| Abstract
| - Artificial neural networks (ANNs) and genetic algorithms (GAs) are applied to the optimal designof a catalyst for propane ammoxidation. The mole percentages of six components of a catalyst(P, K, Cr, Mo, Al2O3/SiO2, and VSb5WSn) are used as inputs, and the activity and the acrylonitrileselectivity serve as the two outputs. This trained optimal linear combination (OLC) network isused to evaluate the yield of new catalyst compositions generated during GA optimization. Thebest yield of acrylonitrile found after GA optimization is 79%, which is higher than the highestyield previously reported (64%). The OLC neural network, using the acrylonitrile yield (i.e.,activity times selectivity) as the output, greatly improves the simulation of the catalyst systemcompared to a simple, single-network architecture. In particular, whereas single-network methodscan all easily reproduce the experimental patterns used for training and validation, the OLC ismarkedly superior for generalizing to novel catalyst patterns.
|
