Computational modelling approach for the optimisation of a pulsed electric field system for liquid foods

The microbial inactivation and specific energy of a pulsed electric field treatment system are dependent on the electric field distribution and treatment time since they have opposite behaviour in relation to these parameters, featuring a problem of multi-objective optimisation. This study proposes a computational methodology capable of providing Pareto optimal solutions for these two objectives, using a coupled electrical-thermal model, solved by COMSOL, which has been integrated to a multi-objective algorithm NSGA-II implemented in MatLab. The simulations were run for a computational design of experiment with the following variables: applied voltage, treatment time and the internal electrode radius (three levels for each one). In the post-processing analysis, the Pareto curves were plotted for two typical microorganisms of grape juice: E. coli and S. aureus, providing a set of solutions in terms of the log of the survival rate versus the specific energy. The methodology enables the decision maker to select the best solution from the Pareto curves as a function of a required microbial inactivation and energy features.