2011 Fiscal Year Final Research Report
Catalyst Design by Support Vector Machine
| Project/Area Number |
22656183
|
|---|
| Research Category |
Grant-in-Aid for Challenging Exploratory Research
|
| Allocation Type | Single-year Grants |
|---|
| Research Field |
Catalyst/Resource chemical process
|
|---|
| Research Institution | Shimane University (2011)
Tohoku University (2010) |
|---|
Principal Investigator |
OMATA Kohji 島根大学, 総合理工学部, 教授 (70185669)
|
|---|
| Project Period (FY) |
2010 – 2011
|
| Keywords | 触媒設計 / ETG反応 / Friedel-Crafts反応 / 元素物性値 / 主成分分析 / k-means法 / サポートベクターマシーン / ガウシアンプロセス回帰 |
|---|
| Research Abstract |
The target of this research is an automated catalyst design by means of physicochemical properties of elements. Such methodology was not established in science of catalysis, and many important catalysts were discovered in accidental processes or by area bombing methodology. Recently non-linear regression methodology often succeeded in finding new catalysts. The regression models were constructed using activityes of a few model catalysts and physicochemical properties in the catalysts. In the present research, some steps in the process of the methodology, such as selections of the physicochemical properties and elements in model catalysts, were automated. Then it was applied to the catalyst design for ethanol-to-gasoline process.
|
