Hybrid Prediction-Driven High-Throughput Sustainability Screening for Advancing Waste-to-Dimethyl Ether Valorization

Area of Science:

• Chemical Engineering
• Environmental Science
• Sustainability Science

Background:

• Assessing novel chemical production sustainability is challenging due to numerous process options and lack of high-throughput screening.
• Developing sustainable methods for waste valorization is crucial for climate preservation.

Purpose of the Study:

• To present a data-driven hybrid prediction framework (ANN-RSM-DOM) for streamlining waste-to-dimethyl ether (DME) upcycling.
• To identify sustainable and cost-effective biorefinery configurations using the organic fraction of municipal solid waste (OFMSW) and sewage sludge (SS).

Main Methods:

• Development of artificial neural networks (ANNs) for in silico waste valorization and ex-ante modeling.
• Utilizing Aspen Plus process flowsheeting, response surface methodology (RSM), and desirability optimization method (DOM) for analysis.
• Sequential application of data-driven hybrid prediction for sustainability pre-screening.

Main Results:

• Identified the importance of targeted waste selection based on elemental composition for DME synthesis.
• Revealed plant configurations achieving climate benefits (-1.241 to -2.128 kg CO2-eq/kg DME) and low production costs (€0.382 to €0.492/kg DME).
• Demonstrated the framework's ability to screen low technological readiness level processes.

Conclusions:

• The hybrid prediction facilitates early-stage process synthesis and design of complex process units.
• Enables simultaneous analysis of qualitative and quantitative variables for sustainable chemical production.
• Supports high-throughput sustainability screening of immature chemical production technologies.