Abstract. In hemodynamics, local phenomena, such as the perturbation of flow pattern in a specific vascular region, are strictly related to the global features of the whole circulation (see, e.g., [L. Formaggia et al., Comput. Vis. Sci., 2 (1999), pp. 75-83]). In [A. Quarteroni, S. Ragni, and A. Veneziani, Comput. Vis. Sci., 4 (2001), pp. 111-124] we have proposed a heterogeneous model, where a local, accurate, three-dimensional description of blood flow by means of the NavierStokes equations in a specific artery is coupled with a systemic, zero-dimensional, lumped model of the remainder of circulation. This is a geometrical multiscale strategy, which couples an initialboundary value problem to be used in a specific vascular region with an initial-value problem in the rest of the circulatory system. It has been successfully adopted to predict the outcome of a surgical operation (see [K. Laganà et al., Biorheology, 39 (2002) Barcelona, Spain, 2000]). However, its interest goes beyond the context of blood flow simulations. In this paper we provide a well-posedness analysis of this multiscale model by proving a local-in-time existence result based on a fixed-point technique. Moreover, we investigate the role of matching conditions between the two submodels for the numerical simulation.Key words. geometrical multiscale models, blood flow simulations, fixed-point techniques
AMS subject classifications. 35M20, 35Q30, 76D03, 65L05, 76Z05PII. S1540345902408482 1. Introduction. In computational fluid dynamics there might be the need for accurately simulating the flow in a subregion of a complex system that can be reasonably represented by a hydraulic network. One may think, for instance, of a system for water supply for which there is the interest of ascertaining the spatial variations of velocity and pressure only in a specific pipeline or in a reservoir. Another instance, which has motivated our investigation, is the blood circulatory system. Here the underlying (low cost) model is based on an analogy with electric circuits and can predict the time evolution of average physical quantities (flow rate and pressure) in the different compartments of the overall system (see, e.g., [15,27]). Yet, there is the interest of carrying out three-dimensional (3D) simulations on a sensible region (e.g., a coronary bypass, a carotid bifurcation, a stented artery, etc.). In these cases, the local investigation requires the specification of boundary conditions on the interface between the region of interest and the remainder of the network. In fact, the significance of numerical results is strictly related to the capability of the numerical device of properly accounting for the exchange of information with the global network. Actually, it is well known that, in the circulatory system, local behavior of blood flow in a specific region is strictly related to the systemic features of the circulation. For