Data acquisition and processing method for satellite micro-vibration ground test

Abstract

The invention belongs to the technical field of vibration precision test, and discloses a data acquisition and processing method for satellite micro-vibration ground test. Comprising five core steps of original vibration signal acquisition, adaptive oversampling, analog-digital hybrid filtering, dynamic extraction recovery and micro-vibration transfer characteristic extraction. Through an analog-to-digital hybrid filtering architecture and in combination with an intelligent oversampling-extraction noise reduction technology, micro-vibration signal measurement with a high signal-to-noise ratio is realized. The method has the main technical advantages that various test items can be completed by using a single-specification sensor; a dynamic parameter adjustment mechanism can automatically optimize a processing strategy according to signal characteristics, and the test adaptability is remarkably improved.

Description

Technical Field

[0001] This invention relates to the field of precision vibration testing technology, and more specifically, to a data acquisition and processing method for ground testing of satellite micro-vibrations. Background Technology

[0002] During the operation of high-resolution Earth observation satellites, the imaging quality directly affects the success or failure of the mission. With the increasing demands on the resolution of satellite optical payloads, micro-vibration interference has become a key factor affecting imaging quality. On the satellite platform, components such as the flywheel assembly, solar array drive mechanism (SADA), refrigerator, and control moment gyroscope (CMG) generate micro-vibrations during operation, typically ranging from 1 to 100 μg. These minute vibrations are transmitted to the optical payloads through the satellite structure; even extremely weak vibrations can lead to a decrease in imaging quality and resolution, severely impacting the satellite's observation capabilities.

[0003] When a satellite is in orbit, the minute vibrations (as low as 1-100 μg) generated by components such as the flywheel assembly, solar array drive mechanism, and refrigerator directly affect the imaging quality of optical payloads. Therefore, the accuracy of micro-vibration testing is crucial. However, traditional testing methods require configuring sensors with various sensitivity specifications (such as 500 mV / g, 1000 mV / g, and 10000 mV / g) for different frequency ranges and vibration magnitudes. This not only significantly increases the cost of testing equipment but also leads to frequent sensor replacements and recalibration during the testing process. In actual testing environments, technicians often need to install hundreds of sensors on a single satellite, requiring reconfiguration, attachment, and calibration every time the testing item changes, which greatly extends the testing cycle. Furthermore, existing technologies exhibit a significant contradiction in measurement accuracy: while high-sensitivity sensors can capture weak vibrations, they often produce saturation distortion due to excessive amplitude during satellite functional testing (such as attitude control testing); while low-sensitivity sensors have a wide range of applications, their insufficient signal-to-noise ratio leads to poor reliability of measurement results when measuring the micro-vibrations of critical optical components. Furthermore, traditional purely analog filtering schemes are limited by hardware implementation complexity, making it difficult to achieve high-order and diverse filtering characteristics, and unable to effectively cope with complex vibration signals and environmental noise interference. In actual satellite integration testing, these problems lead to insufficient reliability of test data, extended testing cycles, and increased testing costs, severely restricting satellite development efficiency and the accuracy of performance verification.

[0004] In view of this, the present invention proposes a data acquisition and processing method for ground testing of satellite micro-vibration to solve the above problems. Summary of the Invention

[0005] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: a data acquisition and processing method for ground testing of satellite micro-vibrations, comprising:

[0006] The original vibration signal of each measuring point in the satellite micro-vibration test is acquired, and the original vibration signal is collected by an accelerometer installed on the satellite structure;

[0007] Based on the frequency distribution characteristics of the original vibration signal and the target analysis bandwidth of the test item, an adaptive oversampling factor is determined, and the original vibration signal is oversampled based on the adaptive oversampling factor to obtain an oversampled vibration signal.

[0008] Based on the signal-to-noise ratio and target analysis bandwidth of the oversampled vibration signal, a multi-level analog-to-digital hybrid filter is constructed to filter the oversampled vibration signal and obtain a filtered vibration signal.

[0009] Based on the oversampling factor of the filtered vibration signal and the target analysis bandwidth, a dynamic extraction and recovery strategy is designed to restore the filtered vibration signal to the target sampling frequency, thereby obtaining the recovered vibration signal.

[0010] Based on the time-domain and frequency-domain characteristics of the recovered vibration signal, the micro-vibration transmission characteristic parameters of the satellite structure are extracted, and the vibration suppression performance of the satellite structure is evaluated based on the micro-vibration transmission characteristic parameters.

[0011] The technical effects and advantages of the data acquisition and processing method for satellite micro-vibration ground testing of this invention are as follows:

[0012] This invention improves the signal-to-noise ratio of the measurement system through a hybrid analog-digital filtering architecture and intelligent oversampling-decimation noise reduction technology. This allows a single-sensitivity accelerometer (1000mV / g) to cover a wide measurement range from microgravity to hundreds of milligravity levels, reducing equipment costs. The combination of adaptive oversampling technology and a multi-level filtering strategy effectively suppresses broadband noise introduced by the test system, lowering the measurement lower limit and enabling the accurate capture and quantification of minute vibration characteristics of the satellite structure. The dynamic decimation recovery strategy, through an intelligent noise reduction algorithm, reduces noise levels while preserving signal integrity, significantly improving the reliability of low-frequency vibration analysis. Furthermore, the closed-loop optimization mechanism of this method dynamically adjusts processing parameters according to satellite structural characteristics and test targets, significantly enhancing adaptability and improving test efficiency. Attached Figure Description

[0013] Figure 1 This is a schematic diagram of a data acquisition and processing method for ground testing of satellite micro-vibration according to the present invention. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] This application provides a data acquisition and processing method for ground testing of satellite micro-vibrations. The execution entities of the method include, but are not limited to: vibration testing equipment, satellite test control system, data acquisition unit, edge computing processor, etc., which can be regarded as general computing nodes of this application. The satellite test control system includes, but is not limited to: test controller, distributed monitoring system, programmable signal processor, and at least one of the following:

[0016] This invention provides a data acquisition and processing method for ground testing of satellite micro-vibrations. By acquiring raw vibration signals from the satellite structure in real time, and based on adaptive oversampling technology and a hybrid analog-digital filtering architecture, combined with a dynamic extraction and recovery strategy, it solves core challenges in micro-vibration signal measurement such as noise suppression and effective signal extraction, achieving high-precision measurement of satellite micro-vibration characteristics. It features a significantly improved signal-to-noise ratio and greatly enhanced testing accuracy, enabling high-precision micro-vibration testing using relatively low-sensitivity sensors, reducing testing costs, and improving testing efficiency.

[0017] In this embodiment of the invention, the detailed implementation steps of a data acquisition and processing method for satellite micro-vibration ground testing include:

[0018] First, the raw vibration signals at each measuring point in the satellite micro-vibration test are acquired using accelerometers mounted on the satellite structure. In actual testing, depending on the testing objective, accelerometers need to be attached to key locations on the satellite structure, such as the location of disturbance sources, sensitive loads, and along the transmission path from the disturbance source to the sensitive load. Preferably, a triaxial accelerometer with a sensitivity of 1000 mV / g and a test frequency range of 0.1 Hz to 1000 Hz is used, with a sensor mass less than 3g to reduce the mass loading effect. The sensor is connected to the input channel of the testing equipment via a high-quality signal cable, and the raw vibration signals directly reflect the vibration characteristics of each measuring point on the satellite structure.

[0019] Based on the frequency distribution characteristics of the original vibration signal and the target analysis bandwidth of the test item, an adaptive oversampling factor is determined, and the original vibration signal is oversampled based on the adaptive oversampling factor to obtain an oversampled vibration signal. Traditional measurement methods typically only satisfy the Nyquist sampling theorem, with the sampling frequency being 2.56 times the analysis frequency. This method, however, significantly improves the signal-to-noise ratio through intelligent oversampling technology. In this implementation scheme, as a preferred option, the adaptive oversampling factor can be selected from 1, 2, 4, 8, 16, 32, 64, 128, etc., with 64 times oversampling being the preferred choice. For example, when the target analysis bandwidth is 1000Hz, the actual sampling frequency is calculated as 1000Hz × 2.56 × 64 = 163840Hz. Sufficient data redundancy is obtained through high-magnification oversampling, laying the foundation for subsequent noise suppression.

[0020] Based on the signal-to-noise ratio and target analysis bandwidth of the oversampled vibration signal, a multi-stage analog-to-digital hybrid filter is constructed to filter the oversampled vibration signal, resulting in a filtered vibration signal. The multi-stage analog-to-digital hybrid filter is one of the core innovations of this invention, combining the advantages of analog and digital filtering and overcoming the limitations of traditional single-filtering methods. Preferably, this method uses an analog high-pass filter with a cutoff frequency of 0.01Hz to remove the DC component, and then uses a digital filter to achieve precise frequency domain selection. For high-resolution satellite micro-vibration testing, the digital part of the analog-to-digital hybrid filter is designed as a low-pass filter with an attenuation of -160dB and a cutoff frequency set to 1280Hz, effectively suppressing high-frequency noise without losing the effective signal.

[0021] Based on the oversampling factor of the filtered vibration signal and the target analysis bandwidth, a dynamic decimation recovery strategy is designed to restore the filtered vibration signal to the target sampling frequency, thus obtaining the recovered vibration signal. The dynamic decimation recovery strategy is a key step in restoring high-sampling-rate data to the standard analysis frequency, achieving a significant improvement in signal quality through intelligent data segmentation and noise averaging. In this implementation scheme, the filtered vibration signal is segmented according to the oversampling factor, and the mean of each segment is calculated to obtain the recovered vibration signal. This process effectively utilizes the random distribution characteristics of noise, suppressing random noise through averaging while preserving the deterministic true vibration signal.

[0022] Based on the time-domain and frequency-domain characteristics of the recovered vibration signal, micro-vibration transmission characteristic parameters of the satellite structure are extracted, and the vibration suppression performance of the satellite structure is evaluated based on these parameters. This step transforms the processed high-quality signal into valuable engineering analysis results, providing crucial information for satellite structure optimization. Depending on the testing objective, FFT analysis, self-power spectral density analysis, cross-power spectral density analysis, and frequency response function calculations are performed on the recovered vibration signal to extract key frequency, amplitude, phase, and other characteristic parameters. A vibration transmission model of the satellite structure is then constructed to evaluate its vibration suppression performance, providing data support for satellite structure optimization and payload performance verification.

[0023] In this embodiment of the invention, the detailed implementation steps for determining the adaptive oversampling factor based on the frequency distribution characteristics of the original vibration signal and the target analysis bandwidth of the test item include:

[0024] A short-time Fourier transform (SFT) is performed on the original vibration signal to obtain its frequency distribution spectrum within different time windows. The SFT is a commonly used time-frequency analysis method that reflects the changes in the signal's spectral characteristics over different time periods. In this implementation, the original vibration signal is windowed using a Hanning window function, with the window length set to 1 / 8 to 1 / 16 of the signal length and a window overlap rate of 50% to obtain sufficient time-frequency resolution. This step provides a comprehensive understanding of the vibration signal's frequency distribution characteristics, laying the data foundation for determining the subsequent oversampling factor.

[0025] Based on the frequency distribution spectrum, the range of frequency components in the original vibration signal whose energy percentage exceeds a preset energy threshold is defined as the effective frequency range. The effective frequency range directly reflects the frequency intervals carrying the main information in the signal and is a key reference for determining the oversampling factor. The statistical process first normalizes the frequency distribution spectrum, calculates the proportion of energy of each frequency component to the total energy, and then accumulates the energy percentages sequentially from low to high frequencies until the preset energy threshold is reached. Preferably, the preset energy threshold is set to 95%, meaning the effective frequency range contains 95% of the signal energy, ensuring that almost all meaningful vibration information is captured.

[0026] The initial oversampling factor is calculated based on the ratio of the effective frequency range to the target analysis bandwidth. The initial oversampling factor is a rough estimate that reflects the basic matching relationship between signal characteristics and analysis requirements. The calculation formula is:

[0027] Initial oversampling factor = target analysis bandwidth / upper limit of effective frequency range × 2;

[0028] This formula ensures that the oversampling frequency is at least twice the effective frequency range of the signal, meeting basic sampling requirements. When the effective frequency range is close to or exceeds the target analysis bandwidth, the initial oversampling factor will be a smaller value; otherwise, a larger value will be used, achieving a reasonable allocation of oversampling resources.

[0029] Based on the signal-to-noise ratio (SNR) of the original vibration signal, the SNR compensation coefficient is obtained. The calculation method for the SNR compensation coefficient is as follows:

[0030] If the signal-to-noise ratio is less than the first signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is the first preset value;

[0031] If the signal-to-noise ratio is greater than or equal to the first signal-to-noise ratio threshold and less than the second signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is the second preset value;

[0032] If the signal-to-noise ratio is greater than or equal to the second signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is the third preset value;

[0033] Among them, the first preset value is greater than the second preset value, and the second preset value is greater than the third preset value.

[0034] The signal-to-noise ratio (SNR) compensation factor is a crucial factor in adjusting the oversampling factor, and differentiated compensation strategies are provided for different SNR conditions. Preferably, the first SNR threshold is set to 20dB, the second SNR threshold is set to 40dB, the first preset value is 2.0, the second preset value is 1.5, and the third preset value is 1.0. This segmented compensation strategy ensures that a higher oversampling factor is used for noise suppression when the SNR is low, while a smaller oversampling factor is used when the SNR is already high to save computational resources, achieving efficient utilization of system resources.

[0035] The product of the initial oversampling factor and the signal-to-noise ratio compensation coefficient is rounded up to obtain the adaptive oversampling factor. This step comprehensively considers signal characteristics and signal-to-noise ratio factors to generate the final oversampling factor decision. The rounding operation ensures that the oversampling factor at least meets the computational requirements. Furthermore, for ease of hardware implementation, the adaptive oversampling factor is usually taken as an integer power of 2, such as 1, 2, 4, 8, 16, 32, 64, 128, etc. This adaptive oversampling strategy can dynamically adjust the oversampling parameters according to signal characteristics, optimizing system resource utilization while ensuring measurement accuracy.

[0036] In this embodiment of the invention, the detailed implementation steps for filtering the oversampled vibration signal by constructing a multi-level analog-to-digital hybrid filter based on the signal-to-noise ratio and target analysis bandwidth of the oversampled vibration signal include:

[0037] An analog high-pass filter is designed, with its cutoff frequency dynamically adjusted based on the lowest effective frequency of the oversampled vibration signal. This filter removes the DC component from the oversampled vibration signal, yielding a filtered analog signal. The analog high-pass filter is the first stage of a multi-stage filtering architecture, primarily addressing the DC bias problem in the sensor output. Unlike traditional designs, this invention's analog high-pass filter does not employ multiple fixed cutoff frequencies but instead uses a single low cutoff frequency, delegating the main filtering task to subsequent digital filtering sections. Preferably, the analog high-pass filter uses an active RC filter structure with a second-order filter and a cutoff frequency preferably set to 0.01Hz, sufficient to filter out the DC component in the sensor output signal and ensure correct signal acquisition by subsequent circuitry.

[0038] The analog filtered signal undergoes range adaptation conditioning to ensure it matches the ADC's measurement range. For the selected ADC, the measurement reference voltage range is 0-5V. Different gain coefficients need to be set according to the different ranges of the test channel (0.1V, 1V, 10V). The traditional formula for calculating the range adaptation gain factor is: , This is the ADC reference voltage. For channel range.

[0039] In this invention, considering the effects of measurement error and DC bias, the actual range is amplified by 8% to ensure usability of the nominal range within a 5% error range. Therefore, the formula for calculating the corrected range adaptation gain is as follows:

[0040] ;

[0041] Based on this formula, the gain is 46.3 for the 0.1V range, 4.63 for the 1V range, and 0.463 for the 10V range. These gain coefficients are achieved by switching different operational amplifier modules controlled by the FPGA, and subsequent measurement results are used to introduce compensation coefficients for precise calibration.

[0042] A digital low-pass filter is designed, with its cutoff frequency determined by the product of the target analysis bandwidth and the adaptive oversampling factor. This digital low-pass filter removes high-frequency noise components from the analog filtered signal, yielding the digital filtered signal. It is the second stage in a multi-stage filtering architecture, responsible for handling high-frequency noise. Preferably, the cutoff frequency is set to 1.28 times the target analysis bandwidth; for example, when the target analysis bandwidth is 1000Hz, the cutoff frequency is set to 1280Hz. To achieve a high attenuation performance of -160dB, the filter employs an FIR structure, selecting different filter configurations depending on the oversampling factor. For the preferred 64x oversampling scheme, two cascaded 8x decimation filters are used, achieving efficient high-frequency noise suppression.

[0043] Design an adaptive bandpass filter. The passband range of the adaptive bandpass filter is dynamically adjusted according to the signal-to-noise ratio of the oversampled vibration signal. The adaptive bandpass filter is used to perform secondary filtering on the digital filtered signal to obtain the filtered vibration signal.

[0044] The method for adjusting the passband range of the adaptive bandpass filter is as follows:

[0045] Calculate the signal-to-noise ratio (SNR) of the digitally filtered signal and record it as the current SNR;

[0046] If the current signal-to-noise ratio is less than the preset signal-to-noise ratio threshold, the passband range is reduced, and the reduction is proportional to the reciprocal of the current signal-to-noise ratio.

[0047] If the current signal-to-noise ratio is greater than or equal to the preset signal-to-noise ratio threshold, the passband range remains unchanged.

[0048] The adaptive bandpass filter, the third stage of a multi-stage filtering architecture, provides fine-grained frequency domain selection capabilities, enabling secondary signal optimization. Preferably, the adaptive bandpass filter employs the Butterworth digital filter dynamic design method, which automatically adjusts the passband range based on signal characteristics. A preset signal-to-noise ratio (SNR) threshold is set to 30 dB. When the SNR falls below this threshold, the passband range is automatically reduced. The reduction is calculated using the following formula:

[0049] Passband range reduction ratio = 0.5 × (30 / current signal-to-noise ratio);

[0050] This dynamic adjustment strategy sacrifices some bandwidth for a higher signal-to-noise ratio when signal quality is poor, while maintaining full-band analysis capability when signal quality is good, achieving an optimal balance between measurement accuracy and bandwidth. The order of the adaptive bandpass filter is dynamically calculated, typically between 8th and 16th order, ensuring a stopband attenuation performance of -40dB. The filtered signal quality is significantly improved, providing a high-quality data foundation for subsequent analysis.

[0051] In this embodiment of the invention, a dynamic decimation and recovery strategy is designed based on the oversampling factor of the filtered vibration signal and the target analysis bandwidth. The detailed implementation steps for restoring the filtered vibration signal to the target sampling frequency include:

[0052] Based on the adaptive oversampling factor, the filtered vibration signal is divided into multiple data segments according to the time series, with each data segment containing a number of sampling points equal to the adaptive oversampling factor. Data segmentation is the first step in the decimation recovery strategy, dividing the high-sampling-rate continuous signal sequence into smaller, more easily processed segments. Taking a preferred 64x oversampling factor as an example, every 64 consecutive sampling points constitute a data segment, corresponding to the same moment at the original sampling rate. The divided data segments preserve the temporal relationship of the signal while providing operational units for subsequent noise reduction processing.

[0053] A weighted average is applied to the sampling points within each data segment. The weights are determined based on the positional distribution of the sampling points within the segment, following a Gaussian distribution, to obtain the initial extracted signal. Weighted averaging is an advanced noise reduction technique that more effectively preserves signal characteristics compared to simple arithmetic averaging. Preferably, a Gaussian weighting function is used. This weighting allocation method assigns higher weights to sampling points closer to the center and lower weights to edge points, effectively suppressing high-frequency noise and aliasing effects that may be introduced during oversampling, thus improving the quality of the extracted signal.

[0054] The initially extracted signal undergoes signal recovery filtering, employing a cascaded finite impulse response (FIR) filter. The order of the FIR filter is dynamically adjusted based on the logarithm of the adaptive oversampling factor to obtain the recovered vibration signal. Signal recovery filtering is the final step in the extraction process, ensuring that the extracted signal retains its smooth and continuous characteristics. Preferably, the signal recovery filter uses an FIR structure, and the filter order is calculated using the following formula:

[0055] Filter order = 2 × log2 (adaptive oversampling factor);

[0056] For example, for a 64x oversampling, the filter order is 2×log2(64)=12. This design ensures that the filter performance matches the oversampling factor, achieving optimal signal recovery. The filter design employs a MATLAB pre-design method, pre-designing a series of filters for different oversampling factors. In practical applications, these filters are selected and cascaded as needed, significantly improving flexibility and efficiency. Through this series of processes, the high-sampling-rate filtered vibration signal is accurately restored to a high-quality signal at the standard analysis frequency, achieving the goal of both noise suppression and preservation of effective information.

[0057] In this embodiment of the invention, the detailed implementation steps for extracting the micro-vibration transmission characteristic parameters of a satellite structure based on the time-domain and frequency-domain characteristics of the recovered vibration signal include:

[0058] Time-domain analysis was performed on the recovered vibration signal to obtain its peak acceleration, root-mean-square acceleration, and vibration duration. Time-domain analysis is the most direct method for feature extraction from vibration signals, quickly reflecting the intensity and duration of the vibration. Peak acceleration reflects the maximum intensity of the vibration, root-mean-square acceleration characterizes the average energy level of the vibration, and vibration duration describes the temporal characteristics of the vibration process. Preferably, peak acceleration is obtained by finding the absolute maximum value of the signal, root-mean-square acceleration is calculated by averaging the squared values ​​of the signal and then taking the square root, and vibration duration is determined by setting a threshold (e.g., 10% of the root-mean-square value) to determine the start and end times of the vibration. These time-domain parameters provide fundamental data for evaluating the vibration response characteristics of satellite structures.

[0059] Frequency domain analysis was performed on the recovered vibration signal to obtain its auto-power spectral density function, cross-power spectral density function, and frequency response function. Frequency domain analysis is a key method for revealing the intrinsic characteristics of vibration, enabling in-depth understanding of the frequency composition and energy distribution of vibration. The auto-power spectral density function reflects the frequency energy distribution of a single-point vibration signal, the cross-power spectral density function characterizes the correlation between vibration signals at different measurement points, and the frequency response function describes the input-output relationship. Preferably, the auto-power spectral density and cross-power spectral density were calculated using the Welch method, with the Hanning window selected as the window function. The window length was 1 / 8 of the data length, and the overlap rate was 50%. The frequency response function was calculated using the H1 estimator, fully considering the influence of measurement noise. These frequency domain parameters provide a powerful tool for in-depth analysis of the dynamic characteristics of satellite structures.

[0060] Based on the autopower spectral density function, the principal frequency and its corresponding amplitude of the recovered vibration signal are extracted and denoted as principal vibration characteristic parameters. These principal vibration characteristic parameters are key indicators describing the main characteristics of the vibration signal, directly reflecting the inherent characteristics and vibration response of the satellite structure. The extraction process first identifies significant peaks in the autopower spectral density function, and then determines the principal frequency and its amplitude based on the peak frequency and amplitude. Preferably, the criteria for determining significant peaks are that the peak height exceeds three times the average spectral value, and the duration bandwidth is not less than five frequency resolution units. The principal vibration characteristic parameters typically contain 3-5 main frequency components, covering the main vibration modes of the satellite structure, providing a core basis for vibration transmission characteristic analysis.

[0061] Based on the cross-power spectral density function and the frequency response function, the transfer function of the recovered vibration signal between different measurement points is calculated and denoted as the transfer characteristic parameter. The transfer characteristic parameter describes the propagation law of vibration in the satellite structure and is an important basis for evaluating vibration isolation and suppression performance. The calculation process is based on the input-output relationship, obtaining the transfer function between different measurement points through the ratio of cross-power spectral density to self-power spectral density, or directly using the frequency response function. Preferably, the transfer function calculation covers all critical paths from the disturbance source point to the sensitive load point, including both amplitude and phase components, with a frequency resolution of not less than 1 Hz, providing detailed data for a comprehensive understanding of the vibration transmission path.

[0062] By combining the principal vibration characteristic parameters with the transmission characteristic parameters, micro-vibration transmission characteristic parameters are obtained. These parameters provide a comprehensive characterization of the satellite structure's vibration properties, incorporating information from both signal characteristics and system transmission. The combination process uses the principal vibration characteristic parameters as node attributes and the transmission characteristic parameters as connection relationships to construct a complete vibration transmission network model. This model visually demonstrates the relationship between the vibration source, transmission path, and response point, revealing the vibration transmission mechanism of the satellite structure and providing a comprehensive basis for subsequent vibration suppression performance evaluation.

[0063] In this embodiment of the invention, the detailed implementation steps for evaluating the vibration suppression performance of a satellite structure based on micro-vibration transmission characteristic parameters include:

[0064] A vibration transmission path model is constructed, with transmission characteristic parameters as edge weights and principal vibration characteristic parameters as node attributes. This model is a network structure representation that intuitively demonstrates the propagation law of vibration within the satellite structure. Preferably, a directed weighted graph structure is used, where nodes represent the measurement points on the satellite structure, edges represent vibration transmission paths, and the weights of the edges are the transmission characteristic parameters of the corresponding paths. The principal vibration characteristic parameters, as attributes of the nodes, describe the vibration characteristics of each measurement point. The model construction employs graph theory algorithms to ensure the integrity and rationality of the model structure, while also considering the differences in transmission characteristics at different frequencies. Typically, separate transmission models are established for 3-5 key frequency points to comprehensively reflect the vibration transmission characteristics of the satellite structure.

[0065] Based on the vibration transmission path model, the vibration transmission attenuation rate from the disturbance source measuring point to the sensitive load measuring point is calculated. The calculation method for the vibration transmission attenuation rate is as follows:

[0066] By multiplying the transmission characteristic parameters along each path in the vibration transmission path model, the transmission gain of each path is obtained.

[0067] The minimum value of the transmission gain of all paths is denoted as the vibration transmission attenuation rate.

[0068] Vibration transmission attenuation rate is a core indicator for evaluating the vibration suppression performance of satellite structures, reflecting the degree of attenuation of disturbances during transmission. The calculation process first identifies all possible paths from the disturbance source measurement point to the sensitive load measurement point. Then, it multiplies the amplitude of the transmission characteristic parameter (transfer function magnitude) along each path to obtain the total transmission gain of the path. Since the goal of vibration suppression is to reduce the vibration energy transmitted to the sensitive load, the minimum value of the transmission gain of all paths is taken as the vibration transmission attenuation rate; the smaller this value, the better the vibration suppression effect. Preferably, a depth-first search algorithm is used for path identification to ensure that all possible transmission paths are found, and the differences in transmission characteristics at different frequencies are considered, with a focus on evaluating the dominant vibration frequency.

[0069] The vibration suppression performance of the satellite structure is evaluated based on the comparison between the vibration transmission attenuation rate and the preset attenuation threshold. The evaluation method is as follows:

[0070] If the vibration transmission attenuation rate is greater than the preset attenuation threshold, the vibration suppression performance of the satellite structure is determined to be unsatisfactory, and the path with the lowest transmission gain in the vibration transmission path model is marked as the optimization target path.

[0071] If the vibration transmission attenuation rate is less than or equal to the preset attenuation threshold, the vibration suppression performance of the satellite structure is deemed to meet the requirements.

[0072] Vibration suppression performance evaluation is the final judgment made on the satellite structure based on test data, and it directly relates to the optimization direction of the satellite design. The preset attenuation threshold is a standard value determined based on satellite mission requirements and payload sensitivity. As a preferred option, for high-resolution optical payloads, the attenuation threshold is usually set to 0.01 (-40 dB), meaning that vibration energy must attenuate by at least 100 times during transmission. When the vibration transmission attenuation rate is greater than the preset threshold, it indicates that the vibration suppression of the satellite structure is insufficient, requiring optimization design. In this case, the path with the lowest transmission gain is marked as the optimization target, providing a clear direction for subsequent structural improvements. When the attenuation rate is less than or equal to the threshold, it indicates that the vibration suppression performance meets the requirements, and the satellite structure design is reasonable. This evaluation method combines quantitative analysis and decision support, providing a scientific basis for the optimized design of the satellite structure.

[0073] In this embodiment of the invention, the method for dynamically adjusting the cutoff frequency of an analog high-pass filter includes:

[0074] The initial cutoff frequency is calculated based on the lowest effective frequency (RDF) of the oversampled vibration signal, and is set to 0.8 times the RDF. This initial cutoff frequency serves as the starting point for the design of the analog high-pass filter parameters, determined initially based on signal characteristics. The RDF is typically obtained through analysis of the signal spectrum to identify the lowest frequency component with significant energy. Preferably, the RDF is defined as the lowest frequency point whose energy exceeds 1% of the total signal energy. Setting the initial cutoff frequency to 0.8 times the RDF ensures that the filter does not truncate the low-frequency components of the effective signal, while filtering out unwanted extremely low-frequency components and DC offset as much as possible. This design philosophy balances signal integrity and filtering effectiveness, providing a reasonable starting point for subsequent dynamic adjustments.

[0075] The initial cutoff frequency is adjusted based on the proportion of the DC component in the oversampled vibration signal. The calculation method for the proportion of the DC component is as follows:

[0076] Perform a Fourier transform on the oversampled vibration signal to obtain its spectrum;

[0077] Calculate the ratio of the energy of the zero-frequency component to the total energy in the spectrum, and denot it as the DC component percentage;

[0078] The DC component ratio is a crucial indicator reflecting the severity of DC bias in a signal and directly influences the parameter selection of a high-pass filter. The calculation process first involves performing a Fast Fourier Transform (FFT) on the acquired raw signal to obtain a complete spectral representation. Then, the ratio of the zero-frequency point (DC component) energy to the total signal energy is calculated to obtain the DC component ratio. This indicator reflects the relative intensity of the DC component in the signal; a higher value indicates a more significant DC bias, requiring a stronger high-pass filter.

[0079] If the proportion of DC component is greater than the preset DC threshold, the initial cutoff frequency will be increased by a factor that is proportional to the square of the proportion of DC component.

[0080] If the proportion of the DC component is less than or equal to the preset DC threshold, the initial cutoff frequency remains unchanged.

[0081] This dynamic adjustment strategy optimizes filter parameters based on the actual characteristics of the signal to ensure optimal filtering performance. Preferably, the preset DC threshold is set to 5%. When the proportion of the DC component exceeds this value, the cutoff frequency needs to be increased to enhance the filtering effect. The formula for calculating the increase is:

[0082] Cutoff frequency up adjustment ratio = (DC component ratio / preset DC threshold)²;

[0083] For example, when the DC component accounts for 10% and the preset DC threshold is 5%, the adjustment ratio is (10% / 5%)² = 4, increasing the initial cutoff frequency by a factor of 4. This nonlinear adjustment strategy employs stronger correction measures for severe DC bias problems, while handling minor issues gently, ensuring an optimal balance between filtering performance and signal integrity. Through this dynamic adjustment mechanism, the analog high-pass filter can adapt to different test conditions and signal characteristics, providing optimal filtering performance.

[0084] In this embodiment of the invention, the method for dynamically adjusting the order of a finite impulse response filter includes:

[0085] The initial filter order is calculated based on the adaptive oversampling factor, and is twice the logarithm of the adaptive oversampling factor. The initial filter order is a crucial parameter in FIR filter design, directly impacting filter performance and computational complexity. Determining the initial order using a logarithmic relationship is a preferred approach to balance performance and efficiency, considering that as the oversampling factor increases, a more complex filter structure is needed to handle a wider frequency range. Preferably, when the adaptive oversampling factor is 64, the initial filter order is calculated as 2 × log2(64) = 12. This design ensures that the filter performance matches the oversampling factor, providing sufficient filtering effect without excessive computational burden.

[0086] Based on the signal-to-noise ratio of the initially extracted signal, adjust the initial filter order as follows:

[0087] If the signal-to-noise ratio of the initially extracted signal is less than the preset extraction signal-to-noise ratio threshold, the initial filter order will be increased, and the increase will be proportional to the reciprocal of the signal-to-noise ratio.

[0088] If the signal-to-noise ratio of the initially extracted signal is greater than or equal to the preset extraction signal-to-noise ratio threshold, then the initial filter order remains unchanged.

[0089] This dynamic adjustment strategy optimizes the filter structure based on signal quality, enhancing filtering capability when signal quality is poor. Preferably, the preset decimation signal-to-noise ratio (SNR) threshold is set to 25 dB. When the SNR falls below this value, the filter order needs to be increased to provide stronger filtering. The formula for calculating the increase is:

[0090] Filter order upscaling factor = Preset decimation SNR threshold / SNR

[0091] For example, when the signal-to-noise ratio is 20dB and the preset threshold is 25dB, the adjustment factor is 25 / 20 = 1.25, and the initial filter order will increase by 25%. This adaptive adjustment strategy provides stronger processing capabilities for low-quality signals while maintaining moderate processing for high-quality signals, ensuring filtering effectiveness while avoiding unnecessary waste of computational resources. Through this dynamic adjustment mechanism, the FIR filter can provide optimal signal recovery performance based on signal characteristics, providing a high-quality data foundation for subsequent analysis.

[0092] In this embodiment of the invention, the method further includes:

[0093] Based on the time-domain characteristics of the recovered vibration signal, a vibration signal anomaly detection strategy is designed to identify abnormal vibration events in the recovered vibration signal.

[0094] The implementation method of the vibration signal anomaly detection strategy is as follows:

[0095] Calculate the local root mean square acceleration of the recovered vibration signal. The local root mean square acceleration is a sequence of root mean square accelerations calculated in units of a preset time window.

[0096] Local root-mean-square acceleration (RMSE) is an important indicator reflecting the time-varying characteristics of vibration intensity and can effectively identify abnormal fluctuations in vibration signals. The calculation process employs a sliding window technique, with the window length typically set between 0.1 and 1 second, adjusted according to the dynamic characteristics of the test object. Preferably, a window overlap rate of 50% is used to ensure sufficiently high temporal resolution to capture rapidly changing vibration events. The RMSE sequence visually reflects the variation of vibration intensity over time, providing fundamental data for anomaly event detection.

[0097] A sliding window analysis was performed on the local root mean square acceleration sequence to obtain the standard deviation of the local root mean square acceleration within each sliding window, which was denoted as the local vibration fluctuation degree.

[0098] Local vibration variability is a key indicator for measuring vibration stability and can effectively identify abrupt changes in vibration modes. The analysis employs a second-order sliding window technique, further applying statistical analysis to the local root mean square (RMS) sequence. The window length is typically 10-20 local RMS data points. The standard deviation calculation reflects the degree of fluctuation in vibration intensity; high variability usually indicates an abnormal change in the system state, possibly related to transient events or mechanical anomalies in the satellite structure.

[0099] If the local vibration fluctuation is greater than the preset fluctuation threshold, it is determined that there is an abnormal vibration event in the recovered vibration signal within the corresponding time window, and the start time, duration and peak acceleration of the abnormal vibration event are recorded.

[0100] Identifying abnormal vibration events is the core step in vibration signal anomaly detection, directly impacting the accuracy and reliability of the detection. Preferably, the preset fluctuation threshold is set to three times the local vibration fluctuation under normal operating conditions. This value is determined through statistical analysis of historical data to ensure the sensitivity and specificity of the detection. When an abnormal event is detected, its key characteristic parameters are automatically recorded, including the start time (relative to the start of the test), duration (the time interval from when the fluctuation exceeds the threshold to when it falls below the threshold), and peak acceleration (the maximum acceleration value during the abnormal event). These parameters comprehensively describe the characteristics of the abnormal event, providing detailed information for subsequent analysis.

[0101] Information about abnormal vibration events is transmitted to the user's computer, and these events are displayed on the user's computer in the form of a timeline.

[0102] The visualization of anomaly information is a crucial aspect of the interaction between the testing system and users, directly impacting the interpretation and application of test results. Ideally, anomaly information is transmitted in real-time to the user's computer via gigabit Ethernet and displayed intuitively in a timeline format within a dedicated software interface. Anomalies are highlighted on the timeline; clicking on a marker reveals detailed parameter information, including event characteristics and waveforms. This intuitive visualization method enables testers to quickly identify and analyze anomalies, assess their potential impact on satellite performance, and provide targeted recommendations for satellite structure optimization.

[0103] This invention achieves high-precision measurement of satellite micro-vibration signals through adaptive oversampling technology, a hybrid analog-digital filtering architecture, and a dynamic decimation recovery strategy. Its innovative features enable dynamic adjustment of test parameters based on signal characteristics and testing requirements, significantly improving the measurement signal-to-noise ratio. It also enables accurate measurement of ultra-weak vibrations using a relatively low-sensitivity sensor (1000mV / g), greatly reducing testing costs and improving testing efficiency.

[0104] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0105] It should be noted that all formulas in this manual are calculated by removing dimensions and taking their numerical values. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0106] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. A data acquisition and processing method for ground testing of satellite micro-vibrations, characterized in that, include: The original vibration signal of each measuring point in the satellite micro-vibration test is acquired, and the original vibration signal is collected by an accelerometer installed on the satellite structure; Based on the frequency distribution characteristics of the original vibration signal and the target analysis bandwidth of the test item, an adaptive oversampling factor is determined, and the original vibration signal is oversampled based on the adaptive oversampling factor to obtain an oversampled vibration signal. Based on the signal-to-noise ratio and target analysis bandwidth of the oversampled vibration signal, a multi-level analog-to-digital hybrid filter is constructed to filter the oversampled vibration signal and obtain a filtered vibration signal. Based on the oversampling factor of the filtered vibration signal and the target analysis bandwidth, a dynamic extraction and recovery strategy is designed to restore the filtered vibration signal to the target sampling frequency, thereby obtaining the recovered vibration signal. Based on the time-domain and frequency-domain characteristics of the recovered vibration signal, the micro-vibration transmission characteristic parameters of the satellite structure are extracted, and the vibration suppression performance of the satellite structure is evaluated based on the micro-vibration transmission characteristic parameters.

2. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 1, characterized in that, The step of determining the adaptive oversampling factor based on the frequency distribution characteristics of the original vibration signal and the target analysis bandwidth of the test item includes: The original vibration signal is subjected to a short-time Fourier transform to obtain the frequency distribution spectrum of the original vibration signal in different time windows; Based on the frequency distribution spectrum, the range of frequency components in the original vibration signal whose energy percentage is greater than a preset energy threshold is statistically analyzed and recorded as the effective frequency range. Calculate the initial oversampling factor based on the ratio of the effective frequency range to the target analysis bandwidth; Based on the signal-to-noise ratio (SNR) of the original vibration signal, a SNR compensation coefficient is obtained. The calculation method for the SNR compensation coefficient is as follows: If the signal-to-noise ratio is less than the first signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is a first preset value; If the signal-to-noise ratio is greater than or equal to the first signal-to-noise ratio threshold and less than the second signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is the second preset value; If the signal-to-noise ratio is greater than or equal to the second signal-to-noise ratio threshold, then the signal-to-noise ratio compensation coefficient is a third preset value; Among them, the first preset value is greater than the second preset value, and the second preset value is greater than the third preset value; The adaptive oversampling factor is obtained by rounding up the product of the initial oversampling factor and the signal-to-noise ratio compensation coefficient.

3. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 1, characterized in that, The step of constructing a multi-level analog-to-digital hybrid filter based on the signal-to-noise ratio and target analysis bandwidth of the oversampled vibration signal to filter the oversampled vibration signal includes: Design an analog high-pass filter, wherein the cutoff frequency of the analog high-pass filter is dynamically adjusted according to the lowest effective frequency of the oversampled vibration signal, and the analog high-pass filter is used to filter out the DC component in the oversampled vibration signal to obtain an analog filtered signal; Design a digital low-pass filter, wherein the cutoff frequency of the digital low-pass filter is determined based on the product of the target analysis bandwidth and the adaptive oversampling factor, and the digital low-pass filter is used to filter out high-frequency noise components in the analog filtered signal to obtain a digital filtered signal; An adaptive bandpass filter is designed, wherein the passband range of the adaptive bandpass filter is dynamically adjusted according to the signal-to-noise ratio of the oversampled vibration signal, and the adaptive bandpass filter is used to perform secondary filtering on the digital filtered signal to obtain the filtered vibration signal; The method for adjusting the passband range of the adaptive bandpass filter is as follows: Calculate the signal-to-noise ratio (SNR) of the digitally filtered signal and record it as the current SNR; If the current signal-to-noise ratio is less than a preset signal-to-noise ratio threshold, the passband range is reduced, and the reduction is proportional to the reciprocal of the current signal-to-noise ratio. If the current signal-to-noise ratio is greater than or equal to a preset signal-to-noise ratio threshold, then the passband range remains unchanged.

4. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 1, characterized in that, The step of designing a dynamic extraction and recovery strategy based on the oversampling factor of the filtered vibration signal and the target analysis bandwidth to restore the filtered vibration signal to the target sampling frequency includes: Based on the adaptive oversampling factor, the filtered vibration signal is divided into multiple data segments according to the time series, and the number of sampling points in each data segment is equal to the adaptive oversampling factor; A weighted average is calculated for the sampling points within each data segment. The weights of the weighted average are determined based on the positional distribution of the sampling points within the data segment, and the weights follow a Gaussian distribution to obtain a preliminary extracted signal. The initially extracted signal is subjected to signal recovery filtering, which employs a cascaded finite impulse response filter. The order of the finite impulse response filter is dynamically adjusted according to the logarithm of the adaptive oversampling factor to obtain the recovered vibration signal.

5. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 1, characterized in that, The step of extracting micro-vibration transmission characteristic parameters of the satellite structure based on the time-domain and frequency-domain characteristics of the recovered vibration signal includes: Time-domain analysis was performed on the recovered vibration signal to obtain the peak acceleration, root mean square acceleration, and vibration duration of the recovered vibration signal; Frequency domain analysis is performed on the recovered vibration signal to obtain its auto-power spectral density function, cross-power spectral density function, and frequency response function. Based on the self-power spectral density function, the dominant frequency and its corresponding amplitude of the recovered vibration signal are extracted and denoted as the dominant vibration characteristic parameters; Based on the cross-power spectral density function and the frequency response function, the transfer function of the recovered vibration signal between different measurement points is calculated and denoted as the transfer characteristic parameter; The micro-vibration transmission characteristic parameters are obtained by combining the principal vibration characteristic parameters with the transmission characteristic parameters.

6. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 5, characterized in that, The evaluation of the vibration suppression performance of the satellite structure based on the micro-vibration transmission characteristic parameters includes: A vibration transmission path model is constructed, wherein the vibration transmission path model uses the transmission characteristic parameters as edge weights and the principal vibration characteristic parameters as node attributes; Based on the vibration transmission path model, the vibration transmission attenuation rate from the disturbance source measuring point to the sensitive load measuring point is calculated. The method for calculating the vibration transmission attenuation rate is as follows: The transmission gain of each path is obtained by multiplying the transmission characteristic parameters along each path in the vibration transmission path model. The minimum value of the transmission gain of all paths is denoted as the vibration transmission attenuation rate. The vibration suppression performance of the satellite structure is evaluated based on the comparison between the vibration transmission attenuation rate and the preset attenuation threshold. The evaluation method is as follows: If the vibration transmission attenuation rate is greater than the preset attenuation threshold, then the vibration suppression performance of the satellite structure is determined to meet the requirements. If the vibration transmission attenuation rate is less than or equal to the preset attenuation threshold, the vibration suppression performance of the satellite structure is determined to be unsatisfactory, and the path with the lowest transmission gain in the vibration transmission path model is marked as the optimization target path.

7. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 3, characterized in that, The method for dynamically adjusting the cutoff frequency of the analog high-pass filter includes: The initial cutoff frequency is calculated based on the lowest effective frequency of the oversampled vibration signal, and the initial cutoff frequency is 0.8 times the lowest effective frequency. The initial cutoff frequency is adjusted based on the proportion of the DC component in the oversampled vibration signal. The method for calculating the proportion of the DC component is as follows: Perform a Fourier transform on the oversampled vibration signal to obtain the spectrum of the oversampled vibration signal; Calculate the ratio of the energy of the zero-frequency component to the total energy in the spectrum, and denot it as the DC component percentage; If the proportion of the DC component is greater than the preset DC threshold, the initial cutoff frequency is increased by a factor that is proportional to the square of the proportion of the DC component. If the proportion of the DC component is less than or equal to the preset DC threshold, the initial cutoff frequency remains unchanged.

8. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 4, characterized in that, The method for dynamically adjusting the order of the finite impulse response filter includes: The initial filter order is calculated based on the adaptive oversampling factor, and the initial filter order is twice the logarithm of the adaptive oversampling factor. Based on the signal-to-noise ratio of the initially extracted signal, the order of the initial filter is adjusted, and the adjustment method is as follows: If the signal-to-noise ratio of the initially extracted signal is less than a preset extraction signal-to-noise ratio threshold, the order of the initial filter is increased, and the increase is proportional to the reciprocal of the signal-to-noise ratio. If the signal-to-noise ratio of the initially extracted signal is greater than or equal to a preset extraction signal-to-noise ratio threshold, then the initial filter order remains unchanged.

9. The data acquisition and processing method for satellite micro-vibration ground testing according to claim 1, characterized in that, The method further includes: Based on the time-domain characteristics of the recovered vibration signal, a vibration signal anomaly detection strategy is designed, which is used to identify abnormal vibration events in the recovered vibration signal. The implementation method of the vibration signal anomaly detection strategy is as follows: Calculate the local root mean square acceleration of the recovered vibration signal, wherein the local root mean square acceleration is a root mean square acceleration sequence calculated in units of a preset time window; A sliding window analysis was performed on the local root mean square acceleration sequence to obtain the standard deviation of the local root mean square acceleration within each sliding window, which was denoted as the local vibration fluctuation degree. If the local vibration fluctuation is greater than the preset fluctuation threshold, it is determined that there is an abnormal vibration event in the recovered vibration signal within the corresponding time window, and the start time, duration and peak acceleration of the abnormal vibration event are recorded. The information of the abnormal vibration event is transmitted to the user's computer, and the abnormal vibration event is marked in the form of a timeline on the user's computer's display interface.

Images