Artificial neural networks assisted catalyst design and optimisation of methane steam reforming

Download files
Access & Terms of Use
open access
Copyright: Arcotumapathy, Viswanathan
Altmetric
Abstract
Hydrogen has diverse industrial applications namely in ammonia synthesis, petroleum refining and methanol production. The advent of fuel cell technology had prompted hydrogen as an efficient source of energy carrier. Catalytic steam reforming of natural gas (primarily methane) supplies half of global hydrogen demand. Since methane is the major component of natural gas, catalytic steam reforming of natural gas is referred as methane steam reforming (MSR). However this process is energy intensive because of industrial operations that are carried out at stem-to-carbon feed ratio (S:C) > 3 to avoid carbon formation. The most cost effective way of optimising this mature technology may be carried out via design of novel catalyst that are highly active at low S:C, thermally stable and resistant to carbon formation. Catalyst design and evaluation for MSR is a multifactorial multi-objective optimisation problem and the absence of well-defined mechanistic relationships between wide ranging input-output variables has stimulated interest in the application of artificial neural network (ANN) for the analysis of the large body of empirical data available. However, single ANN models generally have limited predictive capability and insufficient to capture the broad range of features inherent in the voluminous but dispersed data sources. In this study, a Fibonacci approach to select optimal number of neurons for the ANN architecture followed by a new weighted optimal combination of statistically-derived candidate ANN models in a multi-error sense was employed. The results from the experimental validation was consistent with ANN model prediction and further investigation identified 1 wt% Ce promoted 10 wt% Ni/SBA-15 as an efficient catalyst for MSR. This catalyst displayed excellent activity for MSR with a feed composition of S:C = 1:1 2:1 at atmospheric pressure and reaction temperature 1073 K for 72 h time-on-stream resulting in 92% - 99% methane conversion rates under steady-state conditions. Under these experimental conditions surprisingly, in particular the S:C = 1:1 runs did not exhibit any signs of carbon deposition which may be attributed to 1 wt% promotion of Ce. Moreover, an activation energy of 49.8 kJ mol-1 was obtained for MSR over this catalyst through a macro power-law modelling.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Arcotumapathy, Viswanathan
Supervisor(s)
Lucien, Frank
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2013
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download whole.pdf 3.41 MB Adobe Portable Document Format
Related dataset(s)