Self-Tune is a type of automatic tuning algorithm that continuously optimizes PID terms while a controller is operating. It uses a pattern recognition algorithm, which monitors the control deviation.
The Self-Tune algorithm observes one complete deviation oscillation before calculating a new set of PID values. Successive deviation oscillations cause the values to be recalculated so that the controller converges on optimal control.
When the controller is switched off, these PID terms are stored, and are used as starting values at the next switch on. The stored values may not always be ideal, if for instance the controller is brand new or the application has changed. In these cases, the user can utilize pre-tune to establish new initial values. Self-tune will then fine-tune these values as it monitors any control deviation. Use of continuous self-tuning is not always appropriate.
For example applications which are frequently subjected to artificial load disturbances, for example where an oven door is likely to be frequently left open for extended periods, can lead to incorrect terms being calculated. In addition, because self-tune tunes for PID control, it is not recommended for applications, that require PD, PI or proportional only control.