Harnessing machine learning to decode and optimize bioelectrochemical systems: Principles, progress and future...

Area of Science:

• Bioelectrochemical Systems (BES)
• Machine Learning (ML)
• Environmental Engineering
• Renewable Energy

Background:

• Bioelectrochemical systems (BES) integrate multiple engineering disciplines for energy generation and wastewater treatment.
• The complexity of BES hinders understanding and optimization, necessitating advanced analytical approaches.
• Machine learning (ML) offers powerful tools for pattern recognition and modeling of complex systems like BES.

Purpose of the Study:

• To systematically review the application of ML in BES.
• To identify key application domains, prevalent ML algorithms, and BES subtypes utilizing ML.
• To highlight challenges and future directions for ML deployment in BES research.

Main Methods:

• Systematic literature review of ML applications in BES from 2006 to present.
• Categorization of ML applications into four main domains.
• Analysis of ML algorithm popularity and BES subtype prevalence.

Main Results:

• ML applications in BES surged around 2021, indicating growing research interest.
• Dominant applications include performance prediction and optimization, followed by microbial community analysis, intelligent design, and control.
• Microbial fuel cells (42.4%) and electrochemical biosensors (37.6%) are the most studied BES subtypes using ML.
• Artificial neural networks (30.9%) are the most common ML algorithms, followed by support vector machines (17.3%).

Conclusions:

• ML shows significant potential to advance BES design, operation, and application.
• Current ML models face limitations in transferability due to data scarcity and heterogeneity.
• Future research should focus on expanding data accumulation, diversifying datasets, and developing targeted ML models for broader BES deployment.