An airport pavement condition intelligent identification method and system based on QAR data

Abstract

The application relates to the technical field of airport infrastructure monitoring, in particular to an airport pavement condition intelligent identification method and system based on QAR data. The method comprises the following steps: acquiring QAR data of all airplanes of a target airport, preprocessing the QAR data to obtain an effective data segment of the airplane in a taxiing stage of the target airport; extracting vertical acceleration data in the effective data segment, performing denoising processing on the vertical acceleration data according to airplane model information to obtain purified vertical acceleration data; performing multi-domain feature extraction on the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectrum features; identifying a pavement disease type according to the time-domain statistical features and the frequency-domain spectrum features based on a preset pavement disease identification model, calculating an airport pavement condition evaluation result; and generating a pavement condition report according to the pavement condition evaluation result. The technical scheme solves the problems of the traditional method, such as untimely detection, high cost and dependence on manual work.

Description

Technical Field

[0001] This application relates to the field of airport infrastructure monitoring technology, and in particular to an intelligent identification method and system for airport pavement conditions based on QAR data. Background Technology

[0002] With the rapid development of the civil aviation industry, real-time monitoring of airport runway pavement conditions is crucial for flight safety. Traditional pavement inspection methods mainly rely on periodic manual inspections and specialized inspection equipment, which have problems such as low inspection frequency, high cost, and inability to achieve continuous monitoring, as detailed below.

[0003] Currently, the following methods are mainly used for airport pavement condition assessment:

[0004] First, manual visual inspection: relies on the experience and judgment of the inspectors, is highly subjective, and is difficult to quantify and evaluate.

[0005] Second, specialized detection equipment, such as laser scanners and ground-penetrating radar, is expensive and has limited detection frequency.

[0006] Third, friction coefficient testing: requires specialized test vehicles and cannot be monitored in real time;

[0007] Fourth, pavement embedded or mobile sensor systems: With the development of sensor technology and artificial intelligence, more automated dedicated sensor systems are being applied to pavement monitoring. The advantages of these dedicated systems are their ability to achieve continuous or near-continuous monitoring, acquire detailed data, and improve automation levels. However, their initial investment costs (especially for embedded sensors and dedicated inspection vehicles) are high, and data processing and system maintenance are also issues that need to be considered.

[0008] These traditional methods have several problems in implementation: First, the detection frequency is low, making it impossible to detect changes in pavement conditions in a timely manner; second, the detection cost is high, making it difficult to achieve full coverage monitoring; and finally, there is a lack of objective quantitative evaluation standards, and the evaluation results are greatly affected by personnel experience. Summary of the Invention

[0009] To at least partially overcome the problems of low detection frequency, high cost, and inability to achieve continuous monitoring in related pavement inspection methods, this application provides an intelligent identification method and system for airport pavement conditions based on QAR data.

[0010] The proposed solution is as follows:

[0011] According to a first aspect of the embodiments of this application, a method for intelligent identification of airport pavement conditions based on QAR (Quick Access Recorder) data is provided, comprising:

[0012] Acquire QAR data for all aircraft at the target airport, preprocess the QAR data to obtain valid data segments of the aircraft during the taxiing phase at the target airport;

[0013] Vertical acceleration data is extracted from the valid data segment, and the vertical acceleration data is denoised according to the aircraft type information to obtain purified vertical acceleration data.

[0014] Multi-domain feature extraction was performed on the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectral features;

[0015] Based on the preset pavement defect identification model, the pavement defect type is identified according to the time domain statistical characteristics and the frequency domain spectral characteristics, and the pavement condition assessment results are obtained by computer.

[0016] A pavement condition report is generated based on the pavement condition assessment results.

[0017] Preferably, the QAR data includes: aircraft weight data, vertical acceleration data, lateral acceleration data, longitudinal acceleration data, ground speed data, barometric altitude data, left main wheel status, right main wheel status, nose wheel status, and throttle lever status;

[0018] The QAR data is preprocessed to obtain valid data segments of the aircraft during the taxiing phase at the target airport, including:

[0019] Check the data integrity of QAR data and filter out QAR data with missing values ​​exceeding a preset missing threshold;

[0020] Identifying and processing outliers in QAR data based on empirical normal distribution;

[0021] The landing touchdown time is determined as the start time of the valid data segment based on the time when the left main wheel, right main wheel, and front wheel transition from the air state to the ground state.

[0022] The time when the throttle lever returns to the IDLE position is taken as the end time of the valid data segment;

[0023] Extract valid data segments from the processed QAR data based on the start and end times of the valid data segments;

[0024] The valid data segment is subjected to time synchronization and sampling rate standardization.

[0025] Preferably, the aircraft model information includes: baseline standard deviation, natural vibration frequency, and normal acceleration range;

[0026] The vertical acceleration data is denoised based on the aircraft type information, including:

[0027] Calculate the signal mean of the vertical acceleration data, and remove the DC component based on the signal mean of the vertical acceleration data;

[0028] High-frequency noise in the vertical acceleration data is removed by using a 4th-order Butterworth low-pass filter with a cutoff frequency of 20Hz.

[0029] Design a band-stop filter based on the inherent vibration frequencies in the aircraft model information to remove the aircraft's inherent vibration frequencies from the vertical acceleration data;

[0030] Based on the baseline standard deviation in the aircraft model information, the judgment conditions for the sliding window detection method are constructed, and the sliding window detection method is used to identify and remove abnormal impacts in the vertical acceleration data.

[0031] Based on the normal acceleration range in the model information, remove vertical acceleration data that exceeds the normal acceleration range.

[0032] Preferably, the time-domain statistical features include:

[0033] Mean, standard deviation, maximum, minimum, RMS, skewness, kurtosis, and percentile of vertical acceleration data;

[0034] The frequency domain spectral features include:

[0035] Dominant frequency, power spectral density, energy distribution and spectral peak characteristics of each frequency band;

[0036] The method further includes:

[0037] Outliers in time-domain statistical features and frequency-domain spectral features are identified and processed based on empirical normal distribution.

[0038] Preferably, the method further includes:

[0039] Establish a pavement defect characteristic model database;

[0040] The pavement defect feature model library includes: standard deviation threshold, dominant frequency range, amplitude threshold, and high frequency energy ratio for different pavement defect types;

[0041] The types of defects include: unevenness, misalignment, fracture, settlement, voids, and softening of the base layer;

[0042] Based on a preset pavement defect identification model, the pavement defect type is identified according to the time-domain statistical features and the frequency-domain spectral features, including:

[0043] The time-domain statistical features and the frequency-domain spectral features are matched with the pavement defect feature pattern library to determine pavement defect types.

[0044] Calculate the matching score for various pavement defect types;

[0045] When the matching score of pavement defect type exceeds the preset matching threshold, it is determined that there is a corresponding pavement defect type.

[0046] Preferably, the matching score for various pavement defect types is calculated, including:

[0047] For each type of pavement defect, the first, second, third, and fourth judgment items are executed sequentially. The matching degree score of the pavement defect type in each judgment item is calculated. The sum of the scores of the pavement defect type in all judgment items is taken as the final matching degree score of the pavement defect type.

[0048] The first judgment item includes:

[0049] Determine whether the standard deviation of the vertical acceleration data is greater than the standard deviation threshold for the current pavement defect type;

[0050] If the standard deviation of the vertical acceleration data is greater than the standard deviation threshold of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the first weight, the standard deviation of the vertical acceleration data, and the baseline standard deviation, and then proceed to the next judgment item.

[0051] If the standard deviation of the vertical acceleration data is not greater than the standard deviation threshold of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0052] The second judgment item includes:

[0053] Determine whether the dominant frequency of the vertical acceleration data matches the dominant frequency range of the current pavement defect type;

[0054] If the dominant frequency of the vertical acceleration data matches the dominant frequency range of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the second weight and the first preset coefficient, and then proceed to the next judgment item.

[0055] If the dominant frequency of the vertical acceleration data does not conform to the dominant frequency range of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0056] The third judgment item includes:

[0057] Determine whether the maximum acceleration value of the vertical acceleration data is greater than the amplitude threshold of the current pavement defect type;

[0058] If the maximum acceleration value of the vertical acceleration data is greater than the amplitude threshold of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the third weight, the maximum acceleration value of the vertical acceleration data, and the amplitude threshold of the current pavement defect type, and then proceed to the next judgment item.

[0059] If the maximum acceleration value of the vertical acceleration data is not greater than the amplitude threshold of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0060] The fourth judgment item includes:

[0061] Determine whether the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is greater than a preset outlier proportion threshold;

[0062] If the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is greater than the preset outlier ratio threshold, the score of the current pavement defect type in the current judgment item is calculated based on the fourth weight, the proportion of outliers in the time-domain statistical features and frequency-domain spectral features, and the second preset coefficient, and the next judgment item is executed.

[0063] If the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is not greater than the preset outlier proportion threshold, the score of the current judgment item will be recorded as 0.

[0064] Among them, the first weight is greater than the third weight, which is greater than the second weight, which is greater than the fourth weight.

[0065] Preferably, the computer-aided pavement condition assessment results include:

[0066] Set the deduction items and the maximum deduction value for each deduction item; the deduction items include: fluctuation deduction items, anomaly point deduction items, peak impact deduction items, distribution anomaly deduction items, and high frequency energy deduction items;

[0067] The standard deviation is calculated based on the standard deviation and normal standard deviation of the vertical acceleration data. The fluctuation deduction value is then calculated based on the standard deviation and the maximum deduction value of the fluctuation deduction item.

[0068] Calculate the outlier deduction value based on the proportion of outliers in the time-domain statistical characteristics and frequency-domain spectral characteristics, as well as the maximum deduction value for the outlier deduction item;

[0069] When the maximum acceleration value of the vertical acceleration data is greater than 0.5g, the peak impact deduction value is calculated based on the maximum acceleration value of the vertical acceleration data and the maximum deduction value of the peak impact deduction item.

[0070] Calculate the distribution anomaly deduction value based on the distribution characteristics of the vertical acceleration data and the maximum deduction value of the distribution anomaly deduction item;

[0071] Calculate the high-frequency energy deduction value based on the energy distribution of each frequency band of the vertical acceleration data and the maximum deduction value of the high-frequency energy deduction item;

[0072] The airport pavement condition assessment score is obtained by subtracting the deductions for each item from the preset full score.

[0073] The corresponding pavement condition level is matched based on the airport pavement condition assessment score.

[0074] Preferably, the pavement condition report includes:

[0075] Basic information: airport code, runway number, analysis time, number of flights, and aircraft type distribution;

[0076] Vertical acceleration data statistics: mean, standard deviation, maximum, minimum, RMS, skewness, kurtosis, and percentiles;

[0077] Identified pavement defects: defect type, severity, and affected area;

[0078] Pavement condition assessment: Comprehensive score, condition level, and comparison results with historical data;

[0079] Repair recommendations: repair priority, specific repair measures, and estimated repair costs;

[0080] Trend Analysis: Pavement condition changes and predictive maintenance recommendations.

[0081] Preferably, the method further includes:

[0082] Obtain vertical acceleration data for different models;

[0083] The vertical acceleration data of different models were normalized.

[0084] The standard deviation difference between different models is calculated based on the normalized vertical acceleration data, and the vertical acceleration data of abnormal models are removed.

[0085] Calculate the airport pavement condition assessment results for the same airport pavement for different aircraft types, and perform cross-validation;

[0086] Output the cross-validation results and use these results to determine the accuracy of the airport pavement condition assessment.

[0087] According to a second aspect of the embodiments of this application, an intelligent airport pavement condition recognition system based on QAR data is provided, comprising:

[0088] The data preprocessing module is used to acquire QAR data of all aircraft at the target airport, preprocess the QAR data, and obtain the effective data segments of the aircraft during the taxiing phase at the target airport.

[0089] The data purification module is used to extract vertical acceleration data from the effective data segment, and to perform noise reduction processing on the vertical acceleration data according to the aircraft type information to obtain purified vertical acceleration data.

[0090] The feature extraction module is used to extract multi-domain features from the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectral features;

[0091] The evaluation module is used to identify pavement defect types based on a preset pavement defect identification model, according to the time-domain statistical features and the frequency-domain spectral features, and to evaluate the pavement condition by computer.

[0092] The report generation module is used to generate a pavement condition report based on the pavement condition assessment results.

[0093] The technical solution provided in this application may include the following beneficial effects:

[0094] This technical solution acquires QAR data of all aircraft at the target airport, preprocesses the QAR data to obtain effective data segments of the aircraft during the taxiing phase at the target airport, extracts vertical acceleration signals from these segments, and combines aircraft-specific processing and multi-domain feature extraction to achieve identification of pavement defect types.

[0095] Based directly on QAR data (aircraft-generated data), without relying on additional detection vehicles or equipment, it can automatically collect data with each landing, and every aircraft can participate in monitoring, enabling widespread deployment. Assessment is triggered with each landing, improving response efficiency. It solves the problems of untimely detection, high cost, and reliance on manual labor inherent in traditional methods, demonstrating good engineering feasibility, versatility, and promotional value.

[0096] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0097] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0098] Figure 1 This is a flowchart illustrating an intelligent airport pavement condition recognition method based on QAR data, provided in one embodiment of this application.

[0099] Figure 2 This is a schematic diagram of a process for preprocessing QAR data according to an embodiment of this application;

[0100] Figure 3This is a schematic diagram of a process for denoising vertical acceleration data based on aircraft model information, provided in one embodiment of this application.

[0101] Figure 4 This is a schematic diagram of a process for identifying pavement defect types based on time-domain statistical characteristics and frequency-domain spectral characteristics, provided in one embodiment of this application;

[0102] Figure 5 This is a schematic diagram of the modular structure of an airport pavement condition intelligent recognition system based on QAR data, provided in one embodiment of this application.

[0103] Figure labeling: Data preprocessing module-41; Data cleaning module-42; Feature extraction module-43; Evaluation module-44; Report generation module-45. Detailed Implementation

[0104] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0105] Example 1

[0106] Figure 1 This is a flowchart illustrating an intelligent airport pavement condition recognition method based on QAR data, provided in one embodiment of this application. (Refer to...) Figure 1 A method for intelligent identification of airport pavement conditions based on QAR data, comprising:

[0107] S11: Obtain QAR data for all aircraft at the target airport, preprocess the QAR data, and obtain the effective data segments of the aircraft during the taxiing phase at the target airport.

[0108] S12: Extract vertical acceleration data from the valid data segment, and perform noise reduction processing on the vertical acceleration data according to the aircraft type information to obtain purified vertical acceleration data;

[0109] S13: Perform multi-domain feature extraction on the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectral features;

[0110] S14: Based on the preset pavement defect identification model, identify the type of pavement defect according to the time domain statistical characteristics and frequency domain spectral characteristics, and evaluate the pavement condition using computer.

[0111] S15: Generate a pavement condition report based on the pavement condition assessment results.

[0112] QAR data is flight data extracted from the Quick Access Recorder installed on the aircraft. It records a large number of parameters of the aircraft at various stages of flight, and is particularly suitable for non-accident purposes such as flight performance monitoring, safety assessment, and maintenance analysis. In this embodiment, as shown in Table 1, the QAR data of the aircraft during the landing phase includes: aircraft weight data, vertical acceleration data, lateral acceleration data, longitudinal acceleration data, ground speed data, barometric altitude data, left main wheel status, right main wheel status, nose wheel status, and throttle position.

[0113] Table 1. QAR data required for pavement condition analysis

[0114]

[0115] Preprocess the QAR data, referring to Figure 2 ,include:

[0116] S21: Check the data integrity of QAR data and filter QAR data with missing values ​​exceeding the preset missing threshold;

[0117] S22: Identify and process outliers in QAR data based on empirical normal distribution;

[0118] S23: Determine the landing touchdown time point as the start time point of the valid data segment based on the time point when the left main wheel, right main wheel, and front wheel transition from the air state to the ground state;

[0119] S24: The time when the throttle lever returns to the IDLE position is taken as the end time of the valid data segment;

[0120] S25: Extract valid data segments from the processed QAR data based on the start and end times of the valid data segments;

[0121] S26: Perform time synchronization and sampling rate standardization on the valid data segment.

[0122] Filtering QAR data with missing values ​​exceeding a preset missing threshold can eliminate structurally incomplete data, ensuring the accuracy of the analysis basis.

[0123] The 3σ criterion based on normal distribution can identify and process outliers in QAR data, remove extreme spikes and jumps, reduce noise interference, retain reasonable fluctuation information, which is beneficial for subsequent feature extraction, and significantly improve the quality and usability of the signal before purification.

[0124] The landing touchdown time point is determined by the change in the state of the main wheel / nose wheel (AIR→GROUND) as the starting time point of the valid data segment, and the time point when the aircraft's throttle stick returns to the IDLE position as the ending time point of the valid data segment. This ensures that the extracted segment is an effective vibration segment in which the wheels are in full contact with the pavement and are not affected by thrust.

[0125] Unifying the time axis and adjusting the sampling rate of the effective segment avoids incorrect feature extraction or spectral distortion caused by inconsistent sampling frequencies, providing a unified input format for subsequent model training / inference.

[0126] It should be noted that, referring to Table 2, the aircraft model information includes: baseline standard deviation, natural vibration frequency, and normal acceleration range.

[0127] Table 2 Aircraft Model Information

[0128]

[0129] The vertical acceleration data is denoised based on the aircraft type information. Figure 3 ,include:

[0130] S31: Calculate the signal mean of the vertical acceleration data, and remove the DC component based on the signal mean of the vertical acceleration data;

[0131] S32: High-frequency noise in the vertical acceleration data is removed by using a 4th-order Butterworth low-pass filter with a cutoff frequency of 20Hz;

[0132] S33: Design a band-stop filter based on the inherent vibration frequencies in the aircraft model information to remove the aircraft's inherent vibration frequencies from the vertical acceleration data;

[0133] S34: Construct the judgment conditions for the sliding window detection method based on the baseline standard deviation in the aircraft model information, and use the sliding window detection method to identify and remove abnormal impacts in the vertical acceleration data;

[0134] S35: Based on the normal acceleration range in the model information, remove vertical acceleration data that exceeds the normal acceleration range.

[0135] It should be noted that different aircraft models have significant structural differences, resulting in different vibration frequencies and shaking characteristics. Therefore, in this embodiment, the vertical acceleration data is processed with personalized noise reduction based on the aircraft model information.

[0136] Design a band-stop filter based on the inherent vibration frequency in the aircraft model information; its transfer function is:

[0137] ;

[0138] in, This indicates the inherent vibration frequency of the model. Indicates the sampling frequency. This represents the radius of the pole (taken as 0.9). This represents the z-transformation variable.

[0139] In this embodiment, the sliding window detection method is used to identify abnormal impacts, and the judgment condition is:

[0140] ;

[0141] in, =5 indicates the window size, which is 11 points in front of and behind the current point; This represents the baseline standard deviation in the aircraft model information; This represents the vertical acceleration value at the current moment; This represents the median within the window to the left and right of the current point.

[0142] When using the sliding window detection method to identify abnormal impacts, if an abnormal point is determined, it is replaced with the median of that window.

[0143] It should be noted that time-domain statistical characteristics include:

[0144] Mean, standard deviation, maximum, minimum, RMS, skewness, kurtosis, and percentile of vertical acceleration data;

[0145] Frequency domain spectral characteristics, including:

[0146] Dominant frequency, power spectral density, energy distribution and spectral peak characteristics of each frequency band;

[0147] The method also includes:

[0148] Outliers in time-domain statistical features and frequency-domain spectral features are identified and processed based on empirical normal distribution.

[0149] The extracted time-domain statistical features include:

[0150] Mean: ;

[0151] Standard deviation: ;

[0152] Valid values: ;

[0153] Skewness: ;

[0154] Kuroshi: ;

[0155] Percentiles: P5, P95, P99.

[0156] N represents the number of data points in the vertical acceleration data VRTG.

[0157] In this embodiment, the Fast Fourier Transform is used for frequency domain analysis.

[0158] The extracted frequency domain features include:

[0159] Dominant frequency (representing the frequency component with the strongest energy): f_dom = argmax(PSD(f))

[0160] Energy ratio of each frequency band:

[0161] Low-frequency energy ratio: E_low = ∑PSD(0~2Hz) / ∑PSD(total)

[0162] Intermediate frequency energy ratio: E_mid = ∑PSD(2~10Hz) / ∑PSD(total)

[0163] High-frequency energy ratio: E_high = ∑PSD(>10Hz) / ∑PSD(total)

[0164] The above feature extraction formula is a conventional technique and will not be elaborated here.

[0165] In this embodiment, outliers in time-domain statistical features and frequency-domain spectral features are identified and processed based on empirical normal distribution:

[0166] Using the 3σ principle of normal distribution, outliers exceeding a reasonable range can be identified:

[0167] If |VRTG(t)-μ|>3σ, then this point is an outlier.

[0168] μ is the mean vertical acceleration; σ is the standard deviation.

[0169] Calculate the outlier ratio: R_outlier = N_outlier / N_total × 100%;

[0170] N_outlier represents the number of points that meet the anomaly criteria; N_total represents the total number of points in the entire sampling segment;

[0171] The outlier ratio indicates what percentage of the signal is considered "abnormal vibration".

[0172] Calculate the severity of the anomaly: S_outlier = max(|VRTG_outlier - μ|) / σ;

[0173] VRTG_outlier represents the vertical acceleration value of the point identified as an anomaly;

[0174] The severity of anomalies is used to measure the degree of deviation of the most severe anomaly point. It is the "intensity" index of the most extreme point among the anomalies. The higher the proportion, the more severe the impact may be (such as local cracking or voiding of the pavement).

[0175] It should be noted that the method also includes:

[0176] Establish a pavement defect characteristic model database;

[0177] The pavement defect feature model library includes: standard deviation threshold, dominant frequency range, amplitude threshold, and high-frequency energy ratio for different pavement defect types;

[0178] Types of defects include: unevenness, misalignment, cracking, settlement, voids, and softening of the base layer;

[0179] In this embodiment, a pavement defect feature model library is established, as shown in Table 3.

[0180] Table 3. Characteristics and patterns of pavement defects

[0181]

[0182] Based on a pre-defined pavement defect identification model, pavement defect types are identified according to time-domain statistical characteristics and frequency-domain spectral characteristics, with reference to... Figure 4 ,include:

[0183] Time-domain statistical features and frequency-domain spectral features are matched with pavement defect feature pattern database to determine pavement defect types;

[0184] Calculate the matching score for various pavement defect types;

[0185] When the matching score of pavement defect type exceeds the preset matching threshold, it is determined that there is a corresponding pavement defect type.

[0186] Furthermore, the matching score for various pavement defect types is calculated, including:

[0187] For each type of pavement defect, the first, second, third, and fourth judgment items are executed sequentially. The matching degree score of the pavement defect type in each judgment item is calculated. The sum of the scores of the pavement defect type in all judgment items is taken as the final matching degree score of the pavement defect type.

[0188] The first judgment item includes:

[0189] Determine whether the standard deviation of the vertical acceleration data is greater than the standard deviation threshold for the current pavement defect type;

[0190] If the standard deviation of the vertical acceleration data is greater than the standard deviation threshold of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the first weight, the standard deviation of the vertical acceleration data, and the baseline standard deviation, and then proceed to the next judgment item.

[0191] If the standard deviation of the vertical acceleration data is not greater than the standard deviation threshold of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0192] The second judgment item includes:

[0193] Determine whether the dominant frequency of the vertical acceleration data matches the dominant frequency range of the current pavement defect type;

[0194] If the dominant frequency of the vertical acceleration data matches the dominant frequency range of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the second weight and the first preset coefficient, and then proceed to the next judgment item.

[0195] If the dominant frequency of the vertical acceleration data does not conform to the dominant frequency range of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0196] The third judgment item includes:

[0197] Determine whether the maximum acceleration value of the vertical acceleration data is greater than the amplitude threshold of the current pavement defect type;

[0198] If the maximum acceleration value of the vertical acceleration data is greater than the amplitude threshold of the current pavement defect type, calculate the score of the current pavement defect type in the current judgment item based on the third weight, the maximum acceleration value of the vertical acceleration data, and the amplitude threshold of the current pavement defect type, and then proceed to the next judgment item.

[0199] If the maximum acceleration value of the vertical acceleration data is not greater than the amplitude threshold of the current pavement defect type, the score of the current judgment item is recorded as 0, and the next judgment item is executed;

[0200] The fourth judgment item includes:

[0201] Determine whether the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is greater than a preset outlier proportion threshold;

[0202] If the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is greater than the preset outlier ratio threshold, the score of the current pavement defect type in the current judgment item is calculated based on the fourth weight, the proportion of outliers in the time-domain statistical features and frequency-domain spectral features, and the second preset coefficient, and the next judgment item is executed.

[0203] If the proportion of outliers in the time-domain statistical features and frequency-domain spectral features is not greater than the preset outlier proportion threshold, the score of the current judgment item will be recorded as 0.

[0204] Among them, the first weight is greater than the third weight, which is greater than the second weight, which is greater than the fourth weight.

[0205] In practice, the weighting coefficients are as follows:

[0206] First weight = 0.4, second weight = 0.2, third weight = 0.3, fourth weight = 0.1.

[0207] The first preset coefficient is a fixed value of 0.3, and the second preset coefficient is a fixed value of 20%.

[0208] The score for the current pavement defect type in the current judgment item is calculated based on the first weight, the standard deviation of the vertical acceleration data, and the baseline standard deviation, as follows:

[0209] The score of the first judgment item = the first weight * (standard deviation of vertical acceleration data / baseline standard deviation).

[0210] The score for the current pavement defect type in the current judgment item is calculated based on the second weight and the first preset coefficient, as follows:

[0211] The score for the second judgment item = the second weight * 0.3.

[0212] The score for the current pavement defect type in the current judgment item is calculated based on the third weight, the maximum acceleration value of the vertical acceleration data, and the amplitude threshold of the current pavement defect type, as follows:

[0213] The score of the third judgment item = the third weight * (maximum acceleration value of vertical acceleration data / amplitude threshold of the current pavement defect type).

[0214] The score of the current pavement defect type in the current judgment item is calculated based on the proportion of outliers in the fourth weight, time-domain statistical characteristics, and frequency-domain spectral characteristics, as well as the second preset coefficient, as follows:

[0215] The score for the fourth judgment item = fourth weight * (outlier percentage / 20%).

[0216] The computer-aided pavement condition assessment results include:

[0217] Set the deduction items and the maximum deduction value for each deduction item; the deduction items include: fluctuation deduction items, anomaly point deduction items, peak impact deduction items, distribution anomaly deduction items, and high frequency energy deduction items;

[0218] The standard deviation is calculated based on the standard deviation and normal standard deviation of the vertical acceleration data. The fluctuation deduction value is then calculated based on the standard deviation and the maximum deduction value of the fluctuation deduction item.

[0219] Calculate the outlier deduction value based on the proportion of outliers in the time-domain statistical characteristics and frequency-domain spectral characteristics, as well as the maximum deduction value for the outlier deduction item;

[0220] When the maximum acceleration value of the vertical acceleration data is greater than 0.5g, the peak impact deduction value is calculated based on the maximum acceleration value of the vertical acceleration data and the maximum deduction value of the peak impact deduction item.

[0221] Calculate the distribution anomaly deduction value based on the distribution characteristics of the vertical acceleration data and the maximum deduction value of the distribution anomaly deduction item;

[0222] Calculate the high-frequency energy deduction value based on the energy distribution of each frequency band of the vertical acceleration data and the maximum deduction value of the high-frequency energy deduction item;

[0223] The airport pavement condition assessment score is obtained by subtracting the deductions for each item from the preset full score.

[0224] The airport pavement condition score is used to match the corresponding pavement condition level.

[0225] In this embodiment, the pavement condition assessment score is used to quantify the overall health status of the airport pavement. The numerical range is 0-100, meaning the preset maximum score for the pavement condition assessment score is 100 points. The lower the score, the worse the pavement condition. The pavement condition assessment score is obtained by summing up the "deductions for each item".

[0226] The point deductions are calculated as follows:

[0227] Fluctuation Deduction Items:

[0228] penalty_std=min(30,(σ_measured-σ_baseline) / σ_baseline×30);

[0229] Where penalty_std represents the fluctuation penalty value; σ_measured represents the standard deviation of the vertical acceleration data; σ_baseline represents the baseline standard deviation in the aircraft information; and 30 is the maximum penalty value for the fluctuation penalty item.

[0230] This formula is used to calculate the standard deviation ratio. The larger the deviation, the more unstable the pavement is, and the more points will be deducted.

[0231] Points deduction for anomalies:

[0232] penalty_outlier= min(25,R_outlier×2);

[0233] Where, penalty_outlier represents the penalty value for outliers; R_outlier represents the proportion of outliers; and 25 is the maximum penalty value for outlier penalty items.

[0234] Peak impact deduction items:

[0235] penalty_range = min(20, max(0, (max(|VRTG|) - 0.5) × 20));

[0236] Where penalty_range represents the peak impact penalty value; max(|VRTG|) represents the maximum acceleration value of the vertical acceleration data; and 20 is the maximum penalty value for the peak impact penalty item.

[0237] Deductions for abnormal distribution:

[0238] penalty_distribution = min(15, (|Skewness| + |Kurtosis|) × 2);

[0239] Where, penalty_distribution represents the penalty for distribution anomalies; Skewness represents skewness; Kurtosis represents kurtosis; and 15 is the maximum penalty for distribution anomalies.

[0240] High-frequency energy deduction items:

[0241] penalty_roughness = min(10, E_high × 20);

[0242] Wherein, penalty_roughness represents the high-frequency energy deduction value; E_high represents the high-frequency energy ratio; E_high × 20 means deducting 1 point for every 0.05 increase in the high-frequency energy ratio; 10 is the maximum deduction value for high-frequency energy.

[0243] Finally, the airport pavement condition assessment score is obtained by subtracting the deductions for each item from the preset full score.

[0244] In this embodiment, the score ranges for each pavement condition level are preset:

[0245] Excellent: Score ≥ 90

[0246] Good: 80 ≤ Score < 90

[0247] General: 60 ≤ Score < 80

[0248] Poor: 40 ≤ Score < 60

[0249] Danger: Score < 40.

[0250] Then, the corresponding pavement condition level is matched according to the airport pavement condition assessment score. For example, if the airport pavement condition assessment score is 84, it means that the airport pavement condition is good.

[0251] It should be noted that the pavement condition report includes:

[0252] Basic information: airport code, runway number, analysis time, number of flights, and aircraft type distribution;

[0253] Vertical acceleration data statistics: mean, standard deviation, maximum, minimum, RMS, skewness, kurtosis, and percentiles;

[0254] Identified pavement defects: defect type, severity, and affected area;

[0255] Pavement condition assessment: Comprehensive score, condition level, and comparison results with historical data;

[0256] Repair recommendations: repair priority, specific repair measures, and estimated repair costs;

[0257] Trend Analysis: Pavement condition changes and predictive maintenance recommendations.

[0258] In this embodiment, multi-dimensional information is integrated and output to achieve a closed-loop logic across the entire chain, from underlying data to maintenance recommendations.

[0259] Pavement condition reports are automatically generated and support multiple output formats, greatly improving operation and maintenance efficiency and accuracy.

[0260] In summary, this technical solution obtains QAR data of all aircraft at the target airport, preprocesses the QAR data to obtain effective data segments of the aircraft during the taxiing phase at the target airport, extracts vertical acceleration signals from the data, and combines aircraft-specific processing and multi-domain feature extraction to achieve pavement defect type identification.

[0261] Based directly on QAR data (aircraft-generated data), without relying on additional detection vehicles or equipment, it can automatically collect data with each landing, and every aircraft can participate in monitoring, enabling widespread deployment. Assessment is triggered with each landing, improving response efficiency. It solves the problems of untimely detection, high cost, and reliance on manual labor inherent in traditional methods, demonstrating good engineering feasibility, versatility, and promotional value.

[0262] Example 2

[0263] It should be noted that the method also includes:

[0264] Obtain vertical acceleration data for different models;

[0265] The vertical acceleration data of different models were normalized.

[0266] The standard deviation difference between different models is calculated based on the normalized vertical acceleration data, and the vertical acceleration data of abnormal models are removed.

[0267] Calculate the airport pavement condition assessment results for the same airport pavement for different aircraft types, and perform cross-validation;

[0268] Output the cross-validation results and use these results to determine the accuracy of the airport pavement condition assessment.

[0269] Different aircraft models will have different acceleration responses due to structural differences (such as landing gear position, center of gravity arrangement, etc.). In this embodiment, after normalization processing, the responses of different aircraft models can be classified into a unified evaluation dimension.

[0270] Some flights may cause data deviations due to abnormal landing attitudes, sensor drift, etc. By using standard deviation difference detection and elimination strategies, the interference of abnormal aircraft data on the evaluation results can be identified and eliminated in advance.

[0271] This embodiment introduces a cross-aircraft cross-validation mechanism to improve the consistency and reliability of pavement defect identification. If multiple aircraft types identify consistent defects on the same runway, the reliability of the defect's existence is higher; if multiple aircraft types identify inconsistent defects, it can indicate that the model needs calibration or that the data is incorrect. Through the consistency verification mechanism of the model output, the verification capability and reliability of pavement defect identification results are improved, assisting airports in making more confident decisions.

[0272] Example 3

[0273] An intelligent airport pavement condition recognition system based on QAR data, referring to Figure 5 ,include:

[0274] Data preprocessing module 41 is used to acquire QAR data of all aircraft at the target airport, preprocess the QAR data, and obtain the effective data segments of the aircraft during the taxiing phase at the target airport.

[0275] The data purification module 42 is used to extract vertical acceleration data from the effective data segment, and to perform noise reduction processing on the vertical acceleration data according to the aircraft type information to obtain purified vertical acceleration data.

[0276] The feature extraction module 43 is used to perform multi-domain feature extraction on the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectral features;

[0277] Evaluation module 44 is used to identify pavement defect types based on a preset pavement defect identification model, according to time-domain statistical characteristics and frequency-domain spectral characteristics, and to evaluate pavement condition results using a computer.

[0278] The report generation module 45 is used to generate a pavement condition report based on the pavement condition assessment results.

[0279] This technical solution acquires QAR data of all aircraft at the target airport, preprocesses the QAR data to obtain effective data segments of the aircraft during the taxiing phase at the target airport, extracts vertical acceleration signals from these segments, and combines aircraft-specific processing and multi-domain feature extraction to achieve identification of pavement defect types.

[0280] Based directly on QAR data (aircraft-generated data), without relying on additional detection vehicles or equipment, it can automatically collect data with each landing, and every aircraft can participate in monitoring, enabling widespread deployment. Assessment is triggered with each landing, improving response efficiency. It solves the problems of untimely detection, high cost, and reliance on manual labor inherent in traditional methods, demonstrating good engineering feasibility, versatility, and promotional value.

[0281] It is understood that the same or similar parts in the above embodiments can be referred to each other, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0282] It should be noted that in the description of this application, the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means at least two.

[0283] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the function involved, as will be understood by those skilled in the art to which embodiments of this application pertain.

[0284] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0285] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0286] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0287] The storage media mentioned above can be read-only memory, disk, or optical disk, etc.

[0288] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0289] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. An airport pavement condition intelligent identification method based on QAR data, characterized in that, The method comprises the following steps: obtaining QAR data of all aircrafts in a target airport, preprocessing the QAR data to obtain an effective data segment of the aircrafts in a taxiing phase in the target airport; extracting vertical acceleration data from the effective data segment, and performing denoising processing on the vertical acceleration data according to aircraft model information to obtain purified vertical acceleration data; performing multi-domain feature extraction on the purified vertical acceleration data to obtain time-domain statistical features and frequency-domain spectral features; establishing a pavement disease feature mode library; based on a preset pavement disease identification model, identifying a pavement disease type according to the time-domain statistical features and the frequency-domain spectral features, and calculating an airport pavement condition evaluation result; generating a pavement condition report according to the pavement condition evaluation result; wherein the pavement disease feature mode library comprises standard deviation thresholds, dominant frequency ranges, amplitude thresholds and high-frequency energy ratios of different pavement disease types; the time-domain statistical features comprise mean, standard deviation, maximum, minimum, effective value, skewness, kurtosis and percentile of the vertical acceleration data; the frequency-domain spectral features comprise dominant frequency, power spectral density, energy distribution of each frequency band and spectral peak feature; based on a preset pavement disease identification model, identifying a pavement disease type according to the time-domain statistical features and the frequency-domain spectral features, and calculating an airport pavement condition evaluation result, comprising: matching the time-domain statistical features and the frequency-domain spectral features with the pavement disease feature mode library to match the pavement disease type; calculating the matching degree score of each pavement disease type; when the matching degree score of the pavement disease type exceeds a preset matching degree threshold, it is determined that the corresponding pavement disease type exists; calculating the matching degree score of each pavement disease type, comprising: performing the first judgment item, the second judgment item, the third judgment item and the fourth judgment item on each pavement disease type in turn, calculating the matching degree score of the pavement disease type in each judgment item, and taking the total score of the pavement disease type in all judgment items as the final matching degree score of the pavement disease type; the first judgment item comprises: if the standard deviation of the vertical acceleration data is greater than the standard deviation threshold of the current pavement disease type, calculating the score of the current pavement disease type in the current judgment item according to the first weight, the standard deviation of the vertical acceleration data and the baseline standard deviation, and performing the next judgment item; if not, recording the score of the current judgment item as 0 and performing the next judgment item; the second judgment item comprises: if the dominant frequency of the vertical acceleration data meets the dominant frequency range of the current pavement disease type, calculating the score of the current pavement disease type in the current judgment item according to the second weight and the first preset coefficient, and performing the next judgment item; if not, recording the score of the current judgment item as 0 and performing the next judgment item; the third judgment item comprises: if the maximum acceleration value of the vertical acceleration data is greater than the amplitude threshold of the current pavement disease type, calculating the score of the current pavement disease type in the current judgment item according to the third weight, the maximum acceleration value of the vertical acceleration data and the amplitude threshold of the current pavement disease type, and performing the next judgment item; if not, recording the score of the current judgment item as 0 and performing the next judgment item; the fourth judgment item comprises: If the proportion of abnormal points in the time domain statistical features and the frequency domain spectral features is greater than a preset abnormal proportion threshold, a score of the current pavement disease type in the current judgment item is calculated according to the fourth weight, the proportion of abnormal points in the time domain statistical features and the frequency domain spectral features, and a second preset coefficient, and a next judgment item is executed; otherwise, the score of the current judgment item is recorded as 0. The first weight is greater than the third weight, the third weight is greater than the second weight, and the second weight is greater than the fourth weight.

2. The method of claim 1, wherein, The QAR data includes: aircraft weight data, vertical acceleration data, lateral acceleration data, longitudinal acceleration data, ground speed data, air pressure altitude data, left main wheel state, right main wheel state, front wheel state, and throttle lever state. The QAR data is preprocessed to obtain an effective data segment of the aircraft in the taxiing phase at the target airport, including: checking the data integrity of the QAR data and filtering the QAR data with missing values exceeding a preset missing threshold; identifying and processing abnormal values in the QAR data based on an empirical normal distribution; determining the landing ground contact time point as the starting time point of the effective data segment according to the time points at which the left main wheel, the right main wheel, and the front wheel are switched from the air state to the ground state; taking the time point at which the throttle lever returns to the IDLE position as the termination time point of the effective data segment; extracting the effective data segment from the processed QAR data according to the starting time point and the termination time point of the effective data segment; performing time synchronization and sampling rate standardization processing on the effective data segment.

3. The method of claim 1, wherein, The aircraft model information includes: baseline standard deviation, inherent vibration frequency, and normal acceleration range interval; performing denoising processing on the vertical acceleration data according to the aircraft model information, including: calculating the signal mean value of the vertical acceleration data and removing the direct current component according to the signal mean value of the vertical acceleration data; removing high-frequency noise in the vertical acceleration data through a 4th order Butterworth low-pass filter with a cutoff frequency of 20 Hz; designing a band-stop filter according to the inherent vibration frequency in the model information to remove the inherent vibration frequency of the aircraft in the vertical acceleration data; constructing the judgment condition of the sliding window detection method according to the baseline standard deviation in the model information, and identifying and removing abnormal impacts in the vertical acceleration data using the sliding window detection method; removing the vertical acceleration data that exceeds the normal acceleration range interval according to the normal acceleration range interval in the model information.

4. The method of claim 1, wherein, The method further includes: identifying and processing abnormal points in the time domain statistical features and the frequency domain spectral features based on an empirical normal distribution.

5. The method of claim 4, wherein, The disease types include: unevenness, misalignment, fracture, settlement, void, and base softening.

6. The method of claim 1, wherein, calculating the airport pavement condition assessment result, including: setting deduction items and maximum deduction values for each deduction item; the deduction items include: fluctuation deduction item, abnormal point deduction item, peak impact deduction item, distribution abnormality deduction item, and high-frequency energy deduction item; calculating the fluctuation deduction value according to the standard deviation of the vertical acceleration data and the normal standard deviation, and according to the standard deviation deviation degree and the maximum deduction value of the fluctuation deduction item; calculating the abnormal point deduction value according to the proportion of abnormal points in the time domain statistical features and the frequency domain spectral features, and according to the maximum deduction value of the abnormal point deduction item; and calculating the airport pavement condition assessment result according to the fluctuation deduction value, the abnormal point deduction value, the peak impact deduction value, the distribution abnormality deduction value, and the high-frequency energy deduction value. When the maximum acceleration value of the vertical acceleration data is greater than 0.5g, a peak impact deduction value is calculated according to the maximum acceleration value of the vertical acceleration data and a maximum deduction value of the peak impact deduction item; a distribution anomaly deduction value is calculated according to the distribution characteristics of the vertical acceleration data and a maximum deduction value of the distribution anomaly deduction item; a high-frequency energy deduction value is calculated according to the energy distribution of each frequency band of the vertical acceleration data and a maximum deduction value of the high-frequency energy deduction item; an airport runway surface condition evaluation score is obtained by subtracting the deduction values of the items from a preset full score; a corresponding runway surface condition grade is matched according to the airport runway surface condition evaluation score.

7. The method of claim 6, wherein, The runway surface condition report includes: basic information: airport code, runway number, analysis time, number of flights, and aircraft type distribution; vertical acceleration data statistics: mean, standard deviation, maximum value, minimum value, valid value, skewness, kurtosis, and percentile; identified runway disease types: disease type, severity, and affected area; runway surface condition evaluation: comprehensive score, condition grade, and historical data comparison result; repair recommendations: repair priority, specific repair measures, and estimated repair cost; trend analysis: runway condition change trend and predictive maintenance recommendations.

8. The method of claim 1, wherein, The method further includes: obtaining vertical acceleration data of different aircraft types; normalizing the vertical acceleration data of different aircraft types; calculating the standard deviation difference of different aircraft types according to the normalized vertical acceleration data of different aircraft types, and removing the vertical acceleration data of abnormal aircraft types; calculating the airport runway surface condition evaluation results of different aircraft types on the same airport runway, and performing cross-validation; outputting the cross-validation results and determining the accuracy of the airport runway surface condition evaluation results according to the cross-validation results.

9. An airport pavement condition intelligent identification system based on QAR data, used to implement the airport pavement condition intelligent identification method based on QAR data according to any one of claims 1-8, characterized in that, It includes: a data preprocessing module for obtaining QAR data of all aircraft at a target airport, preprocessing the QAR data, and obtaining valid data segments of aircraft during taxiing at the target airport; a data purification module for extracting vertical acceleration data from the valid data segments, denoising the vertical acceleration data according to the aircraft type information, and obtaining purified vertical acceleration data; a feature extraction module for extracting multi-domain features from the purified vertical acceleration data, obtaining time domain statistical features and frequency domain spectral features; an evaluation module for identifying runway disease types based on a preset runway disease identification model, calculating airport runway surface condition evaluation results according to the time domain statistical features and the frequency domain spectral features; a report generation module for generating a runway condition report according to the runway condition evaluation results.

Images