Alexandru Woinaroschy

An interactive multi-objective optimization framework for sustainable design of bioprocesses

Abstract Interactive multi-objective optimization methods have considerable advantages over the other multi-objective optimization methods and are well suited for biochemical engineering problems. Interactive optimization is a very complex task which requires appropriate software to handle multiple objectives, displaying intermediary results and allowing the decision maker to specify preferences for each iteration. The proposed strategy combines Matlab and SuperPro Designer simulator. This way, the SuperPro Designer simulator benefits from the available toolboxes, computation, and visualization advanced features of Matlab. By linking Matlab and SuperPro Designer simulator, an optimization loop is created. This allows an automatically and repeatedly bidirectional exchange of variables between optimization algorithm from Matlab and SuperPro Designer simulator. In order to fully automate the optimization process, software based on Component Object Module technology, which contains three friendly graphical interfaces, was created. The presented strategy is implemented for a l -lysine feed supplement plant, both economic and environmental objectives being considered.

A sensitivity analysis of the fed‐batch animal‐cell bioreactor with respect to some control parameters

Animal cell culture is widely used in the manufacture of valuable products, and this process is nowadays seeing a rapid expansion. The growth of animal cells is a complex process, because the cells are very sensitive to environmental changes (in, for example, nutrients, pH, temperature, oxygen and osmolarity) during this phase and to the toxic compounds produced by the cell itself. Ammonia and lactate are the two major waste materials of cell culture. They can have inhibitory effects on cell growth and product (monoclonal antibodies among others) formation. In order to model the behaviour of a fed‐batch animal cell bioreactor producing monoclonal antibodies, it is necessary to use a complex kinetic model with optimal operating patterns ensuring high productivities. Good knowledge of such domains of operating parameters, together with the understanding of the response of this rather complex system to small modifications in the working conditions, are essential for on‐line control to improve the quality of product and the yield of an animal cell culture. The present study focuses on the sensitivity analysis of a fed‐batch animal cell bioreactor with respect to some candidate control parameters (substrate set‐point concentrations, feeding time‐step patterns and concentration of feeding solutions), emphasizing the influence of these on the overall performance of the system.

Continuous hybridoma bioreactor: sensitivity analysis and optimal control

Animal cell culture has already established itself as a mature technology able to make a wide range of valuable products, the actual focus being to find the competitive bioreactor design and operating conditions for increasing production. A complex analysis, implying sensitivity calculus and optimal control computation, is done for a system composed of a continuous perfectly mixed bioreactor, with cell recirculation, a cell separator, a mixer and a purge. The bioreactors sensitivity to the control parameters is measured by a new concept, entropic density, developed from the notion of Shannon entropy. An optimization procedure based on a genetic‐algorithms approach is applied for the computation of the inlet flow profile in time, which guarantees optimum monoclonal‐antibody production. Our studies, including the present one, proved that the best approach to obtain high production is to use a hybrid operating sequence: fed‐batch mode followed by the continuous mode.

Optimal control of a continuous perfectly mixed hybridoma bioreactor

Abstract Production of industrial scale quantities of monoclonal antibodies (MAbs) is an expensive task and the current focus is on cutting down the operating costs. One of the pertinent challenges is reducing the formation of large amount of lactate and ammonia (secondary metabolic products of the cellular metabolism of glucose and glutamine), which are waste materials that have also inhibitory effects on cell growth and production rates. It is therefore important to maintain the cells in a physiological state characterized by a minimal production of waste metabolites and a maximum production of biomass and hence antibodies. This goal implies the development of an optimal nutrients supplying strategy which modifies the growth medium in such a way that the cells alter their metabolism to produce as much MAbs as possible, with minimal waste. In the present study, the optimal control of a continous perfectly mixed hybridoma bioreactor with cell recirculation from a separator was sought, using genetic algorithms. The results were compared against the optimal control of a fed batch animal cell bioreactor in order to establish which of the two performs better in similar conditions. An improved result can be obtain by combining the two ways of operating, thus reducing the drawbacks of both methods and increasing the performance of the system.

Divided wall distillation column: Dynamic modeling and control

Abstract The dynamic modeling of the dividing wall distillation column is used to determine optimal startup policies that minimise the time required to reach the imposed steady state operating values in terms of product compositions and flow rates. The problem is resolved by a convenient transformation of the dynamic model in a system of differential equations, avoiding algebraic calculations generally imposed by equilibrium solving, and by using iterative, dynamic programming for the minimization of the startup time. An example referring to the separation of a ternary hydrocarbon mixture is presented. The variables that mostly influence the startup time were found to be the reflux ratio and side-stream flowrate. The optimal policies identified realise a considerable reduction of startup time, up to 70% compared to the startup operation at constant reflux ratio or constant side draw flowrate.

Dynamic Model for the Adsorption in a Multibed Three-Phase Fluidization Column Applied to Wastewater Biological Treatment

It is proposed a dynamic model for adsorption of NH4+ ions from ammonia waters on volcanic tuff in a 10‐bed three‐phase (air – ammonia waters – volcanic tuff) fluidization column. The model consists in the nonstationary material balance differential equations. For each layer the ideal well‐mixing conditions are considered. The effluent ammonia ion concentrations, corresponding to each layer, have been measured at several time values in a laboratory‐scale column. The absolute relative mean error between the calculated and measured values of ammonia ion concentrations into liquid phase for all layers and times is 6.65 %, being in the order of magnitude of experimental errors.

Genetic algorithm optimization of fractional crystallization processes

Abstract This paper advances the optimization of fractional crystallization separation flowcharts applying genetic algorithms (GA), using as example the potassium nitrate separation. All feasible separation sequences are described using the thermodynamic state network model The optimization criterion is the minimization of the sum of flows over the entire network. The results obtained are compared with the solution found by solving the non-linear optimization problem implemented in GAMS.

Optimal control of a hybridoma bioreactor. Changes induced by considering by-products in the objective function

Abstract The main target in the production of monoclonal antibodies (MAbs) is reduction of operating costs. One of the pertinent challenges is improving the yield of MAbs through the reduction of secondary metabolic products. Therefore searching for an optimal nutrients supplying strategy becomes mandatory. This study is about the influence of by-products together with the dead cell concentration upon the performance of a bioreactor for MAbs production. The byproducts and the dead cell concentration are considered in the objective function for optimal control of the system. Three cases were studied: fed batch, continuous and the sequence fed batch - continuous. The optimization procedure was based upon genetic algorithms, which are applied either for the optimal glutamine set point computation for the fed batch operating mode, or the determination of the time inlet flow profile for the continuous mode, both guaranteeing the optimum monoclonal antibody production.

Chapter 1 two level control of the sequence fed batch—continuous hybridoma bioreactor

In the present study a recirculation system for monoclonal antibodies production, operated consecutively fed batch and continuously, was modelled and subjected to optimal control. The optimization procedure uses a two level approach: one regarding the overall process, and two inner ones, concerning the fed batch and continuous steps. The best switch time between the two operating modes was calculated, together with the best control variable profile for each section.