Tuning a PID loop _can_ be done in a more or less heuristic manner -- see
the second half of the article:
http://www.wescottdesign.com/articles/Sampling/pidwophd.html . This sort
of heuristic tuning works when your system can work adequately with a wide
range of tuning constants. Unfortunately, without experience the only way
you can tell if this is so is to give it a whirl.
But there are any number of formal methods for tuning PID controllers
ranging from the traditional frequency response method (see my book,
http://www.wescottdesign.com/actfes/actfes.html , for details) to other
system ID+design methods, to completely algorithmic automatic
on-line adaptive tuning methods.
Most of the methods that put you in the driver's seat not only let you find
the right tuning values, they also give you some sort of metric to predict
performance and system robustness. If you know enough about what you are
controlling it allows you to do your preliminary tuning before the system
is even built, and hence decide if the performance will be adequate or if
you need to change the proposed plant design.
So you can pretty much pick your poison.
Control systems and communications consulting
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.