Computational Modelling of Thermal Dynamics and Fluid Mechanics – what is it and how can it help me?

Computational modelling is powerful tool that can be used to simulate many types of systems and their behaviour. In this article we will focus on computational modelling of thermal dynamics and fluid mechanics. Many engineers are introduced to thermal dynamics and fluid mechanics in school and later work with them while designing processes for manufacturing applications. Unfortunately, the knowledge provided in an undergraduate degree alone is not always adequate to ensure a successful project years later. An incomplete understanding of the fundamentals of thermal dynamics and fluid mechanics is often the root cause of system inefficiencies, poor performance, or the outright failure of a project. 

What is computational modelling?

The physics behind thermal dynamics and fluid mechanics are governed by a series of partial differential equations. On a very basic level computational modelling involves breaking down a proposed system design into small parts (this is called meshing) and applying the governing equations to each individual part. The program simultaneously solves the equations for each element, ensures continuity throughout the system, and checks for adherence to any prescribed boundary conditions inputted by the designer. This allows complex problems to be solved much faster and more reliably than through manual calculations. 

Why use computational modelling?

Not every project requires computational modelling to be solved. In some cases—such as when there is a steady state or simple geometry—problems can be simplified enough that straightforward calculations can be used to get a good estimate of the thermal and air flow conditions of a design.

Unfortunately, there are many applications where the estimates given by these simplified equations are not accurate enough to produce a system that will work. In these cases, the designer must use the fundamental partial differential equations in order to determine the thermal and/or fluid flow characteristics of their system design. Computational modelling is used for projects like these where the designs are too mathematically intense to solve manually and too complex or expensive to simulate experimentally. 

Can I do modelling myself?

Computational modelling software is a powerful tool, however the results are only as valid as the inputs given by the designer. It is very easy to get into a “garbage in, garbage out” situation when attempting to do it yourself. Everything from setting up the initial mesh incorrectly to inputting inappropriate boundary conditions will give spurious results. The real problem occurs when the program outputs invalid results without the designer realizing it. This is especially likely to happen if they do not have much experience in the area being modelled. A person with a background in thermal or fluid dynamics is much more likely to catch nonsensical results and correct the modelling accordingly before an expensive implementation is undertaken.

When should I consider having modelling done?

Not every process involving heat or air requires computational modelling. Here are some general cases when it should be considered:

• If the behaviour of the system before steady state is reached is important
• If the geometry of the system is too complex for simplified equations to be used
• If the distribution of temperature, flow, etc. through the system is important
• If you want to experiment with system or process parameters without the expense of building multiple prototypes

A word of warning

Any computation modelling results bring with them uncertainty. Short of building and testing a system there is no way to be certain of the physical behaviour you’ll see. You should not depend wholly on computational analysis alone; always use it in combination with physical experiments and testing.


Need help determining if your project would benefit from computational modelling? Give your Technical Sales Representative a call to discuss your project.