Neural networks and differential evolution algorithm applied for modelling the depollution process of some gaseous...

Area of Science:

• Environmental Engineering
• Chemical Engineering
• Computational Chemistry

Background:

• Gaseous streams often contain volatile organic compounds like n-hexane.
• Adsorption is a key technique for removing n-hexane from industrial emissions.
• Developing accurate models for dynamic adsorption processes is crucial for efficient depollution.

Purpose of the Study:

• To investigate the depollution of n-hexane from gaseous streams using adsorption.
• To propose and evaluate a novel neuro-evolutionary approach for modeling the dynamic adsorption process.

Main Methods:

• Adsorption experiments were conducted in a fixed bed column using granular activated carbon and two hypercross-linked polymeric resins.
• A modified differential evolution (DE) algorithm combined with neural networks (NNs) and local search optimizers was developed.
• Key features of the DE variant include opposition-based learning, self-adaptive parameter control, and fitness-based mutation.

Main Results:

• The proposed neuro-evolutionary algorithm successfully modeled the n-hexane depollution process.
• The developed model demonstrated superior performance compared to an existing phenomenological model.
• The approach effectively determined optimal neural network parameters for process simulation.

Conclusions:

• The neuro-evolutionary approach offers a powerful and accurate method for modeling dynamic adsorption processes.
• This technique provides a significant advancement over traditional modeling methods for environmental remediation.
• The study highlights the potential of hybrid computational intelligence methods in chemical and environmental engineering.