Plastic Waste Carbon and AI Model Improve Domestic Wastewater Treatment Efficiency

Ilustrasi By AI

FORMOSA NEWS - Researchers from Soegijapranata Catholic University and Sriwijaya University have developed a machine learning–based optimization model that enhances domestic wastewater treatment using activated carbon derived from plastic waste. The study, led by Ikbal Oktaviansyah and published in 2026 in the Formosa Journal of Science and Technology, demonstrates that plastic-based activated carbon can achieve pollutant removal efficiencies of up to 99 percent, with predictive accuracy models reaching an R² value of 0.99. The findings offer a dual solution to two pressing environmental challenges: plastic waste accumulation and declining water quality.

Addressing Two Environmental Crises at Once

Global plastic production now exceeds 400 million tons annually. Much of this waste ends up in landfills, rivers, and oceans. At the same time, domestic wastewater continues to introduce organic pollutants, dyes, oils, pharmaceuticals, and heavy metals into water bodies, reducing dissolved oxygen levels and threatening aquatic ecosystems.

Activated carbon has long been used in wastewater treatment because of its strong adsorption capacity. However, conventional activated carbon is often made from coal or coconut shells, which require energy-intensive processing and can be costly.

The research team explored a circular alternative: converting plastic waste—particularly PET and PVC into activated carbon through pyrolysis and chemical activation. By integrating machine learning into the optimization process, they improved both material performance and operational efficiency.

How the Study Was Conducted

The researchers applied a Systematic Literature Review (SLR) covering peer-reviewed publications from 2021 to 2025 indexed in Scopus, Web of Science, ScienceDirect, and Google Scholar. From an initial pool of 309 articles, they selected 10 high-quality studies for detailed evaluation.

The team examined:

  • Physical and chemical characteristics of plastic-based activated carbon
  • Operational parameters influencing adsorption performance
  • The effectiveness of various machine learning algorithms in predicting adsorption capacity

Instead of conducting new laboratory experiments, the researchers synthesized existing empirical data and applied computational modeling to identify optimal treatment conditions. This approach reduced trial-and-error experimentation and provided data-driven recommendations for wastewater system design.

Key Findings

The study reports strong technical performance for plastic-derived activated carbon:

  • Specific surface area (BET): 800–1,500 m²/g
  • Maximum adsorption capacity: up to 1,086 mg/g
  • Pollutant removal efficiency: 97–99 percent for dyes and oil contaminants
  • Reusability: 70–90 percent efficiency retained after five regeneration cycles
  • Estimated production cost: approximately US$13.75 per kilogram
  • Carbon footprint: 5.92 kg CO₂ per kilogram, lower than many commercial alternatives

Two variables were found to contribute most significantly to system performance:

  • Specific surface area of the adsorbent
  • Initial pollutant concentration

Together, these factors accounted for more than 43 percent of performance variation in adsorption systems.

Machine Learning Achieves High Predictive Accuracy

The integration of machine learning significantly improved predictive modeling and system optimization. The study evaluated several algorithms:

  • Extreme Gradient Boosting (XGBoost): R² between 0.95 and 0.97
  • Gaussian Process Regression (GPR): R² up to 0.99; RMSE 4.29 mg/g
  • Support Vector Regression (SVR): correlation up to 0.9989
  • Artificial Neural Network (ANN): R² between 0.92 and 0.96

Gaussian Process Regression combined with metaheuristic optimization methods such as simulated annealing increased adsorption capacity by 15.40 percent compared to baseline experimental conditions.

Ikbal Oktaviansyah of Soegijapranata Catholic University emphasized that integrating predictive modeling allows engineers to design wastewater systems more precisely. He explained that machine learning enables rapid identification of optimal operational parameters while reducing energy consumption and operational costs.

Real-World Implications

The research has direct implications for environmental management, industry, and public policy.

  1. Wastewater treatment operators can reduce operational costs through data-driven optimization.
  2. Local governments can adopt circular economy strategies by converting plastic waste into high-value environmental materials.
  3. Plastic waste management sectors gain economic incentives for recycling non-biodegradable materials.
  4. Educational institutions and research centers can integrate AI-based modeling into environmental engineering curricula.

With proper sensor integration, treatment plants could operate adaptively, adjusting process conditions in real time based on incoming wastewater quality.

The study positions plastic-derived activated carbon not merely as an alternative material, but as a scalable and sustainable solution aligned with global Sustainable Development Goals, particularly clean water and responsible production.

Advancing Smart Environmental Engineering

By combining material science and artificial intelligence, the research introduces a framework for intelligent wastewater treatment systems. The model supports:

  • Efficient pollutant removal
  • Reduced carbon emissions
  • Lower production and operational costs
  • Improved long-term sustainability

The collaboration between Soegijapranata Catholic University and Sriwijaya University demonstrates how interdisciplinary research can bridge environmental engineering and computational modeling to address complex ecological challenges.

Author Profiles

Ikbal Oktaviansyah, S.T., M.T.
Lecturer and researcher at Soegijapranata Catholic University. His expertise includes environmental engineering, wastewater treatment systems, and computational modeling.

Ridwan Sanjaya, S.T., M.T.
Academic at Soegijapranata Catholic University specializing in process engineering and system optimization.

Budi Setiawan, S.T., M.T.
Researcher in materials engineering and environmental technology at Soegijapranata Catholic University.

Piestie Hawa, S.T., M.T., Ph.D.
Lecturer at Sriwijaya University with expertise in water resource management and adsorption technology.

Source

Oktaviansyah, I., Sanjaya, R., Setiawan, B., & Hawa, P. (2026). Plastic Waste-Based Activated Carbon Adsorption Optimization Model Using Machine Learning to Improve the Quality of Domestic Liquid Waste. Formosa Journal of Science and Technology, Vol. 5(2), 479–502.

Posting Komentar

0 Komentar