I would be interested in hearing about other engineer's experiences with derivative control. What applications has it proved successful? What applications has it resulted in instability?

 

NEVER, EVER use derivative control on solid hydraulic systems.

 

I'm not sure that this is a valid comment. I have
used derivative control on servo-hydraulic systems
with our controllers in several systems. This
includes both very large hydraulic systems and
relatively small ones.

Davis Gentry
Delta Tau Data Systems
www.deltatau.com

 

Why?

 

hi,

control loop parameters are usually set during commissioning. Sometimes its unstable if the one programming it dont have full knowledge of the process and forgot to consider all the factors that would affect it. However, this type of control is very useful in applications that require nearly constant controlled output based on set desired values.

dmr

 

Most of my control engineering work is performed in the North Sea oil & gas industry. After many years of optimising PID controllers on many platforms, I can say that I have never found any benefit in derivative action in the process control world and in fact find that the stability is generally far improved by switching it off! The reason is the noisy nature of the process inputs with frequencies far too low for filtration.

For example, take a look at the trend display on my website:

http://www.contek-systems.co.uk/UTune trend noise.jpg

The blue line on the lower trend is the process variable to a CCC anti-surge controller and clearly shows the unstable nature of the signal. Derivative action would have caused far too much unnecessary movement of the recycle control valve. The controller was very successfully tuned using PI alone.

Regards
John Greene
Process control consultant
Contek Systems Ltd
Aberdeen, UK
http://www.contek-systems.co.uk/Consultancy.htm

 

I have used derivative control to advantage on temperature and pH control loops. A noise-free measurement is critical. If you have to lag the measurement heavily to make derivative control possible you might as well use simple PI control; see John Greene’s comments above. Depending on the relative size of noise and disturbance, a non-linear filter might make derivative control practical.

Derivative loops can be quite vulnerable to variations in process gain. You may need to do ratio control, or some other heat-and-material-balance linearization to stabilize the process that the derivative controller interacts with.

My company provides control strategy designs for many different platforms. If you’re interested contact me at [email protected].