In the metallurgical industryIndustry, Liquid Metal CleanlinessLiquid metal cleanliness Analyser (LiMCA) commercial equipment cannot distinguish between hard particles (e.g., oxides, borides) and deformable particles (e.g., bubbles, molten salts). Therefore, hard particle concentrations can sometimes be grossly overestimated, which reduces the measurement accuracy. In the present study, a mathematical model is developed to distinguish deformable particles from hard particles in an aluminiumAluminium LiMCA system. The deformation of particles under mixed extensional and shear flow is studied by using the conservative level-set (CLS) method. The effects of capillary number and Reynolds number on particle deformationParticle deformation are studied over modest ranges. It can be found that particle deformationParticle deformation and its subsequent dynamics are stimulated by the subtle interplay between surfaceSurface tension, viscous force, and inertial force. Furthermore, a computational basis is given to estimate the influence of particle deformation on electrical resistance pulses (ERP). It is found that ERP features of deformation particles, including the peak magnitude and the pulse width, are different from those of hard particles. Based on the results, the effect of a particle’s deformation and the feasibility to discriminateParticle discrimination it from non-deformable particles in the LiMCA system is evaluated. Finally, an unsupervised classification algorithm balanced iterative reducing and clustering using hierarchies clustering is employed to distinguish the pulsed features between hard particles and bubbles using a self-develop system: OiMPA (Online Micron-sized Particle Analyzer). This study scheme has a guiding and facilitating role in applying the LiMCA system to the industrial online measurement environment.