Asynchron getaktete Gegenstromchromatographie – Prinzip und optimaler Betrieb

Use of Single-Column Models for Efficient Computation of the Periodic State of a Simulated Moving-Bed Process

“…These methods are less robust when handling the minimization or maximization of nonlinear, nonsmooth objective functions. However, if a preliminary parameter study is performed in order to find initial values close to optimum before the optimization problem is solved, these methods require far less function evaluations to converge than stochastic methods. ,− ,, If a gradient-based method is selected, the derivatives are easier to compute (and require less computational effort) from the algebraic solution of the complete discretization method than from the dynamic simulation of the single-column model.…”

“…A detailed SMB process model is computationally expensive. While for exploratory or small parametric studies the computing time is not an issue, that may no longer be the case for rigorous model-based optimization, ,− especially when the number of optimization variables is increased by the extra degrees of freedom in the operating conditions of nonstandard SMB processes. − It is thus apparent that the increasing complexity of the optimization problems addressed in preparative chromatography calls for efficient methods to compute the periodic state. However, with a few exceptions, , little has been published on this issue.…”