Tuning controllers is primarily a scientific process that involves adjusting the time and amount parameters of the controller to match those of the process. Typically, an open-loop test is conducted on the process to determine the necessary parameters, and simple tuning rules based on these parameters have been found to work well for a large proportion of industrial control loops. Alternatively, tuning parameters can be derived from a closed-loop test, but it may not be as comprehensive.
It is important to note that all control loops will cycle if the controller gain is high enough. This cycle’s period, known as the natural period, plays a significant role in determining the loop’s potential performance, with a shorter natural period being more desirable. However, it is paradoxical that the natural period is not determined by the largest lags in the loop but rather by the dead time and smaller lags. In addition, a loop’s potential performance is limited by certain lags, making it challenging to improve performance through tuning.
Tuning rules are designed to achieve reasonably tight control, but this may not always be the objective. Although many loops may not require tight tuning, all loops must be tuned during the controller installation process. It is essential to consider whether retuning a loop is justified, and if the desired improvement can be realistically achieved through tuning before making any tuning adjustments.
It is crucial to recognize and address the typical “gremlins” that can cause loops to behave abnormally. Failure to do so can result in detuning a loop and a loss of trust in the scientific approach. With patience and dedication, learning the methods of tuning controllers can be a rewarding experience, enabling individuals to converse with others and share their experiences. Without a thorough understanding of loop performance, tuning becomes a series of hit-and-miss attempts, which may lead to frustration and suboptimal results.