The Thing

The physics of fluid flow are governed by partial differential equations, known as Navier-Stokes equations. These are both coupled and non-linear, and so have no general analytical solution. They can however be simplified to Euler equations by removing terms that describe viscosity. By further removing terms describing viscosity yields the full potential equations. These can then be linearised to yield the linearised potential equations.

Initially, the only feasible way of generating useful flow data was experimentally. This is no longer the case today. Computational fluid dynamics (CFD) is commonly used to generate numerical solutions to the above outlined equations. There are many advantages to using CFD over collecting data experimentally. One of these is cost savings, as it can be extremely expensive to simulate accurate wind tunnel models or complex geometry. It is also very straightforward to change one or more variable to optimise the component once a suitable model has been generated. CFD codes can also be used to solve inverse engineering problems. In industrial applications, to ensure their accuracy, CFD simulations are regularly backed up with some experimental data.

The basic laws of conservation are the foundations of any CFD code. For a calculation involving a fluid of homogenous composition for example, it is required that energy and mass momentum are both conserved. To satisfy the conservation of any arbitrary physical constant, be it mass, momentum, energy etc., the net flux of property out of the control volume, plus the time rate of change of the property in the control volume, must equal the value of any sources of the property in the control volume.