Spacecraft multichannel telemetry data anomaly detection method based on correlation analysis

Abstract

The invention discloses a spacecraft multichannel telemetry data anomaly detection method based on correlation analysis. The method comprises the following steps: acquiring multichannel telemetry data of a spacecraft; preprocessing the multi-channel telemetry data; the method comprises the following steps: selecting multi-channel telemetry data of an abnormal detection period, determining a time lag relationship among the multi-channel telemetry data based on the selected multi-channel telemetry data, calculating a correlation coefficient between every two channels of telemetry data by using a modified Pearson's correlation coefficient calculation formula, and forming a correlation coefficient matrix; calculating correlation characteristics corresponding to the correlation coefficient matrix; and judging whether the multichannel telemetry data of the spacecraft is abnormal in the current anomaly detection period or not according to the correlation characteristics acquired in the current anomaly detection period and the correlation characteristics acquired by using the telemetry data of the historical detection period when no fault exists. According to the method, anomaly diagnosis can be carried out only through multi-channel telemetry data, and relevance and time hysteresis among the multi-channel telemetry data can be considered at the same time.

Description

Technical Field

[0001] The present invention relates to the field of spacecraft technology, and in particular to a method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis. Background Art

[0002] Spacecraft are key tools for exploring space and orbiting the Earth. They can be used to perform a variety of special tasks, including scientific research, communications, Earth observation, satellite deployment, etc. Since spacecraft work in extreme cosmic environments and usually operate with complex scientific instruments and systems, they are susceptible to various faults. Since failures in spacecraft systems and stand-alone products can cause abnormalities in related parameter data, spacecraft fault diagnosis based on data analysis is currently a common method of spacecraft fault diagnosis. Specifically, on-orbit telemetry data is used to monitor the health status of the spacecraft in real time. If the telemetry data is found to be abnormal, it is judged that the spacecraft may have a fault.

[0003] At present, the conventional methods for detecting anomalies in telemetry data mainly include: manual judgment method, over-limit alarm method, expert knowledge method and model comparison method. The manual judgment method refers to the technicians of each subsystem paying attention to the telemetry data related to their subsystem in real time, and judging whether the system operation status is normal according to the document regulations; the over-limit alarm method refers to setting upper and lower limits for certain important status parameters, and issuing an alarm once the telemetry data exceeds the set threshold; the expert knowledge method refers to professionals building fault rules for each single machine, component and subsystem to form a fault rule library for the entire system. If the telemetry data matches the conditions of the fault rules, a fault alarm is issued; the model comparison method is to build a digital simulation model of the system, and compare the predicted value of the digital simulation model with the actual telemetry data in real time to see if they are consistent. When the deviation between the two exceeds the set threshold, the state is judged to be abnormal.

[0004] In addition, in recent years, with the development of big data and artificial intelligence technology, more and more machine learning methods are used to "mine" the knowledge contained in aerospace big data, and a variety of telemetry data anomaly detection methods based on neural networks have been developed, such as using convolutional neural networks (CNN) to describe high-dimensional telemetry features, using long short-term memory neural networks (LSTM) for telemetry time series prediction, building Transformer models to achieve high-dimensional telemetry encoding and decoding, and using transfer learning frameworks to solve the problem of scarce spacecraft fault label data. The basic concept of the telemetry data anomaly detection method based on neural networks is to automatically extract the features of telemetry based on historical data, learn the functional relationship between the features, and then determine whether the telemetry data is normal.

[0005] However, in the above-mentioned telemetry data anomaly detection methods, the manual judgment method requires technicians to be on duty at a fixed location 24 hours a day, always paying attention to the values ​​of the subsystem telemetry parameters, which consumes a lot of energy of the technicians; and the manual judgment method also has problems such as incomplete monitoring of telemetry measurements and failure to consider the correlation between multi-channel telemetry measurements, and cannot timely discover the signs of faults. Although the over-limit alarm method can realize automatic fault judgment and reduce the workload of technicians to a certain extent, the telemetry data will inevitably introduce noise and bad points during the collection, storage and transmission process, resulting in a high false alarm rate; and the setting of the alarm threshold has a great impact on the judgment result, and it is difficult to find a reasonable threshold to make the missed alarm rate and false alarm rate low; in addition, the over-limit alarm method is generally only for a single telemetry measurement, and does not consider the correlation between multi-channel telemetry measurements. Although the expert knowledge method can realize automatic fault judgment and consider the correlation between multi-channel telemetry measurements, it needs to build a fault judgment rule base in the early stage, and requires engineering and technical personnel to sort out and integrate rules and fault modes, which is a large workload and difficult to cover all states. The model comparison method has the problem of difficulty in building digital models, especially for complex nonlinear spacecraft systems. The construction of digital models is time-consuming and labor-intensive, and the model confidence is not high, making it difficult to truly play a role in the on-orbit operation phase of the spacecraft. The anomaly detection method of telemetry data based on neural networks has the problems of lack of spacecraft fault label data, which makes machine learning lack of sufficient samples, poor interpretability of neural networks, which makes the confidence of fault prediction questionable, and the correlation and impact lag of multi-channel telemetry data, which requires high machine learning modeling capabilities. There are great limitations in practical application. Summary of the invention

[0006] In order to solve some or all of the technical problems existing in the above-mentioned prior art, the present invention provides a method for detecting anomalies in multi-channel telemetry data of a spacecraft based on correlation analysis.

[0007] The technical solution of the present invention is as follows:

[0008] A method for detecting anomaly in multi-channel telemetry data of a spacecraft based on correlation analysis is provided, the method comprising:

[0009] Acquire multi-channel telemetry data from spacecraft;

[0010] Preprocessing the multi-channel telemetry data to ensure that the multi-channel telemetry data are aligned in time and that the telemetry data of each channel has no outliers and missing values;

[0011] Select multi-channel telemetry data of an anomaly detection cycle, determine the time lag relationship between the multi-channel remote sensing data based on the selected multi-channel telemetry data, calculate the correlation coefficient between every two channel telemetry data using a preset modified Pearson correlation coefficient calculation formula, and form a correlation coefficient matrix;

[0012] Calculate the correlation feature corresponding to the correlation coefficient matrix, where the correlation feature corresponding to the correlation coefficient matrix is ​​the sum of squares of all eigenvalues ​​of the correlation coefficient matrix;

[0013] Based on the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current anomaly detection cycle and the correlation characteristics corresponding to the correlation coefficient matrix obtained using the telemetry data of the historical detection cycle when there was no fault, it is determined whether the multi-channel telemetry data of the spacecraft is abnormal in the current anomaly detection cycle.

[0014] In some optional implementations, the preprocessing of the multi-channel telemetry data includes:

[0015] Time alignment of multi-channel telemetry data;

[0016] Outlier removal and missing value filling are performed on each channel telemetry data.

[0017] In some optional implementations, the time lag relationship between the multi-channel remote sensing data is determined by:

[0018] Set the maximum time interval number P;

[0019] For any two channels of telemetry data, set the change of one channel's telemetry data to lag behind the change of the other channel's telemetry data by p time intervals, so that p takes values ​​from -P to P one by one, and based on p after each value taking, use the modified Pearson correlation coefficient calculation formula to calculate the correlation coefficient between the current two channels' telemetry data;

[0020] The time interval number corresponding to the maximum correlation coefficient between the telemetry data of two channels is taken as the time lag relationship between the telemetry data of the two channels, and the time lag relationship between the telemetry data of multiple channels is determined.

[0021] In some optional implementations, the modified Pearson correlation coefficient calculation formula is expressed as:

[0022]

[0023] in, It represents the correlation coefficient between the telemetry data of channel i and the telemetry data of channel j when the change of the telemetry data of channel i lags behind the change of the telemetry data of channel j by p time intervals. If p<0, then If p ≥ 0, then Indicates t 1+|p| Channel i telemetry data at time x i n Indicates t n Channel i telemetry data at the moment, represents the telemetry data of channel j at time t1, Indicates t n-|p| Channel j telemetry data at time, represents the telemetry data of channel i at time t1, Indicates t n-p Channel i telemetry data at the moment, Indicates t 1+p Channel j telemetry data at time, Indicates t n Telemetry data of channel j at time instant.

[0024] In some optional implementations, the two channel telemetry data are set to be channel i telemetry data and channel j telemetry data, respectively, and the time lag relationship between the two channel telemetry data is expressed as:

[0025]

[0026] Among them, p opt Represents the time lag relationship between the telemetry data of two channels, express The p-value at which the maximum value is obtained.

[0027] In some optional implementations, the correlation feature corresponding to the correlation coefficient matrix is ​​expressed as:

[0028]

[0029] Among them, F k represents the correlation feature, λ q represents the qth eigenvalue of the correlation coefficient matrix, and Q represents the number of eigenvalues ​​of the correlation coefficient matrix.

[0030] In some optional implementations, judging whether the multi-channel telemetry data of the spacecraft is abnormal in the current abnormality detection period according to the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current abnormality detection period and the correlation characteristics corresponding to the correlation coefficient matrix obtained using the telemetry data of the historical detection period when there is no fault, includes:

[0031] The correlation characteristics are assumed to obey the normal distribution, and the mean and variance of the normal distribution are estimated based on the correlation characteristics corresponding to the correlation coefficient matrix obtained by using the telemetry data of multiple historical detection cycles when there is no fault and the maximum likelihood method;

[0032] Determine whether the correlation feature corresponding to the correlation coefficient matrix obtained in the current anomaly detection cycle is within the interval If yes, it is determined that there is no abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle; if no, it is determined that there is an abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle;

[0033] in, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the standard deviation of the normal distribution that the correlation characteristic follows.

[0034] In some optional implementations, the mean and variance of the normal distribution obeyed by the correlation feature are solved by the following expressions:

[0035]

[0036] in, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the variance of the normal distribution that the correlation feature follows, {F1,F2,…,F K-1} represents the correlation feature corresponding to the correlation coefficient matrix obtained by using the telemetry data of the historical detection period without faults, F1 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the first historical detection period, F2 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the second historical detection period, and F K-1 Represents the correlation characteristics corresponding to the correlation coefficient matrix obtained in the K-1th historical detection period.

[0037] The main advantages of the technical solution of the present invention are as follows:

[0038] The spacecraft multi-channel telemetry data anomaly detection method based on correlation analysis of the present invention can perform fault and anomaly diagnosis using only the measured data of the spacecraft multi-channel telemetry data, and can simultaneously consider the correlation and time lag between the multi-channel telemetry data. The detection results are highly accurate and reliable, the detection workload is small, and the scope of application is wide. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings described herein are used to provide a further understanding of the embodiments of the present invention and constitute a part of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0040] Figure 1 A flowchart of a method for detecting anomalies in multi-channel telemetry data of a spacecraft based on correlation analysis provided by an embodiment of the present invention;

[0041] Figure 2A schematic diagram of a multi-channel telemetry data structure provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the specific embodiments of the present invention and the corresponding drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0043] The technical solution provided by the embodiments of the present invention is described in detail below with reference to the accompanying drawings.

[0044] refer to Figure 1 The embodiment of the present invention provides a method for detecting anomalies in multi-channel telemetry data of a spacecraft based on correlation analysis, the method comprising the following steps:

[0045] Step 1, obtaining multi-channel telemetry data of the spacecraft;

[0046] In the embodiment of the present invention, the acquired multi-channel telemetry data of the spacecraft includes the multi-channel telemetry data of the spacecraft at the current moment and the multi-channel telemetry data of the spacecraft at the historical moment before the current moment.

[0047] Step 2, preprocessing the multi-channel telemetry data so that the multi-channel telemetry data are aligned in time and the telemetry data of each channel has no outliers and missing values;

[0048] In an embodiment of the present invention, since the on-orbit telemetry data of the spacecraft has the characteristics of multi-channel, time asynchronous, containing outliers, containing missing values, etc., the multi-channel telemetry data is preprocessed so that the multi-channel telemetry data are aligned in time and the telemetry data of each channel are free of outliers and missing values, so as to ensure the accuracy and reliability of subsequent detection results.

[0049] Step 3, selecting multi-channel telemetry data of an anomaly detection cycle, determining the time lag relationship between the multi-channel remote sensing data based on the selected multi-channel telemetry data, and calculating the correlation coefficient between every two channel telemetry data using a preset modified Pearson correlation coefficient calculation formula to form a correlation coefficient matrix;

[0050] In the embodiment of the present invention, the anomaly detection cycle is set according to actual needs. In the process of detecting anomalies in multi-channel telemetry data of a spacecraft, after an anomaly detection cycle has passed, the multi-channel telemetry data of the current anomaly detection cycle can be selected for detection.

[0051] In an embodiment of the present invention, the correlation coefficient matrix is ​​composed of all calculated correlation coefficients, the element in the i-th row and j-th column of the correlation coefficient matrix is ​​the correlation coefficient between the telemetry data of channel i and the telemetry data of channel j, and the element in the i-th row and i-th column of the correlation coefficient matrix is ​​the correlation coefficient between the telemetry data of channel i and the telemetry data of channel i.

[0052] Step 4, calculating the correlation characteristics corresponding to the correlation coefficient matrix;

[0053] In the embodiment of the present invention, the correlation feature corresponding to the correlation coefficient matrix is ​​defined as the sum of squares of all eigenvalues ​​of the correlation coefficient matrix.

[0054] It should be noted that the eigenvalues ​​of the correlation coefficient matrix are determined using conventional mathematical calculation methods, which will not be described in detail here.

[0055] Step 5: Determine whether the spacecraft multi-channel telemetry data is abnormal in the current anomaly detection cycle based on the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current anomaly detection cycle and the correlation characteristics corresponding to the correlation coefficient matrix obtained using the telemetry data of the historical detection cycle when there is no fault.

[0056] In an embodiment of the present invention, a corresponding correlation coefficient matrix can be obtained for each detection cycle, and one correlation coefficient matrix corresponds to a correlation feature. Based on the correlation features corresponding to the correlation coefficient matrix obtained in the current abnormal detection cycle and the correlation features corresponding to the correlation coefficient matrix obtained using telemetry data of historical detection cycles when there is no fault, it is determined whether the multi-channel telemetry data of the spacecraft is abnormal in the current abnormal detection cycle.

[0057] The spacecraft multi-channel telemetry data anomaly detection method based on correlation analysis provided in an embodiment of the present invention can perform fault and anomaly diagnosis using only the measured data of the spacecraft multi-channel telemetry data, and can simultaneously consider the correlation and time lag between the multi-channel telemetry data. The detection results are highly accurate and reliable, the detection workload is small, and the scope of application is wide.

[0058] Furthermore, in an optional implementation of the embodiment of the present invention, preprocessing the multi-channel telemetry data further includes the following steps:

[0059] Step 201, time aligning multi-channel telemetry data;

[0060] Since telemetry data from different channels may use different sensors and signal acquisition circuits, multi-channel telemetry data may be time-asynchronous. Therefore, time alignment is performed on the multi-channel telemetry data.

[0061] In the embodiment of the present invention, a linear interpolation method is used for time alignment.

[0062] Step 202, remove outliers and fill missing values ​​for each channel telemetry data.

[0063] Since telemetry data may be disturbed during the collection, storage and transmission process, which may cause outliers and / or missing values ​​in the final received telemetry data, outliers are removed and missing values ​​are filled in the telemetry data of each channel.

[0064] In the embodiment of the present invention, a rolling statistics method is used to determine outliers, and a nearest neighbor method is used to determine missing values.

[0065] refer to Figure 2 , attached Figure 2 A multi-channel telemetry data structure with m channels after preprocessing is shown. Figure 2 middle, Indicates t N Telemetry data of channel m at the moment.

[0066] In the embodiment of the present invention, by performing time alignment, outlier removal and missing value filling on multi-channel telemetry data, the multi-channel telemetry data can be aligned in time and the telemetry data of each channel can be continuous in time.

[0067] Further, in an optional implementation of the embodiment of the present invention, the time lag relationship between the multi-channel telemetry data is determined by:

[0068] Step 301, setting the maximum time interval number P;

[0069] Step 302: for any two channel telemetry data, set the change of one channel telemetry data to lag behind the change of the other channel telemetry data by p time intervals, so that p takes values ​​from -P to P one by one, and based on p after each value taking, use the modified Pearson correlation coefficient calculation formula to calculate the correlation coefficient between the current two channel telemetry data;

[0070] Step 303, taking the time interval corresponding to the maximum correlation coefficient between the telemetry data of two channels as the time lag relationship between the telemetry data of the two channels, and determining the time lag relationship between the telemetry data of multiple channels.

[0071] In the embodiment of the present invention, the maximum time interval is set according to the actual situation. Specifically, it can be set according to the maximum time lag that may exist between multi-channel remote sensing data to ensure that the product of the maximum time interval and the time difference between two adjacent remote sensing data is not less than the maximum time lag.

[0072] Further, in an embodiment of the present invention, the modified Pearson correlation coefficient calculation formula is expressed as:

[0073]

[0074] in, It represents the correlation coefficient between the telemetry data of channel i and the telemetry data of channel j when the change of the telemetry data of channel i lags behind the change of the telemetry data of channel j by p time intervals. If p<0, then If p ≥ 0, then Indicates t 1+|p| Channel i telemetry data at the moment, Indicates t n Channel i telemetry data at the moment, represents the telemetry data of channel j at time t1, Indicates t n-|p| Channel j telemetry data at time, represents the telemetry data of channel i at time t1, Indicates t n-p Channel i telemetry data at the moment, Indicates t 1+p Channel j telemetry data at time, Indicates t n The telemetry data of channel j at time instant.

[0075] Further, in the embodiment of the present invention, based on the above-defined method for determining the time lag relationship between the multi-channel telemetry data, taking two channel telemetry data as channel i telemetry data and channel j telemetry data as an example, the number of time intervals corresponding to the maximum correlation coefficient between the two channel telemetry data, that is, the time lag relationship between the two channel telemetry data can be expressed as:

[0076]

[0077] Among them, p opt It represents the time interval corresponding to the maximum correlation coefficient between the telemetry data of two channels, that is, the time lag relationship between the telemetry data of two channels. express The p-value at which the maximum value is obtained.

[0078] In the embodiment of the present invention, the time lag relationship between multi-channel telemetry data is determined in the above manner, and the correlation coefficient between the channel telemetry data is calculated using the above-mentioned modified Pearson correlation coefficient calculation formula. This can take into account the possible time lag characteristics between the channel telemetry data, and ensure that the calculated Pearson correlation coefficient can truly reflect the linear correlation between the channel telemetry data.

[0079] Furthermore, in the embodiment of the present invention, the correlation feature corresponding to the correlation coefficient matrix can be expressed as:

[0080]

[0081] Among them, F k represents the correlation feature, λ q represents the qth eigenvalue of the correlation coefficient matrix, and Q represents the number of eigenvalues ​​of the correlation coefficient matrix.

[0082] Further, in an optional implementation of the embodiment of the present invention, judging whether the multi-channel telemetry data of the spacecraft is abnormal in the current abnormality detection period according to the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current abnormality detection period and the correlation characteristics corresponding to the correlation coefficient matrix obtained using the telemetry data of the historical detection period when there is no fault, further includes the following steps:

[0083] Step 501, setting the correlation characteristics to obey the normal distribution, and estimating the mean and variance of the normal distribution based on the correlation characteristics corresponding to the correlation coefficient matrix obtained by using the telemetry data of multiple historical detection cycles when there is no fault and the maximum likelihood method;

[0084] Step 502: determine whether the correlation feature corresponding to the correlation coefficient matrix obtained in the current abnormal detection period is within the interval If yes, it is determined that there is no abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle; if no, it is determined that there is an abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle.

[0085] In an embodiment of the present invention, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the standard deviation of the normal distribution that the correlation characteristic follows.

[0086] Further, in the embodiment of the present invention, the current anomaly detection cycle is set to be the Kth anomaly detection cycle, and the historical detection cycle is from the 1st anomaly detection cycle to the K-1th anomaly detection cycle. The mean and variance of the normal distribution obeyed by the correlation feature are solved by the following expressions:

[0087]

[0088] in, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the variance of the normal distribution that the correlation feature follows, {F1,F2,…,F K-1} represents the correlation feature corresponding to the correlation coefficient matrix obtained by using the telemetry data of the historical detection period without faults, F1 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the first historical detection period, F2 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the second historical detection period, and F K-1 Represents the correlation characteristics corresponding to the correlation coefficient matrix obtained in the K-1th historical detection period.

[0089] In the embodiment of the present invention, by determining whether the multi-channel telemetry data of the spacecraft is abnormal in the current abnormality detection cycle in the above manner, the accuracy and reliability of the detection result can be improved.

[0090] It should be noted that, in this article, relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In addition, "front", "back", "left", "right", "upper" and "lower" in this article are all referenced to the placement state shown in the accompanying drawings.

[0091] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for detecting anomaly in multi-channel telemetry data of spacecraft based on correlation analysis, characterized in that: include: Acquire multi-channel telemetry data from spacecraft; Preprocessing the multi-channel telemetry data to ensure that the multi-channel telemetry data are aligned in time and that the telemetry data of each channel has no outliers and missing values; Select multi-channel telemetry data of an anomaly detection cycle, determine the time lag relationship between the multi-channel remote sensing data based on the selected multi-channel telemetry data, calculate the correlation coefficient between every two channel telemetry data using a preset modified Pearson correlation coefficient calculation formula, and form a correlation coefficient matrix; Calculate the correlation feature corresponding to the correlation coefficient matrix, where the correlation feature corresponding to the correlation coefficient matrix is ​​the sum of squares of all eigenvalues ​​of the correlation coefficient matrix; Based on the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current anomaly detection cycle and the correlation characteristics corresponding to the correlation coefficient matrix obtained using the telemetry data of the historical detection cycle when there was no fault, it is determined whether the multi-channel telemetry data of the spacecraft is abnormal in the current anomaly detection cycle.

2. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 1, characterized in that: The preprocessing of the multi-channel telemetry data includes: Time alignment of multi-channel telemetry data; Outlier removal and missing value filling are performed on each channel telemetry data.

3. The method for detecting anomaly of spacecraft multi-channel telemetry data based on correlation analysis according to claim 1 is characterized in that: The time lag relationship between multi-channel remote sensing data is determined by: Set the maximum time interval number P; For any two channels of telemetry data, set the change of one channel's telemetry data to lag behind the change of the other channel's telemetry data by p time intervals, so that p takes values ​​from -P to P one by one, and based on p after each value taking, use the modified Pearson correlation coefficient calculation formula to calculate the correlation coefficient between the current two channels' telemetry data; The time interval number corresponding to the maximum correlation coefficient between the telemetry data of two channels is taken as the time lag relationship between the telemetry data of the two channels, and the time lag relationship between the telemetry data of multiple channels is determined.

4. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 1 or 3, characterized in that: The calculation formula of the modified Pearson correlation coefficient is expressed as: in, It represents the correlation coefficient between the telemetry data of channel i and the telemetry data of channel j when the change of the telemetry data of channel i lags behind the change of the telemetry data of channel j by p time intervals. If p<0, then If p ≥ 0, then Indicates t 1+|p| Channel i telemetry data at the moment, Indicates t n Channel i telemetry data at the moment, represents the telemetry data of channel j at time t1, Indicates t n-|p| Channel j telemetry data at time, represents the telemetry data of channel i at time t1, Indicates t n-p Channel i telemetry data at the moment, Indicates t 1+p Channel j telemetry data at time, Indicates t n Telemetry data of channel j at time instant.

5. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 4 is characterized in that: Assume that the two channel telemetry data are channel i telemetry data and channel j telemetry data, and the time lag relationship between the two channel telemetry data is expressed as: Among them, p opt Represents the time lag relationship between the telemetry data of two channels, express The p-value at which the maximum value is obtained.

6. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 1, characterized in that: The correlation characteristics corresponding to the correlation coefficient matrix are expressed as: Among them, F k represents the correlation feature, λ q represents the qth eigenvalue of the correlation coefficient matrix, and Q represents the number of eigenvalues ​​of the correlation coefficient matrix.

7. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 1, characterized in that: According to the correlation characteristics corresponding to the correlation coefficient matrix obtained in the current abnormal detection cycle and the correlation characteristics corresponding to the correlation coefficient matrix obtained by using the telemetry data of the historical detection cycle when there is no fault, it is judged whether the multi-channel telemetry data of the spacecraft is abnormal in the current abnormal detection cycle, including: The correlation characteristics are assumed to obey the normal distribution, and the mean and variance of the normal distribution are estimated based on the correlation characteristics corresponding to the correlation coefficient matrix obtained by using the telemetry data of multiple historical detection cycles when there is no fault and the maximum likelihood method; Determine whether the correlation feature corresponding to the correlation coefficient matrix obtained in the current anomaly detection cycle is within the interval If yes, it is determined that there is no abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle; if no, it is determined that there is an abnormality in the multi-channel telemetry data of the spacecraft in the current abnormality detection cycle; in, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the standard deviation of the normal distribution that the correlation characteristic follows.

8. The method for detecting anomalies in spacecraft multi-channel telemetry data based on correlation analysis according to claim 7, characterized in that: The mean and variance of the normal distribution followed by the correlation feature are solved by the following expressions: in, Indicates the mean of the normal distribution that the correlation feature follows, Indicates the variance of the normal distribution that the correlation feature follows, {F1,F2,…,F K-1 } represents the correlation feature corresponding to the correlation coefficient matrix obtained by using the telemetry data of the historical detection period without faults, F1 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the first historical detection period, F2 represents the correlation feature corresponding to the correlation coefficient matrix obtained in the second historical detection period, and F K-1 Represents the correlation characteristics corresponding to the correlation coefficient matrix obtained in the K-1th historical detection period.

Images