. Applying a feed forward error correction to this seems to be the next step forward.

Yes 1000kgf