Many physical and chemical processes involve energy change with rates that depend sensitively on local temperature. Important examples include heterogeneously catalyzed reactions and activated desorption. Because of the multiscale nature of such systems, it is desirable to connect the macroscopic world of continuous hydrodynamic and temperature fields to mesoscopic particle-based simulations with discrete particle events. In this work we show how to achieve real-time measurement of the local temperature in stochastic rotation dynamics (SRD), a mesoscale method particularly well suited for problems involving hydrodynamic flows with thermal fluctuations. We employ ensemble averaging to achieve local temperature measurement in dynamically changing environments. After validation by heat diffusion between two isothermal plates, heating of walls by a hot strip, and by temperature programed desorption, we apply the method to a case of a model flow reactor with temperature-sensitive heterogeneously catalyzed reactions on solid spherical catalysts. In this model, adsorption, chemical reactions, and desorption are explicitly tracked on the catalyst surface. This work opens the door for future projects where SRD is used to couple hydrodynamic flows and thermal fluctuations to solids with complex temperature-dependent surface mechanisms.

The option of varying the molecular mass in multicomponent lattice Boltzmann simulations is being explored. First, results are presented for droplet formation at an aperture in a second immiscible liquid medium in which the difference in density between the two media is achieved by introducing asymmetry in the EOS, via adding particularly intra-component interaction forces in a pseudo-potential LB model. The second application for models with variable molecular masses is a single-phase heterogeneous laminar-flow tubular chemical reactor, where the molecular masses of reactants and products differ. In this application, tuning the molecular mass requires modification of the standard equilibrium distribution function as well as the use of an extended velocity set, in our case D2Q13. The method is validated against analytical solutions for canonical 1-D diffusion-reaction cases. In both the droplet formation study and the chemical reactors, the results of the exploratory 2-D simulations look qualitatively correct.