Studies On Dynamic Characteristics And Dynamic Correction Methods For Wind Tunnel Strain Gauge Balance

With the development of national defense and requirement of energy saving, more and more dynamic tests are required in wind tunnel, and better dynamic characteristic of the balance is also increasingly required in conventional force measurement test. Currently, the strain gauge balance is the most widely used balance in wind tunnel test. However, due to its integrated elastic body design, smaller damp ratio and low natural frequency, the strain gauge balance possesses low dynamic response speed, high overshoot and cross-axis dynamic coupling interference, which lead that the balance cannot meet the requirement of dynamic test. To solve the problem, the dynamic characteristic of the strain gauge balance and its on/off-line dynamic correction methods are studied in this paper, including dynamic calibration experiment, dynamic modeling and dynamic characteristic analyzing, time-domain dynamic compensation, time-domain dynamic decoupling-compensation, frequency-domain dynamic correction, frequency-domain dynamic decoupling-correction, and the development of a real-time dynamic correction system for wind tunnel strain gauge balance, so as to improve the dynamic characteristic of the balance and enhance its dynamic testing capabilities. The specific studies are as follows:(1) Adopt the combination of "load header-balance-support rod" to simulate the actual combination of "model-balance-support rod" applied in the wind tunnel test to study the balance dynamic characteristic, and negative-step unit-load experimental method is adopted for the balance dynamic calibration to obtain the dynamic calibration experimental data and evaluate the time-domain dynamic performance indices of the balance.(2) Introduce output coupling model of multi axis sensor to describe the input/output transfer relation of the wind tunnel strain gauge balance, and adopt the system identification method of output-error model parameter estimation (OE identification method) to identify the balance models, so as to evaluate the measurement bandwidth of the balance and analyze the main restricting factor of the balance dynamic characteristic.(3) Adopt the OE identification method to design the balance dynamic compensator, and develop the real-time dynamic correction system for the wind tunnel strain gauge balance based on digital signal processor (DSP), so as to realize real-time dynamic compensation for the balance to reduce the dynamic error of the balance main channel and accelerate the balance dynamic response speed.(4) Put forward a serial iterative dynamic decoupling-compensation (SIDDC) method for reducing the balance cross-axis dynamic coupling interference and main channel dynamic error synchronously, and ensuring that the static characteristic of the balance remains unchanged. For the method, iterative dynamic decoupling of the balance output is first conducted, then the dynamic compensation is carried out, and last the static relation between the balance input and the dynamic decoupling-compensation result is reconstructed. Especially, Jacobi, G-S and SOR iterative methods are introduced for the dynamic decoupling, and the frequency-dependent relaxation factor based SOR iterative method is proposed to improve the convergence characteristic of the iterative dynamic decoupling.(5) Study the frequency-domain correction (FDC) method for the strain gauge balance to improve the dynamic error correction accuracy by an off-line method. Specifically, the data splicing preprocessing and frequency response function (FRF) interpolation based FRF calculation method, and high/low frequency signal decomposition and windowing approach based FRF calculation method are put forward to overcome the problems of data cycle extension edge and invalid correction function values when calculating the sensor FRF according to its step response calibration data; and then a high/low frequency signal decomposition based FDC method is put forward to overcome the problems of boundary aliasing error and Gibbs phenomenon when conducting FDC for the sensor output signal.(6) Study the frequency-domain decoupling-correction (FDDC) method for the strain gauge balance to improve the decoupling accuracy of the cross-axis interference and the correction accuracy of the main channel dynamic error by an off-line method. Specifically, frequency-domain parameter matrix inversion method is proposed to resolve the key problem of FDDC first; and then, high/low frequency signal decomposition and static reconstruction based FDDC method is put forward to overcome the boundary aliasing error and Gibbs phenomenon when conducting FDDC for the sensor output signal, and ensure that the static characteristic of the sensor remains unchanged.Specifically, a six-component O10bar-shape wind tunnel strain gauge balance is adopted as an example to validate the above studies, and the results show that:due to the balance poor dynamic characteristic, the step response adjust time of the balance main channel is long ranging from a few hundred milliseconds to a few seconds, the maximum overshoot is beyond100%, the dynamic coupling error is also high up to100%, and the entire measurement bandwidth is lower than30Hz; after time-domain SIDDC, the adjust time is shortened to less than15ms (improvement exceeds94%), the overshoot is reduced to less than5%, the dynamic coupling error of the force components is reduced to less than2.28%(improvement exceeds80%), and the dynamic compensation can broaden the measurement bandwidth of the balance to2-3times of the original bandwidth for all the main channels except for Fy-channel; after FDDC, the adjust time is almost shortened to less than3ms (improvement exceeds97%), the overshoot is reduced to less than5%, the dynamic coupling error is reduced to less than2%(improvement exceeds70%for full components and almost90%for force components), and the FDC can broaden the entire measurement bandwidth of the balance to about120Hz.Accordingly, the SIDDC method and FDDC method studied in this paper all can significantly improve the dynamic performance indices of the strain gauge balance, and the dynamic correction accuracy of the FDDC method is obviously better than that of the SIDDC method. This is significant for the off-line analysis of the strain gauge balance output signal which doesn’t require real-time in normal wind tunnel test. Consequently, the dynamic correction methods for the wind tunnel strain gauge balance studied in this paper can significantly improve the dynamic characteristic of the balance and enhance its dynamic testing capabilities, so that it provides the foundation for the strain gauge balance applying in dynamic wind tunnel test.