Integration of Artificial Intelligence into Biogas Plant Operation

A method for predicting methane production from anaerobic digestion of kitchen waste under small sample conditions

Anaerobic digestion (AD) has the great potential to treat organic waste and achieve remarkable results effectively. However, it is very tough to establish an accurate mechanistic model for this process. Data-driven modeling technology has opened a new door to solving this problem. While when the sample set is small, traditional data-driven modeling methods are often powerless. In this paper, an effective method is proposed for data-driven high-precision modeling in small sample scenarios. A time series generative adversarial network (TimeGAN) is first utilized to augment the original high-quality small-sample data collected during the AD methane production. A novel hybrid kernel extreme learning machine (HKELM) is then designed to form a better structure of the data-driven model, whose regularization coefficient C0 is optimized by the sparrow search algorithm (SSA). Finally, this semi-finished model (SSA-HKELM) is trained by the augmented data to form the final mathematical model (TimeGAN-SSA-HKELM) for the AD methane generation process. Comparative experiments of the methane daily production prediction error have verified the effectiveness of the method, which can be extended to other similar small sample data-driven modeling scenarios.

Machine learning methods for the modelling and optimisation of biogas production from anaerobic digestion: a review

Biogas plant operators often face huge challenges in the monitoring, controlling and optimisation of the anaerobic digestion (AD) process, as it is very sensitive to surrounding changes, which often leads to process failure and adversely affects biogas production. Conventional implemented methods and mechanistic models are impractical and find it difficult to model the nonlinear and intricate interactions of the AD process. Thus, the development of machine learning (ML) algorithms has attracted considerable interest in the areas of process optimization, real-time monitoring, perturbation detection and parameter prediction. This paper provides a comprehensive and up-to-date overview of different machine learning algorithms, including artificial neural network (ANN), fuzzy logic (FL), adaptive network-based fuzzy inference system (ANFIS), support vector machine (SVM), genetic algorithm (GA) and particle swarm optimization (PSO) in terms of working mechanism, structure, advantages and disadvantages, as well as their prediction performances in modelling the biogas production. A few recent case studies of their applications and limitations are also critically reviewed and compared, providing useful information and recommendation in the selection and application of different ML algorithms. This review shows that the prediction efficiency of different ML algorithms is greatly impacted by variations in the reactor configurations, operating conditions, influent characteristics, selection of input parameters and network architectures. It is recommended to incorporate mixed liquor volatile suspended solids (MLVSS) concentration of the anaerobic digester (ranging from 16,500 to 46,700 mg/L) as one of the input parameters to improve the prediction efficiency of ML modelling. This review also shows that the combination of different ML algorithms (i.e. hybrid GA-ANN model) could yield better accuracy with higher R2 (0.9986) than conventional algorithms and could improve the optimization model of AD. Besides, future works could be focused on the incorporation of an integrated digital twin system coupled with ML techniques into the existing Supervisory Control and Data Acquisition (SCADA) system of any biogas plant to detect any operational abnormalities and prevent digester upsets.

Estimation of Acetic Acid Concentration from Biogas Samples Using Machine Learning

In a biogas plant, the acetic acid concentration is a major component of the substrate as it determines the pH value, and this pH value correlates with the volume of biogas produced. Since it requires specialized laboratory equipment, the concentration of acetic acid in a biogas substrate cannot be measured on-line. The project aims to use NIR sensors and machine learning algorithms to estimate the acetic acid concentration in a biogas substrate based on the measured intensities of the substrate. As a result of this project, it was possible to determine whether the acetic acid concentration in a biogas substrate is higher or lower than 2 g/l using machine learning models.

Advancing the bioconversion process of food waste into methane: A systematic review

With the continuous rise of food waste (FW) throughout the world, a research effort to reveal its potential for bioenergy production is surging. There is a lack of harmonized information and publications available that evaluate the state-of-advance for FW-derived methane production process, particularly from an engineering and sustainability point of view. Anaerobic digestion (AD) has shown remarkable efficiency in the bioconversion of FW to methane. This paper reviews the current research progress, gaps, and prospects in pre-AD, AD, and post-AD processes of FW-derived methane production. Briefly, the review highlights innovative FW collection and optimization routes such as AI that enable efficient FW valorization processes. As weather changes and the FW sources may affect the AD efficiency, it is important to assess the spatio-seasonal variations and microphysical properties of the FW to be valorized. In that case, developing weather-resistant bioreactors and cost-effective mechanisms to modify the raw substrate morphology is necessary. An AI-guided reactor could have high performance when the internal environment of the centralized operation is monitored in real-time and not susceptible to changes in FW variety. Monitoring solvent degradation and fugitive gases during biogas purification is a challenging task, especially for large-scale plants. Furthermore, this review links scientific evidence in the field with full-scale case studies from different countries. It also highlights the potential contribution of ADFW to carbon neutrality efforts. Regarding future research needs, in addition to the smart collection scheme, attention should be paid to the management and utilization of FW impurities, to ensure sustainable AD operations.

Review on solid-state anaerobic digestion of lignocellulosic biomass and organic solid waste

Sustainable management of organic solid wastes especially the municipal solid waste (MSW) is essential for the realization of various sustainable development goals (SDGs). Resource recovery centric waste processing technologies generate valorizable products to meet the operations and maintenance (O&M) costs while reducing the GHG emissions. Solid-state anaerobic digestion (SSAD) of organic solid wastes is a biomethanation process performed at a relatively higher total solids (TS) loading in the range of 10-45%. SSAD overcomes various limitations posed by conventional anaerobic slurry digesters such as higher degradable matter per unit volume of the bioreactor resulting in a smaller footprint, low freshwater consumption, low wastewater generation, simple upstream and downstream processes, relatively lower operation, and maintenance costs. This review elucidates the recent developments and critical assessment of different aspects of SSAD, such as bioreactor design, operational strategy, process performances, mass balance, microbial ecology, applications, and mathematical models. A critical assessment revealed that the operating scale of SSAD varies between 1000 and 100,000 ts/year at organic loading rate (OLR) of 2-15 g volatile solids (VS)/L·day. The SSAD experiences process failures due to the formation of volatile fatty acids (VFAs), biogas pockets and clogging of the digestate outlet. Acclimatization of microbes accelerates the startup phase, steady-state performances, and the enrichment of syntrophic microbes with 10-50 times greater population of cellulolytic and xylanolytic microbes in thermophilic SSAD over mesophilic SSAD. Experimental limitations in the accurate determination of rate constants and the oversimplification of biochemical reactions result in an inaccurate prediction by the models.

Machine Learning Algorithms for Temperature Management in the Anaerobic Digestion Process

Process optimization is no longer an option for processes, but an obligation to survive in the market in any industry. This argument also applies to anaerobic digestion in biogas plants. The contribution of biogas plants to renewable energy can be increased through more productive systems with less waste, which brings the common goal of minimizing costs and maximizing yields in processes. With the help of data science and predictive analytics, it is possible to take conventional process optimization and operational excellence methods, such as statistical process control and Six Sigma, to the next level. The more advanced the process optimization aspect, the more transparent and responsive the systems. In this study, seven different machine learning algorithms—linear regression, logistic regression, K-NN, decision trees, random forest, support vector machine (SVM) and XGBoost—were compared with laboratory results to define and predict the possible impacts of wide range temperature fluctuations on process stability. SVM provided the best accuracy with 0.93 according to the metric precision of the models calculated using the confusion matrix.

The Effect of Electromagnetic Microwave Radiation on Methane Fermentation of Selected Energy Crop Species

The aim of the present study was to determine how thermal stimulation via electromagnetic microwave radiation impacts the yields of biogas and methane produced by methane fermentation of five selected energy crop species in anaerobic reactors. The resultant performance was compared with that of reactors with conventional temperature control. The highest biogas production capacity was achieved for maize silage and Virginia mallow silage (i.e., 680 ± 28 dm3N/kgVS and 506 ± 16 dm3N/kgVS, respectively). Microwave radiation as a method of heating anaerobic reactors provided a statistically-significantly boost in methane production from maize silage (18% increase). Biomethane production from maize silage rose from 361 ± 12 dm3N/kgVS to 426 ± 14 dm3N/kgVS. In the other experimental variants, the differences between methane concentrations in the biogas were non-significant.

Optimization of the Biomethane Production Process by Anaerobic Digestion of Wheat Straw Using Chemical Pretreatments Coupled with Ultrasonic Disintegration

Biomass is an attractive energy source that can be used for production of heat, power, and transport fuels and when produced and used on a sustainable basis, can make a large contribution to reducing greenhouse gas emissions. Anaerobic digestion (AD) is a suitable technology for reducing organic matter and generating bioenergy in the form of biogas. This study investigated the factors allowing the optimization of the process of biogas production from the digestion of wheat straw (WS). The statistical analysis of the experiments carried out showed that ultrasonic processing plays a fundamental role with the sonication density and solids concentration leading to improved characteristics of WS, reducing particle size, and increasing concentration of soluble chemical oxygen demand. The higher the sonicating power used, the more the waste particles are disrupted. The optimality obtained under mesophilic conditions for WS pretreated with 4% w/w (weight by weight) H2O2 at temperature 36 °C under 10 min of ultrasonication at 24 kHz with a power of 200 W improves the methane yield by 64%.