The prediction of microstructure evolution during passive mixing is of major interest in order to qualify and quantify mixing devices as well as to predict the final morphology of the resulting blend. Direct numerical simulation fails because of the different characteristic lengths of the microstructure and the process itself. Micro-macro approaches could be a valuable alternative but the computational cost remains tremendous. For this reason many authors proposed the introduction of some microstructural variables able to qualify and quantify the mixing process at a mesoscale level. Some proposals considered only the effects induced by the flow kinematics, other introduced also the effects of shape relaxation due to the surface tension and coalescence. The most advanced integrate also the break-up process. However, the derivation of the evolution equations governing the evolution of such microstructural variables needs the introduction of some closure relations whose impact on the computed solution should be evaluated before applying it for simulating complex mixing flows. In this work we consider the Lee and Park's model that considers the flow kinematics, the surface tension, the coalescence and the break-up mechanisms in the evolution of the area tensor. The accuracy of both a quadratic closure and an orthotropic relations will be analyzed in the first part of this work, and then the resulting closed model by using a quadratic closure will be used for simulating complex mixing flows.< Réduire