As part of the ambition to capitalize on Qatar's gas resources we investigated the potentials of gas-to-liquids (GTL) based jet fuels. These fuels consist of similar carbon cuts as oil-derived fuels, yet they are considered ultra-clean because of their extremely low sulfur and aromatics contents. However, because of this difference, GTL based jet fuel requires additional development to meet the strict aviation industry standards. The properties of synthetic fuels are determined by their formulation, where the paraffinic building blocks (normal-, iso-, cyclo-) in varying ratios and carbon numbers impact properties such as freezing point, heat content, density and others; all of which influence the fuel's combustion behavior and emissions. This has led us to develop a model capable of predicting the properties of a given blend of synthetic fuel, in order to tailor jet fuel hydrocarbon structure. Due to highly complex interactions between the constituents of the blend, the resulting properties cannot be estimated easily. We overcame this problem by using artificial neural network (ANN) methodology, which handles statistical data in a way that is particularly useful for non-linear systems. The network is trained using experimental data from our own specialized fuel characterization lab to map the relationship between the jet fuel composition and its properties. Three rounds of experimental studies have been completed and translated into a working model using the ANN. We have subsequently validated this model to predict desired jet fuel blends that satisfy aviation industry standards. More importantly we developed visualization models through unique programs utilizing our supercomputer, which integrated experimental and statistical data. 2D- and 3D-contour plots were generated for visualization of blend compositions vs. property relationships (Figure 1). This allowed us to have a better understanding of component interactions and their effect on fuel characteristics, thereby increasing the market value of such fuels.


Article metrics loading...

Loading full text...

Full text loading...

This is a required field
Please enter a valid email address
Approval was a Success
Invalid data
An Error Occurred
Approval was partially successful, following selected items could not be processed due to error