pH Loop Optimization

Results of PID Control optimization of a pH loop

Een biofarmaceutisch bedrijf heeft DotX gevraagd om:

  • On-site assistentie bij het toepassen van de PID Tuner om 4 regellussen af te stemmen in twee 2000L bioreactoren.
  • On-site training in PID-afstemming met behulp van de DotX PID Tuner-software.

In this post, we explain how we worked, and the amazing results.

The control challenges

The Bioreactor under consideration is a modern high tech piece of equipment with PID loops for pH, Dissolved Oxygen (DO) and temperature. The most challenging to control is DO. This is due to the fact that DO is controlled by a mix of gasses, including oxygen, CO2, and air. The PID output of the DO loop is the input for a schedule of these gasses.

Analysis of the PID loops

We analysed the control schemes, and data of the last batch. From our analysis, we found:

  • All PID control loop tunings could be improved by at least a factor of 2 (in terms of closed-loop response time and damping).
  • There was an issue with the schedule of gases used to control DO: at one point, this schedule was causing unstable loop behavior.
  • It was useful to add low-pass filters to the DO signal to prevent spikes when applying the D-action in the PID controller.

Result with the DO loop

Below, you see measured data of the DO loop before optimisation (collected in a batch before), and after (collected in the batch after changing the PID settings. Clearly, around t= 75 hours, the deviations from target have reduced by more than a factor 10. This is partly due to the changes in the gas schedule, and the improved PID tuning.

Result with the pH loop

Below, you see measured data of the pH loop before optimisation (collected in a batch before), and after (collected in the batch after changing the PID settings. Clearly, the closed loop response time has improved welll over a factor 4, damping has improved, and the deviations from target have reduced by more than a factor 3.

en_USEnglish