41.ソフトコンピューティングと高速スクリーニングによる触媒開発(14) : ニューラルネットワークの網羅計算によるメタノール合成触媒の調製条件と組成の最適化

概要

論文の詳細を見る
Combinatorial approach comprised of high-pressure HTS (high-throughput screening) reactor, artificial neural network (NN) and all-encompassing calculation was applied for the development of catalyst for methanol synthesis. Since the optimum catalyst composition is usually affected by both preparation and reaction conditions, composition of catalyst (Cu-Zn-Al-Sc-B-Zr), calcination temperature (300-360℃) and the amount of precipitant (1.0-2.5times the equivalent to total cations) were optimized simultaneously under pressure (1MPa). In the HTS reactor based on a 96 microplate system, activities of 190 catalysts prepared by randomly determined parameters, were measured to obtain training data for NN. After training, the NN can map the catalyst activity as a function of catalyst composition and parameters for catalyst preparation. The 2.6 million activities of all possible combination of the parameters were predicted by the NN to find the global optimum.
一般社団法人日本エネルギー学会の論文
2003-10-23