Comparative artificial neural network models for predicting kinetic parameters of biomass pyrolysis from the biomass characteristics
Artificial Neural Network (ANN) was proposed to predict kinetic parameters of biomass pyrolysis using simple inputs like proximate or ultimate analysis results. The advantage of predicting via ANN is that it does not require thermogravimetric analysis, unlike the traditional methods. This approach r...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Results in Engineering |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2590123025027689 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Artificial Neural Network (ANN) was proposed to predict kinetic parameters of biomass pyrolysis using simple inputs like proximate or ultimate analysis results. The advantage of predicting via ANN is that it does not require thermogravimetric analysis, unlike the traditional methods. This approach reduces experimental cost and time while maintaining high predictive accuracy. Kinetic parameters in output data for training the ANN models were re-analyzed from thermogravimetric analysis (TGA) results of sixty-three biomasses with different heating rates in the literature by a standardization method based on Coats and Redfern method. Models with three different input sets: proximate results and heating rate (Prx+Hr), ultimate results and heating rate (CHNO+Hr), and combined ultimate and proximate results and heating rate (CHNO+Prx+Hr) were compared by their prediction results. Results showed that CHNO+Hr inputs tended to provide more accurate results (achieving R² up to 0.992 in validation, and 0.961 in verification) than the other input sets, highlighting the superiority of ultimate analysis over proximate analysis for ANN-based prediction. However, Prx+Hr inputs and CHNO+Prx+Hr inputs could also provide highly accurate predictions with a slight deviation. Overall, these ANN models demonstrated a reliable and cost-effective alternative to conventional TGA methods for predicting the kinetic parameters of biomass pyrolysis. |
|---|---|
| ISSN: | 2590-1230 |