Multiphysics simulations are becoming increasingly popular as they represent the complexity of the real world more accurately and include more of the significant phenomena integrated within the simulations. However, the added algorithm complexity can result in less efficient and less stable computations, even in instances when the underlying flow physics is relatively simple. It is proposed to alleviate some of the drawbacks of using multiphysics simulations through implementing a tool that can detect and classify flow physics of interest including turbulent flow, different types of instabilities (for example, Kelvin-Helmholtz rollers, Richtmyer–Meshkov instability, etc.), different types of cavitation, multiphase mixing, and shock waves. The aim is to enable efficient scanning of a target design space using an appropriate modelling complexity for the flow physics detected under specific sets of conditions. This tool can be beneficial in two ways; it allows high-fidelity multiphysics solvers to focus on the flow physics relevant to the problem being simulated at a particular stage of flow development, and it helps identify the most suitable low-fidelity models to include in design studies. The developed tool is tested on the simulation of oleo-pneumatic shock absorber flow, which is a challenging application due to the range of flow physics and complex geometries involved. The conducted simulations range from scale resolving 3D studies to 2-equation dynamic system models, including 2D axisymmetric URANS simulations, which provide a rich dataset to test and calibrate the tool. Further development is envisaged to use training algorithms to progressively improve the prediction of relevant flow physics.