On-Line Steady-State Gain Identification for On-Line Optimization of Multivariable Constrained Chemical Processes.
Date of Award
Doctor of Philosophy (PhD)
Armando B. Corripio
The Amoco Model IV Fluidized Catalytic Cracking Unit (FCCU) (McFarlane et al, 1993) was augmented with Lee & Groves Model (1985) to provide for a simple but realistic product distribution model. This enabled a better and realistic on-line optimization of the FCCU. A Supervisory Multivariable Constrained Optimization (SMCO) algorithm was applied to this augmented model. It was observed that the FCC unit is optimized successfully, without violating any constraints. The parabolic error-penalizing function, incorporated in the algorithm allowed the constraints to asymptotically approach the limits without violating them. The on-line optimization to FCC unit led to maximizing of the feeds, reduction of reactor pressure and maximizing of the reactor temperature. The algorithm handled the benchmark disturbances very well by cutting down or increasing the manipulated variables as needed. An On-line Steady-State Gain Identification (OSGI) method was developed to identify the steady-state gain matrix (SSGM) of any multivariable system. The SSGM is frequently used in multivariable control and/or on-line optimization. This method identifies the steady-state gains with all the controllers in the closed loop and thus can be implemented on-line. The use of the method was demonstrated in identification of SSGM for SMCO application for a CSTR. Two types of reactions were considered: irreversible and reversible reaction. The OSGI identified the drifting SSGM accurately for irreversible reaction. For the reversible reaction, OSGI was also employed to identify the cost-partials along with SSGM, which enabled the SMCO to track the drifting optimum with changing feed rate.
Chitnis, Umesh Kishore, "On-Line Steady-State Gain Identification for On-Line Optimization of Multivariable Constrained Chemical Processes." (1996). LSU Historical Dissertations and Theses. 6326.