We have a distillation column with air-cooling unit. The aim is to control the outlet product temperature. In response to large inertia of the system, we couldn't to control the outlet product temperature. The real change in temperature takes place only after 7 minutes. During this time, PID controller output rises to 100% and the product cools. Could you please help what methods of regulation in this case are suitable?
You have a classic "dead time" case.
The simple way to approach it is by s l o w i n g the PID frequency (make it run once for half or third of the dead time in hand), and use "very small" gain for the TC loop.
Feel free to contact me direct Avihu@HiramEng.com.
Lowering the PID execution frequency is not going to help. The better approach: First try to get quantitative process Knowledge. Make an open loop (step) test and determine the process parameters (gain, time constant and dead time) from the response. With this knowledge you can calculate the controlability ratio CR, the ratio of dead time to time constant. If this is greater than 3 then the std. PID will not be able to deliver satisfactory performance. Alternatives: Use of a Smith Predictor (not really recommended, since outdated) or better Model Based Control.
If the CR is lower than 3 then calculate the PID tuning on the basis of the process parameters using one of the more powerful methods like Astrom, ITAE, Chien-Reswick-Hrones.