Modeling fixed-bed thermo-chemical processes of biomass should be considered as a multi-scale problem. The molecular, particle and reactor level should be considered together in the models. A framework for a multi-scale model for dynamic fixed-bed/moving-bed thermo-chemical conversion processes and the respective numerical solution method is introduced: the Representative Particle Model (RPM). In the RPM approach an intra-particle model is solved for each finite volume element of the reactor. All particles within a finite volume element are assumed to obey the same characteristics as the one for which the intra-particle model is solved, which is why it can be considered as representative. The particle level is then considered, as opposed to quasi-continuous models, that are just appropriate when intra-particle gradients of temperatures and concentrations are negligible. And compared to the Discrete Particle Model (DPM), where a model for each particle in the bed (up to 10(5)-10(6) particles in technical scale fixed-bed gasification) is solved simultaneously with the fluid-phase balances in the reactor domain, the computational effort is significantly smaller. The RPM is applied to fixed-bed pyrolysis and compared to experimental results available in the literature. The RPM is able to predict the experimental results, describing intra-particle gradients, in a much more feasible time compared to the DPM. The importance of intra-particle gradients in fixed-bed pyrolysis is also highlighted, showing differences in the conversion time of more than 30% when the particle size is doubled, from a particle diameter of 1.24-2.48 cm. Therefore the RPM is a feasible way to include the particle level in a multi-scale description of fixed-bed pyrolysis. (C) 2012 Elsevier Ltd. All rights reserved.