In the computational modeling of two-phase flow, many uncertainties are usually faced in simulations and validations with experiments. This has traditionally made it difficult to provide a general method to predict the two-phase flow characteristics for any geometry and condition, even for bubbly flow regimes. Thus, we focus our research on studying in depth the bubbly flow modeling and validation from a critical point of view. The conditions are intentionally limited to scenarios where coalescence and breakup can be neglected, to concentrate on the study of bubble dynamics and its interaction with the main fluid. This study required the development of a solver for bubbly flow with higher resolution level than TFM and a new methodology to obtain the data from the simulation. In Part II, taking profit of the detailed data provided by the CFD-DEM solver presented in Part I, we propose a novel methodology based on virtual sensor probes, to perform a rigorous validation and to investigate the experimental data. The same approximation used for processing the experimental datasets applies to simulation data, then the same assumptions are considered. In this way we can study an extended number of disperse phase variables as bubble velocity, void fraction, interfacial area concentration, mean chord length and distribution, Sauter mean diameter, bubble frequency and missing ratio, in addition to other variables as bubble size distribution or carrier phase velocity and turbulence. Several upward bubbly flow scenarios from datasets of different authors are used to validate the solver using this methodology. Finally, an axial evolution validation is performed including a discussion motivated by the comparison between experiments and the data from the virtual probes. (C) 2017 Elsevier Ltd. All rights reserved.