Comparing Growth Models Dependent on Irradiation and Nutrient Consumption on Closed Outdoor Cultivations of...

Microalgae offer tremendous industrial possibilities for their ability to grow rapidly and capture CO₂ from the atmosphere. The literature contains many models for predicting microalgae growth in lab-scale reactors. However, there exists a gap in the application of these models in outdoor pilot-scale closed photobioreactors. This work proposes a methodology for constructing models for this type of reactor. These models were constructed based on the existing literature, then trained and tested using a dataset of ten cultivations of Nannochloropsis sp. Four models were tested: a model based on a Monod-like equation (Model M); a model based on a Haldane-like equation (Model H); a model based on an exponential equation (Model E); and a model considering both irradiation and the effect of nitrate on the culture using the Droop model (Model D). Model H had the best overall performance, with a global root mean squared error (RMSE) of 0.296 kg^1/2 m^-3/2; Model M and Model E had RMSE values of 0.309 and 0.302, respectively. Model D performed the worst, with an RMSE of 0.413. Future work should involve applying the same methodology to new cultivations of the same or different species and testing more complex models capable of better explaining the data.