On-orbit mass measurement method based on vibration period method mass parameter solution

Abstract

The application discloses an on-orbit mass measurement method based on a vibration period method mass parameter solution, and particularly relates to the technical field of on-orbit mass measurement, wherein the on-orbit mass measurement device needs to sequentially perform no-load characteristic calibration and step loading calibration, and then obtains spring stiffness through a multi-load coupling calibration algorithm to serve as a calibration parameter; vibration data of a measured object is collected to obtain a reference data set, a signal preprocessing algorithm is used to process the reference data set to obtain an effective data set, a double-mode fusion period extraction algorithm is used to obtain period parameters of a characteristic mode set from the effective data set, and a logarithm amplitude operation is used to obtain damping ratio parameters corresponding to the effective data set, and finally, mass is calculated based on the calibration parameter, the period parameter and the damping ratio parameter and cross-validated.

Description

Technical Field

[0001] This invention relates to the field of on-orbit mass measurement technology, and more specifically to an on-orbit mass measurement method based on the vibration period method for calculating mass parameters. Background Technology

[0002] To ensure the long-term stable operation of my country's Tiangong space station, on-orbit health monitoring of astronauts and management of cargo transportation are crucial. In a microgravity environment, astronauts are susceptible to muscle atrophy and fluid redistribution, making changes in body mass a key indicator of health. However, traditional weighing equipment is unusable in space, and existing on-orbit mass measurement devices generally suffer from low accuracy and complex operation, making it difficult to meet the needs of refined medical assessments. Simultaneously, the increased frequency of cargo transport between the space station and Earth places higher demands on the accuracy of cargo package mass measurement, requiring more precise control over the return capsule's mass characteristics.

[0003] To address this issue, based on the principle of spring vibration, this invention designs a novel on-orbit mass measurement device. This device collects vibration signals using built-in force and acceleration sensors to calculate the mass of the measured object. While compact and suitable for on-orbit applications, its measurement signals are susceptible to mechanical vibration, electromagnetic interference, and quantization noise, exhibiting low-frequency and non-stationary characteristics. Traditional signal processing methods struggle to effectively suppress noise. Furthermore, the complex space environment can cause drift in spring stiffness and system damping parameters, further increasing measurement errors. Therefore, a complementary mass calculation algorithm with noise suppression, period extraction, and parameter correction functions is needed to improve measurement accuracy, better serve astronaut health management and cargo handling systems, and promote the development of my country's space program. Summary of the Invention

[0004] To address this, the present invention provides an on-orbit mass measurement method based on the vibration period method for mass parameter calculation. This on-orbit mass measurement device calculates mass parameters based on the vibration period method. Its core principle is to measure the period of free vibration of the object under spring drive, and then solve the mass of the specific moving object by inversely solving the period, thereby solving the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: an on-orbit mass measurement method based on vibration period method for mass parameter calculation, wherein the on-orbit mass measurement device needs to be calibrated by performing no-load characteristic calibration and stepped loading calibration in sequence, and then the spring stiffness is obtained by solving the multi-load coupling calibration algorithm and used as the calibration parameter.

[0006] First, vibration data of the object under test is collected to obtain a reference dataset. The reference dataset is then processed by a signal preprocessing algorithm to obtain a valid dataset. The valid dataset is then processed by a dual-modal fusion period extraction algorithm to obtain the period parameters of the feature mode set. The damping ratio parameters corresponding to the valid dataset are obtained by amplitude logarithmic operation. Finally, the mass is calculated and cross-validated based on the calibration parameters, period parameters, and damping ratio parameters.

[0007] The preferred multi-load coupling calibration algorithm has the following specific steps:

[0008] Step 1.1: Calibration Data Acquisition and Processing:

[0009] Select three sets of precision weights The mass is distributed in an arithmetic sequence and the mass range covers the range of the device.

[0010] Three repeated tests were performed for each calibration load to collect vibration data. Period data was extracted using a signal preprocessing algorithm and a dual-modal fusion period extraction algorithm. The damping ratio was calculated using the logarithmic amplitude decay method to obtain the mass-period-damping ratio dataset. ;

[0011] In the formula: for The corresponding weight set number, For the corresponding group quality, The average period of three repeated tests for the corresponding group. This represents the average damping ratio of three repeated tests for the corresponding group.

[0012] Step 1.2: Inverse parameter solution:

[0013] A parameter identification model was constructed based on the single-degree-of-freedom damped vibration theory for the three sets of calibration data:

[0014] ;

[0015] Set spring stiffness Solve the system of equations using the initial guesses to obtain the parameter solution set. ;

[0016] Step 1.3: Solution set filtering and parameter optimization:

[0017] According to spring stiffness The benchmark value Calculate parameter deviation:

[0018] ;

[0019] Set error tolerance , retain satisfaction The parameters are used to calculate their arithmetic mean as the corrected spring stiffness. The value of .

[0020] In the formula: Spring stiffness as an inverse solution Allowable deviation from spring stiffness reference value The maximum error limit.

[0021] Preferably, the specific steps of the signal preprocessing algorithm are as follows:

[0022] Under the same operating conditions, three vibration excitations were performed, with data continuously collected each time until the vibration amplitude decayed to its initial value. The following steps ensure a complete record of the free decay process of the vibration system:

[0023] Step 2.1: Zero-point calibration:

[0024] The average value of the first 500 sampling points of the original data is calculated to establish an initial zero-point baseline;

[0025] The entire vibration signal data is subtracted from the zero-point reference to eliminate baseline offset caused by sensor zero-point drift, ensuring that no maximum value appears near the zero point and achieving zero-point alignment.

[0026] Step 2.2: Low-pass filtering:

[0027] Based on the low signal-to-noise ratio characteristic of vibration signals, a Butterworth low-pass filter was constructed to remove high-frequency signals. A fourth-order filter structure was adopted, and the cutoff frequency was set to 5Hz to meet the requirement of preserving low-frequency vibration characteristics of the space station.

[0028] Then, zero-phase filtering is performed to eliminate the influence of phase delay on the time-domain waveform and preserve the damped vibration characteristics;

[0029] Step 2.3: Effective segment segmentation using the amplitude threshold method:

[0030] Take the maximum amplitude of the filtered signal Trigger threshold When the signal amplitude exceeds the threshold, it is marked as the start of a valid segment;

[0031] The minimum effective time is set to 0.7 seconds to detect consecutive exceedances of the threshold. Regions whose duration exceeds the minimum valid segment time are considered valid data segments, and adjacent valid segments are merged; the minimum silence time is set to 0.4s, and if the signal remains below the trigger threshold for 0.3 seconds consecutively... This marks the end of the valid segment.

[0032] Retain the valid signals that meet the threshold condition, and divide the original filtered signal into 3 datasets. , , And then store it again for subsequent calculations.

[0033] The preferred steps of the dual-modal fusion period extraction algorithm are as follows:

[0034] Step 3.1: Refined extraction of effective segments:

[0035] For dataset , , Perform adaptive truncation separately:

[0036] The first positive zero-crossing point of the positioning signal is taken as the starting point, and all data before this zero point is truncated to avoid transient oscillations caused by the initial impact;

[0037] Depending on the number of peaks, the backward positioning signal... After each peak, at the zero point, all data after that zero point are truncated to avoid interference from the attenuation segment;

[0038] Extracting the first peak after the first zero point to the second peak Data from the zero-point interval after each wave peak is used only to calculate the period, ensuring that the quasi-steady-state vibration conditions are met, and generating a set of characteristic modes for period extraction. , , ;

[0039] Step 3.2: Dual-modal periodicity detection:

[0040] For feature mode set , , The period was calculated using the time-domain peak detection method and the autocorrelation function method, respectively.

[0041] Set amplitude threshold and the minimum interval between adjacent peaks is , The sampling frequency is used to filter out valid peaks that meet the conditions, and the period is calculated using the time difference between peaks. And calculate the periodic consistency index of the time-domain peak. and peak significance index Comprehensive calculation of credibility ;

[0042] ;

[0043] in, The periodic average, The periodic standard deviation;

[0044] ;

[0045] in, Peak value, The maximum amplitude of the signal is given; the confidence level is calculated as follows:

[0046] ;

[0047] Calculate the autocorrelation function for the feature mode set, detect the significant peak positions of the autocorrelation function, and calculate the period. And calculate the significance index of the secondary peak. Consistency index of decay Comprehensive calculation of credibility ;

[0048] ;

[0049] in, The height of the main peak This is the height of the first secondary peak. This represents the delay time corresponding to the secondary peak.

[0050] ;

[0051] The confidence level is calculated as follows:

[0052] ;

[0053] Step 3.3: Dynamic weight allocation strategy:

[0054] The weights of the results for each period are calculated based on the credibility level. The weighting formula is as follows: ,in ;

[0055] The final cycle is obtained by weighted fusion calculation. ,in The final cycle As a periodic parameter of the corresponding feature mode set.

[0056] (1) This invention addresses the low signal-to-noise ratio characteristics of vibration signals by designing a three-level preprocessing process: zero-point correction to eliminate baseline drift, low-pass filtering to suppress high-frequency noise, and effective segment segmentation based on amplitude thresholds to extract the main vibration modes. This effectively solves the problems of noise interference in the original signal and insufficient accuracy in extracting effective data. The same amplitude threshold screening mechanism accurately separates the characteristic modes reflecting the true motion of the measured object from the original vibration data, avoiding the period misjudgment caused by mode aliasing in traditional decomposition methods.

[0057] (2) To address the problem that traditional signal processing algorithms struggle to accurately extract periodic information, this invention proposes a dual-modal fusion period extraction algorithm. By combining an adaptive standard mode truncation strategy and a dynamic weight allocation mechanism, the period extraction error is reduced to ≤3%, achieving an accuracy improvement of 20%-80% compared to traditional single methods. This significantly enhances the accuracy of extracting motion parameters (damping ratio, period) from vibration data.

[0058] (3) To address the problem of parameter drift in complex environments, this invention designs a multi-load coupling calibration algorithm. Through error tolerance screening and parameter optimization, more accurate spring stiffness parameters can be obtained. This algorithm is then applied to the device's power-on calibration process to achieve parameter self-correction, thereby improving the system's robustness to complex environments. In the unknown mass measurement stage, the designed algorithms are organically combined and a mass cross-validation strategy is used to automate the entire process of data input, algorithm execution, and result output. This simplifies the data processing flow and yields highly efficient and accurate mass measurement results. Attached Figure Description

[0059] Figure 1 A diagram of a mass measurement device based on the vibration principle provided by this invention; Figure 1 (a) in the diagram is the schematic diagram of the device; Figure 1 (b) in the image is a photograph taken during the experiment;

[0060] Figure 2 A schematic diagram of the mass measuring device provided for this invention;

[0061] Figure 3 This is a front view of the on-orbit mass measurement device for the on-orbit mass measurement quality calculation algorithm provided by the present invention;

[0062] Figure 4 This invention provides a sensor installation diagram for an on-orbit mass measurement algorithm.

[0063] Figure 5 The flowchart of the signal preprocessing algorithm for on-orbit mass measurement quality calculation provided by this invention;

[0064] Figure 6 The flowchart of the dual-modal fusion period extraction algorithm for on-orbit mass measurement quality calculation provided by this invention is shown below.

[0065] Figure 7 The flowchart of the multi-load coupling calibration and correction parameter algorithm provided by the present invention is applicable to the on-orbit mass measurement quality calculation algorithm.

[0066] Figure 8 The flowchart illustrates the overall usage of the mass measurement device for the on-orbit mass measurement quality calculation algorithm provided by this invention. Detailed Implementation

[0067] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. 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.

[0068] In 2019, based on the existing technology (Lou Renzhi, Liu Jinsheng, Rui Zehao, et al. Development of a weight measurement device under weightlessness based on the principle of spring oscillator [J]. Aerospace Medicine and Medical Engineering, 2019, 32(06):503-507.DOI:10.16289 / j.cnki.1002-0837.2019.06.006.), a human body mass measurement device that can be used under weightlessness conditions was developed and ground testing was completed. For example... Figure 1 As shown ( Figure 1 (a) is the schematic diagram of the device. Figure 1 (b) is a photograph of the actual experiment. The device consists of a frame, spring, linear optical axis, linear ball bearing, and measurement sensing system. The linear ball bearing and linear optical axis work together to reduce friction and make the reciprocating motion smoother. In ground tests, air buoyancy was used to counteract gravity, simulating a weightless environment for rigid weight measurement experiments and flexible human body measurement experiments. The experimental results show that the measurement error for weights with a mass range of 45~80kg is within ±0.25%, and the measurement error for human body mass of subjects with a mass range of 45~80kg is within ±0.5%.

[0069] The device uses a spring oscillator to drive the object being measured to reciprocate, and the vibration period is based on the simple harmonic motion. The mass of the object being measured can be determined by inverse calculation; the specific formula is as follows: ,in The mass of the object being measured. For the unloaded weight of the seat, This refers to the spring stiffness. The oscillation period can be obtained through no-load calibration of the device. It is obtained by measuring the time difference of the high-level signal output due to the spring vibration blocking the light path by a photoelectric sensor.

[0070] The algorithm of this device does not take into account the effect of the inherent damping of the system, which will affect the accuracy of the mass measurement results.

[0071] The device uses a slotted optical coupler and a light shield to detect the vibration period. If the optical path is contaminated or interfered with by vibration, it will affect the measurement of the vibration period and make it impossible to obtain an accurate period for the algorithm to solve the problem.

[0072] In 2021, based on existing technology (Liu Miao, Liu Jinsheng, Mao Qianmin, et al. Design and implementation of a small mass measuring device under weightlessness [J]. Aerospace Medicine and Medical Engineering, 2021, 34(01):39-43.DOI:10.16289 / j.cnki.1002-0837.2021.01.007.), a device for measuring the mass of small objects (10~700g) under weightlessness was developed. The device consists of a vibration system and a measurement sensing system, and the overall structure is as follows: Figure 2 As shown, the vibration system is based on a cantilever beam structure design. Four leaf springs with identical geometry and physical properties are fixed at one end to a base plate and at the other end to a movable support plate. There is no friction or gap at the connection between the leaf springs and the support plate. The measurement and sensing system consists of a laser displacement sensor and a microcontroller system. The laser displacement sensor measures the displacement of the leaf springs. When the leaf spring is in its equilibrium position, the displacement is zero; when it moves away from the sensor, the displacement is positive, and the sensor outputs a high-level signal; when it moves closer to the sensor, the displacement is negative, and the sensor outputs a low-level signal.

[0073] This method uses a leaf spring as the elastic element and a cantilever beam vibration mode, employing a displacement sensor to obtain the vibration period. By having the leaf spring and the object under test move in a horizontal plane, gravity has no effect on the vibration, simulating weightlessness. This device is based on the principle of free vibration, and the specific calculation formula is as follows: ,in The mass of the object being measured. For the unloaded weight of the seat, The total stiffness of the four leaf springs is determined by (E is the elastic modulus, L is the length of the leaf spring, b is the width of the leaf spring, and h is the thickness of the leaf spring) The vibration period is obtained. The average value was obtained by measuring the time interval between 20 consecutive passes of the leaf spring through the equilibrium position using a laser sensor. Test results show that the measurement error is better than ±0.5g and the repeatability is better than ±0.1g for the measured mass range of 10~700g.

[0074] The device has a narrow measurement range and is only suitable for measuring rigid or symmetrical objects.

[0075] The device is highly sensitive to its installation attitude and must be placed strictly horizontally. However, it is difficult to ensure absolute horizontality in the space environment, which affects the accuracy of the measurement.

[0076] The sheet spring structure of this device may experience stiffness changes during long-term use, and the device algorithm lacks a calibration process for the spring stiffness.

[0077] This invention proposes an on-orbit mass measurement method based on the vibration period method for mass parameter calculation. The on-orbit mass measurement device needs to be calibrated by performing no-load characteristic calibration and stepped loading calibration in sequence. Then, the spring stiffness is obtained by solving the multi-load coupling calibration algorithm and used as the calibration parameter.

[0078] First, vibration data of the object under test is collected to obtain a reference dataset. The reference dataset is then processed by a signal preprocessing algorithm to obtain a valid dataset. The valid dataset is then processed by a dual-modal fusion period extraction algorithm to obtain the period parameters of the feature mode set. The damping ratio parameters corresponding to the valid dataset are obtained by amplitude logarithmic operation. Finally, the mass is calculated and cross-validated based on the calibration parameters, period parameters, and damping ratio parameters.

[0079] like Figure 7 As shown, the specific steps of the multi-load coupling calibration algorithm are as follows:

[0080] Step 1.1: Calibration Data Acquisition and Processing:

[0081] Select three sets of precision weights The mass is distributed in an arithmetic sequence and the mass range covers the range of the device.

[0082] Three repeated tests were performed for each calibration load to collect vibration data. Period data was extracted using a signal preprocessing algorithm and a dual-modal fusion period extraction algorithm. The damping ratio was calculated using the logarithmic amplitude decay method to obtain the mass-period-damping ratio dataset. ;

[0083] In the formula: for The corresponding weight set number, For the corresponding group quality, The average period of three repeated tests for the corresponding group. This represents the average damping ratio of three repeated tests for the corresponding group.

[0084] Step 1.2: Inverse parameter solution:

[0085] A parameter identification model was constructed based on the single-degree-of-freedom damped vibration theory for the three sets of calibration data:

[0086] ;

[0087] Set spring stiffness Solve the system of equations using the initial guesses to obtain the parameter solution set. ;

[0088] Step 1.3: Solution set filtering and parameter optimization:

[0089] According to spring stiffness The benchmark value Calculate parameter deviation:

[0090] ;

[0091] Set error tolerance , retain satisfaction The parameters are used to calculate their arithmetic mean as the corrected spring stiffness. The value of .

[0092] In the formula: Spring stiffness as an inverse solution Allowable deviation from spring stiffness reference value The maximum error limit.

[0093] like Figure 5 As shown, the specific steps of the signal preprocessing algorithm are as follows:

[0094] Under the same operating conditions, three vibration excitations were performed, with data continuously collected each time until the vibration amplitude decayed to its initial value. The following steps ensure a complete record of the free decay process of the vibration system:

[0095] Step 2.1: Zero-point calibration:

[0096] The average value of the first 500 sampling points of the original data is calculated to establish an initial zero-point baseline;

[0097] The entire vibration signal data is subtracted from the zero-point reference to eliminate baseline offset caused by sensor zero-point drift, ensuring that no maximum value appears near the zero point and achieving zero-point alignment.

[0098] Step 2.2: Low-pass filtering:

[0099] Based on the low signal-to-noise ratio characteristic of vibration signals, a Butterworth low-pass filter was constructed to remove high-frequency signals. A fourth-order filter structure was adopted, and the cutoff frequency was set to 5Hz to meet the requirement of preserving low-frequency vibration characteristics of the space station.

[0100] Then, zero-phase filtering is performed to eliminate the influence of phase delay on the time-domain waveform and preserve the damped vibration characteristics;

[0101] Step 2.3: Effective segment segmentation using the amplitude threshold method:

[0102] Take the maximum amplitude of the filtered signal Trigger threshold When the signal amplitude exceeds the threshold, it is marked as the start of a valid segment;

[0103] The minimum effective time is set to 0.7 seconds to detect consecutive exceedances of the threshold. Regions whose duration exceeds the minimum valid segment time are considered valid data segments, and adjacent valid segments are merged; the minimum silence time is set to 0.4s, and if the signal remains below the trigger threshold for 0.3 seconds consecutively... This marks the end of the valid segment.

[0104] Retain the valid signals that meet the threshold condition, and divide the original filtered signal into 3 datasets. , , And then store it again for subsequent calculations.

[0105] like Figure 6 As shown, the specific steps of the dual-modal fusion period extraction algorithm are as follows:

[0106] Step 3.1: Refined extraction of effective segments:

[0107] For dataset , , Perform adaptive truncation separately:

[0108] The first positive zero-crossing point of the positioning signal is taken as the starting point, and all data before this zero point is truncated to avoid transient oscillations caused by the initial impact;

[0109] Depending on the number of peaks, the backward positioning signal... After each peak, at the zero point, all data after that zero point are truncated to avoid interference from the attenuation segment;

[0110] Extracting the first peak after the first zero point to the second peak Data from the zero-point interval after each wave peak is used only to calculate the period, ensuring that the quasi-steady-state vibration conditions are met, and generating a set of characteristic modes for period extraction. , , ;

[0111] Step 3.2: Dual-modal periodicity detection:

[0112] For feature mode set , , The period was calculated using the time-domain peak detection method and the autocorrelation function method, respectively.

[0113] Set amplitude threshold and the minimum interval between adjacent peaks is , The sampling frequency is used to filter out valid peaks that meet the conditions, and the period is calculated using the time difference between peaks. And calculate the periodic consistency index of the time-domain peak. and peak significance index Comprehensive calculation of credibility ;

[0114] ;

[0115] in, The periodic average, The periodic standard deviation;

[0116] ;

[0117] in, Peak value, The maximum amplitude of the signal is given; the confidence level is calculated as follows:

[0118] ;

[0119] Calculate the autocorrelation function for the feature mode set, detect the significant peak positions of the autocorrelation function, and calculate the period. And calculate the significance index of the secondary peak. Consistency index of decay Comprehensive calculation of credibility ;

[0120] ;

[0121] in, The height of the main peak This is the height of the first secondary peak. This represents the delay time corresponding to the secondary peak.

[0122] ;

[0123] The confidence level is calculated as follows:

[0124] ;

[0125] Step 3.3: Dynamic weight allocation strategy:

[0126] The weights of the results for each period are calculated based on the credibility level. The weighting formula is as follows: ,in ;

[0127] The final cycle is obtained by weighted fusion calculation. ,in The final cycle As a periodic parameter of the corresponding feature mode set.

[0128] (1) To address the problem of low signal-to-noise ratio characteristics in the original signal due to interference from non-stationary signals, this invention proposes a signal preprocessing method based on amplitude threshold screening of effective segments. Through zero-point correction, low-pass filtering, and amplitude threshold segmentation identification mechanism, abnormal extreme values ​​at zero points can be effectively avoided, high-frequency noise components are effectively filtered out, and the attenuation characteristics of the main vibration modes are relatively well preserved. Furthermore, it can reliably segment the effective vibration data segments of the experimental group, ensuring that the processed signal meets the accuracy requirements for subsequent dynamic parameter identification.

[0129] (2) To address the problem that a single signal processing algorithm is insufficient to accurately obtain the periodic component, this invention proposes a dual-mode period extraction algorithm that integrates time-domain peak detection and autocorrelation analysis. By extracting data from stable vibration segments and combining it with a dynamic weight allocation mechanism, the advantages of the two detection methods are complemented and their reliability is integrated, effectively improving the measurement accuracy of vibration period under complex working conditions.

[0130] (3) To address the problem of key device parameters drifting under complex environments, this invention proposes a parameter correction algorithm based on multi-load coupling calibration. This algorithm introduces three different sets of load measurement data, extracts periodic features, and solves the problem by combining these with parameter calculations. Based on a preset correction error tolerance (such as...),... This enables real-time correction of parameter drift, eliminates measurement errors caused by inherent parameter drift in the system, and improves the device's adaptability to the environment.

[0131] like Figure 8 As shown, the overall mass measurement process of the device obtained based on the measurement method of the present invention is as follows:

[0132] 1) Device startup calibration parameters

[0133] Step 1: No-load characteristic calibration:

[0134] After the on-orbit mass measurement device is installed, check whether the device can operate normally. If not, the device needs to be repaired and reassembled.

[0135] After confirming that the device is working correctly, maintain an unloaded state and manually apply lateral displacement. Allow it to vibrate freely, and repeatedly collect three sets of vibration data, which are then transmitted to the host computer for processing.

[0136] Step 2: Stepped loading calibration:

[0137] Install a 5kg standard mass block, repeat the operation, and collect data.

[0138] Replace with a 10kg standard mass block, repeat the operation, and collect data.

[0139] Step 3: Multi-load calibration parameters:

[0140] The spring stiffness of the device is solved by applying a multi-load coupling calibration algorithm to the three sets of calibration data. The accurate value is used to complete the initialization of the device's startup parameters.

[0141] If no stiffness value meets the error tolerance, then determine the spring stiffness. There is an error in the parameter measurement and calculation. Please repeat the above process.

[0142] Unknown mass measurement:

[0143] Step 1: Acquisition of vibration data of the object under test:

[0144] Install the object to be tested and ensure that it is fixed in the center of the device.

[0145] Manually apply lateral displacement Collect vibration data until the amplitude decays to The data was collected three times to form the baseline dataset D.

[0146] Step 2: Feature parameter extraction:

[0147] A signal preprocessing algorithm is applied to the benchmark dataset D to preprocess the raw data signal and filter out the valid dataset based on the amplitude threshold. , , .

[0148] For dataset , , The dual-modal fusion period extraction algorithm is applied to extract the feature modality set. , , Extracting the period , , The damping ratio parameter for each dataset is calculated using the logarithmic average of the amplitude ratios of n consecutive wave peaks. , , .

[0149] (15)

[0150] Step 3: Quality calculation and cross-validation:

[0151] Based on calibration parameters and the corresponding dataset , Calculate the dataset separately quality , obtain mass set .

[0152] According to the theory of damped vibration, the period of vibration is... The relationship with the system's inherent parameters is as follows:

[0153] (16)

[0154] In the formula: This represents the natural frequency of a damped system. This represents the system's natural frequency under undamped conditions. The damping ratio represents the ratio of the system's damping to its critical damping. Due to spring stiffness and the total mass of the moving object Decide:

[0155] (17)

[0156] In the formula: For system damping, it can be determined by the damping ratio. and total weight and spring stiffness Sure, Represents the total weight of a moving object, including the mass of the device itself. and the mass of the object being measured The specific expression is as follows:

[0157] (18)

[0158] Damping ratio The calculation is performed using the logarithmic decay method, which is based on measuring the attenuation rate of the vibration amplitude before and after the wave peaks. The damping ratio is calculated using the following expression:

[0159] (19)

[0160] (20)

[0161] In the formula: Represents the logarithmic decay rate. It is the amplitude of the first peak. It is the amplitude of the (n+1)th peak.

[0162] Among them, the mass of the device body It is known and can be obtained through ground measurements. The system damping ratio. This is also known and can be calculated from the amplitude attenuation ratio of vibration data. And the spring stiffness... These are the main unknown parameters, which need to be calibrated using a multi-load coupling calibration method. Finally, by extracting the vibration period and the magnitudes of various parameters in the system, the mass of the moving component can be deduced, and then the mass of the measured object can be obtained by subtracting the mass of the device itself.

[0163] By combining the above relationships, we can derive the expression for the mass of the object being measured:

[0164] (twenty one)

[0165] Quality cross-validation:

[0166] Calculate the relative error of the mass set

[0167] (twenty two)

[0168] , Similarly. If , , Output final quality:

[0169] (twenty three)

[0170] If any relative error exceeds 5%, discard the mass group with excessively large relative error, and use the remaining relative error. The final mass is calculated using the mass group:

[0171] (twenty four)

[0172] If the relative errors of all quality groups are If the data collected for the corresponding working condition is deemed abnormal, the data is discarded, and the data for that working condition is collected again to calculate the quality.

[0173] Figure 3 As an on-orbit mass measurement device based on an on-orbit mass measurement algorithm, the seat structure 1 is in contact with the object being measured and is connected to the base structure 3 through a linear optical axis and a linear bearing; the base structure 3 is fixed to the ground to achieve overall fixation of the device; the electromagnetic damping structure 2 is fixed to the seat structure 1 through a central guide rod, and generates stable and controllable electromagnetic damping after being energized, which is used to regulate the period and amplitude attenuation of vibration (this function is not used in this invention); the bidirectional spring structure 4 is installed between the seat structure 1 and the base structure 3 to overcome the nonlinear change in stiffness during the extension process of the compression spring and ensure that the spring is always within a stable stiffness range.

[0174] Figure 4For sensors suitable for on-orbit mass measurement algorithms, an electromagnetic pressure sensor 51 is installed between the seat structure 1 and the central guide rod to measure the magnitude of the damping force generated by the electromagnetic damping structure 2 (this function is not used in this invention). A spring pressure sensor 52 is installed between the seat structure 1 and the sensor tray 53 to measure the spring pressure; the collected pressure dataset serves as the reference dataset to be processed in this invention. The sensor tray 53 is fixed to the seat structure 1 via the spring pressure sensor 52 and moves with the object being measured.

[0175] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. An on-orbit mass measurement method based on the vibration period method for mass parameter calculation, characterized in that: The on-orbit mass measurement device needs to be calibrated by performing no-load characteristic calibration, stepped loading calibration, and then the spring stiffness is obtained by solving the multi-load coupling calibration algorithm and used as the calibration parameter. First, vibration data of the object under test is collected to obtain a reference dataset. The reference dataset is then processed by a signal preprocessing algorithm to obtain a valid dataset. The valid dataset is then processed by a dual-modal fusion period extraction algorithm to obtain the period parameters of the feature mode set. The damping ratio parameters corresponding to the valid dataset are obtained by amplitude logarithmic operation. Finally, the mass is calculated and cross-validated based on the calibration parameters, period parameters, and damping ratio parameters.

2. The on-orbit mass measurement method based on vibration period method for mass parameter calculation according to claim 1, characterized in that: The specific steps of the multi-load coupling calibration algorithm are as follows: Step 1.1: Calibration Data Acquisition and Processing: Select three sets of precision weights The mass is distributed in an arithmetic sequence and the mass range covers the range of the device. Three repeated tests were performed for each calibration load to collect vibration data. Period data was extracted using a signal preprocessing algorithm and a dual-modal fusion period extraction algorithm. The damping ratio was calculated using the logarithmic amplitude decay method to obtain the mass-period-damping ratio dataset. ; In the formula: for The corresponding weight set number. For the corresponding group quality, The average period of three repeated tests for the corresponding group. This represents the average damping ratio of three repeated tests for the corresponding group; Step 1.2: Inverse parameter solution: A parameter identification model was constructed based on the single-degree-of-freedom damped vibration theory for the three sets of calibration data: ； Set spring stiffness Solve the system of equations using the initial guesses to obtain the parameter solution set. ; Step 1.3: Solution set filtering and parameter optimization: According to spring stiffness The benchmark value Calculate parameter deviation: ； Set error tolerance , retain satisfaction The parameters are used to calculate their arithmetic mean as the corrected spring stiffness. The value; where: Spring stiffness for inverse solution Allowable deviation from spring stiffness reference value The maximum error limit.

3. The on-orbit mass measurement method based on vibration period method for mass parameter calculation according to claim 1, characterized in that: The specific steps of the signal preprocessing algorithm are as follows: Under the same operating conditions, three vibration excitations were performed, with data continuously collected each time until the vibration amplitude decayed to its initial value. The following steps ensure a complete record of the free decay process of the vibration system: Step 2.1: Zero-point calibration: The average value of the first 500 sampling points of the original data is calculated to establish an initial zero-point baseline; The entire vibration signal data is subtracted from the zero-point reference to eliminate baseline offset caused by sensor zero-point drift, ensuring that no maximum value appears near the zero point and achieving zero-point alignment. Step 2.2: Low-pass filtering: Based on the low signal-to-noise ratio characteristic of vibration signals, a Butterworth low-pass filter was constructed to remove high-frequency signals. A fourth-order filter structure was adopted, and the cutoff frequency was set to 5Hz to meet the requirement of preserving low-frequency vibration characteristics of the space station. Then, zero-phase filtering is performed to eliminate the influence of phase delay on the time-domain waveform and preserve the damped vibration characteristics; Step 2.3: Effective segment segmentation using the amplitude threshold method: Take the maximum amplitude of the filtered signal Trigger threshold When the signal amplitude exceeds the threshold, it is marked as the start of a valid segment; The minimum effective time is set to 0.7 seconds to detect consecutive exceedances of the threshold. Regions whose duration exceeds the minimum valid segment time are considered valid data segments, and adjacent valid segments are merged; the minimum silence time is set to 0.4s, and if the signal remains below the trigger threshold for 0.3 seconds consecutively... This marks the end of the valid segment. Retain the valid signals that meet the threshold condition, and divide the original filtered signal into 3 datasets. , , And then store it again for subsequent calculations.

4. The on-orbit mass measurement method based on vibration period method for mass parameter calculation according to claim 1, characterized in that: The specific steps of the dual-modal fusion period extraction algorithm are as follows: Step 3.1: Refined extraction of effective segments: For dataset , , Perform adaptive truncation separately: The first positive zero-crossing point of the positioning signal is taken as the starting point, and all data before this zero point is removed to avoid transient oscillations caused by the initial impact; Depending on the number of peaks, the backward positioning signal... After each peak, at the zero point, all data after that zero point are truncated to avoid interference from the attenuation segment; Extracting the first peak after the first zero point to the second peak Data from the zero-point interval after each wave peak is used only to calculate the period, ensuring that the quasi-steady-state vibration conditions are met, and generating a set of characteristic modes for period extraction. , , ; Step 3.2: Dual-modal periodicity detection: For feature mode set , , The period was calculated using the time-domain peak detection method and the autocorrelation function method, respectively. Set amplitude threshold and the minimum interval between adjacent peaks is , The sampling frequency is used to filter out valid peaks that meet the conditions, and the period is calculated using the time difference between the peaks. And calculate the periodic consistency index of the time-domain peak. and peak significance index Comprehensive calculation of credibility ; ； in, The periodic average, The periodic standard deviation; ； in, Peak value, The maximum amplitude of the signal is given; the confidence level is calculated as follows: ； Calculate the autocorrelation function for the feature mode set, detect the significant peak positions of the autocorrelation function, and calculate the period. And calculate the significance index of the secondary peak. Consistency index of decay Comprehensive calculation of credibility ; ； in, The height of the main peak This is the height of the first secondary peak. This represents the delay time corresponding to the secondary peak. ； The confidence level is calculated as follows: ； Step 3.3: Dynamic weight allocation strategy: The weights of the results for each period are calculated based on the credibility level. The weighting formula is as follows: ,in ; The final cycle is obtained by weighted fusion calculation. ,in The final cycle As a periodic parameter of the corresponding feature mode set.

Images