Long-endurance aircraft task health degree assessment method

Abstract

The invention discloses a long-endurance aircraft task health degree assessment method, and relates to the technical field of digital simulation, and the method comprises the following components: S1, a data acquisition and trajectory difference quantification step, S2, a data forward processing step, S3, a data standardization processing step, S4, an index weight determination step, and S5, a TOPSIS health degree assessment and sorting step. According to the method, ideal flight state data and multiple groups of actual flight state data of a flight task are collected, the difference between the actual state and the ideal state is quantified by using a dynamic time warping distance, and data forward processing, standardization processing and fusion of an objective weight determined by an entropy weight method and a user-defined weight are combined; the method achieves the comprehensive and precise evaluation of the health state of the long-endurance aircraft task, not only considers the multi-dimensional state indexes of the aircraft, but also improves the evaluation accuracy and adaptability through the dynamic adjustment of the weight and the standardization range.

Description

Technical Field

[0001] This invention relates to the field of digital simulation technology, specifically to a method for assessing the health of long-endurance aircraft missions. Background Technology

[0002] With the rapid development of aerospace technology, long-endurance aircraft have shown broad application prospects in many fields such as military reconnaissance, environmental monitoring, and communication relay. However, during the execution of complex missions, the health status of long-endurance aircraft is affected by a variety of factors, including but not limited to environmental changes, equipment aging, and operational errors. These factors may pose a threat to the performance and safety of the aircraft.

[0003] Traditional methods for assessing the health status of long-endurance aircraft often rely on single indicators or simple weighting, which have significant limitations. First, single-indicator assessments cannot comprehensively reflect the overall health status of the aircraft and easily overlook changes in certain key indicators. Second, while simple weighting considers multiple indicators, the weight allocation is often based on experience or fixed values, lacking dynamic adjustment capabilities and making it difficult to adapt to the assessment needs of different mission phases and environmental conditions. In addition, traditional methods often lack effective data fusion and preprocessing mechanisms when dealing with multi-source heterogeneous data, leading to inaccurate and unreliable assessment results. Finally, traditional assessment methods lack real-time mission optimization and self-optimization capabilities, failing to adjust flight parameters or optimize assessment models in a timely manner based on assessment results, thus limiting further improvements in aircraft performance.

[0004] In view of the shortcomings of traditional long-endurance aircraft health status assessment methods, this invention proposes a method for assessing the mission health of long-endurance aircraft, which is of particular importance. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a method for assessing the mission health of long-endurance aircraft. This method collects ideal flight state data and multiple sets of actual flight state data, quantifies the difference between the actual and ideal states using dynamic time warping distance, and combines data forwarding, standardization, and the fusion of objective weights determined by entropy weighting with user-defined weights to achieve a comprehensive assessment of the aircraft's health status. Furthermore, this method possesses real-time mission optimization and self-optimization capabilities, enabling timely adjustments to flight parameters and optimization of the assessment model based on the assessment results. This improves the accuracy and adaptability of the assessment, providing strong support for the safe and efficient execution of missions by long-endurance aircraft.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: a method for assessing the mission health of a long-endurance aircraft, the specific steps of which are as follows: S1. Data Acquisition and Trajectory Difference Quantification Steps: Acquire ideal flight state data and multiple sets of actual flight state data for the flight mission. Each set of data includes 12 evaluation indicators: longitude, latitude, altitude, pitch angle, roll angle, yaw angle, linear velocity U, linear velocity V, linear velocity W, pitch angular velocity, roll angular velocity, and yaw angular velocity. Through dynamic time-normalized distance calculation, quantify the differences between each indicator in each set of actual flight state data and the corresponding indicator in the ideal flight state, forming an actual-ideal trajectory difference matrix. S2. Data forwarding process steps: Determine the type of each column of indicators in the difference matrix, adopt corresponding processing methods for different types of indicators, and convert all indicators into a positive evaluation standard where the larger the indicator value, the closer the flight status is to the ideal state, thus obtaining a forwarding data matrix. S3. Data standardization processing steps: Standardize the positive data matrix to eliminate the influence of the dimensions of each indicator, so that all indicators are in a uniform and comparable range, and obtain a standardized data matrix. S4. Steps for determining indicator weights: The objective weight vector is calculated using the entropy weight method. Combined with the user-inputted custom weight vector based on task characteristics, the proportion of the two types of weights is adjusted by the fusion coefficient to obtain the final weight vector. Among them, the user-inputted custom weights need to be validated for legality. If the validation conditions are not met, automatic normalization is performed. S5, TOPSIS Health Assessment and Ranking Steps: Construct a weighted standardized matrix based on the standardized data matrix and the final weight vector to determine the positive and negative ideal solutions; calculate the distance from each set of actual flight data to the positive and negative ideal solutions, and obtain the comprehensive health score through the distance ratio; sort the comprehensive scores from high to low and output the health ranking results.

[0007] Furthermore, the specific steps for data acquisition and trajectory difference quantification in S1 are as follows: Data acquisition phase: Ideal flight state data is acquired from the aircraft mission planning system. This data represents the optimal state parameters preset for the entire process, and the sampling frequency is consistent with the actual flight data. Actual flight state data is acquired through multiple types of sensors on board the aircraft. Among them, longitude and latitude data are acquired by a high-precision GNSS positioning module, altitude data is acquired by the fusion of laser altimeter and pressure sensor, pitch angle, roll angle, and yaw angle are calculated by a combination of three-axis gyroscope and three-axis accelerometer, linear velocity U / V / W is acquired by the fusion of Doppler radar and inertial measurement unit (IMU) data, and pitch angular velocity, roll angular velocity, and yaw angular velocity are directly output by the three-axis gyroscope. Data preprocessing stage: outlier removal and data alignment are performed on the collected actual flight data; DTW distance calculation stage: For each set of preprocessed actual flight data, DTW distance is calculated with the ideal flight data according to the index dimensions. Before the calculation, the length of the two types of data is unified. In the final output difference matrix, each row corresponds to the difference value of 12 indexes of a set of actual flight data, and each column corresponds to the difference value of the same index in different sets of actual data.

[0008] Furthermore, the specific implementation process of the data forwarding process in S2 includes: Indicator type determination: The type of each indicator is determined by the preset indicator attribute library. Among the 12 indicators corresponding to DTW distance, the DTW distances for longitude, latitude, altitude, pitch angle, roll angle, yaw angle, linear velocity U / V / W, pitch angular velocity, roll angular velocity, and yaw angular velocity are all very small indicators. Only the altitude indicator is set as an interval indicator in a specific mission scenario. The indicator type needs to be selected before the evaluation according to the mission type. Forwarding of extremely small indicators: For extremely small indicators, the reverse mapping method is used. First, the maximum value of the indicator in the difference matrix is ​​calculated, and then each difference value is subtracted from the maximum value to obtain the forwarding result. This ensures that the larger the forwarded value, the smaller the corresponding original difference value. Forwarding of intermediate indicators: For height indicators set as intervals, first determine the optimal interval [L, R], then calculate the deviation of each difference value from the interval. If the difference value is within [L, R], the forwarding result is set to 1; if the difference value is less than L, the forwarding result is 1-(L-difference value) / (max(L-all difference values ​​less than L)); if the difference value is greater than R, the forwarding result is 1-(difference value-R) / (max(all difference values ​​greater than R-R)), ensuring that the larger the forwarded value, the closer the difference value is to the optimal interval. Forwarding verification: The rationality of the processed forwarding data matrix is ​​verified. The verification standard is that all forwarded values ​​are in the range of [0, 1]. If there are values ​​that exceed the range, a linear scaling method is used to map them to the interval of [0, 1] to avoid the interference of extreme values ​​on subsequent evaluation.

[0009] Furthermore, the specific implementation process of data standardization processing in S3 includes: Standardization method selection: The dynamic range standardization method is adopted. This method dynamically adjusts the standardization range in combination with the flight phase of the aircraft. The flight phase is divided into takeoff phase, climb phase, cruise phase and landing phase. Each phase corresponds to a different index fluctuation range threshold, which is obtained by statistical analysis of historical flight data. For example, the altitude fluctuation threshold in the cruise phase is ±0.5km and the altitude fluctuation threshold in the takeoff phase is ±2km. Standardization calculation: For each indicator in the positive data matrix, first determine the fluctuation range of the indicator based on the current flight phase, and then calculate the standardization result using the formula standardization value; if there is no historical data reference for the current flight phase, the mean-standard deviation standardization method is used by default, that is, standardized value = (positive value - mean of positive data) / standard deviation of positive data; Standardization correction: For negative values ​​or values ​​greater than 1 that appear after standardization, a truncation method is used to process them, setting negative values ​​to 0 and values ​​greater than 1 to 1, ensuring that the standardization results are all in the range [0, 1]. At the same time, the number of corrections is recorded. If the number of corrections for a certain indicator accounts for more than 5% of the total data volume, the user is prompted to check whether there are any abnormalities in the collection or positiveization process of that indicator.

[0010] Furthermore, in S4, the entropy weight method calculates the objective weight vector using a dynamic entropy weight calculation formula, specifically: the first... Weight of each indicator The calculation formula is: ,in For the first The dynamic difference coefficient of each indicator , The dynamic entropy value is calculated using the following formula: , For the first Group of data The weight of each indicator is calculated using the following formula: , , This is a minimum value, used to avoid... When it is 0 Meaningless; The number of sets of actual flight data. For the first The task relevance of each indicator, with a value range of [0, 1], is determined by the task scenario library. This is the correlation adjustment coefficient, with a value range of [0, 1], which is set by the user according to the evaluation accuracy requirements.

[0011] Furthermore, the validity verification and normalization processing of the user-defined weight vector in S4 adopts a weight-layered verification-correction technique. The specific implementation process includes: Validity layered verification: First, check whether the user-inputted custom weights are valid positive numbers. If there are non-positive numbers, the system automatically prompts the user to confirm the rationality of the non-positive number. If the user confirms retention, it is replaced with a preset reasonable minimum value. If the user chooses to modify, the input is received again. Then, calculate the sum of all weights and determine whether it meets the preset weight sum standard of the evaluation system. If it does not meet the standard, the normalization process is automatically triggered. Deviation correction normalization: First, extract the objective weights calculated by the entropy weight method as a reference benchmark. Then, compare the deviation between the user weights and the objective weights. User weights with deviations less than a preset threshold are directly retained. User weights with deviations exceeding the preset threshold are appropriately compressed and adjusted according to the deviation ratio. Finally, recalculate the sum of all adjusted weights and scale them to the standard weight sum range required by the evaluation system to generate a valid custom weight vector that is suitable for the task characteristics.

[0012] Furthermore, in step S4, the final weight vector obtained by adjusting the proportions of the two types of weights through the fusion coefficient adopts a dynamic fusion formula, specifically: final weight The calculation formula is: ,in The dynamic fusion coefficient is calculated using the following formula: ,in The initial fusion coefficient, The flight duration corresponding to the current assessment moment is obtained by the aircraft's timing module. The total task duration is obtained from the task planning system. It is a dynamic adjustment factor with a value range of [0, 0.5], determined by the task complexity.

[0013] Furthermore, in S5, the weighted normalization matrix constructed based on the normalized data matrix and the final weight vector adopts a time-sensitive weighting formula, specifically: the weighted normalization matrix... Group of data Elements of each indicator The calculation formula is: ,in For the first Group of data The standardized values ​​of each indicator are derived from the standardized data matrix in step 3. For the first The final weights of each indicator come from the final weight vector in step 4. For the first The timeliness coefficient of the data set.

[0014] Furthermore, in S5, the comprehensive health score obtained through the distance ratio adopts a trend-corrected scoring formula, specifically: The... Overall health score of the group data The calculation formula is: ,in For the first The Euclidean distance from the set of data to the positive ideal solution. For the first The Euclidean distance from the set of data to the negative ideal solution. For the first Health trend coefficient of group data This is the trend influence coefficient, with a value range of [0, 0.3], which is set by the user according to their risk preference.

[0015] Furthermore, the process also includes a supplementary step of task optimization based on the health ranking results output in step 5. The specific implementation process includes: Health Level Classification: Health is classified into five levels based on comprehensive score: Excellent, Good, Qualified, Warning, and Abnormal. Each level corresponds to a preset task handling strategy. Abnormal Data Source Tracing: For flight data at the warning or abnormal level, the influence weight of each indicator on health is analyzed, and the top three most influential key indicators are selected. Possible causes and inspection suggestions for the abnormality of key indicators are automatically output. Task Parameter Optimization Suggestions: Combining key indicators and health trend changes, adjustment suggestions for flight control system parameters are generated to adapt to different types of flight control task requirements. Evaluation Model Iteration: The evaluation results are periodically compared with actual task fault records to calculate the evaluation accuracy. If the accuracy does not meet the preset standard, the fusion coefficient or trend influence coefficient is automatically adjusted to achieve self-optimization of the evaluation model and improve the accuracy of subsequent evaluations.

[0016] Compared with existing technologies, this method for assessing the mission health of long-endurance aircraft has the following advantages: I. This method collects ideal flight state data and multiple sets of actual flight state data for flight missions, uses dynamic time warping distance to quantify the difference between actual and ideal states, and combines data forwarding, standardization processing, and the fusion of objective weights determined by the entropy weight method with user-defined weights to achieve a comprehensive and accurate assessment of the health status of long-endurance aircraft missions. This assessment method not only considers the multi-dimensional status indicators of the aircraft, but also improves the accuracy and adaptability of the assessment by dynamically adjusting the weights and standardization range.

[0017] Second, this assessment method not only provides health ranking results, but also optimizes tasks based on the ranking results, including health level classification, abnormal data tracing, task parameter optimization suggestions, and assessment model iteration. This real-time task optimization capability enables the aircraft to adjust parameters in a timely manner to deal with potential health problems during mission execution. At the same time, by periodically comparing the assessment results with actual mission failure records, the parameters of the assessment model are automatically adjusted, realizing the self-optimization of the assessment model and improving the accuracy and reliability of subsequent assessments.

[0018] Other advantages, objectives and features of the invention will be set forth in part in the description which follows, and in part will be apparent to those skilled in the art from the following examination or study, or may be learned from the practice of the invention. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the accompanying drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0020] Figure 1 A flowchart illustrating a method for assessing the mission health of long-endurance aircraft. Figure 2 A flowchart for determining the weights of indicators in a long-endurance aircraft mission health assessment method; Figure 3 This is a flowchart illustrating the TOPSIS health assessment breakdown, a method for assessing the health of long-endurance aircraft missions. Detailed Implementation

[0021] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0022] Example 1 This embodiment describes a border reconnaissance mission performed by a certain type of long-endurance reconnaissance aircraft. The mission lasts for 12 hours and requires maintaining a stable flight attitude and speed throughout to ensure reconnaissance accuracy. Figure 1As shown, ideal flight state data is obtained from the aircraft mission planning system and includes 12 indicators: longitude, latitude, altitude, pitch angle, roll angle, yaw angle, linear velocity U, linear velocity V, linear velocity W, pitch angular velocity, roll angular velocity, and yaw angular velocity. The sampling frequency is set to 10Hz, consistent with the sampling frequency of actual flight data. Actual flight state data is collected by multiple types of sensors on the aircraft: longitude and latitude data are obtained by a high-precision GNSS positioning module; altitude data is obtained by fusion of laser altimeter and pressure sensor; pitch angle, roll angle, and yaw angle are obtained by combination of three-axis gyroscope and three-axis accelerometer; linear velocity U / V / W is obtained by fusion of Doppler radar and inertial measurement unit data; and pitch angular velocity, roll angular velocity, and yaw angular velocity are directly output by the three-axis gyroscope.

[0023] During the data collection process, 20 sets of actual flight status data were acquired simultaneously. For each set of data, outlier removal and data alignment were performed first. Then, according to the indicator dimension, the DTW distance was calculated for each set of preprocessed actual flight data and ideal flight data. Before the calculation, the length of the two types of data was unified to 120,000 data points by interpolation. Finally, a 20-row, 12-column actual-ideal trajectory difference matrix was formed. Each row corresponds to the difference value of 12 indicators in a set of actual flight data, and each column corresponds to the difference value of the same indicator in the 20 sets of actual data.

[0024] The type of each indicator is determined by a pre-set indicator attribute library. Since this reconnaissance mission requires high stability within a specific range to ensure the efficiency of the reconnaissance equipment, the altitude indicator is set as a range-type indicator, and the DTW distance of the remaining 11 indicators are all extremely small indicators. For the 11 extremely small indicators, a reverse mapping method is used for forwarding: first, the maximum value of each indicator in the difference matrix is ​​calculated, and then the maximum value is subtracted from each difference value to obtain the corresponding forwarding result. For interval-type height indicators, the optimal interval [8000m, 8500m] is determined based on task requirements. The deviation of each difference value from this interval is calculated: if the difference value is within [8000m, 8500m], the positiveization result is set to 1; if the difference value is less than 8000m, the positiveization result is calculated as 1-(8000-difference value) / (max(8000-all difference values ​​less than 8000m)); if the difference value is greater than 8500m, the positiveization result is calculated as 1-(difference value-8500) / (max(all difference values ​​greater than 8500m-8500)). After positiveization processing, the data matrix is ​​checked for rationality to ensure that all positiveization values ​​are within the range of [0, 1]. Among them, the positiveization results of the height indicators of 3 sets of data exceed the range. The linear scaling method is used to map them to the interval of [0, 1], and finally the positiveized data matrix is ​​obtained.

[0025] Dynamic range standardization was employed to process the forward data matrix, with the standardization range dynamically adjusted according to the flight phases of the mission. Specifically, the fluctuation thresholds for each indicator were set to larger values ​​during takeoff and landing, and smaller values ​​during climb and cruise, to adapt to the flight characteristics of different phases. For each indicator in the forward data matrix, the corresponding fluctuation range was determined based on the current flight phase, and the standardization result was calculated using the dynamic range standardization formula. Since historical data was available for reference from similar border reconnaissance missions performed by this aircraft, the default mean-standard deviation standardization method was not used throughout the process. After standardization, the data was corrected using a truncation method, setting negative values ​​to 0 and values ​​greater than 1 to 1. Statistical analysis showed that the number of corrections for all indicators did not exceed 5% of the total data volume, eliminating the need for user prompts to check the data collection or forwarding process, ultimately resulting in the standardized data matrix.

[0026] The objective weight vector is calculated using the dynamic entropy weight calculation formula, which is as follows: ,in For the first The dynamic difference coefficient of each indicator , The dynamic entropy value is calculated using the following formula: , For the first Group of data The weight of each indicator is calculated using the following formula: , , It is the minimum value; The number of sets of actual flight data. For the first The task relevance of each indicator As the correlation adjustment coefficient, the dynamic difference coefficient of the j-th indicator is calculated first. The dynamic entropy value avoids logarithmic calculation abnormalities by introducing a minimum value ε. Then, the objective weights of 12 indicators are obtained by combining the task correlation and correlation adjustment coefficient λ of each indicator. The user inputs a custom weight vector based on the characteristics of the reconnaissance mission, focusing on ensuring the stability of the course and speed. The weights of yaw angle and linear velocity U / V / W are set to higher values, and the weights of the other indicators are set to normal values.

[0027] The system performs layered validity checks on custom weights: First, it checks each input weight to ensure they are all valid positive numbers, without prompting the user for confirmation. Then, it calculates the total weights; if the sum does not match the system's preset weight sum standard, it automatically triggers a normalization process. Using the objective weights calculated by the entropy weight method as a reference, it compares the deviations between the user weights and the objective weights. For the three indicators whose deviations exceed a preset threshold, the weights are appropriately compressed and adjusted according to the deviation ratio. Finally, all adjusted weights are scaled to the standard weight sum range to generate a valid custom weight vector, such as... Figure 2As shown, the weight ratio of the two types is adjusted through a dynamic fusion formula, which is: ,in The dynamic fusion coefficient is calculated using the following formula: ,in The initial fusion coefficient, The flight duration corresponding to the current assessment time. Total task duration The initial fusion coefficient is a dynamic adjustment factor. Set to 0.5, dynamic adjustment factor Set to 0.2, based on the flight duration corresponding to the current assessment time. Total task duration The ratio of the values ​​is used to calculate the dynamic fusion coefficient α(t), and finally the final weight vector of the 12 indicators is obtained.

[0028] like Figure 3 As shown, based on the standardized data matrix and the final weight vector, a weighted standardization matrix is ​​constructed using a time-sensitive weighting formula, specifically: the weighted standardization matrix contains the first... Group of data Elements of each indicator The calculation formula is: ,in For the first Group of data The standardized value of each indicator, For the first The final weight of each indicator, For the first The timeliness coefficient of the data set, where the timeliness coefficient τ of the i-th data set is... i The timeliness coefficient for recently collected data is set to 1.2, and that for earlier collected data to 1.0, based on the interval between data collection and the current assessment time, to highlight the impact of recent flight conditions on health. Positive and negative ideal solutions are determined, and the Euclidean distance from each set of actual flight data to the positive and negative ideal solutions is calculated. The overall health score is calculated using a trend-corrected scoring formula, where the health trend coefficient θ... i The data θ is determined based on the trend of the differences between the three sets of data before and after each set, showing a stable trend. i Set to 0.1, the data θ shows a deteriorating trend. i Set the threshold to -0.1 and the trend influence coefficient η to 0.2. Sort the data by comprehensive score from highest to lowest and output the health ranking results. The 15th group of data has the highest comprehensive score, and the 3rd group of data has the lowest comprehensive score.

[0029] The overall health score is divided into five levels: Excellent (0.9-1.0), Good (0.8-0.9), Qualified (0.7-0.8), Warning (0.6-0.7), and Abnormal (≤0.6). In this embodiment, out of 20 data sets, 3 sets are Excellent, 8 sets are Good, 6 sets are Qualified, 3 sets are Warning, and 0 sets are Abnormal. The abnormality sources of the 3 sets of data at the Warning level are traced, and the influence weight of each indicator on the health score is analyzed. The top 3 key indicators with the greatest influence are selected: yaw angle, altitude, and linear velocity U. The possible causes of abnormalities in key indicators and inspection suggestions are automatically output. Based on the key indicators and the trend of health score changes, flight control system parameter adjustment suggestions are generated to adapt to the stability requirements of reconnaissance missions. The evaluation results are compared with actual mission fault records periodically to calculate the evaluation accuracy. If the accuracy does not meet the preset standard, the dynamic fusion coefficient or trend influence coefficient is automatically adjusted to achieve iterative optimization of the evaluation model.

[0030] Example 2 This embodiment describes a long-endurance aircraft performing a mountainous material delivery mission. The mission lasted 8 hours, and the flight area had complex terrain, requiring frequent adjustments to flight attitude to avoid mountain obstacles. This necessitated high precision in longitude and latitude, as well as adaptability to pitch and roll angles. Figure 1 As shown, the ideal flight state data is obtained from the aircraft mission planning system and includes 12 evaluation indicators. The sampling frequency is 8Hz, which is consistent with the sampling frequency of the actual flight data. The actual flight state data is collected by corresponding sensors: longitude and latitude are obtained by a high-precision GNSS positioning module; altitude is obtained by the fusion of laser altimeter and pressure sensor; pitch angle, roll angle, and yaw angle are obtained by the combination of three-axis gyroscope and three-axis accelerometer; linear velocity U / V / W is obtained by the fusion of Doppler radar and inertial measurement unit data; and pitch angular velocity, roll angular velocity, and yaw angular velocity are directly output by the three-axis gyroscope. Fifteen sets of actual flight status data were collected. During the preprocessing stage, the focus was on outlier removal and data alignment. For each set of preprocessed actual flight data, the DTW distance was calculated between the data and the ideal flight data according to the indicator dimensions. Before the calculation, the length of the two types of data was unified to 57,600 data points by zero padding. Finally, a 15-row, 12-column actual-ideal trajectory difference matrix was formed. Each row corresponds to the difference values ​​of 12 indicators in a set of actual data, and each column corresponds to the difference values ​​of the same indicator in the 15 sets of actual data.

[0031] The type of each indicator is determined by a pre-set indicator attribute library. This material delivery mission does not require maintaining a specific altitude range, but only needs to approach the ideal altitude. Therefore, the DTW distance of all 12 indicators is a very small indicator. For all very small indicators, a reverse mapping method is used for forwarding: first, the maximum value of each indicator in the difference matrix is ​​calculated, and then the maximum value is subtracted from each difference value to obtain the forwarding result. After forwarding, a rationality check is performed to ensure that all forwarding values ​​are within the range of [0, 1]. Among them, the forwarding results of the pitch angle and roll angle indicators of two sets of data exceed the range. A linear scaling method is used to map them to the range of [0, 1], and finally the forwarding data matrix is ​​obtained.

[0032] The dynamic range standardization method was used to process the forward data matrix, and the standardization range was adjusted according to the flight phase. Considering the terrain limitations during takeoff and landing in mountainous areas, the fluctuation range threshold for indicators was set to a relatively large value, while the fluctuation threshold for the cruise and climb phases was set to a normal value. For each indicator in the forward data matrix, the standardization result was calculated after determining the fluctuation range based on the current flight phase. Since this aircraft was performing a material delivery mission in this area for the first time, there was no historical data reference for the landing phase. The mean-standard deviation standardization method was used by default to calculate the standardized values ​​of the indicators for this phase. After standardization, the data was corrected using a truncation method, setting negative values ​​to 0 and values ​​greater than 1 to 1. The proportion of all indicator correction times to the total data volume did not exceed 5%, so there was no need to prompt the user to check the data collection or forwarding process, thus obtaining the standardized data matrix.

[0033] The objective weight vector is calculated using a dynamic entropy weight calculation formula. A minimum value ε is introduced to avoid logarithmic calculation anomalies. Combining the task relevance and relevance adjustment coefficient λ of each indicator, the objective weights of 12 indicators are obtained. Users input custom weight vectors based on the characteristics of the material delivery task, emphasizing longitude, latitude, pitch angle, and roll angle. These four indicators are assigned higher weights. A layered validity check is performed on the custom weights: first, all input weights are checked to be valid positive numbers; then, the total weight is calculated. If it does not conform to the system's preset standard, a normalization process is triggered. Using the objective weights as a reference, the deviation between the user weights and the objective weights is compared. The weights of two indicators with deviations exceeding a preset threshold are compressed and adjusted proportionally to the deviation. Finally, the adjusted weights are scaled to the standard total range to generate a valid custom weight vector. Figure 2 As shown, the weight ratio of the two categories is adjusted by a dynamic fusion formula. The initial fusion coefficient α0 is set to 0.6, and the dynamic adjustment factor β is set to 0.3. The dynamic fusion coefficient α(t) is calculated based on the ratio of the current flight time t to the total mission time T, and finally the final weight vector of 12 indicators is obtained.

[0034] like Figure 3As shown, based on the standardized data matrix and the final weight vector, a weighted standardized matrix is ​​constructed using a timeliness-weighted formula, where the timeliness coefficient τ of the i-th data set is... i Based on the distance between the data collection time and the delivery node, the timeliness coefficient of the three data sets closest to the delivery node is set to 1.3, and the timeliness coefficient of the remaining data sets is set to 1.0, highlighting the importance of flight status in the approaching delivery phase. Positive and negative ideal solutions are determined, and the Euclidean distance from each set of actual flight data to the positive and negative ideal solutions is calculated. The overall health score is calculated using the trend correction scoring formula, and the health trend coefficient θ is determined. i The data θ that shows an optimization trend is determined based on the changing trend of the differences between the two sets of data before and after each set. i Set to 0.15, the data θ shows a stable trend. i Set the value to 0, and the trend influence coefficient η to 0.25. Sort the data from highest to lowest based on the overall score, and output the health ranking results. The 12th group of data has the highest overall score, and the 5th group of data has the lowest overall score.

[0035] The overall health score is divided into five levels: Excellent (0.9-1.0), Good (0.8-0.9), Qualified (0.7-0.8), Warning (0.6-0.7), and Abnormal (≤0.6). In this embodiment, out of 15 data sets, 1 set is Excellent, 6 sets are Good, 5 sets are Qualified, 3 sets are Warning, and 0 sets are Abnormal. The abnormality sources of the 3 sets of data at the Warning level are traced, and the top 3 key indicators with the greatest impact are selected: longitude, latitude, and pitch angle. Possible causes of abnormalities and inspection suggestions are automatically output. Combined with the key indicators and the trend of health changes, flight control system parameter adjustment suggestions are generated to adapt to the terrain adaptation needs of mountain material delivery. The evaluation results are compared with the actual mission fault records regularly to calculate the evaluation accuracy. If the accuracy does not meet the preset standard, the dynamic fusion coefficient or trend influence coefficient is automatically adjusted to achieve iterative optimization of the evaluation model.

[0036] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for assessing the mission health of a long-endurance aircraft, characterized in that, The specific steps of this method are as follows: S1. Data Acquisition and Trajectory Difference Quantification Steps: Acquire ideal flight state data and multiple sets of actual flight state data for the flight mission. Each set of data includes 12 evaluation indicators: longitude, latitude, altitude, pitch angle, roll angle, yaw angle, linear velocity U, linear velocity V, linear velocity W, pitch angular velocity, roll angular velocity, and yaw angular velocity. Through dynamic time-normalized distance calculation, quantify the differences between each indicator in each set of actual flight state data and the corresponding indicator in the ideal flight state, forming an actual-ideal trajectory difference matrix. S2. Data forwarding process steps: Determine the type of each column of indicators in the difference matrix, adopt corresponding processing methods for different types of indicators, and convert all indicators into a positive evaluation standard where the larger the indicator value, the closer the flight status is to the ideal state, thus obtaining a forwarding data matrix. S3. Data standardization processing steps: Standardize the positive data matrix to eliminate the influence of the dimensions of each indicator, so that all indicators are in a uniform and comparable range, and obtain a standardized data matrix. S4. Steps for determining indicator weights: The objective weight vector is calculated using the entropy weight method. Combined with the user-inputted custom weight vector based on task characteristics, the proportion of the two types of weights is adjusted by the fusion coefficient to obtain the final weight vector. Among them, the user-inputted custom weights need to be validated for legality. If the validation conditions are not met, automatic normalization is performed. S5, TOPSIS Health Assessment and Ranking Steps: Construct a weighted standardized matrix based on the standardized data matrix and the final weight vector to determine the positive and negative ideal solutions; calculate the distance from each set of actual flight data to the positive and negative ideal solutions, and obtain the comprehensive health score through the distance ratio; sort the comprehensive scores from high to low and output the health ranking results.

2. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The specific steps for data acquisition and trajectory difference quantification in S1 are as follows: Data acquisition phase: Ideal flight state data is acquired from the aircraft mission planning system. This data represents the optimal state parameters preset for the entire process, and the sampling frequency is consistent with the actual flight data. Actual flight state data is acquired through multiple types of sensors on board the aircraft. Among them, longitude and latitude data are acquired by a high-precision GNSS positioning module, altitude data is acquired by the fusion of laser altimeter and pressure sensor, pitch angle, roll angle, and yaw angle are calculated by a combination of three-axis gyroscope and three-axis accelerometer, linear velocity U / V / W is acquired by the fusion of Doppler radar and inertial measurement unit (IMU) data, and pitch angular velocity, roll angular velocity, and yaw angular velocity are directly output by the three-axis gyroscope. Data preprocessing stage: outlier removal and data alignment are performed on the collected actual flight data; DTW distance calculation stage: For each set of preprocessed actual flight data, DTW distance is calculated with the ideal flight data according to the index dimensions. Before the calculation, the length of the two types of data is unified. In the final output difference matrix, each row corresponds to the difference value of 12 indexes of a set of actual flight data, and each column corresponds to the difference value of the same index in different sets of actual data.

3. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The specific implementation process of data forwarding in S2 includes: Indicator type determination: The type of each indicator is determined by the preset indicator attribute library. Among the 12 indicators corresponding to DTW distance, the DTW distances for longitude, latitude, altitude, pitch angle, roll angle, yaw angle, linear velocity U / V / W, pitch angular velocity, roll angular velocity, and yaw angular velocity are all very small indicators. Only the altitude indicator is set as an interval indicator in a specific mission scenario. The indicator type needs to be selected before the evaluation according to the mission type. Formalization of extremely small indicators: For extremely small indicators, the reverse mapping method is used. First, the maximum value of the indicator in the difference matrix is ​​calculated, and then the maximum value is subtracted from each difference value to obtain the formalized result. For intermediate-type indicators, positive transformation is performed as follows: For height indicators set as interval types, first determine the optimal interval [L, R], then calculate the deviation of each difference value from the interval. If the difference value is within [L, R], the positive transformation result is set to 1; if the difference value is less than L, the positive transformation result is 1-(L-difference value) / (max(L-all difference values ​​less than L)); if the difference value is greater than R, the positive transformation result is 1-(difference value-R) / (max(all difference values ​​greater than R-R)). Forwarding verification: The rationality of the processed forwarding data matrix is ​​verified. The verification standard is that all forwarded values ​​are in the range of [0, 1]. If there are values ​​that are out of range, the linear scaling method is used to map them to the interval of [0, 1].

4. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The specific implementation process of data standardization processing in S3 includes: Standardization method selection: The dynamic range standardization method is adopted. This method dynamically adjusts the standardization range in combination with the flight phase of the aircraft. The flight phase is divided into takeoff phase, climb phase, cruise phase and landing phase, and each phase corresponds to a different index fluctuation range threshold. Standardization calculation: For each indicator in the positive data matrix, first determine the fluctuation range of the indicator based on the current flight phase, and then calculate the standardization result using the formula standardization value; if there is no historical data reference for the current flight phase, the mean-standard deviation standardization method is used by default. Standardization correction: For negative values ​​or values ​​greater than 1 that appear after standardization, a truncation method is used to process them, setting negative values ​​to 0 and values ​​greater than 1 to 1, ensuring that the standardization results are all in the range [0, 1]. At the same time, the number of corrections is recorded. If the number of corrections for a certain indicator accounts for more than 5% of the total data volume, the user is prompted to check whether there are any abnormalities in the collection or positiveization process of that indicator.

5. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The S4 method for calculating the objective weight vector uses a dynamic entropy weight calculation formula, specifically: the first... Weight of each indicator The calculation formula is: ,in For the first The dynamic difference coefficient of each indicator , The dynamic entropy value is calculated using the following formula: , For the first Group of data The weight of each indicator is calculated using the following formula: , , It is the minimum value; The number of sets of actual flight data. For the first The task relevance of each indicator This is the correlation adjustment coefficient.

6. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The validity verification and normalization of the user-defined weight vector in S4 adopts a weight-layered verification-correction technique. The specific implementation process includes: Validity layered verification: First, check whether the user-inputted custom weights are valid positive numbers. If there are non-positive numbers, the system automatically prompts the user to confirm the rationality of the non-positive number. If the user confirms retention, it is replaced with a preset reasonable minimum value. If the user chooses to modify, the input is received again. Then, calculate the sum of all weights and determine whether it meets the preset weight sum standard of the evaluation system. If it does not meet the standard, the normalization process is automatically triggered. Deviation correction normalization: First, extract the objective weights calculated by the entropy weight method as a reference benchmark. Then, compare the deviation between the user weights and the objective weights. User weights with deviations less than a preset threshold are directly retained. User weights with deviations exceeding the preset threshold are appropriately compressed and adjusted according to the deviation ratio. Finally, recalculate the sum of all adjusted weights and scale them to the standard weight sum range required by the evaluation system to generate a valid custom weight vector that is suitable for the task characteristics.

7. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, In step S4, the final weight vector is obtained by adjusting the proportion of the two types of weights through the fusion coefficient using a dynamic fusion formula, specifically: Final Weight The calculation formula is: ,in The dynamic fusion coefficient is calculated using the following formula: ,in The initial fusion coefficient, The flight duration corresponding to the current assessment time. Total task duration It is a dynamic adjustment factor.

8. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, In S5, the weighted normalization matrix constructed based on the normalized data matrix and the final weight vector adopts a time-sensitive weighting formula, specifically: the weighted normalization matrix... Group of data Elements of each indicator The calculation formula is: ,in For the first Group of data The standardized value of each indicator, For the first The final weight of each indicator, For the first The timeliness coefficient of the data set.

9. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The health score obtained from the distance ratio in S5 uses a trend-corrected scoring formula, specifically: Overall health score of the group data The calculation formula is: ,in For the first The Euclidean distance from the set of data to the positive ideal solution. For the first The Euclidean distance from the set of data to the negative ideal solution. For the first Health trend coefficient of group data This represents the trend influence coefficient.

10. The method for assessing the mission health of a long-endurance aircraft according to claim 1, characterized in that, The process also includes a supplementary step of task optimization based on the health ranking results output in step 5. The specific implementation process includes: Health Level Classification: Health is classified into five levels based on comprehensive score: Excellent, Good, Qualified, Warning, and Abnormal. Each level corresponds to a preset task handling strategy; Abnormal Data Source Tracing: For flight data at the warning or abnormal level, the influence weight of each indicator on health is analyzed, and the top three most influential key indicators are selected. Possible causes and inspection suggestions for the abnormal key indicators are automatically output; Task Parameter Optimization Suggestions: Combining key indicators and health trend changes, adjustment suggestions for flight control system parameters are generated to adapt to different types of flight control task requirements; Evaluation Model Iteration: The evaluation results are periodically compared with actual task fault records to calculate the evaluation accuracy. If the accuracy does not meet the preset standard, the fusion coefficient or trend influence coefficient is automatically adjusted.

Images