Machine Learning Applications in Biogas and Methane Production: A Bibliometric Analysis

Biogas processes play an important role in the disposal of organic waste. However, these processes are difficult to control because they are highly sensitive and variable. A lot of work has been done to date in order to eliminate this problem. With the development of technology and artificial intelligence, the spread of “Autonomous” systems has become widespread in the control of anaerobic processes as in many other fields. The Anaerobic Digestion Model No. 1 (ADM1) developed by the International Water Association (IWA) has been adopted as the standard model for the AD process since 2002. With the development of this model, Simple Regression Tree (SRT), Probabilistic Neural Networks (PNN), Artificial Neural Networks (ANN), Gradient Boosted Tree (GBT), Linear Regression (LR), Tree Ensemble Regression (TER), Random Forest Regression (RFR), Polynomial Regression (PR), Fuzzy Logic (FL), Adaptive Network-Based Fuzzy İnference System (ANFIS), Different ML algorithms such as Support Vector Machine (SVM), Particle Swarm Optimization (PSO), Genetic Algorithm (GA) Developing Data-Driven Models (DDDV), Deep neural network (DNN) have been used in various studies and tried to perform process optimization, real-time monitoring, disturbance detection and parameter estimation. In this study, the data obtained by using the Bibliometrix package and Biblioshiny package through the R programming language in the R-Studio programme were evaluated. For this purpose, a total of 80 articles in the field of ‘Machine Learning’ in the Web of Science (WoS) database between 2012-2024 in the fields of ‘Biogas Production’, ‘Methane Production’ and ‘Anaerobic Digestion’ processes were accessed and evaluated. As a result of the evaluations, the development of ML models in biogas processes was determined and recommendations were presented.