Dynamic Load Control

Dear Sir,

I am working an a Delta’s RMC Motion controllers.
Presently I am working on hydraulic Test Bench used to test the endurance testing for servo actuators.
Here the test set up is as bellow -
one is the main Loading Jack controlled by servo Valve, and the other one is the Test Actuator.
The test requirement is to apply a constant load using the Loading Jack on the Test Actuator which is moving with some constant speed (ex: 25mm/sec) and so on…

Presently I am not able to apply a constant load when the test actuator is moving.

Here I am attaching the plots for your reference.

If anyone has some kind of algorithm or any one worked on these kind of application, then please guide me in these regards,

Thanking you

Arun Bandekar.
Dynamic Load Control.ppt (201 KB)

Controlling pressure and speed at the same time is tricky. Because it takes a flow of oil to move the system where it doesn’t if you controlling pressure or force. Do you know how fast the system is moving? This is important.

I just got through doing a similar project where the goal is maintain a force on wheels as they moved up and down over ruts. I had to estimate the speed.

There is another possible problem. There must be sufficient resistance to motion. If there is no resistance there will be no force.

thanks for the reply…

Dear Sir,

Thanks for the reply.

There are basically two test to be done.

  1. Apply a constant Load on a actuator moving with a constant speed of 25mm/Sec.
  2. Apply a constant load on a actuator following a sine wave of 3mm amplitude at 1 Hz frequency.

If you know the speeds then you can generate a feed forward or bias offset that anticipates how far open the valve must be instead of waiting for the integrator to wind up. For instance the feed forward value to move may be 10% or 1 volt for every inch per second. If you are moving at one inch per second the bias should be ramped to 1 volt. Like wise with the sine wave. A 1 HZ sine wave with an amplitude of 1 inc needs to have a bias offset that changes with the target velocity. The voltage would need to go as high as 2*PI volts. Due to the rod on one side of the cylinder the bias in the reverse direction will be different.

Dear Sir,

Thanks for your support.
I tries as per your instruction but there is not significant improvement in the system.
Is there any thing else to be done in the system?


What are the gain values now? If you post a real plot file, then we can see the gain values included in the plot.

I think you may need to use the Output Filter and more differential gain. The differential gain can cause lots of high-frequency oscillation, which is just beginning to occur on the plot. The Output Filter can filter out that high-frequency oscillation, aloowing the differential gain to do its work. The Output Filter is a low-pass filter, so a lower value filters more than a higher value. 0 is disabled. Start with a high value, say 100, and work your way down. You should be able to increase the differential gain and proportional gain as you decrease the Output Filter.