This paper shows that an electric are furnace off-gas system can provide va
luable manipulated variables for feedback control which can improve furnace
efficiency and contribute to safety in the workplace. Model predictive con
trol (MPC) is used to illustrate this concept using practically motivated c
ontrol objectives. An initial verification of a non-linear furnace model wi
th plant data is shown. The design of MPC controllers for the furnace is di
scussed and results are shown by way of simulation. Evaluation of the final
controller against traditional manual operation is done, and the setpoint
tracking capability of the controller is tested. (C) 2000 Elsevier Science
Ltd. All rights reserved.