[01] M. A. Mustafa, Department of Chemical Engineering, Faculty of Engineering, University of Khartoum, Khartoum, Sudan.

An important amount of work exists on the topic of optimal operation and control of batch distillation. Although previous efforts are still based on the assumption of an accurate process model being available, there will always remain the challenge of practical applications. Reinforcement Learning (RL) has been recognized already as a particularly suitable framework for optimizing batch process operation however no successful application to batch distillation has been reported. Thus, this paper presents RL as an automatic learning approach to batch distillation. Incremental batch to batch learning is demonstrated, for a range of case studies starting from a small initial training data set and with no assumption of prior knowledge of VLE. Furthermore the robustness of the RL algorithm, towards uncertainty in process variables, is demonstrated.

Reinformcent Leanring, Batch Processes, Optimisation

[01] Diwekar U. M. Batch distillation: Simulation, optimal design and control. Carneige Mellon University, Pittsburg, Pennsylvania; 1995. [02] Luyben W. L. Practical Distillation Control. Van Nostrand Reinhold. New York; 1992. [03] Bonny L., Domenech S., Floquet P., Pibouleau L. Recycling of slop cuts in multicomponent batch distillation. Comput. Chem. Eng. 1994; 18: S75-S79. [04] Mujtaba I. M. and Macchietto S. An optimal recycle policy for multicomponent batch distillation. Comput. Chem. Eng. 1992; 16: S273-S280. [05] Sorensen E. Alternative ways of operating a batch distillation column. Institution of Chemical Engineers Symposium Series 1997; 142: 643-652. [06] Farhat S., Czernicki M., Pibouleau L., Domenech S. Optimization of multiple-fraction batch distillation by nonlinear programming. AIChE Journal 1990; 36: 1349-1360. [07] Converse A. O., Gross G. D. Optimal distillate-rate policy in batch distillation. Ind. Eng. Chem. Fund. 1963; 2: 217-221. [08] Keith F. M., Brunet. Optimal operation of a batch packed distillation column. Canadian J. Chem. Eng. 1971; 49: 291-294. [09] Coward I. The time optimal problem in binary batch distillation. Chem. Eng. Sci. 1967; 22: 503-516. [10] Mayur D. N., Jackson R. Time optimal problems in batch distillation for multicomponent mixtures columns with hold-up. Chem. Eng. J. 1971; 2: 150-163. [11] Egly H., Ruby N., Seid B. Optimum design and operation of batch rectification accompanied by chemical reaction. Comput. Chem. Eng. 1979; 3: 169-174. [12] Hansen T. T., Jorgensen S. B. Optimal control of binary batch distillation in tray or packed columns. Chem. Eng. J. 1986; 33: 151-155. [13] Mujtaba I. M., Hussain M. A. Optimal operation of dynamic processes under process-model mismatches: Application to batch distillation. Comput. Chem. Eng. 1998; 22: S621-S624. [14] Kerhof L. H., Vissers H. J. M. On the profit of optimum control in batch distillation. Chem. Eng. Sci. 1978; 33: 961-970. [15] Logsdon J. S., Diwekar U. M., Biegler L. T. On the simultaneous optimal design and operation of batch distillation columns. Chem. Engi. Res. Des. 1990; 68: 434-444. [16] Mujtaba I. M., Macchietto S. Efficient optimization of batch distillation with chemical reaction using polynomial curve fitting. Ind. Eng. Chem. Res. 1997; 36: 2287-2295. [17] Cressy D. C., Nabney I. T., Simper A. M. Neural control of a batch distillation. Neural Computing and Applications 1993; 1: 115-123. [18] Stenz R., Kuhn U. Automation of a batch distillation column using fuzzy and conventional control. IEEE Transactions on Control Systems Technology 1995; 3: 171-176. [19] Wilson J. A., Martinez E. C. Neuro-fuzzy modeling and control of a batch process involving simultaneous reaction and distillation. Comput. Chem. Eng. 1997; 21: S1233-S12. [20] Barolo M., Cengio P. D. Closed-loop optimal operation of batch distillation columns. Comput. Chem. Eng. 2001; 25: 561-569. [21] Kim Y. H. Optimal design and operation of a multi-product batch distillation column using dynamic model. Chem. Eng. Process. 1999; 38: 61-72. [22] Lopes M. M., Song T. W. Batch distillation: Better at constant or variable reflux? Chem. Eng. Process: Process Intensification 2010; 49: 1298-1304. [23] Pommier S., Massebeuf S., Kotai B., Lang P., Baudouin P., Floquet P., Gerbaud V. Heterogeneous batch distillation processes: Real system optimization. Chem. Eng. Process: Process Intensification 2008; 48: 408-419. [24] Martinez E. C., Pulley R. A., Wilson J. A. Learning to control the performance of batch processes. Chem. Eng. Res. Des. 1998a; 76: 711-722. [25] Martinez E. C., Wilson J. A. A hybrid neural network first principles approach to batch unit optimization. Comput. Chem. Eng. 1998b; 22: S893-S896. [26] Mustafa M. A., Wilson J. A. Application of Reinforcement Learning to Batch Distillation, The Sixth Jordan International Chemical Engineering Conference, 12-14 March 2012, Amman, Jordan, 117-126. [27] Sutton R. S., Barto A. G. Reinforcement Learning: An Introduction, The MIT Press, Cambridge, Massachusetts, London, UK; 1998. [28] Martinez E. C., Wilson J. A., Mustafa M. A. An incremental learning approach to batch unit optimization. The 1998 IChemE Research Event, Newcastle; 1998c. [29] Bellman R. Dynamic Programming, Princeton University, Press, Princeton, New Jersey; 1957. [30] Carling A. Introducing Neural Networks, SIGMA Press, UK; 1992. [31] Baird L. C., Klopf A. H. Reinforcement Learning with High-dimensional Continuous Actions, Technical Report WL-TR-93-1147, Wright Laboratory, Wright Patterson Air Force Base; 1993. [32] Smoker E. H. Analytical determination of plates in fractionating columns. Trans AICHE 1938; 34: 165. [33] Tolliver T. L., Waggoner R. C. Approximate solutions for distillation rating and operating problems using the smoker equations. Ind. Eng. Chem. Fundam 1982; 21: 422-427. [34] Jafarey A., Douglas J. A. McAvoy T. J. Short-Cut Techniques for Distillation Column Design and Control. 1. Column Design. Ind. Eng. Chem. Process Des. Dev. 1979; 18; 2: 197–202.