Why a Self Tuning Controller?

The transfer function of SBHS is assumed as

$$<tex2html_comment_mark>33 \Delta T = \frac {K_p}{\tau_ps+1} \Delta H + \frac {K_f}{\tau_fs+1} \Delta F$$ (9.1)

$\Delta$T: Temperature change

$\Delta$F: Fan input change

$\Delta$H: Heater input change

The values of $K_p$, $K_f$, $\tau_s$ and $\tau_f$ can be found by conducting step test experiments. Using these values, the parameters ($K_c$, $\tau _i$ and $\tau _d$) of the PID controller can be defined using methods like Direct Synthesis of Ziegler Nichols Tuning. However, when the apparatus is used in over a large range of temperature, the values of the plant parameters ($K_p$, $K_f$, $\tau_s$ and $\tau_f$) may change. The new values would give new values of PID controller parameters. However, in a conventional PID controlled system, the parameters $K_c$,$\tau _i$ and $\tau _d$ are defined beforehand and are not changed when the system is working. Therefore, we might have a situation in which the PID controller is working with unsuitable values that may not give the desired performance.Therefore, it becomes necessary to change or update the values of the PID parameters so that the plant gives the optimum performance.