I am using LabVIEW and NI DAQ cards. I am operating a heater using PID controller in LabVIEW. How can I tune the value of the gains so that I can get an excellent response from the heater? By excellent response, I mean the system should get to the set temperature in reasonable time, not too fast and not too slow.
Well, i am using a 200 W heater and am using copper as conducting media. I think i have to do it manually which is a bit cumbersome process.
Simplest method is Zigler-Nichols method. Refer to the book Modern Control Engineering by K. Ogata. Also refer to many papers available in RG on tuning of PID controller.
here is a good starting point resource:
http://www.controleng.com/single-article/tuning-thermal-pid-loops/aebf3cf5326fbf2db40450e202903301.html
In addition to Florians excellent response I would add, that addition of feedforward node usually helps in getting better performance. You should already have a clue what power settings work for you well to obtain some temperatures. Add this values to output and you will be able to work on lower P values to get reasonable settling times.
Temperature systems are normally modeled as first order systems, i.e. k/(s+a). You can empirically get the values of k and a.
After determining these values, you can analytically derive the PID gains depending on your specifications. Normally, a PI controller will do the work. To have a faster response, you can add a D gain.
The closed loop system model is given in this link.
http://www.chegg.com/homework-help/questions-and-answers/model-temperature-control-system-large-test-chamber-shown-assume-d-s-0-door-chamber-remain-q5440642
However, this is continuous-time modeling, but, nonetheless, before you proceed with digital control, it is still better to understand the continuous-time modeling.
Dear Nauman Khan,
For this system I would recommend an experimental First order plus dead-time model, the parameters could be determined by the step response and then you can tune PID Controllers with some Integral Performance Criteria. A good reference could be the book:
Principles and Practices of Automatic Process Control by Carlos A. Smith, Armando B. Corripio
Regards,
Alfonso Gómez
Matlab is very strong in field of tuning kp , ki & kd however its automatic tuning requires knowledge about a lot of other terms like integrator windup etc........
But manual tuning isn't too complicated problem in matlab, once you get the step response, you can easily manually tune. Infact I manually tuned for DC motor speed control. The only thing is getting Transfer Function of Heater.......Is LabView a requirement????
Matlab even can give you the transfer function of your heater, there are two methods:
Tuning is not a big issue , its easy once you know transfer function.
A heater is a very slowly process. So, an on-off command with hysteresis and a PI element may be recommended.
i also would like to suggest to use an On-Off method. else ur heater cn control it heating, going for a PID would be nice learning experience.
According to Ang Jia Yi Constantin Volosencu answers I would like to reccomend you to use an On-Off algorithm as a good element of a tuning relay procedure, described by Astrom and Hagglund in e.g. http://www.controleng.com/search/search-single-display/relay-method-automates-pid-loop-tuning/4a5774decc.html.
I think it is a nice solution which I tested in practical 2th order processes with delay.
A procedure is relatively simply to implement and safe during experiments with real time plants. Some problem is a moment of switich from On-Off algorithm to PID algorithm, but it is possible to apply then some heuristic function to minimize a problem of excessive difference between present control value from OnOff algortihm to the future PID control signal.
- Badges
- Science topic
- Similar topics
- Engineering
- Mechanical Engineering