The rapid identification of unknown objects by their thermo-fluid flow field signature is becoming increasingly more important. In this work, a machine-learning framework is developed that efficiently simulates and adapts object geometries in order to match the thermo-flow field signature generated by an unknown object, across a time series of voxel-frames. In order to achieve this, a thermo-fluid model is developed, based on the Navier–Stokes equations and the first law of thermodynamics, using a voxel rendering of the system, which is rapidly solved with a voxel-tailored, temporally-adaptive, iterative solution scheme. This voxel-framework is then combined with a genomic-based machine-learning algorithm to develop a digital-twin (digital-replica) of the system that can run in real-time or faster than the actual physical system. Numerical examples are provided to illustrate the framework.