Schémas numériques compacts basés sur le résidu en maillage irrégulier pour les équations de Navier-Stokes en compressible

This work is devoted to an approach which makes use of the residual vanishing at steady-state tobuild 2,d or 3d_order space-accurate numerical schemes for the Euler and Navier-Stokes equations, which aresunple, compact, robust and efficient. In the first part of this memoir, the design principles of this approach are recalled in a fuiite-differenceframework and the 3d order residual-based scheme is proved to compare very favourably with a conventional 3dorder approach on a set of model problems. Next, the residual-based scheme is extended in a rigorous way onirregular Cartesian grids so as to preserve its accuracy order. In the second part of this study, we propose a fmite-volume extension of the residual-based scheme thatguarantees, in a simple way, 2 ,d order accuracy on highly irregular grids. This accuracy-preserving extension ismade possible thanks to the compactness of the residual-based dissipation. The very good accuracy of this FVversion is demonstrated for inviscid and viscous flows over airfoils and a wing. In the last part of this work, a residual-based scheme is developed for unsteady flow problems within a dual-time framework. In this case, the residual used to build the scheme contains the derivative with respect to the physicaltime and vanishes at steady-state on a fictitious time. A 2d order time-accurate and 3d order space-accurate residual-based scheme is then devised, which yields better solution than a conventional directional scheme both for a model advection problem and a complex unsteady viscous flow. The implementation of a new organisation is a way to answer the pressure that an external environment firm may experience. We decide to manage this work as a project. The mainpoint of this project will be a new organisation design and implementation. We link design phase with an anticipated risk project management approach. In order to change we decide to propose an original approach which is an alternation between breaking change and continuous improvement change. We also identify the need for helping change adaptation people. All of our proposals are born of a work in a pharmaceutical firm. This research leads to the creation of a model based on three processes: project management, strategic management and people management. In order to answer the need for controlling theses processes, we propose a control method. This method is composed by five phases created from our model and by eight principles coming from our fieldwork.