CENG 471 - Process Control Lab

Spring 2000

PART II: CONTROLLER FIELD TEST AND CONCLUSIONS

• Check your best PI and PID controllers (i.e. the controllers that minimize the amount of off-spec product) designed for the first-order plus time delay model) on the experimental setup. How does your predicted performance compare with the actual one?
  
  
• Use a trial-and-error procedure to improve (if possible) the performance of your controller.
  • Start with a controller design of your choice (i.e. a set of Kc, tI and tD).
  • Make up to three changes in the controller parameters and test your designs. Your report must justify your choice of parameters in this trial-and-error process and attempt to explain why things worked (or didnŐt work).
    
    
• Critically analyze your results making sure that you provide detailed answers to the following questions:
  • Which design gave you the best control of the 3 CSTR system?
  • Which controller designs failed and why?
  • Were the experimental results consistent with SIMULINK predictions?
  • Did PID controllers behave better than PI controllers? Why?
  • Which approach seems better for designing PID controllers?

  
  Use tables of performance results (e.g. integral performance criteria) and plots to fully justify your answers to these questions. Make any additional comments that you think could help designers of feedback controllers like this one.

Reminder on Honor Code Policy: "The students may talk to each other, the labbies or the instructor about any assignment in the course that is not specifically designated as pledged. This assistance is limited to the discussion of the problem. Consulting another student's solution (even from a previous class on the same subject) is prohibited, and submitted solutions may not be copied from any source."