His postdoctoral research has branched out to investigate how high order numerical schemes can be applied in fluid flow modelling. He has also been involved in the application of laser techniques, such as PIV (Particle Image Velocimetry) to better understand the main characteristics of fluid flow.

His PhD thesis was focused on the application of the porosity distributed resistance model in non-structured computational mesh for accidental explosion modelling. The research evaluated how the parameterisation of small scale objects in terms of porosities affect the generation and dissipation of turbulence and how it is related to the wrinkling of the flame area. The CFD code developed on the framework of the research project was also used in the development of accidental response surfaces for probabilistic explosion studies.

His master dissertation addressed the application the application of the Set Covering Problem (SCP) combined with computational fluid dynamics to solve the problem of optimisation of gas detectors. The problem comprised the calculation of the optimal number of detectors as well as the optimal location as far as the covering problem is concerned. The 0-1 integer problem was solved using the Balas Algorithm leading to a computer code, namely Optimi. Optimi relies on the input from CFD results. Later the developed code was applied in various designs, particularly in Petrobras (South America) and Sevan Marine (North Sea).

Sávio Vianna is a university professor of transport phenomena and computational methods at University of Campinas (Unicamp), teaching both undergraduate and graduate students.

He has previously supervised 12 master students and five scientific undergraduate students. At the moment, his research group is formed by four PhD students, two of them are working with experimental fluid dynamics and the other two are working in the development of PFS (Porosity Flow Solver).

It was not until the Buncefield accident in 2005 that chemical and process engineers started to discuss the possibility that detonations could indeed occur in storage tank farms and chemical process areas. Up to that point, the deflagration mode was the preferred choice to model such types of accidents.

It turns out that it is not clear how the reacting flow switches from the deflagration regime to the detonation regime. In fact, the DDT (deflagration to detonation transition) phenomena is an area of active research. Alongside the detonation comes the shock wave that might well be described as a discontinuity.

The level of turbulence, particularly enhanced by the small scale geometry, plays a role of paramount importance in reacting flows. However, a computational mesh that takes into account all geometrical details of a process plant is not feasible within the time scale demanded in any engineering design.

In this context, the porosity approach and the higher order numerical scheme can help the modelling of such phenomena in an efficient manner. The parameterisation of the geometry combined with a coarser mesh (due to a more accurate numerical scheme) can lead to computational tools that can help to describe and understand such class of fluid flows, particularly those where an experimental setup is prohibitively expensive.

Along these lines, arises another very interesting problem. The numerical diffusion. It poses an enormous challenge in the modelling of fluid flow in micro-scale. In such scales, the importance of diffusion is enhanced as it is the main mixing mechanism. In microfluidic, the Reynolds numbers are very small and therefore turbulence is hardly observed. Mixing is driven by diffusion.

Although this aspect of the flow simplifies the CFD modelling, on the other hand, diffusion is difficult to account for with the level of accuracy required in the small scale flow. A possible "remedy" could be the utilisation of higher order numerical schemes, such WENO (Weighted Essentially non-oscillatory). It is thought that it might help to suppress the numerical diffusion issue. Here again, an appropriate computational tool can be of paramount importance during the design and conception of microfluidic devices which can be used in various areas of bioengineering and genetics.

The diffusion process is also important in many aspects of fluid flow in bio systems, as for example, the conducting system of the human lung. The computational modelling of fluid flow in the airways is not only import for drug delivery but also to calculate the rate of deposition of particulates on the mucous walls of the bio channels. The application of open source codes (such as OpenFoam) allows for the modification of the source code that eases the verification of how higher order schemes behave in this type of modelling. Another important point is the low cost of the research as there are no software license fees.

As a strategy within the context of fluid flow, his research group dedicates efforts to be successful in external grant funding, by combining the know-how acquired in previous successfully approved funding grants. Other than the traditional research fundings, the utilisation of crowdfunding is seriously considered. Particularly due to the fact that initial investment in theoretical research is not research is not severely high.

Sávio Vianna's research group also strive for successful partnership with multi centres where it was possible to develop professional relationship during the last years of research. Collaboration between the chemical engineering department at Unicamp and other departments, which share the same research interests are certainly more than welcome.