Multibody Advanced Airship for Transport

Funder

Organizational Unit

Publications

Development of an open source software tool for propeller design in the MAAT project

Publication . Morgado, João Paulo Salgueiro; Silvestre, Miguel Ângelo Rodrigues; Marques, José Carlos Páscoa

This thesis presents the development of a new propeller design and analysis software capable of adequately predicting the low Reynolds number performance. JBLADE software was developed from QBLADE and XFLR5 and it uses an improved version of Blade Element Momentum (BEM) theory that embeds a new model for the three-dimensional flow equilibrium. The software allows the introduction of the blade geometry as an arbitrary number of sections characterized by their radial position, chord, twist, length, airfoil and associated complete 360º angle of attack range airfoil polar. The code provides a 3D graphical view of the blade, helping the user to detect inconsistencies. JBLADE also allows a direct visualization of simulation results through a graphical user interface making the software accessible and easy to understand. In addition, the coupling between different JBLADE modules avoids time consuming operations of importing/exporting data, decreasing possible mistakes created by the user. The software is developed as an open-source tool for the simulation of propellers and it has the capability of estimating the performance of a given propeller geometry in design and off-design operating conditions. The current development work was focused in the design of airship propellers. The software was validated against different propeller types proving that it can be used to design and optimize propellers for distinct applications. The derivation and validation of the new 3D flow equilibrium formulation are presented. This 3D flow equilibrium model accounts for the possible radial movement of the flow across the propeller disk, improving the performance prediction of the software. The development of a new method for the prediction of the airfoil drag coefficient at a 90 degrees angle of attack for a better post-stall modelling is also presented. Different post-stall methods available in the literature, originally developed for wind turbine industry, were extended for propeller analysis and implemented in JBLADE. The preliminary analysis of the results shows that the propeller performance prediction can be improved using these implemented post-stall methods. An inverse design methodology, based on minimum induced losses was implemented in JBLADE software in order to obtain optimized geometries for a given operating point. In addition a structural sub-module was also integrated in the software allowing the estimation of blade weight as well as tip displacement and twist angle changes due to the thrust generation and airfoil pitching moments. To validate the performance estimation of JBLADE software, the propellers from NACA Technical Report 530 and NACA Technical Report 594 were simulated and the results were checked against the experimental data and against those of other available codes. The inverse design and structural sub-module were also validated against other numerical results. To verify the reliability of XFOIL, the XFOIL Code, the Shear Stress Transport k-ω turbulence model and a refurbished version of k-kl-ω transition model were used to estimate the airfoil aerodynamic performance. It has been shown that the XFOIL code gives the closest prediction when compared with experimental data, providing that it is suitable to be used in JBLADE Software as airfoil’s performance estimation tool. Two different propellers to use on the MAAT high altitude cruiser airship were designed and analysed. In addition, the design procedure and the optimization steps of the new propellers to use at such high altitudes are also presented. The propellers designed with JBLADE are then analysed and the results are compared with conventional CFD results since there is no experimental data for these particular geometries. Two different approaches were used to obtain the final geometries of the propellers, since, instead of using the traditional lift coefficient prescription along the blade, the airfoil’s best L3/2/D and best L/D were used to produce different geometries. It was shown that this new first design approach allows the minimization of the chord along the blade, while the thrust and propulsive efficiency are maximized. A new test rig was developed and used to adequately develop and validate numerical design tools for the low Reynolds numbers propellers. The development of an experimental setup for wind tunnel propeller testing is described and the measurements with the new test rig were validated against reference data. Additionally, performance data for propellers that are not characterized in the existing literature were obtained. An APC 10”x7” SF replica propeller was built and tested, providing complementary data for JBLADE validation. The CAD design process as well as moulds and propeller manufacture are also described. The results show good agreement between JBLADE and experimental performance measurements. Thus it was concluded that JBLADE can be used to design and calculate the performance of the MAAT project high altitude cruiser airship propellers.

2016Doctoral thesis

Open access

Development of computational methodologies for turbulence transitional flow prediction

Publication . Oliveira, Rui Vizinho de; Marques, José Carlos Páscoa; Silvestre, Miguel Ângelo Rodrigues

Turbulence transition modelling is still, albeit the past developments, an active research area of interest for various industry sectors. Its modelling can range from RANS based closures to full DNS computations. The former approach is of course the most feasible simulation methodology. Therefore, RANS based transition models have been developed for industry use. These, range from empirically correlated transition models to physics based phenomenological transition closures. Implementation and validation of these models resulted in a deeper understanding of the processes by which RANS based closures are able to predict turbulence transition onset. The research presented herein on the speci c type of physics in which the transition models are based resulted in an accuracy improvement of an existing turbulence transition closure, the k-kl-!. Additionally, upon gaining a deeper understanding on the role of the pre-transitional ow region, a new turbulence transition model was devised. This is based on a never before applied concept of pre-transitional turbulent vortex deformation due to mean ow shear. This will induce the appearance of a small pre-transitional turbulent viscosity on the edge of the laminar boundary layer. The induced viscosity is a result from the predicted small negative pre-transitional u0v0 values. Although experimentally veri ed, up until now, no model has ever been able to predict this turbulent feature based on a mechanical analogy. The transition V-model was then coupled to a turbulence model, the Spalart-Allmaras closure, resulting in the V-SA transition model. This was validated for a wide range of ow conditions and multiple geometries. It is concluded that the mechanical analogy based closure is a feasible concept with a promising future. Although the developed V-SA turbulence transition model is simple, it is able to predict complex transition phenomenon.

2016Doctoral thesis

Open access