Related Experiment Videos
Interfacial mass transfer in randomly packed towers: a confident correlation for environmental applications
S Piché1, B P Grandjean, I Iliuta
1Department of Chemical Engineering & CERPIC, Laval University, Québec, Canada.
Environmental Science & Technology
|January 5, 2002
Summary
This study compiles mass-transfer data for pollution abatement, finding limitations in existing correlations. An artificial neural network (ANN) model is proposed for improved prediction of mass-transfer coefficients in packed towers.
Area of Science:
- Chemical Engineering
- Environmental Engineering
- Mass Transfer Operations
Background:
- Accurate prediction of volumetric mass-transfer coefficients is crucial for designing packed towers used in water and air pollution abatement.
- Existing correlations, such as Onda (1968) and Billet and Schultes (1993), have limitations in accuracy and applicability.
- A comprehensive database of 2675 measurements was compiled to evaluate these correlations.
Purpose of the Study:
- To critically assess existing correlations for mass-transfer coefficients in packed towers.
- To develop a more accurate and reliable method for predicting mass-transfer coefficients.
- To improve the design of packed towers for pollution control applications.
Main Methods:
- Literature data compilation for volumetric mass-transfer coefficients (kLa(w), KLa(w), kGa(w), kGa(w)).
- Cross-examination and performance evaluation of Onda and Billet and Schultes correlations.
- Development and application of an Artificial Neural Network (ANN) model to predict dimensionless Sherwood numbers.
- Implementation of a reconciliation procedure using ANN predictions and two-film theory.
Main Results:
- Identified limitations in accuracy and application range of existing mass-transfer correlations.
- Developed a single ANN model predicting the Sherwood number (ShL/G) based on six dimensionless groups.
- Achieved an absolute average relative error of 22.1% and a standard deviation of 21.1% for the ANN-based predictions.
- Demonstrated that ANN predictions align with physical evidence reported in the literature.
Conclusions:
- Artificial Neural Network modeling offers a superior approach for predicting mass-transfer coefficients in packed towers compared to traditional correlations.
- The developed ANN model and reconciliation procedure provide enhanced accuracy for designing pollution abatement systems.
- This work contributes a valuable tool for optimizing packed tower performance in environmental engineering applications.