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.

Facile Preparation of Durable and Eco-Friendly Superhydrophobic Filter with Self-Healing Ability for Efficient Oil/Water Separation.

The superhydrophobic feature is highly desirable for oil/water separation (OWS) operation to achieve excellent separation efficiency. However, using hazardous materials in fabricating superhydrophobic surfaces is always the main concern. Herein, superhydrophobic filters were prepared via an eco-friendly approach by anchoring silica particles (SiO2) onto the cotton fabric surface, followed by surface coating using natural material-myristic acid via a dip coating method. Tetraethyl orthosilicate (TEOS) was used in the synthesis of SiO2 particles from the silica sol. In addition, the impact of the drying temperature on the wettability of the superhydrophobic filter was investigated. Moreover, the pristine cotton fabric and as-prepared superhydrophobic cotton filters were characterised based on Fourier-transform infrared (FTIR) spectroscopy, X-ray photoelectron spectroscopy (XPS), scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM-EDX) and contact angle (CA) measurement. The superhydrophobic cotton filter was used to perform OWS using an oil-water mixture containing either chloroform, hexane, toluene, xylene or dichloroethane. The separation efficiency of the OWS using the superhydrophobic filter was as high as 99.9%. Moreover, the superhydrophobic fabric filter also demonstrated excellent durability, chemical stability, self-healing ability and reusability.

Development of ß-cyclodextrin crosslinked citric acid encapsulated in polypropylene membrane protected-µ-solid-phase extraction device for enhancing the separation and preconcentration of endocrine disruptor compounds.

Endocrine disruptor compounds (EDCs) such as plasticisers, surfactants, pharmaceutical products, personal care products and pesticides are frequently released into the environmental waters. Therefore, a sensitive and environmentally friendly method is entailed to quantify these compounds at their trace level concentrations. This study encapsulated the ß-cyclodextrin crosslinked with citric acid in a polypropylene membrane protected-µ-solid phase extraction (BCD-CA µ-SPE) device for preconcentrating the EDCs (triclosan, triclocarban, 2-phenylphenol, 4-tert-octylphenols and bisphenol A) in real water samples before the analysis by high-performance liquid chromatography. FT-IR and TGA results indicated that BCD-CA was successfully synthesised with the formation of ester linkage (1078.33 cm-1) and O-H stretching from carboxylic acid (3434.70 cm-1) with higher thermal stability as compared with native CD with the remaining weight above 72.1% at 500 °C. Several critical parameters such as the sorbent loading, type and amount of salts, extraction time, sample volume, sample pH, type and volume of desorption solvents and desorption time were sequentially optimised and statistically validated. Under the optimum condition, the use of BCD-CA µ-SPE device had manifested good linearity (0.5-500 µg L-1) with the determination of the coefficient range of 0.9807-0.9979. The p-values for the F-test and t-test (6.60 × 10-8 - 1.77 × 10-5) were lesser than 0.05 and low detection limits ranging from 0.27 to 0.84 µg L-1 for all studied EDCs. The developed technique was also successfully applied for EDC analyses in four distinct real water samples, namely, wastewater, river water, tap water and mineral water, with good EDCs recoveries (80.2%-99.9%), low relative standard deviations (0.1%-3.8%, n = 3) with enrichment factor ranging from 9 to 82 folds. These results signified the potential of the BCD-CA µ-SPE device as an efficient, sensitive, and environmentally friendly approach for analyzing EDCs.

Evaluation of potential feedstock for biogas production via anaerobic digestion in Malaysia: kinetic studies and economics analysis.

As the population increases, energy demands continue to rise rapidly. In order to satisfy this increasing energy demand, biogas offers a potential alternative. Biogas is economically viable to be produced through anaerobic digestion (AD) from various biomass feedstocks that are readily available in Malaysia, such as food waste (FW), palm oil mill effluent (POME), garden waste (GW), landfill, sewage sludge (SS) and animal manure. This paper aims to determine the potential feedstocks for biogas production via AD based on their characteristics, methane yield, kinetic studies and economic analysis. POME and FW show the highest methane yield with biogas yields up to 0.50 L/g VS while the lowest is 0.12 L/g VS by landfill leachate. Kinetic study shows that modified Gompertz model fits most of the feedstock with R 2 up to 1 indicating that this model can be used for estimating treatment efficiencies of full-scale reactors and performing scale-up analysis. The economic analysis shows that POME has the shortest payback period (PBP), highest internal rate of return (IRR) and net present value (NPV). However, it has already been well explored, with 93% of biogas plants in Malaysia using POME as feedstock. The FW generation rate in Malaysia is approximately 15,000 tonnes per day, at the same time FW as the second place shows potential to have a PBP of 5.4 years and 13.3% IRR, which is close to the results achieved with POME. This makes FW suitable to be used as the feedstock for biogas production.

Enzymatic pretreatment to enhance anaerobic bioconversion of high strength wastewater to biogas: A review.

Oil and grease, carbohydrate, protein, and lignin are the main constituents of high strength wastewaters such as dairy wastewater, cheese whey wastewater, distillery wastewater, pulp and paper mill wastewater, and slaughterhouse wastewaters. These constituents have contributed to various operational problems faced by the high-rate anaerobic bioreactor (HRAB). During the hydrolysis stage of anaerobic digestion (AD), these constituents can be hydrolyzed. Since hydrolysis is known to be the rate-limiting step of AD, the overall AD can be enhanced by improving the hydrolysis stage. This can be done by introducing pretreatment that targets the degradation of these constituents. This review mainly focuses on the biological pretreatment on various high-strength wastewaters by using different types of enzymes namely lipase, amylase, protease, and ligninolytic enzymes which are responsible for catalyzing the degradation of oil and grease, carbohydrate, protein, and lignin respectively. This review provides a summary of enzymatic systems involved in enhancing the hydrolysis stage and consequently improve biogas production. The results show that the use of enzymes improves the biogas production in the range of 7 to 76%. Though these improvements are highly dependent on the operating conditions of pretreatment and the types of substrates. Therefore, the critical parameters that would affect the effectiveness of pretreatment are also discussed. This review paper will serve as a useful piece of information to those industries that face difficulties in treating their high-strength wastewaters for the appropriate process, equipment selection, and design of an anaerobic enzymatic system. However, more intensive studies on the optimum operating conditions of pretreatment in a larger-scale and synergistic effects between enzymes are necessary to make the enzymatic pretreatment economically feasible.

Performance and kinetic evaluation of an integrated anaerobic-aerobic bioreactor in the treatment of palm oil mill effluent.

This work presents the evaluation of biokinetic coefficients for a novel integrated anaerobic-aerobic bioreactor (IAAB) at different organic loading rates (OLRs) (10.5-22.5 g COD/L per day) treating palm oil mill effluent. The overall efficiencies of the chemical oxygen demand (COD) and biochemical oxygen demand (BOD) were more than 99% for OLR up to 19.5 g COD/L day with biogas production containing 48-64% of methane. The effluent quality remained stable (BOD < 80 mg/L) and satisfied with the discharge limit. Stover-Kincannon model was the most appropriate model to estimate the performance of anaerobic compartment of IAAB, while Monod model was best suited for describing the aerobic compartment. ABBREVIATIONS: BOD: biochemical oxygen demand; COD: chemical oxygen demand; DO: dissolved oxygen; GLSS: gas-liquid-solid separator; MLSS: mixed liquor suspended solid; MLVSS: mixed liquor volatile suspended solid; POME: palm oil mill effluent; RAS: returned activated sludge; SBR: sequencing batch reactor; SVI: sludge volume index; TA: total alkalinity; TN: total nitrogen; TS: total solids; TSS: total suspended solids; UASB: upflow anaerobic sludge blanket; UASFF: upflow anaerobic sludge blanket fixed film; VFA: volatile fatty acid; VSS: volatile suspended solids.

Start-up, steady state performance and kinetic evaluation of a thermophilic integrated anaerobic-aerobic bioreactor (IAAB).

Thermophilic treatment of palm oil mill effluent (POME) was studied in a novel integrated anaerobic-aerobic bioreactor (IAAB). The IAAB was subjected to a program of steady-state operation over a range of organic loading rate (OLR)s, up to 30 g COD/L day in order to evaluate its treatment capacity. The thermophilic IAAB achieved high chemical oxygen demand (COD), biochemical oxygen demand (BOD) and total suspended solids (TSS) removal efficiencies of more than 99% for OLR up to 18.5 g COD/L day. High methane yield of 0.32 LCH(4) (STP)/g COD(removed) with compliance of the final treated effluent to the discharge limit were achieved. This is higher than that of the mesophilic system due to the higher maximum specific growth rate (µ(max)) of the thermophilic microorganisms. Besides, coupling the model of Grau second order model (anaerobic system) with the model of Monod (aerobic system) will completely define the IAAB system.

Biological treatment of anaerobically digested palm oil mill effluent (POME) using a Lab-Scale Sequencing Batch Reactor (SBR).

The production of highly polluting palm oil mill effluent (POME) has resulted in serious environmental hazards. While anaerobic digestion is widely accepted as an effective method for the treatment of POME, anaerobic treatment of POME alone has difficulty meeting discharge limits due to the high organic strength of POME. Hence, subsequent post-treatment following aerobic treatment is vital to meet the discharge limits. The objective of the present study is to investigate the aerobic treatment of anaerobically digested POME by using a sequencing batch reactor (SBR). The SBR performance was assessed by measuring Chemical Oxygen Demand (COD), Biochemical Oxygen Demand (BOD) and Total Suspended Solids (TSS) removal as well as Sludge Volume Index (SVI). The operating pH and dissolved oxygen concentrations were found to be 8.25-9.14 and 1.5-6.4 mg/L, respectively, throughout the experiment. The experimental results demonstrate that MLVSS, OLR and sludge loading rate (SLR) play a significant role in the organic removal efficiency of SBR systems and therefore, further investigation on these parameters was conducted to attain optimum SBR performance. Maximum COD (95-96%), BOD (97-98%) and TSS (98-99%) removal efficiencies were achieved at optimum OLR, SLR and MLVSS concentration ranges of 1.8-4.2 kg COD/m(3)day, 2.5-4.6 kg TSS/m(3)day and 22,000-25,000 mg/L, respectively. The effluent quality remained stable and complied with the discharge limit. At the same time, the sludge showed good settling properties with average SVI of 65. It is envisaged that the SBR process could complement the anaerobic treatment to produce final treated effluent which meets the discharge limit.