CFD MODELLING FOR THE CHARACTERIZATION OF FAST-PYROLYSIS BIO-OIL (FPBO) ATOMIZATION

Fast pyrolysis bio-oil (FPBO) is considered a good substitute for conventional fuel oils in several application, such as boilers, furnaces, and gas turbines. This bio-oil has different physical and chemical properties compared to the one of petroleum and other bio-oils. In fact, as it has been widely demonstrated in many studies, this replacing needs some modifications of the already existing technology to be adapted to work with FPBO. The aim of this thesis is to investigate how to set up a CFD model of a nozzle with FPBO, and to analyse the average size of the spray particles for future applications in gas turbines. To achieve this purpose an initial literature review has been done to have an overview on the studies already done about FPBO and its characteristics. Particular attention was paid to works focused on the evolution of spray performance of FPBO. Most of them showed that, due to the uncommon properties of FPBO, the atomization of this fuel with pressure atomizers is not so efficient, but different results have been achieved with twin-fluid nozzles. For this reason, as starting point for the creation of a CFD model was chosen the experimental work of Sangsig Yun et al., who obtained, using an air blast nozzle, a FPBO spray with Sauter mean diameter (SMD) of 50 microns. For the setting up of the CFD model, it has been necessary to select a surrogate and to define its physical and chemical properties. In parallel a 3D model has been developed in order to simplify the original nozzle used in the experimental work. The CFD model has been created thanks to the innovative software CONVERGE, the strength of this software is the capability to create an autonomous mesh, reducing in this way the total computational time. Once the CFD model was ready the results have been post-processed by Tecplot, another common software. Finally, to validate the results performed by the simulation, they have been compared with the one obtained in the experimental work of Sangsig Yun et al., showing a good similarity.