Model-based optimization of the operation procedure of emulsification

Abstract

Emulsions are widely encountered in the food and cosmetic industry. The first food we consume is an emulsion, namely breast milk. Other common emulsions are mayonnaise, dressings, skin creams and lotions. Equipment often used for the production of oil-in-water emulsions in the food industry consists of a stirred vessel in combination with a colloid mill and a circulation pipe. Within this set-up there are two main variations: i) Configuration I where the colloid mill acts like a shearing device and at the same time as a pump. This configuration is used in the majority of the production facilities, and ii) Configuration II where the shearing and pumping action are not coupled. The operation procedure for obtaining a certain predefined emulsion quality is often established based on experience (best practice). This is most probably time-consuming (e.g. large experimental efforts for new developed products) and it is also unclear if the process is operated at its optimum (e.g. in minimum time). An other drawback is that there is no feedback during the production process. Hence, it is not possible to deal with disturbances acting on the process. A possible consequence is that, at the end of the production process, the product quality specifications are not met and the product has to be classified as off-spec. In order to be able to enlarge the efficiency of the production processes and to shorten the time to market of new products - and therewith create an advantage over competition - it is necessary to overcome these limitations of the current operation procedure. In the work reported a first step is set into this direction. A model describing the droplet size distribution (DSD) and the emulsion viscosity as function of the time was developed and several off-line optimization studies were performed. The model comprises several fit parameters and experiments were performed in order to estimate the values of these parameters. A number of additional experiments were performed to compare the simulated results with the measurements (model validation). The results of the parameter estimation and the model validation show that the simulated results are qualitatively in good agreement with the measurement data. Given the overall performance of the model it is expected that the model quality is sufficient to render practical relevant optimization results. Although the optimization studies have been performed for a model emulsion, small scale equipment and are not yet experimentally validated, the results of this work strongly suggest that it is indeed possible to minimize the production times and to shorten the product development times for new products. This overall conclusion is based on the following observations: 1) The optimization results show that it is beneficial to produce emulsions with Configuration II: - Configuration II allows the production of emulsions with a bi-modal DSD. No operation procedure was found for the production of such an emulsion in Configuration I. - The production of emulsions in Configuration II is always at least as fast as in Configuration I. 2) The followed approach allows to calculate: * If an emulsion with a certain, predefined, DSD and emulsion viscosity can be produced. * How the process should be controlled in order to produce such an emulsion. * How the process should be controlled to produce this emulsion in minimal time. 3) The optimization results show that it is possible to produce emulsions with: * A bi-modal DSD. * Less oil while maintaining a similar DSD and value of the emulsion viscosity evaluated at a shear rate of 10 1/s by adapting only the operation procedure. Hence, the addition of extra stabilizers is not considered. This offers possibilities for the production of a broader range of emulsion products and could direct product development in a new direction. Based on this, it is worthwhile and therefore recommended to expand this research work in the direction of industrial emulsions.