Distributed parameter state estimation and optimal feedback control--An experimental study in two space dimensions

Both optimnl and suboptimal distributed parameter state estimation algorithms were applied in real time to a heated cylindrical ingot. Following this, deterministic and stochastic feedback controllers were implemented. Detailed studies of the effects of variables such as sensor choice, measurement error, model accuracy, controller and estimator parameters, and computational approximations on the control system performance are reported. The results show that the control scheme is robust to measurement noise and performs well even with only a single thermocouple sensor provided to the state estimator. Real-time computational requirements were modest and indicate that practical industrial implementation of such state estimation and control schemes is feasible.