Mohamed Abourida

This file contains the survey instrument developed to support the empirical validation of the BiogAI-RMF framework for wastewater treatment and biogas production. The questionnaire examines risk factors, effective risk management, artificial intelligence adoptability, performance outcomes, and the perceived usability and integration of the proposed framework. It was designed for professionals working in wastewater treatment, anaerobic digestion, biogas production, engineering, operations, maintenance, health and safety, environmental management, regulation, risk management, and digital technologies. Items use a five-point Likert scale. The instrument supports descriptive analysis, reliability and validity assessment, partial least squares structural equation modelling, regression analysis, and fuzzy logic modelling. The survey was administered electronically through Microsoft Forms, participation was voluntary, and responses were collected anonymously. The instrument is provided to support methodological transparency, reproducibility, and future research in AI-enabled risk management for wastewater and biogas systems.

This dataset contains nine anonymised transcripts generated from semi-structured interviews conducted to support the qualitative validation of the BiogAI-RMF framework. Participants represented relevant operational, technical, managerial, safety, environmental, regulatory and digital perspectives associated with wastewater treatment and biogas production. The interviews explored current risk-management practices, AI readiness, data quality, governance, explainability, user trust, implementation barriers and the practical integration of the proposed framework. The transcripts represent the qualitative source data prior to formal coding and thematic analysis. Direct identifiers were removed and participants were assigned anonymous identification codes. Any minor transcription corrections were limited to improving readability without altering the intended meaning of participants’ responses.

This file contains the cleaned, coded and analysis-ready survey dataset used for the empirical validation of the BiogAI-RMF framework. The dataset was derived from the original Microsoft Forms export and includes anonymised responses from professionals working across wastewater treatment, biogas production, engineering, risk management, health and safety, environmental management, regulation and digital technologies. Data preparation included consistency checks, removal of unusable or incomplete responses where applicable, standardisation of variable names, coding of categorical responses and conversion of Likert-scale responses into numerical values. The file supports descriptive analysis, reliability and validity assessment, partial least squares structural equation modelling, regression analysis and fuzzy logic modelling. No directly identifying personal information is included.

This file contains the semi-structured interview protocol developed to support the qualitative validation of the BiogAI-RMF framework. The protocol was designed to explore participants’ professional experiences and perspectives concerning risk management in wastewater treatment and biogas production, organisational readiness for artificial intelligence, data availability and quality, governance, explainability, user trust, implementation barriers, and the practical integration of the proposed framework. The questions were informed by the study’s conceptual model, survey findings and wider literature. The semi-structured format ensured consistency across interviews while allowing relevant issues to be explored in greater depth. Interviews were conducted with professionals possessing relevant operational, technical, managerial, regulatory, safety or digital expertise.

This file contains the original survey-response dataset exported directly from Microsoft Forms for the empirical validation of the BiogAI-RMF framework. The dataset includes anonymised responses collected from professionals with relevant experience in wastewater treatment, biogas production, engineering, risk management, health and safety, environmental management, regulation, and digital technologies. The file is provided in its original exported structure and has not been cleaned, recoded, transformed, or statistically processed. It is included to maintain a transparent record of the collected data and to enable comparison with the accompanying cleaned and coded dataset. Direct personal identifiers have been removed or excluded before publication.

•First combined systematic + bibliometric review on AI-risk in biogas systems.•Maps global trends (2015–2025) in AI-driven risk management at WWTPs.•Identifies research clusters linking AI, safety, and circular-economy goals.•Highlights transferable AI practices from oil & gas to waste-to-energy plants.•Proposes XAI-based framework to enhance transparency and risk governance.
This review presents the first combined systematic and bibliometric review synthesising artificial intelligence (AI)-driven approaches to risk management in biogas production within wastewater treatment plants (WWTPs), with emphasis on decision-optimisation and operational safety. Seven academic databases: Scopus, Web of Science, IEEE Xplore, ScienceDirect, SpringerLink, Taylor & Francis, and Google Scholar were systematically searched from 2015 to March 2025, and screening followed PRISMA 2020 guidelines. Of 3,716 retrieved records, 109 studies met the inclusion criteria. Bibliometric mapping (VOSviewer) and qualitative synthesis identified five thematic clusters: (i) biogas process safety, (ii) IoT integration and renewable-energy, (iii) optimisation and supply-chain resilience, (iv) AI-driven decision-support frameworks, and (v) advanced machine-learning techniques. The analysis reveals a marked increase in publications since 2020, reflecting a shift from conceptual modelling toward applied digital risk solutions. Europe and China remain leading contributors, although collaboration networks are fragmented and methodological heterogeneity persists. Full-scale validation of AI models in operational WWTP-based biogas plants remains limited, with most studies relying on laboratory experiments, simulations, or pilot-scale data. Constraints include publication bias, database coverage, English-language restrictions, inconsistent performance metrics, and limited access to long-term Supervisory Control and Data Acquisition (SCADA) Systems datasets. The review demonstrates that AI-driven methods have significant potential to improve safety, operational efficiency, and regulatory assurance in biogas facilities. However, achieving practical and scalable implementation will require rigorous multi-site validation, standardised evaluation indicators, integration of explainable AI, and alignment with plant-level risk-governance frameworks.

Biogas production within wastewater treatment plants (WWTPs) provides renewable energy recovery but also introduces complex process-safety challenges arising from the interplay of human, organisational, and technical factors. This study systematically evaluates how these dimensions interact to influence process-safety outcomes in UK wastewater-based biogas facilities, addressing a gap where prior safety research has largely prioritised the examination of technical and operational failures in the petrochemical and chemical sectors. A mixed-methods design was adopted, integrating a PRISMA-guided literature review, a structured national survey of industry professionals (n = 90), and analysis of historical incident reports (n = 63). Triangulated quantitative and qualitative data were examined using correlation and regression analyses to identify interdependencies among risk drivers. Findings reveal that gas-leak risk shows the strongest correlation with general site risks (r = 0.857, p < 0.001), followed by human and organisational factors (HOFs) (r = 0.768, p < 0.001) and process factors (r = 0.733, p < 0.001). The regression model (R-2 = 0.754) confirms that site-level governance (e.g., maintenance discipline, leadership visibility, and resource sufficiency) exerts the greatest influence on incident probability. These findings demonstrate that technical safeguards alone are insufficient without robust organisational and behavioural integration within Process Safety Management Systems (PSMS). The study demonstrates that HOFs directly shape safety outcomes in WWTPs, presenting a novel empirical contribution to literature, emphasising predictive maintenance, competence-based training, proactive reporting, and AI-enabled monitoring within PSMS as enablers of safer and more sustainable biogas recovery in wastewater operations.