Machine Learning Methods Modeling Carbohydrate-Enriched Cyanobacteria Biomass Production in Wastewater Treatment Systems

One-stage production of carbohydrate-enriched microalgae biomass in wastewater is a promising option to obtain biofuels. Understanding the interaction water quality parameters such as nutrients, carbon, internal carbohydrates, and microbial composition culture crucial for efficient operation viable large-scale cultivation. Bioprocess models are an essential tool studying simultaneous effect complex factors on carbohydrate accumulation, optimizing process, reducing operational costs. In this sense, we use dataset obtained from empirical model that analyzed accumulation carbohydrates single process (simultaneous growth accumulation) real wastewater. experiment, there were no ideal conditions (limiting nutrient conditions), but rather these limitations guaranteed by operating (hydraulic retention times/nutrient or carbon loads). Thus, integrates 18 variables affected not only carbohydrates. The directly influences Therefore, paper analyzes artificial intelligence (AI) algorithms develop forecast treatment systems. Carbohydrates modeled using five methods: (1) Artificial Neural Networks (ANNs), (2) Convolutional (CNN), (3) Long Short-Term Memory Network (LSTMs), (4) K-Nearest Neighbors (kNN), (5) Random Forest (RF)). AI methods allow learning how several components interact if their combinations work faster than building physical experiments over same period time. After comparing models, CNN-1D best results with MSE (Mean Squared Error) = 0.0028. This result shows adequately approximates system’s dynamics.