An automated least-squares parameter identification technique and optimal PID control

This thesis presents the result of a study on the PID control theory and application. A so-called automated least squares parameter identification algorithm is developed and implemented with a HP desktop computer. This technique can automatically model most stable SISO processes in the first-order-lag-plus-delay (FOLPD) or first-order-lag (POL) form based on the least-squares curve-fitting criterion. The POL and POLPD model related PID controller tuning techniques are reviewed and improved based on the commonly preferred control performance specification: the 0.707 damping ratio criterion. Simulations of tuned PID control systems relating to the optimal models and controller settings are performed. Very interesting and encouraging results are obtained.