Stefanie Tömmers

• The ripening of 6.0 million tonnes of imported dessert bananas each year in the European Union is an extensive but not well analysed process in the food industry. Data and information about the biochemistry and the technical realisation of the banana ripening process is missing. The control of the highly complex processes depends on the personal knowledge and experience of the ripening master. For a better understanding of the industrial process, an experimental banana ripening plant was developed, equipped with sensitive measuring technology and connected to a process control system. The produced data of the 4-day to 7-day ripening processes, carried out based on industrial standards and objectives, provides additional and new information about the food technological process. Based on the generated data and knowledge, a mathematical model of the banana ripening process was developed. Characteristic biochemical reactions, like the degradation of starch to soluble sugars and their further oxidation, the change of colour pigments and the ethene production, were described with the help of kinetics specified in related food technological and biotechnological processes. The plant effects and process characteristics were realised with the help of mass and energy balances. The simulation results confirm with the measured values with accuracy below 10%. With the help of the model significant banana ripening process variables, like the starch concentration in the pulp, the peel colour, the mass of bananas, the concentration of carbon dioxide, oxygen, water and ethene in the process air as well as the banana and process air temperature, are simulated and can be used for the process evaluation. Furthermore, the model can be used for the prediction of future process courses and thus help to ensure targeted ripening quality. The model was used to enhance the process control of industrial ripening procedures. For this purpose, the Open-Loop-Feedback-Optimal controller was applied for an adaptive and optimised banana ripening process control. Only two significant parameters, the starch oxidation speed factor and the colour change factor, were fitted to accurate describe the running process. A simplified model was used to calculate an optimised control function with the target of an optimal storage temperature the desired banana ripening degree. The adaption to varying process characteristics and changing ripening times was successfully applied in one ripening experiment. The model-based adaptive control can be used for high quality but practical banana ripening process control.