Multi-framework numerical model of the interaction between drilling fluid and drill-string

Abstract

The need for drilling in deeper depths and different conditions drives the development of technologies associated with geoenergy sources. This process is defined by a tighly woven multi-physics environment where there is a multi-phase flow, constituted by a polydisperse suspension on a non-Newtonian fluid, and an extremely slender rotating structure, which is prone to several nonlinear vibration phenomena. In this work, a multi-framework simulation model is developed through several iterations – in each, a new depth of the problem is added and analysed. This framework aims to combine the effects of the drill-string vibration, with the multiphase flow approaching a monolithic method, which models the transport properties with the structural dynamics where the couplings are able reciprocate the interactions. Ultimately, a combination of the lattice-Boltzmann and discrete element methods is used for the multiphase flow, together with either the finite element or lumped parameter schemes, which estimates the drill-string vibration in a limited section. However, throughout this work, special detail is given to isolated couplings. First, the interaction between fluid and particles is analysed, followed by the fluid-structure behaviour of the flow and the drill-string. In sequence, the grinding method is explored without the use of the fluid phase. Finally, a coupled framework is explored with the flow, the cuttings and the drill-string vibration. Initially, the two-dimensional flow inside the drill-pipes is analysed as a Venturi-like configuration. Through this, the fluid solver and the hypotheses are analysed as to better understand its characteristics, such as compressibility limits, boundary condition stability, and the implications of the two-dimensional simplifications on the development of turbulence. Next, the discrete element model is included to the solver and the behaviour of the suspension is analysed in both annular configuration as in a free-fall setup. In the latter, both Newtonian and shear-thinning fluids are simulated and this way, the particle’s ability to influence the flow is qualitatively evaluated – the formation of agglomerates drastically affects the flow. With the lattice-Boltzmann method defined for moving boundary problems, several simulation setups are conducted to estimate the effect of hydraulic forces on the drill-string. Whereas most of the flow solver is kept unchanged, the drill-string dynamic is modelled with two different approaches. Initially, the dynamic is prescribed. This way, functions that mimics common vibration problems are used as inputs for the moving boundary and the forces are directly calculated as the surrounding flow is obtained. Next, a drill-string section is modelled through the finite element method and, for each element, a fluid cross-section is set. This way, a two-way coupled system is obtained. Finally, the three-dimensional flow problem is solved around a moving tool-joint. To do so, the movement of the structure is once again prescribed. After the first three-dimensional simulation batch, the comminution process is modelled. At this stage, only the drill-string and the particles are considered, whereas the flow is assumed to be uniform. With this, the capabilities of the numerical model for grinding at the tool-joints is assessed – the stability of the method and the implications, such as the mass conservation of the system. Finally, several simulation are conducted to evaluate both quantitatively and qualitatively the cuttings transport properties at the drill-pipes. To this end, a polydispersed cuttings configuration is assumed for both horizontal and vertical drilling. Afterwards, different cases for monodispersed transport are simulated in a deviated well configuration, as to evaluate how inclination is relevant to effectively carry the cuttings. Those results are shown and explained throughout this work in several sections as well as the literature review and the theoretical background that based them. To accomplish this, the text is organized in the following way: In Chapter 1, an overview of the publications that were originated from this work is done, linking each publication in a logical thread. In Chapter 2 an introduction is provided, where details on the problem at hand is described as well as the motivation for this research. This is followed by a literature review in Chapter 3, where publications that were relevant to this work are presented. In Chapter 4, the different models are developed and, in Chapter 5, the results obtained are shown. In sequence, Chapter 6 contains a discussion of the previous chapter. Finally, the Chapter 7 contains the conclusions of the work with possible future works.