Focus of our research
The noise simulation of helicopters with regard to high-speed and blade-vortex-interaction noise is based on the acoustic analogy of Ffowcs Williams and Hawkings, for which a high-resolution CFD calculation is evaluated.
The double-rotor counter-rotating propfan is similar in terms of the simulation technology. CROR is attractive in terms of consumption, but also extremely loud, so that intelligent noise reduction measures are required.
Next-generation flow solvers allow the application of high order methods on hybrid grids to achieve high accuracies with refined physical models at reasonable effort.
For validation purposes, IAG owns an instrumented model helicopter with a take-off mass of almost 5 kg. In addition to much more expensive wind tunnel measurements, valuable measurement data for different conditions in free flight can be obtained with this model helicopter.
Power requirements and aeroacoustics are becoming equally important for future rotor developments. The optimal design of ground plan and profiling is therefore essential.
A central component of the simulation is the consideration of the elastic deformation of the rotor blades (fluid-structure coupling). For this we use beam models with different levels of detail (Euler-Bernoulli, Timoshenko beams, ...).
For comparability with wind tunnel measurements or flight tests, trimming to the specified forces and moments is of crucial importance.
The dynamic behavior of the helicopter, in particular the rotor blades, is calculated using CSD (Computational Structure Dynamics) methods. The codes "HOST" of Airbus Helicopters and the commercial code "CAMRADII" are used. The dynamic information is needed for the aerodynamic simulation of a helicopter, since the position and deformation of the blade has a direct effect on the flow around it.
The simulation of helicopters generally poses a threefold problem, since it does not only consist of the calculation of the flow itself, but also of the elastic reaction of the rotor blades and the simultaneous maintenance of the desired trim condition.
The trim state of a rotor is a selection of characteristic rotor parameters, such as the total thrust generated and the roll and pitch moments introduced into the fuselage via the rotor mast. Other common parameters are the amplitudes of the blade movements. The purpose of trimming is to bring these parameters to given values. In principle, the number of parameters which can be trimmed is limited by the number of control options on the rotor. If, for example, three control inputs of a swash plate (i.e. the collective and the two cyclically changed blade angles) are available, these can be used to bring the three parameters thrust, roll and pitch torque of the rotor to the desired values. In the case of a so-called active rotor, where e.g. small flaps at the rotor blades can be controlled, further parameters could be considered in the trimming.
Fields of applications
By adhering to specified values for the trimmable parameters, it is ensured that the simulated rotor operates in a defined operating state. This is important for a comparison with an experiment, since the flow-mechanical phenomena and also the rotordynamic processes only then behave in a similar way. Although it would be conceivable to take the swashplate position directly from the experiment without trimming, the forces and moments generated by the rotor are the more suitable quantities to guarantee the similarity of simulation and experiment. This is due to possibly different deformation of the blades with reduced elasticity models and also due to a disturbing influence of the wind tunnel walls.
A further field of application for trimming are simulations for the design of rotors, where the rotor power is to be predicted in the design phase. In order to compare the performance of different designs, they are evaluated using an identical trim state. This allows to determine whether the modifications to the rotor, including their various coupling effects, actually lead to an increase in performance in a realistic operating state, which are often difficult to understand.
Free flight trim
The so-called free flight trim also falls within the scope of preliminary design and performance determination. It is characterized by the fact that not only the rotor control, but also the flight attitude of the complete helicopter is used as variable quantity. This is necessary if a helicopter is to be simulated in free flight instead of a fixed model of a wind tunnel test. In the case of free flight, the balance of forces applies to the complete helicopter: The aerodynamic forces on the rotor and fuselage balance the weight of the helicopter, the moments cancel each other out. As a rule, this equilibrium cannot be achieved in the wind tunnel due to the clamping of the model. Since the power requirement of the helicopter and the aerodynamic load of various components depend on the flight attitude, trimming the attitude angles enables more accurate predictions to be made about these values.
The trimming is methodically implemented through an iterative procedure. A helicopter aeromechanics program provides a prediction for the values of the control inputs, which is tested in a subsequent CFD simulation. Depending on the deviations from the desired trim state, a new proposal is then calculated and another CFD simulation is performed. These steps are repeated until all trim parameters have converged to the desired values.
The main task of the helicopter and aeroacoustics working group is the simulation of aerodynamics. The CFD (Computational Fluid Dynamics) code "FLOWer" of the DLR, which was significantly extended and expanded at the IAG, is primarily used for this purpose. In the past, the DLR code "TAU" was also used for such investigations. In addition, a DG method is being developed for the future: Discontinous Galerkin (DG)
Latest flow models such as Detached Eddy Simulation (DES) or even Large Eddy Simulation (LES) require a significantly improved accuracy of the flow solution in order to achieve the desired improvements and therefore increase the requirements on the code, which can only be achieved with disproportionately high effort. In addition, geometries that can no longer be economically networked with structured grids are now routinely investigated.
Therefore, next generation flow solvers allow the application of high order methods on hybrid grids to achieve high accuracies with refined physical models at reasonable effort. Of great interest are discontinous Galerkin methods, a combination of finite elements with reduced continuity requirements and finite volumes. They allow arbitrary orders on freely selectable - even highly irregular or stretched - networks and are characterized by excellent parallelizability and high locality, which makes them very attractive for current and future computer architectures.
In contrast to many other research codes, our flow solver SUNWinT (Stuttgart University Numerical Wind Tunnel) is clearly focused on problems relevant to engineering applications. This requires for example an efficient implementation, an inclusion of RANS and DES models as well as the usability of any two- and three-dimensional grids. With the last generation, comparatively simple problems such as
- Profile flow NACA0012 laminar and transonic
- Wing M6
- Cavity transient flow
- Flow around a sphere with separation
can be processed successfully. After a currently ongoing restructuring, the efficiency will be significantly improved and will also allow the simulation of a rotor flow through the installation of grid movements.
The main sources of noise at the helicopter are the main rotor and the tail rotor, at least in the outside area (in the cabin, the gear noise plays a certain role). Particularly in the encased version (Fenestron), the tail rotor is usually of secondary importance, but operates in a significantly higher and thus more unpleasant frequency range.
At the main rotor, noise is generated by the turbulent flow (broadband noise), similar to most other bodies surrounded by flows, and in particular by highly unsteady surface pressure fluctuations. Apart from the "normal" transient pressure changes caused by changing inflow conditions and angle of attack at each revolution, fluctuations occur mainly in two flight situations, namely in fast forward flight and in slow inclination flight (landing approach).
HSI - High Speed Impulsive Noise
In high-speed flight, circumferential and forward speeds overlap strongly at the leading blade and result in local supersonic flow, which is terminated by a shock wave. As this superimposition effect changes continuously during the rotation, these shock waves are created from about 50° azimuth and reduced again beyond 140° - combined with strong pressure fluctuations at the blade surface. The temporal correlation of these shock wave phenomena over the blade length means that this form of pressure fluctuation is essentially radiated forward as sound. This effect is particularly strong at highly loaded rotors with relatively few blades, symptomatical for example for the Bell UH-1 with two-blade rotor.
BVI - Blade Vortex Interaction Noise
In general, edge vortices develop at the tip of the blade, similar to the rigid wing aircraft, due to the generation of lift. Due to the induced velocity through the rotor disc, these are normally transported downwards. In forward flight they are transported spirally backwards.
If the sink rate is comparable to the induced speed, the blade tip vortex and the rotor blades lie in the same plane, so that the following blades can hit the vortices of the preceding ones. At certain azimuth angles (in the range of 45° and 315°) the vortices are then "cut open" across the blade, which leads to strong pressure fluctuations at the blade due to the vertical velocity induced by the vortex (upwind/downwind) shortly before or after the vortex core. This is associated with a corresponding increase and decrease in the effective angle of attack and thus in the lift, which is closely linked to the pressure distribution on the blade.
A more detailed analysis of the temporal sequence over the length of the blade shows that these vortex-induced pressure fluctuations are mainly radiated obliquely towards the front and bottom, i.e. in the direction of the landing point.