Biodiesel as an alternative fuel has shown the potential to mitigate the exceeding energy demands in automotive sector. In this scenario, third-generation biodiesel production from microalgae emerged as a clean and reliable method. In the current work, growth conditions for freshwater algae collected from Punjab University, Lahore, Pakistan were optimized. The collected samples were grown in a glass photobioreactor utilizing BG-11 and Basel Bold Medium (BBM). The optical density (OD) at 688 nm was recorded for 15 days, at the light/dark cycle of 12/24 and 16/24, and at two distinct temperatures of 25^oC and 28 ^oC. The developed quadratic prediction model of OD using Response Surface Methodology (RSM) and diagnostic study showed the reasonable agreement between actual and predicted results. Moreover, RSM predicted results were incorporated for creating 12 different Artificial Neural Network Models (ANNMs) with hidden layer neurons in range of 9-20. The ANNM with 5 neurons in hidden layer and TRAINCGP training function showed best statistical performance with least percentage contribution to mean square error (MSE) chart (1.24%) and highest correlation coefficient (0.99005). Finally, the OD results were numerically optimized using multi objective optimization which designated 13^th day, photoperiod of 16/24, temperature of 25 ^oC and BG11 media for maximum OD of 2.862. The optimization identified result was experimentally validated and absolute percentage error (APE) was 3.92%. Thus, the integrated use of RSM statistical results and ANN modelling could be effectively used for prediction and optimization of Chlorella Voulgaris growth for maximizing biodiesel yield.

Keywords- Artificial Neural Network, Biodiesel, Chlorella Voulgaris, Microalgae, Response Surface Methodology