The Minimized Power Geometric model: An analytical mixing model for calculating polyphase rock viscosities consistent with experimental data

Abstract

AbstractHere we introduce the Minimized Power Geometric (MPG) model which predicts the viscosity of any polyphase rocks deformed during ductile flow. The volumetric fractions and power law parameters of the constituting phases are the only model inputs required. The model is based on a minimization of the mechanical power dissipated in the rock during deformation. In contrast to existing mixing models based on minimization, we use the Lagrange multipliers method and constraints of strain rate and stress geometric averaging. This allows us to determine analytical expressions for the polyphase rock viscosity, its power law parameters, and the partitioning of strain rate and stress between the phases. The power law bulk behavior is a consequence of our model and not an assumption. Comparison of model results with 15 published experimental data sets on two‐phase aggregates shows that the MPG model reproduces accurately both experimental viscosities and creep parameters, even where large viscosity contrasts are present. In detail, the ratio between experimental and MPG‐predicted viscosities averages 1.6. Deviations from the experimental values are likely to be due to microstructural processes (strain localization and coeval other deformation mechanisms) that are neglected by the model. Existing models that are not based on geometric averaging show a poorer fit with the experimental data. As long as the limitations of the mixing models are kept in mind, the MPG model offers great potential for applications in structural geology and numerical modeling.