Integration of large-scale experiments and numerical simulations for the calibration of friction laws in volcanic conduit flows

The eruptive columns of explosive volcanism are fed by a gas–particle conduit flow, which characteristics determine the eruptive regime and are important for assessing the hazard of active volcanoes. In this paper, by means of the combined use of large-scale experiments and numerical modeling, a study on some parameters of the gas–particle conduit flow of explosive eruptions is carried out. A 1D two-phase non-homogeneous Eulerian–Eulerian model has been developed for checking the influence of some crucial quantities: interphase drag, particle–wall friction and particle shape factor. Hundreds of different parameters combinations are tested and used for the simulation of controlled experimental runs. The parameter combination that best fits the whole set of experiments, including both column collapses and convective plumes, results into an average error of about 10%. A further analysis has been carried out to determine the sensitivity of solutions to model parameters. The choice of the interphase drag does not influence dramatically the solution, except for highly concentrated flows. The particle shape factor severely affects gas and particle velocities. The influence of various particle–wall friction laws, which were originally obtained in pneumatic engineering, is thoroughly investigated, as the suitability of these laws has never been proved in volcanology. A detailed parametric analysis allowed the re-calibration of two of these laws, which are now specifically tailored for the case of highly concentrated conduit flows that feed collapsing columns, and dilute flows that feed convective plumes.

► Steady 1D non-homogeneous numerical model of gas–particle conduit flow.
► Investigation on the influence of different particle–wall friction factor laws.
► Recalibration of some friction laws for fitting different experimental scenarios.
► Validation of multiphase conduit flow numerical models via large-scale experiments.