A Systematic Review of Machine-Learning Solutions in Anaerobic Digestion

Abstract

The use of machine learning (ML) in anaerobic digestion (AD) is growing in popularity and improves the interpretation of complex system parameters for better operation and optimisation. This systematic literature review aims to explore how ML is currently employed in AD, with particular attention to the challenges of implementation and the benefits of integrating ML techniques. While both lab and industry-scale datasets have been used for model training, challenges arise from varied system designs and the different monitoring equipment used. Traditional machine-learning techniques, predominantly artificial neural networks (ANN), are the most commonly used but face difficulties in scalability and interpretability. Specifically, models trained on lab-scale data often struggle to generalize to full-scale, real-world operations due to the complexity and variability in bacterial communities and system operations. In practical scenarios, machine learning can be employed in real-time operations for predictive modelling, ensuring system stability is maintained, resulting in improved efficiency of both biogas production and waste treatment processes. Through reviewing the ML techniques employed in wider applied domains, potential future research opportunities in addressing these challenges have been identified.

Estimation of some AM2 model parameters from a derived empirical logistic function of methane production

“…On the other side, because of the nonlinearity of the anaerobic digestion processes and the sensitivity to their parameters and operating conditions, the simple traditional empirical-driven models may not ensure efficient performance for a generalized prediction of biogas production (Amran et al, 2024;Rutland, 2023). Therefore, some models inspired from Artificial Intelligence and modern computation techniques have recently emerged as alternatives providing better performances for biogas estimation, prediction, and even for real-time control and monitoring of bioreactors (Ling et al, 2024;Swami et al, 2023).…”

“…These facts make them not largely generalizable to other bioreactors operating possibly under different conditions. In addition, most applications are still lab-based and performed in batch flow rather than continuous flow bioreactors (Onu et al, 2023;Rutland, 2023). Finally, ML models are interesting but are still under progress and the comparison of their performances is not yet sufficiently mature (Amran et al, 2024;Ling et al, 2024).…”

Justifying the prediction of major soil nutrients levels (N, P, and K) in cabbage cultivation

“…Notably, the application of the Tan Sig (Hyperbolic Tangent Sigmoid) transfer function during the training procedure contributed to achieving these favorable outcomes [ 2 ]. The utilization of tansig , known for its ability to model complex relationships, enhanced the DNN's capacity to capture intricate patterns in the data, thereby facilitating the attainment of superior performance metrics, which are essential indicators of the model's effectiveness in learning and generalization [ 3 , 4 ].…”

ADToolbox: Incorporating Metagenomics Data for Improved Prediction of Anaerobic Digestion Dynamics

“…Machine learning (ML) models can capture the relationship between inputs and outputs of AD by learning directly from the experimental data provided to the model without the necessity of knowing the underlying kinetic equations governing the biochemical reactions of the system. Although multiple studies have shown the predictive power of ML algorithms in AD or similar environments [30], [31], [32], [33], [34], their generality, given the complexity of the underlying microbiome is questionable, especially, in the absence of plenty of data for training the models which is usually the case in AD studies [35], [36]. To incorporate the high-dimensional metagenomics features, gene or taxa abundances, into ML models that aim to predict AD metabolism, one needs to find features that exist in a lower-dimensional space.…”