OPTIMAL AVERAGING LEVEL CONTROL FOR THE TENNESSEE EASTMAN PROBLEM

This paper focuses on an industrial test problem, the Tennessee Eastma n problem. The Tennessee Eastman process is moved to three operation m odes with 90/10, 50/50 and 10/90 product G/H ratios. The base control system proposed in an earlier paper can properly control the process a t all the three operation modes. Optimal averaging level control is ad ded to the base control to improve the control of product flow during random and step feed disturbances. A new implementation of the optimal feedback averaging control algorithms (McDonald et al., 1986) is pres ented in order to deal with random disturbances. Simulation results sh ow that control of product Bow rate is significantly improved by using the optimal feedback averaging level control at all three operation m odes.