A Systematic Review of Machine-Learning Solutions in Anaerobic Digestion.

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.

Microbial fuel cells and their electrified biofilms.

Bioelectrochemical systems (BES) represent a wide range of different biofilm-based bioreactors that includes microbial fuel cells (MFCs), microbial electrolysis cells (MECs) and microbial desalination cells (MDCs). The first described bioelectrical bioreactor is the Microbial Fuel Cell and with the exception of MDCs, it is the only type of BES that actually produces harvestable amounts of electricity, rather than requiring an electrical input to function. For these reasons, this review article, with previously unpublished supporting data, focusses primarily on MFCs. Of relevance is the architecture of these bioreactors, the type of membrane they employ (if any) for separating the chambers along with the size, as well as the geometry and material composition of the electrodes which support biofilms. Finally, the structure, properties and growth rate of the microbial biofilms colonising anodic electrodes, are of critical importance for rendering these devices, functional living 'engines' for a wide range of applications.

Microbial Fuel Cell Based Thermosensor for Robotic Applications.

On the roadmap to building completely autonomous artificial bio-robots, all major aspects of robotic functions, namely, energy generation, processing, sensing, and actuation, need to be self-sustainable and function in the biological realm. Microbial Fuel Cells (MFCs) provide a platform technology for achieving this goal. In a series of experiments, we demonstrate that MFCs can be used as living, autonomous sensors in robotics. In this work, we focus on thermal sensing that is akin to thermoreceptors in mammalian entities. We therefore designed and tested an MFC-based thermosensor system for utilization within artificial bio-robots such as EcoBots. In open-loop sensor characterization, with a controlled load resistance and feed rate, the MFC thermoreceptor was able to detect stimuli of 1 min directed from a distance of 10 cm causing a temperature rise of ∼1°C at the thermoreceptor. The thermoreceptor responded to continuous stimuli with a minimum interval of 384 s. In a practical demonstration, a mobile robot was fitted with two artificial thermosensors, as environmental thermal detectors for thermotactic application, mimicking thermotaxis in biology. In closed-loop applications, continuous thermal stimuli were detected at a minimum time interval of 160 s, without the need for complete thermoreceptor recovery. This enabled the robot to detect thermal stimuli and steer away from a warmer thermal source within the rise of 1°C. We envision that the thermosensor can be used for future applications in robotics, including as a potential sensor mechanism for maintaining thermal homeostasis.

Electrosynthesis, modulation, and self-driven electroseparation in microbial fuel cells.

Microbial electrosynthesis (MES) represents a sustainable platform that converts waste into resources, using microorganisms within an electrochemical cell. Traditionally, MES refers to the oxidation/reduction of a reactant at the electrode surface with externally applied potential bias. However, microbial fuel cells (MFCs) generate electrons that can drive electrochemical reactions at otherwise unbiased electrodes. Electrosynthesis in MFCs is driven by microbial oxidation of organic matter releasing electrons that force the migration of cationic species to the cathode. Here, we explore how electrosynthesis can coexist within electricity-producing MFCs thanks to electro-separation of cations, electroosmotic drag, and oxygen reduction within appropriately designed systems. More importantly, we report on a novel method of in situ modulation for electrosynthesis, through additional "pin" electrodes. Several MFC electrosynthesis modulating methods that adjust the electrode potential of each half-cell through the pin electrodes are presented. The innovative concept of electrosynthesis within the electricity producing MFCs provides a multidisciplinary platform converting waste-to-resources in a self-sustainable way.

Neural Networks Predicting Microbial Fuel Cells Output for Soft Robotics Applications.

The development of biodegradable soft robotics requires an appropriate eco-friendly source of energy. The use of Microbial Fuel Cells (MFCs) is suggested as they can be designed completely from soft materials with little or no negative effects to the environment. Nonetheless, their responsiveness and functionality is not strictly defined as in other conventional technologies, i.e. lithium batteries. Consequently, the use of artificial intelligence methods in their control techniques is highly recommended. The use of neural networks, namely a nonlinear autoregressive network with exogenous inputs was employed to predict the electrical output of an MFC, given its previous outputs and feeding volumes. Thus, predicting MFC outputs as a time series, enables accurate determination of feeding intervals and quantities required for sustenance that can be incorporated in the behavioural repertoire of a soft robot.

From the lab to the field: Self-stratifying microbial fuel cells stacks directly powering lights.

The microbial fuel cell (MFC) technology relies on energy storage and harvesting circuitry to deliver stable power outputs. This increases costs, and for wider deployment into society, these should be kept minimal. The present study reports how a MFC system was developed to continuously power public toilet lighting, with for the first time no energy storage nor harvesting circuitry. Two different stacks, one consisting of 15 and the other 18 membrane-less MFC modules, were operated for 6 days and fuelled by the urine of festival goers at the 2019 Glastonbury Music Festival. The 15-module stack was directly connected to 2 spotlights each comprising 6 LEDs. The 18-module stack was connected to 2 identical LED spotlights but going through 2 LED electronic controller/drivers. Twenty hours after inoculation the stacks were able to directly power the bespoke lighting system. The electrical energy produced by the 15-module stack evolved with usage from ≈280 mW (≈2.650 V at ≈105 mA) at the beginning to ≈860 mW (≈2.750 V at ≈300 mA) by the end of the festival. The electrical energy produced by the LED-driven 18-module stack increased from ≈490 mW at the beginning to ≈680 mW toward the end of the festival. During this period, illumination was above the legal standards for outdoor public areas, with the 15-module stack reaching a maximum of ≈89 Lx at 220 cm. These results demonstrate for the first time that the MFC technology can be deployed as a direct energy source in decentralised area (e.g. refugee camps).

A new method for urine electrofiltration and long term power enhancement using surface modified anodes with activated carbon in ceramic microbial fuel cells.

This work is presenting for the first time the use of inexpensive and efficient anode material for boosting power production, as well as improving electrofiltration of human urine in tubular microbial fuel cells (MFCs). The MFCs were constructed using unglazed ceramic clay functioning as the membrane and chassis. The study is looking into effective anodic surface modification by applying activated carbon micro-nanostructure onto carbon fibres that allows electrode packing without excessive enlargement of the electrode. The surface treatment of the carbon veil matrix resulted in 3.7 mW (52.9 W m-3 and 1626 mW m-2) of power generated and almost a 10-fold increase in the anodic current due to the doping as well as long-term stability in one year of continuous operation. The higher power output resulted in the synthesis of clear catholyte, thereby i) avoiding cathode fouling and contributing to the active splitting of both pH and ions and ii) transforming urine into a purified catholyte - 30% salt reduction - by electroosmotic drag, whilst generating - rather than consuming - electricity, and in a way demonstrating electrofiltration. For the purpose of future technology implementation , the importance of simultaneous increase in power generation, long-term stability over 1 year and efficient urine cleaning by using low-cost materials, is very promising and helps the technology enter the wider market.

Complete Microbial Fuel Cell Fabrication Using Additive Layer Manufacturing.

Improving the efficiency of microbial fuel cell (MFC) technology by enhancing the system performance and reducing the production cost is essential for commercialisation. In this study, building an additive manufacturing (AM)-built MFC comprising all 3D printed components such as anode, cathode and chassis was attempted for the first time. 3D printed base structures were made of low-cost, biodegradable polylactic acid (PLA) filaments. For both anode and cathode, two surface modification methods using either graphite or nickel powder were tested. The best performing anode material, carbon-coated non-conductive PLA filament, was comparable to the control modified carbon veil with a peak power of 376.7 µW (7.5 W m-3) in week 3. However, PLA-based AM cathodes underperformed regardless of the coating method, which limited the overall performance. The membrane-less design produced more stable and higher power output levels (520-570 µW, 7.4-8.1 W m-3) compared to the ceramic membrane control MFCs. As the final design, four AM-made membrane-less MFCs connected in series successfully powered a digital weather station, which shows the current status of low-cost 3D printed MFC development.

Urine in Bioelectrochemical Systems: An Overall Review.

In recent years, human urine has been successfully used as an electrolyte and organic substrate in bioelectrochemical systems (BESs) mainly due of its unique properties. Urine contains organic compounds that can be utilised as a fuel for energy recovery in microbial fuel cells (MFCs) and it has high nutrient concentrations including nitrogen and phosphorous that can be concentrated and recovered in microbial electrosynthesis cells and microbial concentration cells. Moreover, human urine has high solution conductivity, which reduces the ohmic losses of these systems, improving BES output. This review describes the most recent advances in BESs utilising urine. Properties of neat human urine used in state-of-the-art MFCs are described from basic to pilot-scale and real implementation. Utilisation of urine in other bioelectrochemical systems for nutrient recovery is also discussed including proofs of concept to scale up systems.

Multidimensional Benefits of Improved Sanitation: Evaluating 'PEE POWER®' in Kisoro, Uganda.

With 2.3 billion people around the world lacking adequate sanitation services, attention has turned to alternative service provision models. This study suggests an approach for meeting the sanitation challenge, especially as expressed in Sustainable Development Goal 6.2, using a toilet technology system, such as Pee Power® that generates electricity using diverted urine as a fuel. A field trial was carried out in a girls' school in Kisoro, Uganda, where the generated electricity was used to light the existing toilet block. The trial was evaluated in terms of social acceptability and user experience using a multidimensional assessment protocol. The results of our assessment show that users felt safer when visiting the toilets at night. Lights provided from the technology also helped with the perceived cleanliness of the toilets. The technology was well accepted, with 97% of the respondents saying that they liked the idea of the Pee Power® technology and 94% preferring it over other facilities on site. This shows how the technology helps meet SDG target 6.2, with its particular focus on vulnerable populations.