P.181 Machine learning for classification of cognitive impairment according to antipsychotic dopamine receptor occupancy and illness severity in late-life schizophrenia

Abstract

Hydrodesulfurization of heavy oil is usually operated in a trickle bed reactor. Analysis and design of this reactor require modeling both at the macro-reactor scale and micro-catalyst scale. The paper is focused on the catalyst-level modeling and takes into consideration complex catalyst shapes, wettability, and spatial variation in diffusivity due to catalyst pore plugging. Furthermore, the effects of mass transfer from dry and wet zones around the partially wetted catalyst surface are included. Finite element method is used for two-dimensional analysis as it is versatile and can handle complex shapes. The catalyst-level model includes an updating of effective diffusivity with time, and hence the lifetime of a complex shaped catalyst can be predicted. Long-time operation shows that thickness of plugged pore is large and also non-uniform across the two-dimensional catalyst particle with more plugging near the catalyst surface. In addition, the reaction regime changes with time, the fresh catalyst initially being in a diffusion-resistance-free regime and changing to a strong-diffusion-resistance regime at longer operating times. The computational method can be used for optimization of the operating protocol and to design optimum pore size and shape of the catalyst, although these details are not addressed in this paper.