We present a novel recurrence CFD (rCFD) method for the efficient simulation of passive scalar transport in pseudo-periodic flows. rCFD transfers full CFD(-DEM) simulations into a purely Lagrangian description of scalar transport by evaluating trajectories of fluid tracers and discrete particles. These trajectories are subsequently converted into cell to cell shift operations, resulting in data shift pattern on a computational grid. Finally, a recurrence process, based on the statistics of the full CFD(-DEM) simulation, controls the sequence of consecutive frames of data shift patterns. First, we applied rCFD to species transport in single-phase large eddy simulations, yielding excellent agreement with corresponding full CFD simulations. Second, we addressed heat transfer in a bubbling fluidized bed, yielding acceptable agreement with corresponding CFD-DEM simulations. In both cases, simulation times were reduced by more than four orders of magnitude without losing spatial resolution, resulting in faster than real-time simulations. Subsequently, we discuss existing model limitations such as (i) process periodicity, (ii) passiveness of transport, (iii) data storage requirements and (iv) physical diffusion and sketch possible future remedies. Concluding, we consider rCFD a promising bridging technology between full CFD simulations and online process prediction. (C) 2018 Elsevier Ltd. All rights reserved.