Blade optimization of unmanned helicopter for variable rotor speed

Abstract

Helicopter unmanned aerial vehicle systems are preferred in many areas due to their vertical landing, take-off and hovering capabilities. In addition, manned helicopters are converted to unmanned systems and used for activities such as cargo and surveillance. Helicopter rotor systems operate at a constant rpm and change the angle of the blades to create the required forces and moments. However, operating the rotor at the ideal speed for the flight condition is one of the concepts studied to increase performance and efficiency. This thesis aims to improve the performance of a helicopter unmanned aerial vehicle rotor system with variable rotor speed and to optimize the rotor blade vibration in the determined rotor speed range. It is aimed to reduce the power required by reducing the speed of the rotor system. In addition, by creating an optimization process, it is aimed to avoid vibration problems due to reducing the rotor speed. Within the scope of the thesis, first of all, literature research was carried out about variable rotor speed and rotor blade optimization. In line with the researches, it was decided to model the rotor blade with the Variational Asymptotic Method instead of using three-dimensional elements in commercial finite element analysis software such as Ansys, Abaqus and Nastran. The Variational Asymptotic Method is used for the analysis of slender structures such as blades, rotor blades and wind turbines. This method, developed by Berdichevski, transforms a three-dimensional problem into a two-dimensional cross-sectional analysis and a one-dimensional nonlinear beam analysis problem. The modelling effort and computational cost are reduced by using this modeling method. A finite element model of the blade cross-section was created to perform cross-sectional analysis for the rotor blade. In order to create the finite element model of the blade section, PreVABS input files were created using cross-section design parameters. In these input files, the section profile, geometric dimensions, material layout and material properties are defined. By using the input files prepared in this way, the section finite element model was obtained parametrically with the PreVABS software. This finite element model was solved with VABS software. As a result of the analysis with VABS, the cross-sectional mass and stiffness properties were calculated. After the cross-sectional analysis, the rotor blade is modeled as a flexible beam with Dymore, a finite element based multibody dynamics software. In this study, a simple model was created with Dymore to be included in an optimization process. Only the blade is modeled as flexible, other components of the rotor are assumed as rigid. The dominant force acting on the rotor blades is the centrifugal force. In addition, with the increase in rotor speed, the centrifugal force also increases, causing the natural frequencies to change. The blade natural frequencies should be calculated in the rotor operating speed range. In this direction, the variation of natural frequencies according to rotor speed should be examined by creating a fan diagram. Centrifugal force was applied to the blade flexible beam model created with Dymore in 10\% increments, and natural frequencies were obtained according to the rotor speed and a fan diagram was created. Thus, the rotor blade analysis was carried out in two stages by cross-section and beam analysis. The main objective of this study is to increase the performance by changing the rotor speed. For this purpose, the rotor power required should be calculated. In this study, rotor power equations derived from the blade element momentum theory are used to estimate the rotor power. Using these equations, rotor power estimation was performed under different load and forward flight speed conditions. By reducing the rotor speed, the drag force of the rotor profile can also be reduced. However, as the rotor speed decreases, the lifting force also decreases. The angle of attack of the blades can be increased to provide the required lifting force. However, the increase in angle of attack is limited due to the stall limit of the blade. To avoid stall, reduction of rotor speed should be limited and the flight conditions should be determined to be applied reduced rotor speed. In this study, considering the torque and power curve of the engine, it was decided to limit the reducing of the rotor speed by 20\%. In this study, it is assumed that the rotor system will operate at two different rotor speeds, which are the baseline rotor speed and 20\% reduced rotor speed. In this direction, power required calculations were carried out for both the baseline rotor speed and the reduced rotor speed. Calculations were performed under different gross weight and forward flight speed conditions to determine the flight conditions which reduced rotor speed is applicable. According to the power estimation results, the flight conditions in which the rotor speed can be reduced by 20\% have been determined. It is concluded that reducing the rotor speed is not applicable in low and high speed forward flight conditions. However, it has been observed that the power required can be reduced by reducing the rotor speed in the cruising speed condition. It has also been observed that as the gross weight increases, reduced rotor speed can be applied in a more limited range of flight speeds. When the percentages of the power required reduction is examined, it is observed that it is the highest at low gross weight and cruising speed. As a result, in this flight condition, it is possible to reduce the power required by about 10\% by reducing the rotor speed by 20\%. In the final stage, it is aimed to reduce vibrations by creating an optimization process. There are different approaches to reduce vibration in rotors. In rotor optimization studies, it is aimed to reduce vibration by reducing the forces and moments on the rotor hub or by tuning the natural frequencies. In this thesis, it is aimed to calculate natural frequencies and separate them from rotor excitations in the optimization process. Sequential Quadratic Programming(SQP), which is a gradient-based algorithm, was preferred as the optimization algorithm. By choosing a gradient-based algorithm, it is aimed to reach the result quickly and with low computational cost. SQP requires low number of function calculations than other gradient-based algorithms. The gradient of functions were calculated numerically by using finite difference method. The optimization aim was determined as increasing the difference between natural frequencies and rotor operating frequency multiples (rotor harmonics). For this purpose, the difference between each natural frequency and the closest rotor harmonic is calculated. The smallest of these differences is defined as the optimization objective function. In this study, the objective function and constraints are calculated for both baseline and 20\% reduced rotor speed. Geometric properties and location of the nonstructural mass were used as optimization design variables. The cross-section dimensions, spanwise and chordwise location of the nonstructural mass are defined as the design variable. Bounds are defined for these design variables, taking into account the geometric limits. In addition, constraints for optimization are defined. Centrifugal force and non-structural mass are constrained to satisfy performance and structural strength criteria with optimized blade design. In addition, optimized blade design, a constraint is defined for the difference between the natural frequencies and it is aimed to avoid undesired vibrations due to coupling of natural frequencies. As a result of the optimization, the difference between the blade natural frequencies and rotor harmonics has been increased by about 48\%. In addition, the nonstructural mass and centrifugal force are also reduced. According to the result of the optimization, optimized design have better vibration and performance characteristic than baseline rotor blade design.