Advancements in Waste‐to‐Energy Conversion Using Pyrolysis and Artificial Intelligence

ABSTRACT Traditional pyrolysis systems commonly operate under fixed process parameters, limiting their adaptability to variations in feedstock characteristics and operating conditions. This study investigates the integration of Artificial Intelligence (AI), specifically an Artificial Neural Network (ANN), into waste‐to‐energy pyrolysis to improve energy utilization and bio‐oil yield. Experimental data were obtained from a laboratory‐scale fixed‐bed pyrolysis reactor using different feedstock types, temperatures, residence times, and specific energy consumption levels. A feedforward ANN with two hidden layers was developed to predict bio‐oil yield, using a 70:15:15 training, validation, and testing data partition. Model performance was evaluated using the coefficient of determination ( R 2 ), mean absolute error (MAE), and root mean square error (RMSE), and benchmarked against conventional linear regression. The ANN demonstrated superior predictive performance, achieving an R 2 of approximately 0.665 on the testing data set, with MAE and RMSE of 0.049% and 0.061%, respectively, compared with an R 2 of approximately 0.531 for linear regression. The findings indicate that AI‐assisted pyrolysis can provide a more flexible and data‐driven approach for predicting bio‐oil yield, reducing specific energy consumption, and supporting optimization of waste‐to‐energy conversion processes.