Modeling selective CO oxidation over Pt based catalysts using modular neural network
240th ACS National Meeting and Exposition, Boston, MA, Amerika Birleşik Devletleri, 22 - 26 Ağustos 2010, (Tam Metin Bildiri)
- Yayın Türü: Bildiri / Tam Metin Bildiri
- Basıldığı Şehir: Boston, MA
- Basıldığı Ülke: Amerika Birleşik Devletleri
- Boğaziçi Üniversitesi Adresli: Evet
Özet
The selective CO oxidation over Pt/Al2O3 catalysts was modeled using modular neural networks. The catalyst preparation and operational variables were processed in different modules in the same network to model CO conversion so that these groups of variables could be used together but treated differently since both are important but different. It was found that modular neural networks are quite successful in predicting the experimental results (figure 1), and they are superior to the monolithic networks.