Liquid temperature prediction in bubbly flow using ant colony optimization algorithm in the fuzzy inference system as...

Area of Science:

• Chemical Engineering
• Computational Science
• Artificial Intelligence

Background:

• Chemical reactors require accurate simulation for process optimization.
• Traditional simulation methods like computational fluid dynamics (CFD) can be computationally intensive.
• Hybrid models combining first-principle and AI offer a promising alternative.

Purpose of the Study:

• To develop and evaluate a novel hybrid model for chemical reactor simulation.
• To integrate ant colony optimization (ACO) with fuzzy logic for enhanced prediction capabilities.
• To compare the proposed model's performance against other AI-based methods.

Main Methods:

• A 2D bubble column reactor (BCR) was simulated using CFD.
• A two-stage approach was employed: ACO for data learning and fuzzy logic for prediction.
• The developed Ant Colony Optimization Fuzzy Inference System (ACOFIS) model was validated against CFD results.

Main Results:

• The ACOFIS model accurately predicted temperature distribution in the BCR.
• ACO efficiently learned complex input-output relationships from CFD data.
• The model demonstrated comparable accuracy and prediction capability to GAFIS and PSOFIS with significantly reduced computation time.

Conclusions:

• The hybrid ACO-fuzzy logic model provides an efficient and accurate alternative to complex CFD simulations.
• Swarm intelligence, specifically ACO, can effectively predict chemical processes.
• Optimizing tuning parameters like CIR is crucial for achieving cost-effective and accurate model performance.