ARTIFICIAL NEURAL NETWORK SOFT SENSOR FOR PREDICTING A DRY MASS OF SPIRULINA MAXIMA IN PHOTOBIOREACTOR

The real-time knowledge of dry mass in bioprocess cultures is a very important measure for decision-making on the process. Common analysis during the growth of microorganisms, like microalgae, bacteria or yeast, takes a long time until obtaining final dry mass values. Such difficulty inspired this study, which through an artificial neural network with eight inputs and ten neurons hidden layer, was able to predict the dry biomass behavior with precision of R² = 0.9935 in a photobioreactor.