In this paper we present the development of a Fast Boundary-Domain Integral Method for the solution of fluid flow and heat transfer problems. Adaptive cross approximation of integral matrices is performed on a hierarchically decomposed matrix structure. We couple the developed flow simulation algorithm with stochastic modelling of input parameters. In order to assess the influence of input parameters on the simulation results, we employ the stochastic collocation method as a wrapper around the deterministic code. We apply the developed numerical tool for the simulation of flow and heat transfer of nanofluids.