Phenomenological-Based Model and Control Proposal for Sunflower Oil Leaching.

In the past, some alcohols were considered as possible solvents to extract oils from collets. But eventually, oil yields and some economic aspects favored the selection of hexane, an inflammable petroleum-derived solvent. Nowadays, it is known that this solvent can be emitted during extraction, being necessary its recovery because due to its characteristic as air pollutant since it can react with other pollutants to form ozone and photochemical oxidants. Ethanol is a good option to be studied as an alternative to mentioned petroleum-derived solvents. Ethanol is cheap and can be produced by fermentation from a large variety of biological materials using simple technology, and therefore labeled as “natural” or “biorenewable”. Due to the lower selectivity of ethanol towards triglycerides, during the extraction process other compounds such as phosphatides, polyphenols, pigments and soluble sugars are extracted jointly. The aim of this work is to improve the performance of sunflower oil leaching using ethanol by including the model information into the controller algorithm. A phenomenological-based semiphysical model is deduced for the leaching process and used for a model-based controller. A comparison between the proposed controller and the common PID controller is presented. The results indicate that the model-based controller represents an alternative to improve the process operation, increasing the response speed of the system, rejecting disturbances and thus allowing the equipment to operate at its desired extraction point, avoiding losses of the desired product (oil retained in the collet).