Operational envelopes for batch processes

Authors

  • Nouri J. Samsatli,

    1. Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, U.K.
    Search for more papers by this author
  • Mona Sharif,

    1. Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, U.K.
    Search for more papers by this author
  • Nilay Shah,

    Corresponding author
    1. Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, U.K.
    • Centre for Process Systems Engineering, Imperial College of Science, Technology and Medicine, London SW7 2BY, U.K.
    Search for more papers by this author
  • Lazaros G. Papageorgiou

    1. Dept. of Chemical Engineering, University College London, London WC1E 7JE, U.K.
    Search for more papers by this author

Abstract

Batch processes are often subjected to a number of sources of variability, which are frequently overlooked due to operating policy implementation. In many pharmaceutical and speciality chemical processes, which are predominantly batch in nature, a high proportion of the operating procedure is implemented manually, which can lead to significant variability. The effect of this variability is typically neglected at all stages in process development and even when full-scale production begins. The result of any variability can have a dramatic effect on some process stages (particularly sensitive ones such as crystallization), and it is therefore important to understand how variability in any of the operating variables affects the overall process. Ideally, one would like to design processes that are less sensitive to variability and specify precisely the limits of any allowable variability. A method for designing batch processes using “operating envelopes” that specify process operation in terms of a region within which the operating variables should be maintained to guarantee feasible and profitable operation of the process is presented. The problem is formulated as a multiscenario dynamic optimization problem, and an efficient hierarchical solution procedure is presented. Finally, the applicability of the procedure is demonstrated by using two example problems.

Ancillary