Loss Minimisation in Dynamic Food Processes

Sammanfattning: The overall aim of the present work was to reduce losses during product changes or during rinsing in continuous food processing, with particular attention to dairy plants. The concept of sensor fusion of common process instruments was implemented as a tool to monitor the milk and fat concentrations in a dairy processing line. Special interest was drawn to the dynamic processes such as start-up, rinsing and product change-overs. These dynamic processes are causes for loss of raw material and a better control can efficiently decrease this wastage. The three process instruments used were an optical instrument, a conductivity meter and a density meter. The system was indirectly calibrated; partly from literature data, partly from sets of water-milk mixtures of varying fat content and temperature. A numerical evaluation algorithm evaluating the overdetermined equation system and allowing the derivation of a measure of uncertainty was written in Matlab. The method used on dynamic measurement data was shown to predict the milk in water concentration within 2% and the fat concentration within 0.1%. The relations between conductivity, temperature, fat content and degree of dilution of milk were studied with respect to the validity of Kohlrausch's law and the Debye-Hückel-Onsager theory. It was shown that the Debye-Hückel-Onsager theory could correctly describe the relationship between concentration and temperature for concentrations up to 25% milk in water. To be able to use the conductivity meter as one of the sensors predicting the milk and fat concentrations, an empirical relation was derived that adequately described the conductivity as a function of temperature and degree of dilution in the temperature range from 2 ? 70°C and the milk fraction range from 0 ? 1 milk in water. In order to decrease the loss of yoghurt, its rheological properties during pipe flow were studied by cross correlation of dual plane Electrical Resistance Tomography (ERT). The use of two ERT planes allowed the determination of the local velocity in the whole cross section of the pipe. The results showed that at low velocities, i.e. up to 0.25 m s-1, yoghurt could flow as a plug. In addition to the experimental evaluation of yoghurt rinsing with water by ERT, Computational Fluid Dynamics (CFD) was used as a tool to predict the three dimensional rinsing behaviour of yoghurt, taking into account its rheological properties and higher density. Various flow velocities and rheological properties of the yoghurt were simulated and an overall agreement of experimental and calculated results was achieved. The simulations showed that high flow velocity, high yield stress of the yoghurt and wall slip decreased the mixing zone between water and yoghurt and therefore also the loss of yoghurt.