Modeling and optimization of biomass productivity of Chlorella vulgaris using response surface methodology, analysis of variance and machine learning for carbon dioxide capture.

This study explores bioremediation's effectiveness in reducing carbon emissions through the use of microalgae Chlorella vulgaris, known for capturing carbon dioxide and producing biomass. The impact of temperature and light intensity on productivity and carbon dioxide capture was investigated, and cultivation conditions were optimized in a photobioreactor using response surface methodology (RSM), analysis of variance (ANOVA), and deep neural networks (DNN). The optimal conditions determined were 28.74 °C and 225 µmol/m2/s with RSM, and 29.55 °C and 226.77 µmol/m2/s with DNN, closely aligning with literature values (29 °C and 225 µmol/m2/s). DNN demonstrated superior performance compared to RSM, achieving higher accuracy due to its capacity to process larger datasets using epochs and batches. The research serves as a foundation to further in this field by demonstrating the potential of utilizing diverse mathematical models to optimize bioremediation conditions, and offering valuable insights to improve carbon dioxide capture efficiency in microalgae cultivation.