Manufacturing processes in bioreactors are initially developed at small scales. These are usually operated in homogeneous regime. However, the flow regime may go through transition during scale-up to production size and become heterogeneous. Measurements of mass transfer rates, bubble populations and flow structures at large scale are often infeasible. Therefore, this paper proposes a modeling approach which can be used to calculate holdup and mass transfer at large scale in order to estimate local conditions in production-size bioreactors. The methodology includes the estimation of flow field by CFD followed by calculations using a compartmental model in order to include physics, gas phase population balances and reactions to the system that would be computationally very demanding to calculate with CFD. Different models for coalescence and drag terms in high volume fraction systems are tested. It is shown that flooding of impellers causes unstable conditions and worsens the mass transfer capability of reactors. However, homogeneous conditions with no flooding cause oxygen limitation in the reactor and an increase in the aeration rate is required, which subsequently leads to heterogeneous conditions. Comparison to results with ideal mixing assumption shows the importance of considering local conditions and emphasizes the strength of the compartmental model.