Radial Basis Function Neural Networks-Based Modeling of the Membrane Separation Process: Hydrogen Recovery from Refinery Gases

Membrane technology has found wide applications in the petrochemical industry, mainly in the purification and recovery of the hydrogen resources. Accurate prediction of the membrane separation performance plays an important role in carrying out advanced process control （APC）. For the first time, a soft-sensor model for the membrane separation process has been established based on the radial basis function （RBF） neural networks. The main performance parameters, i.e, permeate hydrogen concentration, permeate gas flux, and residue hydrogen concentration, are estimated quantitatively by measuring the operating temperature, feed-side pressure, permeate-side pressure, residue-side pressure, feed-gas flux, and feed-hydrogen concentration excluding flow structure, membrane parameters, and other compositions. The predicted results can gain the desired effects. The effectiveness of this novel approach lays a foundation for integrating control technology and optimizing the operation of the gas membrane separation process.

Simultaneous recovery of carbon and sulfur resources from reduction of CO_2 with H_2S using catalysts

An approach to the simultaneous reclamation of carbon and sulfur resources from CO2 and H2S has been proposed and effectively implemented with the aid of catalysts. A brief thermodynamic study reveals the potential of direct reduction of CO2 with H2S（15：15 mol% balanced with N2） for selective production of CO and elemental sulfur. The experiments carried out in a fixed-bed flow reactor over the temperature range of 400–800 °C give evidence of the importance of the employment of catalysts. Both the conversions of the reactants and the selectivities of the target products can be substantially promoted over most catalysts studied. Nevertheless, little difference appears among their catalytic performance. The results also prove that the presence of CO2 can remarkably enhance H2S conversion and the sulfur yield in comparison with H2S direct decomposition. A longtime reaction test on Mg O catalyst manifests its superior durability at high temperature（700 °C） and huge gas hourly space velocity（100,000 h-1）. Free radicals initiated by catalysts are supposed to dominate the reactions between CO2 and H2S.