This study presents a hybrid Aspen Plus–Neural Network (NN) framework to model, optimize, and analyze the hydrothermal liquefaction (HTL) process for converting microalgae into biocrude oil. A validated Aspen Plus simulation established baseline performance metrics: 45 % biocrude oil yield (30.43 MJ/kg HHV), 67.99 % energy efficiency (with by-products), 0.477 kg CO2e/kg biocrude oil, and a net minimum biocrude oil selling price (net MBSP) of $11.48/GGE. Three neural network models were developed to predict energy, emissions, and economic outcomes, achieving average R2 scores of 0.93, 0.92, and 0.92, respectively. Bayesian optimization using NN surrogates identified conditions to maximize energy efficiency to 82.42 %, emphasizing biomass concentration (30 %), biocrude oil yield (60 %), and by-product energy recovery. Key trade-offs between efficiency, emissions, and costs were analyzed, highlighting feedstock cost and biocrude oil yield as dominant economic drivers. The hybrid approach demonstrates superior computational efficiency over traditional methods, enabling rapid multi-objective optimization. This work advances HTL process design by integrating rigorous simulation with machine learning, offering actionable insights for sustainable biofuel production.
