如同真实的驾驶: 在风洞测试中仿真赛车的动态性能
大多数赛车的风洞测试中模型都是固定不变的. 通过一个创新的操纵器，仿真赛车的实际路况进行空气动力学测试成为可能.
在空气动力学优化过程中，赛车模型的测试必须能捕捉到瞬间的空气动力学影响.
新车辆的空气动力学性能的改善通常在风洞中进行. 缩小的模型固定在风洞的空气动力学装备上. 传感器安装在模型内来获取力，力矩以及在模型表面的压力等.
空气动力学采用测试和数字化的特性曲线来改进设计. 在风洞测试中，一般只测量动态特性的平均值如，上/下压力和拉力等。但是在 瑞士 Emmen RUAG 航天空气动力学中心的测试中显示，仅使用空气动力学平均值是不完全正确的.
Fig. 1: LeMans 汽车模型 (30 %) 在汽车风洞中。 At no time during this test do the wheels touch the model; they are held in the correct place by special wheel arms at the side.
波动通常是自然风的扰动或者是由于风的原因产生的测量物体的运动引起的，由此而产生瞬间值和平均值的巨大变化. 这些瞬间值对于赛车的设计改良有很大的帮助，如保持前后轮力的分布在安全的范围内，以及获得可靠的操作特性等.
Just like running the car over the real track
To realistically simulate vehicle movements in the wind tunnel, vertical movements of several millimeters in amplitude are induced to the wheel axles, at frequencies up to more than 20 Hz. A hydraulic "shaker" with a control unit has been developed to produce such motions on the wind tunnel model. In the model, the shaker is located beneath the weighing sensor – the balance – that in turn is fixed at the vertically aligned model support protruding through the roof of the model. Separation of aerodynamic forces from the often considerably greater inertial forces caused by the above described movements poses a particular challenge in this measurement task, as the balance itself can only account for the sum of these forces and moments. To accomplish this partitioning with sufficient precision, the recording of about 100 measurement channels must be perfectly synchronized and analyzed. High-quality measurement data at sampling rates from 400 Hz upwards are the basis for this task.
A balance determines all the forces and moments acting on the model
The six aerodynamic forces and moments (drag, side force, lift/downforce, rolling moment, pitching moment and yawing moment) that act on the test object during the wind tunnel test are determined with a special six-component balance integrated into the shaker. Providing excellent quality of signal conditioning and amplifier, the balance meets stringent demands for accuracy at minimal dimensions and extreme stiffness.
Fig. 2: The "shaker" model movement unit built into the vehicle model, with model mounting
Strain gages are applied as full Wheatstone bridges on the balance measuring beams. In order to cope with temperature variations the wiring is thermally compensated. The measuring beams under the action of the load only minimally deform in the elastic range and produce output signals in the strain gages proportional to the applied load.
The correlation between these electrical signals and the actual load is determined by a balance calibration procedure carried out subsequent to fabrication. This procedure takes place with calibrated mass weights and a calibration device connected to the balance offering various pivot options.
Fig. 3: A selection of multi-component balances, as used in wind tunnel testing. The smallest balance weighs 0.25 kg, the largest 112 kg.
By attaching the weights to these pivot points on the model side of the balance, pure forces or a combination of a force and up to two moments can be applied on the balance.
The calibration matrix – defining the relationship between load values and electrical strain gage values – is derived by a regression algorithm that takes into account all the well defined load cases which that were applied during the calibration process. The inverted matrix – defining the relationship between electrical values and load values – allows the calculation of the applied forces and moments from the electrical measurement signals as is needed during measurement.
For some years, Ruag Aerospace has relied for data acquisition in its wind tunnels on HBM’s MGCplus technology. By choosing different amplifier types and combining the amplifiers with the proper connection boards, it is possible to meet the demands for accuracy (divided into accuracy classes) as well as for a wide variety of possible sensor types. For each wind tunnel at the Center Aerodynamics, the hardware components have been integrated in a flexible and easily used complete system in a cabinet with various monitoring devices.
Fig. 4: Data acquisition system with MGCplus technology: Front (on the left) and back with connection boards (on the right)
For practical reasons, it was decided to use standard modular enclosure systems on rollers with five MGCplus housings, with the amplifier displays and control buttons at the front and standard connection boards with sockets for the power, measurement signal and data lines at the back.
Drivers for communication between the hardware and the company’s own master computer and data analysis software also had to be created and adapted to the specific measurement tasks. Finally, a special calibration system/procedure was developed for the measurement system, for flexible – and for the expensive wind tunnel test environment most important – quick on-site use that allowed to trace back all the different system components to national standards.
Motion greatly influences aerodynamics
In first wind tunnel test campaigns with the moving model, motion trajectories were initiated at frequencies up to 10 Hz, for 20 second periods, in order to search for transient aerodynamic phenomena. The movements took the front and rear axles of the vehicle model up to 4 mm out of the standard position, the front and rear axle being either fully in phase, 180° out of phase or the rear axle being kept still.
A sampling rate of 400 Hz was specified for these tests by weighing up the timing resolution and the amount of data this would produce. Once data is available from a wind tunnel test with time-resolved measurement, there are many possible ways to evaluate it.
In the initial phase, non-linear filtering is applied, in order to reduce the noise component of the individual measurements. Then the inertial forces are calculated and subtracted from the total forces, for which signals from the acceleration transducers at selected model positions are consulted. With the calculated result a number of relevant stability problems caused by the interplay of aerodynamics with the chassis dynamics can be analyzed in detail and used to characterize the drivability or even the safety of the vehicle configuration at hand. This method of measurement allows the optimization of the vehicle not just to the lab-like conditions in the wind tunnel, but also, by simulating realistic vehicle movements, to real situations that occur on the track.
Fig. 5: Comparison of static and time-resolved measurements for a simplified a LeMans race car
Figure 5 shows the time varying downforce coefficient (negative lift) of a simplified model of a LeMans racing car moved at 10 Hz (Fig. 1) compared to static measurements. The green curve describes the movement when, in this case, the front and rear axles move up and down in phase at an amplitude of 4 mm.
As is usual for race cars, there was very little ground clearance for this test, so that the nose of the model approached to within about 2 mm of the ground. The red curve describes the downforce measurements with a fixed model. To derive this curve, the model was taken to every one of the given positions, was fixed in position and the average downforce level was established; in the configuration shown, there proves to be only a slight dependency between the downforce coefficient and the height above ground.
The situation is completely different when measurement and analysis are time-resolved and relate to the moving model (blue curve). The enormous differences from the mean value indicate that the movement causes vast fluctuations in aerodynamic force and moment, which under certain circumstances can greatly influence the stability of the vehicle.
Claus Zimmermann is the measurement technology team leader, Peter Aschwanden is the “shaker” project manager and Werner Häberli is the supervisor of the electronics lab at the Center Aerodynamics at Ruag Aerospace.
