Intelligent interpretation method and system for launch vehicle flight data and attitude stability analysis

Abstract

The application discloses a launch vehicle flight data intelligent interpretation and attitude stability analysis method and system, relates to the launch vehicle flight data processing and attitude analysis technical field, wherein the launch vehicle flight data intelligent interpretation and attitude stability analysis method comprises the following steps: standardizing preprocessing of multi-source data in the flight process of a launch vehicle to obtain a standardized flight data set; obtaining fusion attitude feature data based on the standardized flight data set, and constructing a dynamic self-adaptive interpretation threshold to complete intelligent interpretation of the fusion attitude feature data; calculating an attitude stability margin value by fusing the fusion attitude feature data, the dynamic self-adaptive interpretation threshold and a dynamic correction coefficient of an attitude real-time change rate, so as to realize quantitative grading and determination of the flight attitude stability of the launch vehicle. The application can improve the accuracy of attitude analysis and the timeliness of early warning, realize fine active control of the attitude, and effectively guarantee the flight attitude safety and controllability of the launch vehicle.

Description

Technical Field

[0001] This application relates to the field of launch vehicle flight data processing and attitude analysis technology, and in particular to a method and system for intelligent interpretation of launch vehicle flight data and attitude stability analysis. Background Technology

[0002] The flight attitude of a launch vehicle is affected by a combination of factors, including engine operation, atmospheric disturbances, and sensor measurement errors. Attitude stability directly determines the success or failure of a flight mission. Real-time intelligent interpretation of multi-source flight data, accurate analysis of attitude stability, and early warning are the core requirements for launch vehicle flight control.

[0003] However, existing technologies related to launch vehicle attitude analysis have many shortcomings: First, the fusion of multi-source flight data adopts a fixed weight strategy and does not dynamically adjust based on sensor health and real-time flight conditions, resulting in insufficient accuracy and robustness of the fused attitude feature data. Second, attitude interpretation uses static thresholds or single-factor dynamic thresholds, which cannot adapt to the multiple influences of flight phases, environmental disturbances, and data fluctuations, leading to a high rate of misjudgment and missed judgments. Third, attitude stability determination relies solely on static deviation thresholds, ignoring dynamic attitude change trends and lacking the ability to provide early warnings of attitude instability in the future, only enabling post-event alerts, which is insufficient to meet the requirements of high-precision and high-reliability attitude control. Fourth, the attitude stability grading rules are coarse, and the early warning and control strategies are not precisely matched with the degree of attitude stability, failing to achieve refined and proactive attitude control.

[0004] To address one or more of the problems existing in the above-mentioned technologies, it is urgent to develop a new intelligent interpretation method and system for launch vehicle flight data and attitude stability analysis, so as to improve the accuracy of launch vehicle attitude analysis, the foresight of early warning, and the precision of control. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for intelligent interpretation of launch vehicle flight data and attitude stability analysis, which can improve the accuracy of attitude analysis and the timeliness of early warning, realize refined active control of attitude, and effectively ensure the safe and controllable flight attitude of launch vehicles.

[0006] To achieve the above objectives, this application provides a method for intelligent interpretation and attitude stability analysis of launch vehicle flight data, comprising the following steps: S1: Acquire multi-source data during the launch vehicle flight process, and perform standardized preprocessing on the multi-source data to obtain a standardized flight dataset with timestamp alignment, unified dimensions, and preservation of real-time dynamic characteristics; wherein, the multi-source data includes at least: inertial measurement data, flight status data, environmental perception data, and sensor health data; S2: Based on the multi-dimensional coupling relationship between the inertial measurement data, flight status data, and sensor health data in the standardized flight dataset, dynamically calculate the adaptation weight to obtain fused attitude feature data adapted to the real-time flight status of the launch vehicle; S3: Construct a dynamic adaptive interpretation threshold based on the fused attitude feature data, as well as the flight status data and environmental perception data in the standardized flight dataset, and complete the intelligent interpretation of the fused attitude feature data based on the dynamic adaptive interpretation threshold; S4: Calculate the attitude stability margin value through the fused attitude feature data, the dynamic adaptive interpretation threshold, and the dynamic correction coefficient of the real-time attitude change rate, and realize the quantitative classification and judgment of the flight attitude stability of the launch vehicle based on the attitude stability margin value.

[0007] As described above, the sub-steps for standardizing and preprocessing multi-source data to obtain a standardized flight dataset with timestamp alignment, unified dimensions, and preserved real-time dynamic characteristics are as follows: S11: Perform time synchronization processing on the multi-source data, aligning data from different sensors and acquisition sources to the same time axis, so that each data corresponds to the other at the same point in time, resulting in time-synchronized data; S12: Perform outlier removal processing on the time-synchronized data, removing invalid data with jumps, missing values, or values ​​exceeding reasonable ranges, resulting in outlier-removed data; S13: Perform data denoising processing on the outlier-removed data, reducing the impact of random noise on the data's authenticity, resulting in denoised data; S14: Perform normalization processing on the denoised data, mapping data with different dimensions and numerical ranges to the same numerical interval, achieving unified dimensions, resulting in a standardized flight dataset.

[0008] As above, the sub-step of step S2 is: S21: Extract from the standardized flight dataset The system includes inertial measurement data, flight status data, and sensor health data at various times. The inertial measurement data includes at least: attitude information, angular velocity, and linear acceleration collected by various sensors; the flight status data includes at least: flight phase information and trajectory tilt angle; and the sensor health data includes at least: sensor health coefficients and noise variance of various sensors. S22: Based on the multi-dimensional coupling relationship between the inertial measurement data, flight status data, and sensor health data, calculate the multi-dimensional coupling dynamic fusion weights. S23: Based on the multi-dimensional coupling dynamic fusion weights, perform weighted fusion of the inertial measurement data to obtain... Fusion pose feature data at different times .

[0009] As mentioned above, among them, Fusion pose feature data at different times The expression is:

[0010] ;

[0011] in, for Time of the first Multi-dimensional coupling dynamic fusion weights for inertial measurement-like data. , This represents the total number of categories of inertial measurement data. for Time of the first Standardized values ​​for inertial measurement data.

[0012] As above, the sub-step of step S3 is: S31: Extract from the standardized flight dataset The corresponding flight status data and environmental perception data at each moment; wherein, the flight status data includes at least flight phase information, and the environmental perception data includes at least dynamic pressure and wind speed; S32: Obtain multi-frame continuously fused attitude feature data within a preset sliding time window, and calculate based on the multi-frame continuously fused attitude feature data... The variance of the fused attitude feature data at each moment; S33: Based on The flight status data, environmental perception data, fused attitude feature data, and fluctuation variance corresponding to each moment are used to construct... Dynamic adaptive interpretation threshold at any time S34: will Fusion pose feature data at different times and Dynamic adaptive interpretation threshold at any time To make a comparison, if ,determination Normal; if ,determination An anomaly is identified and an anomaly information is marked. The anomaly information includes at least the anomaly time, the anomaly value, and the flight phase and environmental conditions corresponding to the anomaly time.

[0013] As above, the sub-step of step S4 is: S41: Based on Fusion pose feature data at different times Dynamic adaptive interpretation threshold And the dynamic correction coefficient of the real-time attitude change rate, to obtain Attention stability margin value at time step; S42: Based on preset hierarchical rules, for S43: Analyze the attitude stability margin value at any given time to determine the trend quantification classification of the launch vehicle's current flight attitude stability; S44: Based on the preset matching execution rules, determine and execute the attitude stability early warning and control strategy corresponding to the trend quantification classification of the launch vehicle's current flight attitude stability.

[0014] As shown above, the preset grading rule is: when When the second-level threshold is greater than or equal to the first-level threshold, the trend-quantified classification of the launch vehicle's current flight attitude stability is determined as Level 1, which represents both static and trend stability; when the second-level threshold is less than or equal to the first-level threshold, the stability is determined as Level 1. When the threshold value is less than the first-level boundary threshold, the trend-based quantitative classification of the launch vehicle's current flight attitude stability is determined to be Level 2. Level 2 indicates static stability, but the trend remains to be observed; when the threshold value is less than or equal to the third-level boundary threshold... When the threshold of the second grading is less than the threshold, it is judged as the third level. The third level indicates that the static situation is basically stable, but the trend is slightly deteriorating; when When the threshold is less than the third-level boundary threshold, it is determined to be at the fourth level, which indicates a high risk in static / trend conditions; where the first-level boundary threshold > the second-level boundary threshold > the third-level boundary threshold; for The attitude stability margin value at time t.

[0015] As shown above, the first grade boundary threshold is 0.6, the second grade boundary threshold is 0.4, and the third grade boundary threshold is 0.2.

[0016] As shown above, the preset matching execution rules are as follows: When the trend quantification classification of the current flight attitude stability of the launch vehicle is at level one, the corresponding attitude stability warning and control strategy is: do not trigger the warning, and maintain normal flight attitude interpretation and monitoring status; when the trend quantification classification of the current flight attitude stability of the launch vehicle is at level two, the corresponding attitude stability warning and control strategy is: trigger a low-level warning and continue to track. The trend of attitude stability margin values ​​at any given moment and in subsequent moments is monitored, with a focus on the persistence of attitude changes. When the trend of the launch vehicle's current flight attitude stability is classified as Level 3, the corresponding attitude stability warning and control strategy is to trigger a medium-level warning to remind operators to closely monitor attitude changes and assess the risk of attitude deterioration. When the trend of the launch vehicle's current flight attitude stability is classified as Level 4, the corresponding attitude stability warning and control strategy is to trigger a high-level warning so that relevant systems or operators can take timely attitude adjustment and fault diagnosis measures to ensure flight safety.

[0017] This application also provides a system for intelligent interpretation and attitude stability analysis of launch vehicle flight data, including a processor and a memory; wherein, the memory stores a computer program, and when the computer program is executed by the processor, it implements the above-mentioned method for intelligent interpretation and attitude stability analysis of launch vehicle flight data. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0019] Figure 1 A schematic diagram of the structure of an embodiment of a launch vehicle flight data intelligent interpretation and attitude stability analysis system;

[0020] Figure 2 A flowchart of one embodiment of a method for intelligent interpretation and attitude stability analysis of launch vehicle flight data. Detailed Implementation

[0021] 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, 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.

[0022] like Figure 1 As shown, this application provides a launch vehicle flight data intelligent interpretation and attitude stability analysis system, including: a processor and a memory; wherein, the memory stores a computer program, and when the computer program is executed by the processor, it implements the following launch vehicle flight data intelligent interpretation and attitude stability analysis method.

[0023] like Figure 2As shown, this application provides a method for intelligent interpretation of launch vehicle flight data and attitude stability analysis, including the following steps:

[0024] S1: Acquire multi-source data during the launch vehicle's flight process, and perform standardized preprocessing on the multi-source data to obtain a standardized flight dataset that is timestamped, dimensionally consistent, and retains real-time dynamic characteristics; among which, the multi-source data includes at least: inertial measurement data, flight status data, environmental perception data, and sensor health data.

[0025] Specifically, the inertial measurement data refers to the data collected by the launch vehicle's inertial measurement unit (IMU) and positioning module, including: roll angle, pitch angle, yaw angle and corresponding angular velocity collected by the gyroscope, axial linear acceleration of the rocket body collected by the accelerometer, and real-time latitude, longitude, altitude and flight speed data collected by the GPS / BeiDou positioning module.

[0026] Flight status data are key parameters reflecting the rocket's flight process, specifically including: flight altitude, flight Mach number, trajectory inclination, engine operating status (such as start / stop signals), interstage separation status (such as separated / not separated indicators), and orbital insertion status-related parameters.

[0027] The environmental perception data is the monitoring data of the external environment of the rocket body, specifically including: atmospheric dynamic pressure, real-time tangential / normal wind speed, atmospheric density and rocket body surface temperature data.

[0028] Sensor health data includes the self-test and operating status data of each sensor, specifically including: sensor operating status codes (such as normal / fault indicators), data transmission packet loss rate, deviation between measured and theoretical values, sensor self-test pass results, and power supply voltage stability data.

[0029] In the standardization preprocessing process of step S1, only the format of the multi-source data is normalized and invalid data is removed. The timestamp correlation and real-time dynamic fluctuation characteristics of each data are not changed, ensuring that subsequent steps can be dynamically calculated based on the real data of each time period.

[0030] Furthermore, as an example, the sub-steps for performing standardization preprocessing on multi-source data to obtain a standardized flight dataset that is timestamped, has unified units, and retains real-time dynamic characteristics are as follows:

[0031] S11: Perform time synchronization processing on multi-source data, aligning data from different sensors and acquisition sources to the same time axis, so that each data corresponds to the other at the same point in time, and obtain time-synchronized data.

[0032] The specific implementation method of step S11 can be selected according to the system hardware configuration of the launch vehicle, the acquisition accuracy requirements and real-time requirements, including but not limited to conventional time synchronization processing methods, such as: unified timestamp alignment, linear interpolation synchronization, and clock synchronization correction.

[0033] S12: Perform outlier removal on the time-synchronized data. After removing invalid data such as jumps, missing data, and data exceeding the reasonable range, the data after removing outliers is obtained (i.e., the retained valid real data).

[0034] The specific implementation method of step S12 can be selected according to the system hardware configuration of the launch vehicle, the acquisition accuracy requirements and real-time requirements, including but not limited to conventional outlier removal methods, such as threshold judgment, sliding window verification and data rationality comparison.

[0035] S13: After removing outliers, perform data denoising to reduce the impact of random noise on the data's authenticity, and obtain the denoised data.

[0036] The specific implementation method of step S13 can be selected according to the system hardware configuration of the launch vehicle, the acquisition accuracy requirements and real-time requirements, including but not limited to conventional data denoising methods, such as: moving average filtering, median filtering and first-order hysteresis filtering.

[0037] S14: Normalize the denoised data to map data with different dimensions and numerical ranges to the same numerical interval. After unifying the dimensions, a standardized flight dataset is obtained.

[0038] The specific implementation method of step S14 can be selected according to the system hardware configuration of the launch vehicle, the acquisition accuracy requirements and real-time requirements, including but not limited to conventional data normalization methods, such as linear normalization and interval mapping normalization.

[0039] S2: Based on the multi-dimensional coupling relationship between inertial measurement data, flight status data and sensor health data in the standardized flight dataset, the adaptation weight is dynamically calculated to obtain fused attitude feature data that adapts to the real-time flight status of the launch vehicle.

[0040] Furthermore, the sub-steps of step S2 are as follows:

[0041] S21: Extracting from standardized flight datasets The data includes inertial measurement data, flight status data, and sensor health data at any given time. The inertial measurement data includes at least the attitude information, angular velocity, and linear acceleration collected by various sensors. The flight status data includes at least the flight phase information and trajectory tilt angle. The sensor health data includes at least the sensor health coefficients and noise variance of various sensors.

[0042] Furthermore, the specific implementation method of step S21 can be selected according to the system hardware configuration of the launch vehicle, the data reading accuracy requirements and real-time requirements, including but not limited to the reading method by timestamp index, the reading method by data type field filtering or the reading method by storage address address.

[0043] S22: Calculate the multi-dimensional coupling dynamic fusion weight based on the multi-dimensional coupling relationship among inertial measurement data, flight status data, and sensor health data.

[0044] Specifically, the multi-dimensional coupling relationship among inertial measurement data, flight status data, and sensor health data refers to the inherent objective correlation between these three data during the flight of a launch vehicle. This correlation is obtained through conventional methods in this field, such as historical flight data statistical analysis, ground test data fitting, and empirical modeling.

[0045] Furthermore, the expression for the multi-dimensional coupled dynamic fusion weights is:

[0046] ;

[0047] Furthermore, the multi-dimensional coupled dynamic fusion weights satisfy the normalization condition:

[0048] ;

[0049] in, for Time of the first Multi-dimensional coupling dynamic fusion weights for inertial measurement-like data. , This represents the total number of categories of inertial measurement data. for Time of the first Noise variance of sensor-like devices; for Time of the first Sensor health coefficient for this type of sensor; for Time of the first Noise variance of sensor-like devices ; for Time of the first Sensor health coefficient for this type of sensor; According to Time of the first The sensor health coefficient and noise variance of the sensor type are determined. The coupling coefficient between the health status and noise at any given moment; According to The attitude information and trajectory inclination angle are determined at any given moment. Flight condition adaptability coefficient at any given time.

[0050] Specifically, When the value is 3, the inertial measurement data includes three categories: attitude information, angular velocity, and linear acceleration.

[0051] The specific method for determining this can be selected based on the actual application scenario, for example, it can be based on... Time of the first The sensor health coefficient and noise variance of the sensor type are determined by numerical fitting, empirical calculation formulas, or table lookup.

[0052] The specific method for determining this can be selected based on the actual application scenario, for example, it can be based on... The moment attitude information and trajectory inclination angle are determined through experimental fitting, empirical assignment, or table lookup.

[0053] The multi-dimensional coupled dynamic fusion weights obtained by dynamically weighting inertial measurement data, flight status data, and sensor health data can be adaptively adjusted in real time according to sensor health, noise variance, and flight phase information. Multi-factor collaborative correction is achieved through the health and noise coupling coefficient and the flight condition adaptation coefficient, while satisfying the weight normalization condition. Compared with the traditional fixed weight and simple weighting method, this significantly improves the accuracy and robustness of inertial measurement results, and can better adapt to the complex operating conditions of the launch vehicle throughout the entire flight cycle. Moreover, the algorithm is simple and easy to implement in engineering.

[0054] S23: Based on multi-dimensional coupled dynamic fusion weights, inertial measurement data is weighted and fused to obtain... Fusion pose feature data at different times .

[0055] Furthermore, Fusion pose feature data at different times The expression is:

[0056] ;

[0057] in, for Time of the first Multi-dimensional coupling dynamic fusion weights for inertial measurement-like data. , This represents the total number of categories of inertial measurement data. for Time of the first Standardized values ​​for inertial measurement data.

[0058] Specifically, It is the first Inertial measurement data in The value obtained after standardizing the original measurement data at time t. The standardization process can be, but is not limited to, extreme value standardization and mean-variance standardization.

[0059] S3: Construct a dynamic adaptive interpretation threshold based on the fused attitude feature data, as well as the flight status data and environmental perception data in the standardized flight dataset, and complete the intelligent interpretation of the fused attitude feature data based on the dynamic adaptive interpretation threshold.

[0060] Furthermore, the sub-steps of step S3 are as follows:

[0061] S31: Extracting from standardized flight datasets The flight status data and environmental perception data corresponding to each moment; among them, the flight status data includes at least flight phase information, and the environmental perception data includes at least dynamic pressure and wind speed.

[0062] Furthermore, a timestamp matching method is used to extract data from the standardized flight dataset. The flight status data and environmental perception data corresponding to each moment, but not limited to the timestamp matching method.

[0063] Specifically, the timestamp matching method can be either exact timestamp matching or nearest neighbor timestamp matching, but it is not limited to either exact timestamp matching or nearest neighbor timestamp matching.

[0064] S32: Obtain multi-frame continuously fused attitude feature data within a preset sliding time window, and calculate based on the multi-frame continuously fused attitude feature data... The fluctuation variance corresponding to the attitude feature data is fused at all times.

[0065] Specifically, the length of the preset sliding time window and the specific calculation method for the fluctuation variance are determined according to the engineering application requirements. This application preferably adopts the sliding variance calculation method, specifically: selecting a sliding time window of length L, extracting all fused attitude feature data within the sliding time window, calculating its sample variance, and using this sample variance as the current... The fluctuation variance corresponding to the attitude feature data is fused at all times.

[0066] S33: Based on The flight status data, environmental perception data, fused attitude feature data, and fluctuation variance corresponding to each moment are used to construct... Dynamic adaptive interpretation threshold at any time .

[0067] Furthermore, Dynamic adaptive interpretation threshold at any time The expression is:

[0068] ;

[0069] in, The attitude baseline threshold is determined by the launch vehicle design specifications; According to The flight phase information is determined in the flight status data corresponding to the given time. Flight phase coefficient at any given moment; For the reason Dynamic pressure at any moment and wind speed The fitted The coupling coefficient between dynamic pressure and wind speed at any given time; For the reason Fusion pose feature data at different times and volatility variance Jointly determined The correlation correction coefficient of the fused data at any given time.

[0070] Specifically, for different flight phases, corresponding flight phase coefficient values ​​are pre-set to form a flight phase coefficient table, based on... The flight phase information in the flight status data corresponding to the given time is used to query the flight phase coefficient table to obtain... . This is used to enable the dynamic adaptive interpretation threshold to adapt to the attitude change characteristics of different flight phases, avoiding the problem of inaccurate interpretation of a single threshold in both the violent maneuvering phase and the smooth taxiing phase.

[0071] The values ​​of the flight stage coefficients corresponding to each flight stage are determined comprehensively based on the model design specifications of the launch vehicle, the statistical regularity of telemetry historical data, and engineering experience. They can be updated and improved in a timely manner according to model iteration and flight test data accumulation, and are pre-fixed in the flight stage coefficient table.

[0072] The flight stage coefficients vary depending on the flight stage. For example, the flight stage coefficient for takeoff is 1.2, for boost stage it is 1.0, for stage separation stage it is 1.5, for taxiing stage it is 0.8, and for orbit insertion stage it is 0.6.

[0073] Among them, dynamic pressure and wind speed The fitting method is determined based on the characteristics of the launch vehicle model and the engineering application requirements. This application preferably adopts linear fitting or piecewise fitting.

[0074] Dynamic pressure and wind speed The fitting relationship is determined based on the flight test data, aerodynamic simulation data, and engineering experience of the launch vehicle, and can be updated and optimized as data accumulates. This application preferably uses the calculated... The value of the coupling coefficient between dynamic pressure and wind speed at any given time is constrained to be within the range of 0.7 to 1.3.

[0075] By using the coupling coefficient between dynamic pressure and wind speed, the intensity of environmental disturbance is incorporated into the calculation of the dynamic adaptive judgment threshold in real time. This allows the dynamic adaptive judgment threshold to be adjusted according to the magnitude of external disturbances, thereby improving the judgment stability and anti-interference capability under strong disturbance conditions.

[0076] Among them, the fusion of pose feature data and volatility variance The joint calculation method is determined based on the data fluctuation characteristics and engineering interpretation requirements. This application preferably adopts a linear correction or piecewise correction method.

[0077] Fusion of pose feature data and volatility variance The correction relationship is determined based on the statistical regularity of telemetry historical data, the fluctuation characteristics of attitude data, and engineering experience. It can be updated and optimized as experimental data accumulates. This application preferably uses the calculated... The value of the linkage correction coefficient for the fusion data at any given time is constrained to be within the range of 0.85 to 1.15.

[0078] The real-time data linkage correction coefficient can finely adjust the dynamic adaptive interpretation threshold according to the stability and volatility of the fused data itself, so that the dynamic adaptive interpretation threshold is deeply matched with the current data quality, further reducing the false alarm rate and improving the interpretation accuracy.

[0079] Compared to traditional static thresholds and single-factor dynamic thresholds, the dynamic adaptive interpretation threshold of this application introduces a triple-coupling correction of flight phase information, environmental disturbances, and fused data fluctuations. This enables the threshold to be adaptively adjusted in real time according to flight conditions, environmental conditions, and data quality, thereby improving the matching degree between the threshold and the actual flight state, reducing the false and false judgment rates, and enhancing the reliability and robustness of interpretation.

[0080] S34: Will Fusion pose feature data at different times and Dynamic adaptive interpretation threshold at any time To make a comparison, if ,determination Normal; if ,determination An anomaly is identified and an anomaly information is marked. The anomaly information includes at least the anomaly time, the anomaly value, and the flight phase and environmental conditions corresponding to the anomaly time.

[0081] Specifically, the abnormal information is used to provide a basis for subsequent attitude stability analysis and fault tracing.

[0082] Furthermore, the flight phase and environmental condition information should include at least the following: flight phase name, dynamic pressure value, and wind speed value.

[0083] S4: By integrating attitude feature data, dynamic adaptive interpretation thresholds, and dynamic correction coefficients of real-time attitude change rate, the attitude stability margin value is calculated, and based on the attitude stability margin value, the quantitative classification and determination of the flight attitude stability of the launch vehicle is realized.

[0084] Furthermore, the sub-steps of step S4 are as follows:

[0085] S41: Based on Fusion pose feature data at different times Dynamic adaptive interpretation threshold And the dynamic correction coefficient of the real-time attitude change rate, to obtain The attitude stability margin value at time t.

[0086] Furthermore, The expression for the attitude stability margin at time t is:

[0087] ;

[0088] in, for The attitude stability margin value at time t. The value range is [0,1]; for Fusion pose feature data at each moment; for Dynamic adaptive interpretation threshold at any given time; for The dynamic correction coefficient for the real-time rate of change of attitude at any given moment. This is a dynamic correction coefficient for attitude trend; for The amount of change in the fused pose feature data at any given time. , This is the fused pose feature data from the previous sampling time. The system sampling time interval; The duration of the attitude change.

[0089] Specifically, The closer the value is to 1, the better the static stability and trend stability. The closer it is to 0, the greater the risk of static over-limit or trend deterioration in the attitude.

[0090] in, The value is determined based on the following: the greater the real-time rate of attitude change and the longer the duration of attitude change, the better. The smaller the value of , the smaller the real-time rate of attitude change and the shorter the duration of attitude change, then The closer the value is to 1.0, the better. In this embodiment, The preferred value is when the posture does not change significantly or continuously. The threshold is set to 1.0; 0.8 is used when there is a slight but continuous change in attitude, 0.6 when there is a significant but continuous change in attitude, and 0.5 when there is a substantial but continuous change in attitude. The determination of the degree of attitude change is based on pre-set thresholds for slight change, significant change, and substantial change, determined by comparing the absolute value of the real-time attitude change rate with each preset threshold: if the absolute value of the real-time attitude change rate is less than the slight change threshold, it is considered no significant but continuous change; if the absolute value of the real-time attitude change rate is greater than or equal to the slight change threshold but less than the significant change threshold, it is considered a slight but continuous change; if the absolute value of the real-time attitude change rate is greater than or equal to the significant change threshold but less than the substantial change threshold, it is considered a significant but continuous change; if the absolute value of the real-time attitude change rate is greater than or equal to the substantial change threshold, it is considered a substantial but continuous change. The specific values ​​of the slight change threshold, significant change threshold, and substantial change threshold can be pre-calibrated based on the launch vehicle model, flight conditions, inertial navigation system measurement accuracy, and data from previous flight tests. These values ​​can be dynamically updated and adaptively corrected during flight based on the current real-time flight status and attitude change characteristics to improve the accuracy and adaptability of the attitude change degree determination. In this embodiment, the threshold for slight change is preferably 0.05° / s, the threshold for significant change is preferably 0.2° / s, and the threshold for substantial change is preferably 0.5° / s.

[0091] in, The specific value can be set according to the system hardware sampling rate. In this application, it is preferably 0.01s~0.05s.

[0092] in, The specific value can be determined in real time based on the actual attitude change process. Preferably, in this application, timing begins when the absolute value of the attitude change rate is greater than a preset slight change threshold, and stops when the attitude change rate falls below the slight change threshold. The resulting timing length is the [time value]. .

[0093] This application The attitude stability margin value at any given time is characterized by the ratio of the fused attitude feature data to the dynamic adaptive interpretation threshold, representing the degree of static stability. A dynamic correction coefficient for the real-time attitude change rate is introduced to reflect the impact of the speed and duration of attitude change. Compared with the existing method of using only static thresholds for stability determination, this method can simultaneously reflect the magnitude of static attitude deviation and dynamic change trend, improving the accuracy and timeliness of stability determination. It avoids misjudgment and omission caused by ignoring attitude change trend, and is more suitable for the dynamic change scenarios during the launch vehicle flight process.

[0094] S42: Based on preset hierarchical rules, The attitude stability margin value at any given time is analyzed to determine the trend and quantitative classification of the current flight attitude stability of the launch vehicle.

[0095] Furthermore, the preset grading rules are as follows:

[0096] when When the value is greater than or equal to the first grade boundary threshold, the trend-quantified classification of the current flight attitude stability of the launch vehicle is determined to be the first grade, which represents both static and trend stability.

[0097] When the second grade boundary threshold ≤ When the threshold of the first classification is reached, the trend of the stability of the current flight attitude of the launch vehicle is classified as the second level. The second level indicates static stability, but the trend needs to be observed.

[0098] When the third-level boundary threshold is ≤ When the threshold of the second grading is less than the threshold, it is judged as the third level. The third level indicates that the static situation is basically stable, but the trend is slightly deteriorating.

[0099] when When the threshold of the third level is exceeded, it is determined to be at the fourth level, which indicates a high risk in static / trend conditions.

[0100] Among them, the first level boundary threshold > the second level boundary threshold > the third level boundary threshold; for The attitude stability margin value at time t.

[0101] Specifically, the values ​​of the first, second, and third grade boundary thresholds can be determined by comprehensive calibration based on the launch vehicle flight test data, attitude control accuracy requirements, and system interpretation strategies. They can also be adjusted adaptively according to actual flight conditions. In this application, the preferred values ​​are: the first grade boundary threshold is 0.6, the second grade boundary threshold is 0.4, and the third grade boundary threshold is 0.2.

[0102] The preset grading rules in this application enable trend-based quantitative control of the flight attitude stability of launch vehicles. Compared with the traditional grading method based solely on static thresholds, this method can more accurately reflect short-term attitude fluctuations and gradual trends, thereby improving the comprehensiveness and reliability of stability assessment.

[0103] S43: Based on the preset matching execution rules, determine and execute the attitude stability early warning and control strategy corresponding to the trend quantification classification of the current flight attitude stability of the launch vehicle.

[0104] Furthermore, the preset matching execution rules are pre-defined rules that correspond one-to-one with each trend-based quantitative level.

[0105] The preset matching execution rules are as follows:

[0106] When the trend-based quantitative classification of the current flight attitude stability of the launch vehicle is at level one (both static and trend stability), the corresponding attitude stability warning and control strategy is: do not trigger the warning, and maintain normal flight attitude interpretation and monitoring status.

[0107] When the trend-based quantitative classification of the launch vehicle's current flight attitude stability is at level two (static stability, trend to be observed), the corresponding attitude stability early warning and control strategy is: trigger a low-level warning and continuously track the status. The trend of attitude stability margin values ​​at different times and subsequently is studied, with a focus on the persistence of attitude changes.

[0108] When the trend quantification classification of the current flight attitude stability of the launch vehicle is level three (statically basically stable, with a slight deterioration in trend), the corresponding attitude stability warning and control strategy is: trigger a medium-level warning to remind operators to focus on monitoring attitude changes and assess the risk of attitude deterioration.

[0109] When the current flight attitude stability trend of the launch vehicle is classified as Level 4 (static / high-risk trend), the corresponding attitude stability warning and control strategy is to trigger a high-level warning so that relevant systems or operators can take timely attitude adjustment and fault diagnosis measures to ensure flight safety.

[0110] Furthermore, after step S4, step S5 is also included: based on the attitude stability margin values ​​at multiple consecutive moments, the instability trend of the launch vehicle's flight attitude in future periods is predicted and an early warning is given, so as to realize the early prediction and active control of attitude instability, and further improve the safety and controllability of the launch vehicle's flight attitude.

[0111] Specifically, the future time period is based on the current time. Starting from a certain point, a prediction time interval is projected into the future; the specific value of this prediction time interval is determined based on the actual flight conditions, system response speed, and prediction accuracy requirements, and is preferably 0.5s to 3s in this application.

[0112] Furthermore, the sub-steps of step S5 are as follows:

[0113] S51: In terms of time (i.e., the current time) is the end time, and the following is selected: A continuous time period of preset duration is used to calculate the attitude stability margin value at each moment within the continuous time period. All attitude stability margin values ​​are arranged in chronological order to form a sequence of attitude stability margin values ​​for multiple consecutive moments.

[0114] Specifically, the attitude stability margin value at each moment within the continuous time period can be calculated using the expression for the attitude stability margin value, so it will not be elaborated further here.

[0115] S52: The sequence of attitude stability margin values ​​over multiple consecutive time periods and... The current flight condition data at any given moment is input into the preset instability trend early warning model, which then predicts the attitude stability margin value of the launch vehicle at each moment in the future.

[0116] Specifically, an instability trend early warning model is obtained by pre-collecting and utilizing historical flight data of the launch vehicle, historical attitude stability margin value sample data, and corresponding flight condition characteristics. The specific type of the preset instability trend early warning model can be determined according to the actual prediction accuracy requirements and system hardware conditions. In this application, a deep learning prediction model or a time series prediction model is preferred.

[0117] S53: Based on a preset warning threshold, the attitude stability margin value of the launch vehicle at various times within a future period is judged; if the attitude stability margin value at any time within a future period is lower than the preset warning threshold, a first-level advanced warning is triggered; if the attitude stability margin value within a future period decreases continuously for a preset number of consecutive decreases over time, and the magnitude of each decrease exceeds a preset change magnitude threshold, a second-level advanced warning is triggered; if the attitude stability margin value at any time within a future period is lower than the preset warning threshold, and the attitude stability margin value within a future period decreases continuously for a preset number of consecutive decreases over time, and the magnitude of each decrease exceeds a preset change magnitude threshold, a third-level advanced warning is triggered.

[0118] Specifically, the preset warning threshold is determined based on the launch vehicle model, flight phase information, attitude control accuracy requirements, and current flight conditions. The preset warning threshold can be a fixed threshold or a dynamic threshold that is adaptively updated according to the flight conditions and attitude stability change trends.

[0119] As an example, when the preset warning threshold is a fixed threshold, the preferred value in this application is 0.3.

[0120] As another embodiment, when the preset warning threshold adopts a dynamic threshold, the expression of the preset warning threshold is: preset warning threshold = reference threshold × flight condition coefficient × attitude stability change coefficient, where the reference threshold is a fixed value pre-calibrated based on the model test data of the launch vehicle and the flight control requirements; the flight condition coefficient is a dynamic value, which is calculated in real time from the current flight condition; the attitude stability change coefficient is dynamically updated according to the attitude stability change trend.

[0121] The preset number of consecutive decreases can be set according to the attitude control response speed, data sampling frequency and flight control accuracy requirements. In this application, it is preferred that the preset number of consecutive decreases is 3 or more.

[0122] The preset change amplitude threshold is determined based on the launch vehicle attitude control accuracy requirements, data sampling frequency, and attitude stability change tolerance. In this application, the preferred value of the preset change amplitude threshold is 0.02.

[0123] Furthermore, the execution content corresponding to the first-level advanced warning is as follows: issue a warning signal indicating deteriorating attitude stability, and record the current flight status and attitude stability margin value data; the execution content corresponding to the second-level advanced warning is as follows: issue an attitude stability anomaly warning signal, simultaneously record flight data and initiate real-time monitoring of attitude control parameters; the execution content corresponding to the third-level advanced warning is as follows: issue a high-risk attitude instability warning signal, simultaneously record all flight data, initiate emergency attitude control intervention preparation, and remind the operator to intervene and handle the situation.

[0124] The beneficial effects achieved by this application are as follows:

[0125] (1) The intelligent interpretation and attitude stability analysis method and system for launch vehicle flight data of this application can improve the accuracy and robustness of multi-source data fusion. Specifically, it can dynamically calculate the multi-dimensional coupling dynamic fusion weight based on the multi-dimensional coupling relationship of inertial measurement data, flight status data and sensor health data. The multi-dimensional coupling dynamic fusion weight can be adaptively adjusted according to sensor health, noise variance and real-time flight conditions. The fused attitude feature data obtained is adapted to the complex conditions of the entire flight cycle of the launch vehicle and the data accuracy is higher.

[0126] (2) The intelligent interpretation and attitude stability analysis method and system for launch vehicle flight data of this application can improve the accuracy of attitude interpretation and anti-interference ability. Specifically, it can construct a dynamic adaptive interpretation threshold that integrates flight phase, environmental disturbance and data fluctuation triple coupling correction. The dynamic adaptive interpretation threshold can be adjusted in real time according to flight conditions, environmental conditions and data quality, effectively reducing the misjudgment and omission rate under strong disturbance conditions.

[0127] (3) The intelligent interpretation and attitude stability analysis method and system of launch vehicle flight data of this application can realize the comprehensive quantitative judgment of attitude stability. Specifically, it integrates the static attitude deviation and dynamic change trend to calculate the attitude stability margin value, establishes trend-based quantitative grading rules, realizes the static and trend dual-dimensional judgment of attitude stability, accurately reflects the short-term fluctuation and gradual trend of launch vehicle attitude, and avoids the judgment deviation caused by ignoring the trend.

[0128] (4) The intelligent interpretation and attitude stability analysis method and system for launch vehicle flight data of this application can realize advanced early warning and refined control of attitude instability. Specifically, based on the attitude stability margin value at multiple consecutive moments, the instability trend prediction and early warning of the flight attitude of the launch vehicle in the future period can be made, realizing advanced prediction and active control of attitude instability. At the same time, a dual control system is formed by combining the preset hierarchical rules and the corresponding matching execution rules to realize refined and active control of attitude, which greatly improves the safety and controllability of the launch vehicle flight attitude.

[0129] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the scope of protection of this application is intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application. Obviously, those skilled in the art can make various alterations and variations to this application without departing from the spirit and scope of this application. Thus, if these modifications and variations of this application fall within the scope of protection of this application and its equivalents, this application also intends to include these modifications and variations.

Claims

1. A method for intelligent interpretation and attitude stability analysis of launch vehicle flight data, characterized in that, Includes the following steps: S1: Acquire multi-source data during the launch vehicle's flight process, and perform standardized preprocessing on the multi-source data to obtain a standardized flight dataset that is timestamped, dimensionally consistent, and retains real-time dynamic characteristics; wherein, the multi-source data includes at least: inertial measurement data, flight status data, environmental perception data, and sensor health data; S2: Based on the multi-dimensional coupling relationship between inertial measurement data, flight status data and sensor health data in the standardized flight dataset, the adaptation weight is dynamically calculated to obtain fused attitude feature data that adapts to the real-time flight status of the launch vehicle. S3: Construct a dynamic adaptive interpretation threshold based on the fused attitude feature data, as well as the flight status data and environmental perception data in the standardized flight dataset, and complete the intelligent interpretation of the fused attitude feature data based on the dynamic adaptive interpretation threshold. S4: By integrating attitude feature data, dynamic adaptive interpretation threshold, and dynamic correction coefficient of real-time attitude change rate, the attitude stability margin value is calculated, and based on the attitude stability margin value, the quantitative classification and judgment of the flight attitude stability of the launch vehicle is realized. The sub-steps of step S4 are as follows: S41: Based on Fusion pose feature data at different times Dynamic adaptive interpretation threshold And the dynamic correction coefficient of the real-time attitude change rate, to obtain The attitude stability margin value at time t; S42: Based on preset hierarchical rules, The attitude stability margin value at any given time is analyzed to determine the trend and quantitative classification of the current flight attitude stability of the launch vehicle. S43: Based on the preset matching execution rules, determine and execute the attitude stability early warning and control strategy corresponding to the trend quantification classification of the current flight attitude stability of the launch vehicle; The preset grading rules are as follows: when When the value is greater than or equal to the first grade boundary threshold, the trend-quantified classification of the current flight attitude stability of the launch vehicle is determined to be the first grade, which represents both static and trend stability. When the second grade boundary threshold ≤ When the first level boundary threshold is reached, the trend quantification classification of the current flight attitude stability of the launch vehicle is determined to be the second level. The second level indicates static stability, but the trend needs to be observed. When the third-level boundary threshold is ≤ When the value is less than the second-level threshold, it is classified as the third level. The third level indicates that the static situation is basically stable, but the trend is slightly deteriorating. when When the threshold of the third grade is less than the third grade boundary, it is judged as the fourth grade, which indicates high static / trend risk; Among them, the first level boundary threshold > the second level boundary threshold > the third level boundary threshold; for The attitude stability margin value at time t.

2. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 1, characterized in that, The sub-steps for standardizing and preprocessing multi-source data to obtain a standardized flight dataset that is timestamped, has uniform units, and retains real-time dynamic characteristics are as follows: S11: Perform time synchronization processing on multi-source data, aligning data from different sensors and acquisition sources to the same time axis, so that each data corresponds to each other at the same point in time, and obtain time-synchronized data; S12: Perform outlier removal on the time-synchronized data. After removing invalid data such as jumps, missing data, and data exceeding the reasonable range, the data after removing outliers is obtained. S13: Perform data denoising on the data after removing outliers to reduce the impact of random noise on the data authenticity, and then obtain the denoised data. S14: Normalize the denoised data to map data with different dimensions and numerical ranges to the same numerical interval. After unifying the dimensions, a standardized flight dataset is obtained.

3. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 1, characterized in that, The sub-steps of step S2 are: S21: Extracting from standardized flight datasets The data includes inertial measurement data, flight status data, and sensor health data at any given time. The inertial measurement data includes at least the attitude information, angular velocity, and linear acceleration collected by various sensors. The flight status data includes at least the flight phase information and trajectory tilt angle. The sensor health data includes at least the sensor health coefficients and noise variance of various sensors. S22: Calculate the multi-dimensional coupling dynamic fusion weight based on the multi-dimensional coupling relationship among inertial measurement data, flight status data, and sensor health data; S23: Based on multi-dimensional coupled dynamic fusion weights, inertial measurement data is weighted and fused to obtain... Fusion pose feature data at different times .

4. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 3, characterized in that, Fusion pose feature data at different times The expression is: ； in, for Time of the first Multi-dimensional coupling dynamic fusion weights for inertial measurement-like data. , This represents the total number of categories of inertial measurement data. for Time of the first Standardized values ​​for inertial measurement data.

5. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 1, characterized in that, The sub-steps of step S3 are as follows: S31: Extracting from standardized flight datasets The flight status data and environmental perception data corresponding to each moment; among them, the flight status data includes at least flight phase information, and the environmental perception data includes at least dynamic pressure and wind speed; S32: Obtain multi-frame continuously fused attitude feature data within a preset sliding time window, and calculate based on the multi-frame continuously fused attitude feature data... The fluctuation variance corresponding to the fused attitude feature data at any time; S33: Based on The flight status data, environmental perception data, fused attitude feature data, and fluctuation variance corresponding to each moment are used to construct... Dynamic adaptive interpretation threshold at any time ; S34: Will Fusion pose feature data at different times and Dynamic adaptive interpretation threshold at any time To make a comparison, if ,determination Normal; if ,determination An anomaly is identified and an anomaly information is marked. The anomaly information includes at least the anomaly time, the anomaly value, and the flight phase and environmental conditions corresponding to the anomaly time.

6. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 1, characterized in that, The first-level boundary threshold is 0.6, the second-level boundary threshold is 0.4, and the third-level boundary threshold is 0.

2.

7. The intelligent interpretation and attitude stability analysis method for launch vehicle flight data according to claim 1, characterized in that, The default matching rules are: When the trend-quantified classification of the current flight attitude stability of the launch vehicle is at the first level, the corresponding attitude stability warning and control strategy is: do not trigger the warning, and maintain the normal flight attitude interpretation and monitoring status. When the trend-based quantitative classification of the launch vehicle's current flight attitude stability is at level two, the corresponding attitude stability early warning and control strategy is: trigger a low-level warning and continuously track the status. The trend of attitude stability margin values ​​at time points and subsequently, with a focus on the persistence of attitude changes; When the trend-quantified classification of the current flight attitude stability of the launch vehicle is level three, the corresponding attitude stability warning and control strategy is: trigger a medium-level warning to remind operators to focus on monitoring attitude changes and assess the risk of attitude deterioration. When the trend-quantified classification of the current flight attitude stability of the launch vehicle is at level four, the corresponding attitude stability early warning and control strategy is to trigger a high-level early warning so that relevant systems or operators can take timely attitude adjustment and fault diagnosis measures to ensure flight safety.

8. A system for intelligent interpretation of launch vehicle flight data and attitude stability analysis, characterized in that, include: Processor and memory; The memory stores a computer program, which, when executed by the processor, implements the intelligent interpretation and attitude stability analysis method for launch vehicle flight data as described in any one of claims 1-7.

Images