Petroleum sludge is one of the most persistent byproducts of crude oil refining, posing a significant environmental problem due to its complex composition of hydrocarbons, polycyclic aromatic hydrocarbons (PAHs), and heavy metals. This paper examined the ecological toxicology of petroleum sludge at the Warri Refining and Petrochemical Company (WRPC), Delta State, Nigeria, through empirical, computational, and biological analyses, coupled with the Systems Theory of Environmental Toxicology. The primary objective was to describe the sludge composition, assess the human and ecological risks, and develop artificial intelligence (AI)-driven predictive models to enhance environmental management. Unlike previous refinery toxicology studies that focus solely on chemical characterization or risk estimation, this study uniquely integrates field data, quantitative risk assessment, and multi-model AI prediction to address the lack of predictive environmental intelligence in refinery-impacted ecosystems. Sludge pits, storage tanks, and effluent ponds yielded a total of 30 samples of sludge, which were collected and analysed by GC-MS and ICP-MS to determine PAHs and heavy metals, respectively. Quantitative risk assessment was conducted in accordance with the guidelines of the USEPA, focusing on the aspects of the hazard index, carcinogenic risk, and ecological risk (PERI). In contrast, AI models such as the Random Forest (RF), Support Vector Machine (SVM), and Artificial Neural Network (ANN) were employed in the study of predictive risk mapping. The analysis showed very high contaminant concentrations (TPH: 215,400mg kg-1; Cr: 78mg kg-1; Pb: 42mg kg-1; Benzo[a]pyrene: 42.8mg kg-1), as well as a considerable level of health hazards (HI: 3.2 -4.5; CR: 1.7×10-3 -2.7×10-3). The ANN model proved to be more accurate in its predictive capacity (R² = 0.96), with TPH, Cr, Pb, and benzo[a]pyrene emerging as the primary risk drivers. The paper finds that the Warri Refinery ecosystem is a highly hazardous area that requires timely remediation. It suggests monitoring with AI, as well as sludge stabilisation and bioaugmentation with indigenous hydrocarbon-degrading microorganisms, to alleviate toxicity and support Sustainable Development Goals 3, 9, 11, and 13.

Keywords: Petroleum Sludge Toxicology, Artificial Intelligence Modelling, Quantitative Risk Assessment (QRA), Heavy Metal–Hydrocarbon Synergy, Environmental Systems Theory, Predictive Ecotoxicology, Human Health Risk Mapping, Microbial Bioremediation Potential, GIS-Based Contamination Analysis, Sustainable Refinery Management.

Acha, S., Dave, U.J., John, A., Isangadighi, G.E., Islam, M., & Ifeanyi, M.E. (2025). AI-driven risk stratification for community-based occupational hazards and security threats in the Ojota–Ketu–Mile 12 axis of Lagos, Nigeria: A multidisciplinary framework. European Journal of Applied Science, Engineering & Technology, 3(4): 231–241.

Dragoi, E.N., Godini, K., & Koolivand, A. (2021). Modeling of oily sludge composting process using artificial neural networks and differential evolution: Prediction of removal of petroleum hydrocarbons and organic carbon. Environmental Technology & Innovation, 21: 101338.

Hasan, A.M., Kamal, R.S., Farag, R.K., & Abdel-Raouf, M.E. (2024). Petroleum sludge formation and its treatment methodologies: A review. Environmental Science and Pollution Research, 31(6): 8369–8386.

Hu, G., Li, J., Zhang, X., & Li, Y. (2017). Investigation of waste biomass co-pyrolysis with petroleum sludge using response surface methodology. Journal of Environmental Management, 192: 234–242.

Huang, Q., Han, X., Mao, F., Chi, Y., & Yan, J. (2014). A model for predicting solid particle behavior in petroleum sludge during centrifugation. Fuel, 117: 95–102.

Isangadighi, G.E., & Udeh, J.A. (2023). Emergencies, preparedness, and management: A case study of Nigeria. World Safety Organization, 32(2): 38.

Isangadighi, G.E., Mathew, P., Islam, M., Obahor, G., Offordum, A., & Momoh, P.O. (2025). Machine learning-based multivariate risk stratification framework for assessing the combined burden of occupational, infectious, and non-communicable diseases in Bagega (Zamfara) and Shiroro (Niger), Nigeria. Journal of Clinical Practice and Medical Research, 1(1): 24–29.

Isangadighi, G.E., Obahor, G., Ekanem, M.E., Judith, U.O., Paul, C.J., Adedamola, S.B., & Ozohili, L.I. (2024). Machine learning-based prediction of bioaccumulation and ecotoxicity of emerging contaminants in aquatic ecosystems. Journal of Systematic, Evaluation and Diversity Engineering.

Isangadighi, G.E., Orhuebor, E.N., Essien, U.B., Matthew, P., Islam, M., Ahanor, E., & Obahor, G. (2024). Machine learning-enhanced multidisciplinary assessment of petroleum hydrocarbon pollution: Biochemical, microbial, toxicological, and environmental perspectives. Journal of Science Innovation and Technology Research.

Johnson, O.A., & Affam, A.C. (2019). Petroleum sludge treatment and disposal: A review. Environmental Engineering Research, 24(2): 191–201.

Lee, H., Chang, S., Kim, S.K., & Kwon, J.H. (2017). Fugacity analysis of polycyclic aromatic hydrocarbons between microplastics and seawater. Ocean Science Journal, 52(1): 43–55.

Ogwu, M.C., Izah, S.C., Sawyer, W.E., & Amabie, T. (2025). Environmental risk assessment of trace metal pollution: A statistical perspective. Environmental Geochemistry and Health, 47(4): 94.

Orhuebor, E.N., Isangadighi, G.E., Akpololohor, C.A., & Akinboboye, R. (2025). Data-driven techno-economic and environmental assessment of compressed natural gas (CNG) vehicles as a source of transportation in Nigeria. European Journal of Applied Science, Engineering and Technology, 3(5): 142–152.

Roy, A., Dutta, A., Pal, S., Gupta, A., Sarkar, J., Chatterjee, A., & Kazy, S.K. (2018). Biostimulation and bioaugmentation of native microbial community accelerated bioremediation of oil refinery sludge. Bioresource Technology, 253: 22–32.

Wang, Y., Dong, B., Fan, Y., Hu, Y., Zhai, X., Deng, C., & Dai, X. (2019). Nitrogen transformation during pyrolysis of oilfield sludge with high polymer content. Chemosphere, 219: 383–389.

Source of Funding:

The authors declare that no specific grant from any funding agency in the public, commercial, or not-for-profit sector was received for this research. The authors fully funded all analyses, laboratory work, and computational modelling as part of their institutional research requirements.

Competing Interests Statement:

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Consent for publication:

All authors consent to the publication of this manuscript and approve its submission to the target journal.

Authors' contributions:

E.N. Orhuebor – Study design, sampling, laboratory analysis, manuscript drafting. U.B. Essien – Toxicological interpretation, risk assessment modelling, manuscript revision. U.N. Matthew – Chemical analysis, ICP–MS interpretation, field coordination. A.G. Essiet – Epidemiological analysis, environmental health interpretation. I. Isah - GIS mapping and spatial modelling. L.I. Ozohili – Microbial analysis and metagenomic profiling. E.A. Debekeme – GIS mapping and spatial modelling. N.L. Iwalehin – AI modelling, machine learning algorithm development. I.M. Udofia – Statistical analysis and data validation. All authors reviewed and approved the final manuscript.

Ethical Approval and Consent to Participate:

This study did not involve human or animal subjects. All environmental sampling and laboratory procedures complied with institutional and national ethical guidelines. Ethical considerations, including plagiarism avoidance, data integrity, and responsible conduct of research, were strictly observed.

Availability of data and materials:

The datasets generated and analysed during the current study are available from the corresponding author upon reasonable request. Additional AI codes, GIS layers, and raw laboratory data can be provided subject to institutional permission.

Institutional Review Board Statement:

Not applicable for this study.

Informed Consent:

Not applicable for this study.

Declaration of Originality:

The authors affirm that this manuscript is an original work that has not been published previously and is not under consideration for publication elsewhere. All sources used have been cited correctly in accordance with academic standards.

Research Transparency Statement:

All methods, analytical procedures, statistical tools, AI models, and data processing workflows have been fully disclosed in this manuscript to promote transparency and reproducibility.

Acknowledgements:

The authors acknowledge the cooperation of the laboratory technologists, environmental sampling officers, JessieGie Team and academic supervisors whose support contributed to the successful completion of this study.