Course Outline
← ChE 578: Dynamic Optimization (Optimal Control) of Chemical Processes
- ChE 578 Dynamic Optimization (Optimal Control) of Chemical Processess COURSE OUTLINE 2001/2002 Fall Semester 1999-2001 Catalog Data: ChE 578: Optimization of chemical processes under unsteady-state operation. Application of the optimal control theory and numerical algorithms to lumped-parameter and distributed-parameter chemical engineering systems. Iterative and simultaneous solution of differential/algebraic optimization problems. Nonlinear-model-based predictive control of chemical engineering systems Prerequisite: Consent of the instructor Goals: This course is designed to give seniors/graduates in chemical engineering an ability to formulate optimization problems related to chemical engineering systems operating at unsteady-state, to develop dynamic optimization computer programs, and to compare various dynamic optimization algorithms. Instructor: Prof. Dr. Uğur Akman Textbook: not required, photocopies of instructor's notes will be made available to students Reference Textbooks: Geering H.P., Optimal Control with Engineering Applications, Springer, 2007 Naidu D.S., Optimal Control Systems, CRC, 2003 Bryson A.E., Applied Linear Optimal Control - Examples and Algorithms, Cambridge, 2002 Betts J.T., Practical Methods for Optimal Control Using Nonlinear Programming, SIAM, 2001 Edgar T.F. and Himmelblau D.M., Optimization of Chemical Processes, McGraw-Hill Book Co., 1988 Lewis F.R., Optimal Control, John-Wiley & Sons, 1986 Ray W.H., Advanced Process Control, McGraw-Hill, 1981 Kirk D.E., Optimal Control Theory - An Introducton, Prentice-Hall, 1970 Bryson A.E. and Ho Y.-C., Applied Optimal Control: Optimization, Estimation, and Control, Blaisdell Publishing Co., 1969 Denn M.M., Optimization by Variational Methods, McGraw-Hill, 1969 Schechter R.S., The Variational Method in Engineering, McGraw-Hill, 1969 Prerequisites: matrix algebra, eigenvalue analysis dynamic chemical engineering systems computer programming, differential equations theory & algorithms of analytical static optimization (ChE 477) Topics: 1. STATIC OPTIMIZATION: review of matrix algebra, static optimization without constraints, static optimization with equality constraints, static optimization with inequality constraints, Lagrange multipliers, Hamiltonian, numerical solution techniques (4 classes) 2. MATHEMATICAL MODELS FOR PROCESS DYNAMICS: linear multivariable lumped models, time-domain and transform-domain models, nonlinear lumped models, distributed-parameter models, controllability, stabilizability, normality (4 classes) 3. OPTIMAL CONTROL THEORY & CALCULUS OF VARIATIONS: introduction, The Calculus of Variations: several dependent variables, boundary conditions, some special cases (6 classes) 4. OPTIMAL-CONTROL OF CONTINUOUS LUMPED SYSTEMS: more on calculus of variations, solution of the general continuous lumped-parameter optimal control problem, some state variables specified at a fixed terminal time, some state variables unspecified at initial time, minimum-time problems, tracking problem, path constraints, integral constraints, control variable equality constraints, equality constraints on functions of control and state variables, equality constraints on state variables, interior-point constraints, inequality constraints on control variables, inequality constraints on functions of control and state variables, inequality constraints on functions of state variables, constrained-input problems, continuous-time linear quadratic regulator, fixed final state and open-loop control, free final state and closed-loop control, steady-state closed-loop control and suboptimal feedback, regulators with functions of final state fixed, linear minimum-time problem, optimal linear-quadratic feedback control of nonlinear systems, perturbation control (12 classes) 5. NUMERICAL SOLUTION OF OPTIMAL CONTROL PROBLEMS: control-vector iteration, control-vector parameterization, neighboring-extremal algorithm, strategies for simultaneous solution and optimization of differential-algebraic systems, feasible-path or sequential NLP methods, infeasible-path or simultaneous NLP methods, method of weighted-residuals (orthogonal collocation) discretization of optimal control problems (9 classes) 6. MODEL-BASED PREDICTIVE CONTROL: linear-model predictive control (MPC), nonlinear-model predictive control (NMPC), Newton-type model-predictive control (4 classes) Computer Usage: use of FORTRAN (required) and MATLAB in homework and project assignments. Grading: Approximate breakdown is as follows (may vary by e.g., ± 5%): Exam .......... 25% Final Exam .... 25% Homework ...... 25% Projects ...... 25% ___________________ 100%