A pseudo 1-D model of an industrial NOx storage and reduction (NSR) catalyst has been developed under isothermal conditions. The Pt/Ba proximity or the presence of barium in different chemical forms such as BaO and BaCO3 can result in a wide range of storage rates. To represent these multiple storage sites, an idealized storage particle with a spherical configuration is proposed where the outer layer is assumed to represent the sites with the highest storage capacities/rates while subsequently deeper layers represent the sites with relatively lower storage capacities. It is further assumed that the NOx species first react with the particles in the outer layer, followed by the particles in subsequent deeper layers. To model the progressively decreasing storage rates, a shrinking core approach is used with a diffusivity coefficient, decreasing nonlinearly along the particle depth. The degree of nonlinearity depends on the complexity of the catalyst and therefore, depending on the catalyst, being studied, one can calibrate this function instead of quantifying different types of storage sites. During cyclic operation starting with a completely regenerated catalyst, model predictions for the NOx concentration profiles are in a good agreement with the experimental data during both the transient and steady cycle-to-cycle regimes. It is also shown that with limited amount of reductant, the regeneration of storage sites takes place only in the front portion of the catalyst up to a certain length and the remaining part of the catalyst is only involved in the storage reactions. After a certain number of cycles, the back of the catalyst reaches its saturation limit for storing NOx and from then on, only the amount of catalyst being regenerated in the front part of the catalyst during the rich phase participates in storage. (C) 2010 Elsevier B.V. All rights reserved.