A Binary Hyperbox Classifier Model for Hydrogen Storage in Metal Hydrides
Despite being recognized as a key component towards reducing global GHG emissions, many challenges remain throughout the hydrogen value chain. Hydrogen storage is one critical aspect, since proper storage is essential for its wide scale deployment. One storage option is the chemisorption of hydrogen...
Guardado en:
| Autores principales: | , , , , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
AIDIC Servizi S.r.l.
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/68b9ef69da14401b9194913382f18ad3 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Despite being recognized as a key component towards reducing global GHG emissions, many challenges remain throughout the hydrogen value chain. Hydrogen storage is one critical aspect, since proper storage is essential for its wide scale deployment. One storage option is the chemisorption of hydrogen in metals and intermetallic alloys (i.e., metal hydrides). This approach can potentially address issues on safety, infrastructure, and cost which currently hinder the transition toward a hydrogen economy. The only suitable method for determining appropriate hydrogen storage materials is through experimentation which is resource intensive. There are many parameters that can affect the hydrogen storage capacity of a material; databases on previously investigated materials can be used to narrow down future investigation to the most promising candidates. Machine learning (ML) techniques can be employed to determine how different properties predict for hydrogen storage capacity. ML can use previously compiled data to generate a classification model that can be utilized for determining and predicting a material's viability for hydrogen storage. This paper uses the hyperbox ML technique to generate interpretable decision rules to predict if a metal hydride is a good candidate for hydrogen storage. A case study which specifically focuses on complex and magnesium hydrides is used to demonstrate this approach. The generated decision model had a false positive rate of 22.0 % and false negative rate of 36.8 %. |
|---|