Lumped parameter models are a popular approach for the modelling of the vocal tract. They are well suited to synthesize vowels, their main advantages being their relative simplicity and low computation cost. However, difficulties arise in simulation when trying to reproduce consonants, as their corresponding configurations do not always comply with the quasi-static approximation used to derive these models. In this communication we introduce a simple fluid mechanical model that consistently describes the powerbalanced interaction between the moving walls of the vocal tract and the fluid dynamics, expressed in the port-Hamiltonian systems framework. The capacities of the model are demonstrated through simulations using a numerical scheme that preserves the power balance at discrete time. We then compare this model to the well known lumped parameter model of [Maeda, 1982] using equivalent electrical circuits which proves to be useful to understand the underlying structure that governs the exchange of power between components. At last we discuss how we intend to extend the aforementioned model to include more complex phenomena while preserving the structure and passivity provided by the port-Hamiltonian framework.