Version 1
: Received: 20 September 2024 / Approved: 23 September 2024 / Online: 23 September 2024 (14:14:42 CEST)
How to cite:
Vaiyapuri, T. Interpretable Regression Models for Predicting and Optimizing Hydrogen Production in Co-Gasification Process. Preprints2024, 2024091756. https://doi.org/10.20944/preprints202409.1756.v1
Vaiyapuri, T. Interpretable Regression Models for Predicting and Optimizing Hydrogen Production in Co-Gasification Process. Preprints 2024, 2024091756. https://doi.org/10.20944/preprints202409.1756.v1
Vaiyapuri, T. Interpretable Regression Models for Predicting and Optimizing Hydrogen Production in Co-Gasification Process. Preprints2024, 2024091756. https://doi.org/10.20944/preprints202409.1756.v1
APA Style
Vaiyapuri, T. (2024). Interpretable Regression Models for Predicting and Optimizing Hydrogen Production in Co-Gasification Process. Preprints. https://doi.org/10.20944/preprints202409.1756.v1
Chicago/Turabian Style
Vaiyapuri, T. 2024 "Interpretable Regression Models for Predicting and Optimizing Hydrogen Production in Co-Gasification Process" Preprints. https://doi.org/10.20944/preprints202409.1756.v1
Abstract
Renewable energy is essential to environmental sustainability and ranks high among the United Nations strategic goals. In this context, as opposed to conventional biomass gasification, co-gasification emerges as a promising technological pathway to effectively combine the distinct benefits of various gasification feedstocks in order to generate hydrogen-rich syngas. The goal of achieving net-zero emissions has further raised the demand for biomass co-gasification. However, the complicated reactions involved in co-gasification process pose challenge in optimizing the process parameters for increased productivity and performance. To the author’s knowledge, there is no consensus regarding the most suitable machine learning (ML) based regression models for optimizing biomass-plastics co-gasification process, as no prior research has examined the effectiveness of different regression models for this process. Further, the practical application of ML models is adversely affected by their black box nature and lack of interpretability. To address these gaps, the objective of the currect research is two-fold: Firstly, to model the co-gasification process using seven different ML algorithms. Secondly, to develop an evaluation framework and assess model interpretability to select an ideal model for co-gasification interpretation and optimization. The support vector regression(SVR) model demonstrated the most reliable predictive performance. The study employed learning curve analysis and cross validation analysis to substantiate the model predictive performance and exhibit its potential to generalize. Beyond performance analysis, Shapely additive explanation, (SHAP) is first explored in the realm of cogasification to interpret the prediction of the studied models from global and local perspective. The research findings are expected to provide valuable insights to better understand and optimize biomass-plastics cogasification process for H2-rich syngas production.
Computer Science and Mathematics, Artificial Intelligence and Machine Learning
Copyright:
This is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
