A method for detecting temperature anomaly of main reflector of intersatellite link antenna

Abstract

A method for detecting temperature anomalies in the main reflector of an intersatellite link antenna is proposed. This method first constructs a time-varying description of the heat flux outside the orbital space environment, providing a reference for establishing relative thermal energy parameters for the main reflector. Compared to directly analyzing the temperature parameters, the relative thermal energy parameters incorporate the effects of heat flux caused by space radiation, enabling more accurate characterization of temperature anomalies. Furthermore, the method employs a training-free Euler state network for anomaly feature extraction, significantly reducing computational energy consumption and memory requirements. This improves computational efficiency and saves computing power, making it feasible to deploy high-performance temperature anomaly detection devices on satellites.

Description

Technical Field

[0001] The present invention relates to the field of aerospace technology, and in particular to a method for detecting temperature anomaly of a main reflector of an intersatellite link antenna. Background Art

[0002] With my country's high-density satellite launches, the number of in-service satellites has increased dramatically, posing significant challenges to satellite monitoring and maintenance. Temperature monitoring of the main reflector (MR) of the intersatellite link antenna (ISL) is a crucial component. Due to the wide range of temperature variations, the MR is periodically heated and cooled by variations in external space radiation, resulting in significant temperature fluctuations. Even at the same moment, the temperature distribution on the satellite is uneven. To accurately assess the MR's temperature, it is essential to accurately analyze and calculate the external space heat flux to which the satellite is exposed. Furthermore, the harsh space environment significantly degrades the performance of the MR's temperature-measuring resistor (TMR). However, this degradation does not affect the normal operation of other components. Therefore, compensation for anomalies caused by this degradation is necessary. Furthermore, current algorithms for long-term time series modeling suffer from large parameter counts and high energy consumption, making them unsuitable for satellites with low computing power and power requirements.

[0003] The Euler state network is a new type of recursive neural network developed from the state echo network. The feature representation of this type of network is driven by an untrained internal dynamic system, which has obvious advantages in tasks that require long-term memory modeling. In addition, in terms of time series classification performance, it can match (or even exceed) the accuracy level of trainable recursive neural networks (such as recurrent neural networks RNN, long short-term memory neural networks LSTM, gated recurrent networks GRU and other mainstream timing networks), while allowing up to 100 times the computing time and energy consumption to be saved. Summary of the Invention

[0004] The purpose of the present invention is to provide a method for detecting temperature anomaly of a main reflector of an intersatellite link antenna.

[0005] The purpose of the present invention is achieved through such technical solution, and the specific steps are as follows:

[0006] 1) Obtain temperature data of the antenna main reflector through a sensor;

[0007] 2) Calculate the relative thermal energy parameters under the current temperature data;

[0008] 3) Analyze the relative thermal energy parameters through the Euler state network model to determine the degree of abnormality of the current temperature;

[0009] 4) Monitor sensor data and go to step 1) if there is new temperature data.

[0010] Preferably, the specific steps of calculating the relative thermal characterization energy parameter under the current temperature data in step 2) are as follows:

[0011] 2-1) Construct a time-varying description of the external heat flow in the orbital space environment;

[0012] 2-2) forming a binary discrete temperature sequence based on the acquired temperature of the antenna main reflector, and performing equal-time fusion on the binary discrete temperature sequence to obtain a first processing result;

[0013] 2-3) compensating for performance degradation on the first processing result to obtain a second result;

[0014] 2-4) Based on the time-varying description of the heat flow outside the orbital space environment and combined with the second processing result, the relative thermal characterization energy parameter, i.e., the third processing result, is calculated.

[0015] Preferably, the specific steps of analyzing the relative thermal characterization energy parameters by the Euler state network model in step 3) to determine the current temperature anomaly degree are as follows:

[0016] 3-1) Construct an Euler state network model, initialize the Euler state network and tune its hyperparameters;

[0017] 3-2) Extracting long-term time series features based on the Euler state network model according to the third processing result to obtain a fourth processing result;

[0018] 3-3) Calculate the degree of temperature anomaly based on the processing results.

[0019] Preferably, the construction of the time-varying description of the external heat flow of the orbital space environment in step 2-1) includes establishing a solar radiation external heat flow characterization model, establishing an earth sunlight reflection heat flow characterization model, and establishing an earth infrared radiation heat flow characterization model.

[0020] Preferably, the solar radiation external heat flow characterization model is:

[0021] Where q1 represents the external heat flux from solar radiation, It represents the average solar absorption ratio of the satellite in one orbital period, usually taken as 0.3, S represents sunlight, φ1 represents the solar radiation angular coefficient, A s-p It represents the projection area of ​​the satellite's outer surface perpendicular to the direction of sunlight. where τ s represents the sunshine time, τ c Indicates the operating time of one orbital period of the satellite.

[0022] The model for characterizing the heat flow reflected by the Earth's sunlight is established as follows:

[0023] Where q2 represents the heat flux reflected by the Earth's sunlight, ρ represents the average albedo, which is usually taken as 0.3, and A E-S Indicates the solar radiation angle coefficient, A s-p It represents the surface area of ​​the earth directly exposed to the sun, and φ2 represents the angular coefficient of sunlight reflection of the earth;

[0024] The model for characterizing the earth's infrared radiation heat flow is established as follows:

[0025] Where q3 represents the earth's infrared radiation heat flow, represents the average infrared emissivity of the satellite surface, A s represents the satellite's surface area, and φ3 represents the Earth's infrared radiation angular coefficient.

[0026] Preferably, in step 2-2), the acquired antenna main reflector temperature is formed into a temperature binary discrete sequence, and the temperature binary discrete sequence is subjected to equal time fusion to obtain the first processing result. The specific method is as follows:

[0027] Collecting the low-end temperature resistance data and the high-end temperature resistance data of the antenna main reflector, uniformly sampling the measured temperature data with a sampling period of T, and obtaining the measured temperature binary discrete sequence;

[0028] By s t =f(x 1t ,x 2t ,w1,w2) performs equal time fusion on the binary discrete sequence of the antenna main reflector temperature, where s t Characterizes the temperature of the main reflector of the intersatellite link antenna at time t, x 1t ,x 2t They are the low-end temperature and high-end temperature measured by the main reflector of the intersatellite link antenna at time t, w1 and w2 represent the fusion weight coefficients, respectively, and s t The time series set {s0,s1,s2,...,s n} is the first processing result;

[0029] in, When the temperature low-end resistance data reaches the upper limit T1, w1 takes the value of 0, and when the temperature high-end resistance data reaches the lower limit T2, w2 takes the value of 0.

[0030] Preferably, the specific method of performing performance degradation compensation on the first processing result in step 2-3) to obtain the second result is as follows:

[0031] By s t '=s t-δ(d) performs performance degradation compensation on the first processing result, where δ(d) represents the performance degradation model of the temperature measuring resistor, and s t 'The time series set {s0',s1',s2',...,s n '} is the second processing result;

[0032] Among them, δ(d) is the empirical model of resistor performance degradation Where d represents the service life of the temperature measuring resistor, c1 represents the degradation coefficient, and c2 represents the time coefficient. It represents the initial phase of the resistance performance, and its value is greater than or equal to 1. ε(d) represents the linear degradation term of the performance.

[0033] Preferably, the specific formula for calculating the relative thermal characterization energy parameter, i.e., the third processing result, in step 2-4) is as follows:

[0034] where Q r is the relative thermal energy parameter, which is a discrete sequence about time, κ is the total modulation coefficient, κ1, κ2, κ3 are the solar radiation heat flux modulation coefficient, the earth's sunlight reflection heat flux modulation coefficient and the earth's infrared radiation heat flux modulation coefficient respectively.

[0035] Preferably, the specific method of constructing the Euler state network model, initializing the Euler state network and adjusting the hyperparameters in step 3-1) is as follows:

[0036] From the interval [-w r ,w r ] randomly initialize a matrix W∈R from a uniform distribution on N×N The value of w r Is a positive recursive scaling factor, and then set the reservoir matrix W h =WW T ;

[0037] From [-w x ,w x The input weight matrix W is randomly initialized from the uniform distribution on x The value of w x is a positive input scaling factor;

[0038] From [-w b ,w b The value of the bias vector b is randomly initialized from a uniform distribution on ], where w b is a positive deviation scale coefficient.

[0039] Preferably, the specific method of extracting long time series features based on the Euler state network model according to the third processing result in step 3-2) to obtain the fourth processing result is as follows:

[0040] Traverse each element of the third processing result set in turn, and input it into the Euler state network for recursive calculation to obtain the output state set {h1,h2,h3,...,h n}, and the last output state h n As the output feature, that is, the fourth processing result.

[0041] Preferably, the specific method for calculating the degree of temperature anomaly based on the processing results in step 3-3) is as follows:

[0042] Collect the temperature normal Euler state network output state set {h1',h2',h3',...,h n '}, the sample set obeys the X distribution, which represents the characteristic distribution of the normal temperature state;

[0043] The center of the Euler state network output feature space, that is, the expectation of the distribution, is estimated by the elements of the set. Calculate, where p represents the probability, and takes the value p=1-β, where β is the confidence level;

[0044] pass The eccentricity of the characteristic vector output by the Euler state network is calculated, and the eccentricity is the degree of temperature anomaly.

[0045] Due to the adoption of the above technical solution, the present invention has the following advantages:

[0046] 1. The present invention constructs a time-varying description of the heat flow outside the orbital space environment, providing a reference for establishing the relative thermal characterization energy parameter of the main reflector of the intersatellite link antenna. Compared with directly analyzing the temperature parameters of the main reflector of the intersatellite link antenna, the relative thermal characterization energy parameter incorporates the heat flow effect caused by space radiation, and can more accurately characterize the abnormal temperature of the main reflector of the intersatellite link antenna.

[0047] 2. The present invention uses an Euler state network that does not require training to extract abnormal features, which greatly reduces computing energy consumption and has low requirements for resources such as memory, thereby improving computing efficiency and saving computing power, making it feasible to deploy high-performance temperature anomaly detection devices on satellites.

[0048] Other advantages, objectives, and features of the present invention will be described in part in the following description and, in part, will be apparent to those skilled in the art upon examination of the following description or may be learned from practice of the present invention. The objectives and other advantages of the present invention may be realized and obtained through the following description and claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] The accompanying drawings of the present invention are described below.

[0050] Figure 1 It is a schematic diagram of the process of the present invention;

[0051] Figure 2 This is a waveform diagram of sensor monitoring data according to an embodiment of the present invention;

[0052] Figure 3 A discrete sampling sequence diagram of sensor monitoring data according to an embodiment of the present invention;

[0053] Figure 4 This is a simulation diagram of the heat flow outside the orbital space environment according to an embodiment of the present invention;

[0054] Figure 5 This is a sensor monitoring data fusion sequence diagram according to an embodiment of the present invention;

[0055] Figure 6 A comparison diagram of sensor performance degradation compensation according to an embodiment of the present invention;

[0056] Figure 7 This is a relative thermal characterization energy parameter distribution diagram of an embodiment of the present invention;

[0057] Figure 8 A temperature characteristic graph extracted from the Euler state network model according to an embodiment of the present invention;

[0058] Figure 9 is an abnormality degree curve diagram of an embodiment of the present invention;

[0059] Figure 10 Schematic diagram of the device of the present invention. DETAILED DESCRIPTION

[0060] The present invention will be further described below with reference to the accompanying drawings and examples.

[0061] A method and device for detecting abnormal temperature of the main reflector of an intersatellite link antenna, such as Figure 1 and Figure 10 As shown, the specific steps are:

[0062] Step 11, constructing a time-varying description of the heat flow outside the orbital space environment;

[0063] Step 12, obtaining a binary discrete sequence of measured temperatures of the main reflector of the intersatellite link antenna;

[0064] Step 13, performing equal time fusion on the measured temperature binary discrete sequence to obtain a first processing result;

[0065] Step 14: performing performance degradation compensation on the first processing result to obtain a second processing result;

[0066] Step 15, based on the time-varying description of the heat flow outside the orbital space environment and combined with the second processing result, calculate the relative thermal characterization energy parameter, i.e., the third processing result;

[0067] Step 16: construct an Euler state network model and tune hyperparameters;

[0068] Step 17: extracting long-term temperature features based on the Euler state network model to obtain a fourth processing result;

[0069] Step 18: Calculate the degree of temperature anomaly based on the fourth processing result;

[0070] In this embodiment, a time-varying description of the external heat flow in the orbital space environment is constructed; a binary discrete sequence of measured temperatures of the main reflector of an intersatellite link antenna is obtained; the binary discrete sequence of measured temperatures is subjected to equal-time fusion to obtain a first processing result; performance degradation compensation is performed on the first processing result to obtain a second processing result; a relative thermal characterization energy parameter (i.e., a third processing result) is calculated based on the time-varying description of the external heat flow in the orbital space environment and the second processing result; an Euler state network model is constructed and hyperparameters are adjusted; long-term time series features are extracted based on the Euler state network model to obtain a fourth processing result; and the degree of temperature anomaly is calculated based on the fourth processing result. Constructing the time-varying description of the external heat flow in the orbital space environment provides a reference for establishing the relative thermal characterization energy parameter of the main reflector of the intersatellite link antenna. Compared to directly analyzing the temperature parameters of the main reflector of the intersatellite link antenna, the relative thermal characterization energy parameter incorporates the heat flow effects caused by space radiation and can more accurately characterize temperature anomalies of the main reflector of the intersatellite link antenna. In addition, an Euler state network that does not require training is used to characterize the temperature state of the main reflector of the intersatellite link antenna, which greatly reduces computing energy consumption and requires less resources such as memory, thereby improving computing efficiency and saving computing power.

[0071] Step 111, establishing a solar radiation external heat flow characterization model;

[0072] Step 112, establishing a model for characterizing the heat flow reflected by the Earth's sunlight;

[0073] Step 113: Establishing a model for characterizing the earth's infrared radiation heat flow.

[0074] Step 111 includes, including:

[0075] Where q1 represents the external heat flux from solar radiation, It represents the average solar absorption ratio of the satellite in one orbital period, usually taken as 0.3, S represents sunlight, φ1 represents the solar radiation angular coefficient, A s-p It represents the projection area of ​​the satellite's outer surface perpendicular to the direction of sunlight. where τ s represents the sunshine time, τc Indicates the operating time of one orbital period of the satellite.

[0076] Step 112 includes, among others:

[0077] Where q2 represents the heat flux reflected by the Earth's sunlight, ρ represents the average albedo, which is usually taken as 0.3, and A E-S Indicates the solar radiation angle coefficient, A s-p It represents the surface area of ​​the earth directly exposed to the sun, and φ2 represents the angular coefficient of sunlight reflection from the earth.

[0078] Step 113 includes, including:

[0079] Where q3 represents the earth's infrared radiation heat flow, represents the average infrared emissivity of the satellite surface, A s represents the satellite's surface area, Represents the Earth's infrared radiation angle coefficient.

[0080] In this embodiment, constructing a time-varying description of the heat flow outside the orbital space environment provides a reference basis for establishing a temperature characterization of the main reflector of the intersatellite link antenna. Compared with directly analyzing the temperature parameters of the main reflector of the intersatellite link antenna, incorporating the heat flow influence caused by space radiation can more accurately characterize the abnormal temperature of the main reflector of the intersatellite link antenna.

[0081] Step 12 includes: obtaining the measured temperature low-end resistance data and the measured temperature high-end resistance data; uniformly sampling the measured temperature data with a sampling period of T to obtain the measured temperature binary discrete sequence.

[0082] In this embodiment, the temperature data measured from the main reflector of the intersatellite link antenna of a certain satellite includes three fields, represented as {Date, TA10, TA11}, where Date represents the sampling time, that is, the timestamp, and the sampling interval is 1 minute. TA10 is the high-end temperature measurement data. When the temperature value is lower than the lower limit T2, that is, -76.126, it exceeds the high-end temperature resistance measurement range. TA11 is the low-end temperature measurement data. When the temperature is higher than the upper limit T1, that is, 1.0862, it exceeds the low-end temperature resistance measurement range.

[0083] Step 13 may include: performing equal time fusion on the binary discrete sequence of measured temperatures.

[0084] In this embodiment, the data in the TA10 field and the data in the TA11 field are strongly correlated. Therefore, the two fields are considered to be fused to facilitate subsequent modeling and analysis. However, considering that the temperature measuring resistor may fail during certain temperature-out-of-bounds periods, for example, when the temperature value is lower than the lower bound T2, i.e., -76.126, TA10 is invalid data. Similarly, when the temperature is higher than the upper bound T1, i.e., 1.0862, TA11 is invalid data. In view of this, the numerical fusion of the two fields cannot be achieved by simple linear superposition. Therefore, this embodiment provides an optional conditional weighted fusion strategy, which is achieved by:

[0085] where s t Characterizes the temperature of the main reflector of the intersatellite link antenna at time t, x 1t ,x 2t Here, w1 and w2 represent the low- and high-end temperatures of the intersatellite link antenna main reflector measured at time t, respectively. The conditional fusion approach states that when the low-end temperature measurement reaches the upper bound T1, w1 is set to 0. Similarly, when the high-end temperature measurement reaches the lower bound T2, w2 is set to 0. This conditional fusion method ensures that the resulting numerical sequence more accurately represents the temperature of the intersatellite link antenna main reflector.

[0086] Step 14 may include: t '=s t -δ(d) performs performance degradation compensation on the first processing result, where δ(d) represents the performance degradation model of the temperature measuring resistor, and s t 'The time series set {s0',s1',s2',...,s n '} is the second processing result.

[0087] Empirical model of performance degradation of this type of temperature measuring resistor Where d represents the service life of the temperature measuring resistor, c1 represents the degradation coefficient, and c2 represents the time coefficient. It represents the initial phase of the resistor performance, which is generally set to 1 and cannot be less than 1. ε(d) represents the linear degradation term of the performance, which is usually in the form of a linear function. The slope is related to the actual resistor model.

[0088] The performance degradation compensation is performed on the first processing result, thereby avoiding the incorrect evaluation of the temperature condition of the main reflector of the intersatellite link antenna due to the influence of the performance degradation of the temperature measuring resistor itself, and improving the accuracy of anomaly detection.

[0089] Step 15 includes: calculating a relative thermal characterization energy parameter, i.e., a third processing result, based on a time-varying description of the heat flow outside the orbital space environment and in combination with the second processing result, by:

[0090] where Qr is the relative thermal energy parameter, which is a discrete sequence about time, κ is the total modulation coefficient, κ1, κ2, κ3 are the solar radiation heat flux modulation coefficient, the earth's sunlight reflection heat flux modulation coefficient and the earth's infrared radiation heat flux modulation coefficient respectively.

[0091] Step 16 includes:

[0092] Step 161: Establish the Euler state equation h(t)=F(h(t-1),x(t)), where h(t) represents the output state of the storage pool at time t, h(t-1) represents the output state of the storage pool at time t-1, and x(t) represents the input at time t.

[0093] Step 162, initializing the Euler state network;

[0094] Step 163: Euler state network model hyperparameter tuning.

[0095] Step 161 includes: F(h(t-1),x(t))=h(t)+ε·sigmoid((W h -γI)h(t-1)+W x x(t)+b), where W h , W x Represent the reservoir matrix, ε represents the step size, γ is the diffusion coefficient used to stabilize forward propagation, and b represents the bias. It is worth noting that in the above Euler state equation, the sigmoid nonlinear transformation function is optional, such as the tanh nonlinear transformation function.

[0096] Step 162 includes:

[0097] Step 1621, from the interval [-w r ,w r ] randomly initialize a matrix W∈R from a uniform distribution on N×N The value of w r Is a positive recursive scaling factor, and then set the reservoir matrix W h =WW T ;

[0098] Step 1622, from [-w x ,w x The input weight matrix W is randomly initialized from the uniform distribution on x The value of w x is a positive input scaling factor;

[0099] Step 1623, from [-w b ,w b The value of the bias vector b is randomly initialized from a uniform distribution on ], where w bis a positive deviation scale coefficient.

[0100] Step 163 includes: adjusting w r ,w x ,w b The value of is chosen to balance the contributions of different terms in the recursive transformation of the Euler state network. In this embodiment, empirical methods are used for parameter tuning. In addition, classical hyperparameter optimization methods such as particle swarm optimization, differential evolution algorithm, genetic algorithm, and Bayesian optimization can also be used for parameter tuning to achieve better model performance.

[0101] In this embodiment, the Euler state equation is constructed, its state parameter matrix is ​​randomly initialized, and the hyperparameters are tuned. It is worth noting that the stability of the storage pool is due to W h The antisymmetric structure of the network is achieved without scaling its spectral radius to ensure stability, as is required in traditional state echo networks (ESNs). Furthermore, after initializing the state parameter matrix, this method eliminates the need for training, as is required for RNNs, LSTMs, and GRUs, saving significant computational power.

[0102] Step 17 includes: traversing each element of the third processing result of the time period to be detected in turn, and inputting it into the Euler state network for recursive calculation to obtain the output state set {h1,h2,h3,...,h n}, and the last output state h n As the output feature, that is, the fourth processing result.

[0103] Step 18 may include:

[0104] Step 181, calculating the center of the Euler state network output feature space;

[0105] Step 182: Calculate the eccentricity of the current Euler state network output feature vector, that is, the degree of temperature anomaly.

[0106] Step 181 includes:

[0107] Step 1811, collect the temperature normal Euler state network output state set {h1',h2',h3',...,h n '}, the sample set obeys the X distribution, which represents the characteristic distribution of the normal temperature state;

[0108] Step 1812, estimate the center of the Euler state network output feature space from the set elements, that is, the expectation of the distribution, by Calculate, where p represents the probability, and the value is p = 1-β, where β is the confidence level and is usually set to 0.05.

[0109] Step 182 includes:

[0110] pass Calculate the eccentricity of the output feature vector of the Euler state network. Note that the variables in this formula are all N-dimensional vectors, where N is the number of neurons in the reservoir. To express the eccentricity metric as a single value, modulo the eccentricity vector can be used. Alternatively, based on business needs, threshold selection can be used to convert the quantitative anomaly metric (eccentricity) into a qualitative indicator (normal or abnormal).

[0111] Example:

[0112] Sensor monitoring data waveform diagram Figure 2 As shown in the figure, the simulation data comes from the temperature sensor telemetry data of the main reflector of the intersatellite link antenna of a certain satellite in 2012; the data processing simulation is based on Python 3.8, specifically using the scientific computing library NumPy for data processing, and Matplotlib for visualization. Figure 2 As shown, the monitoring data consists of a binary sequence and changes on a daily basis, but the extreme points fluctuate and the distribution of anomalies has a large span, manifesting as sudden anomalies and slow-changing anomalies.

[0113] The discrete sampling sequence of sensor monitoring data is as follows: Figure 3 As shown, the sampling interval is 1 minute; visualization is achieved using the scatter function in the matplotlib library; as Figure 3 As shown in the figure, the discretized monitoring data series has obvious stratification in spatial distribution, and the abnormal and significant features are concentrated in the middle layer.

[0114] The simulation diagram of heat flow outside the orbital space environment is as follows Figure 4 As shown, the sampling interval is 1 minute; visualization is achieved using the heatmap function in the seaborn library; as Figure 4 As shown in the figure, the changes in the external heat flux of the orbital space environment are also periodic in time. In fact, there is a strong causal relationship between the changes in the sensor monitoring data and the changes in the external heat flux of the orbital space environment, that is, the changes in the sensor monitoring data are largely determined by the changes in the external heat flux of the orbital space environment.

[0115] Sensor monitoring data fusion sequence diagram is as follows Figure 5 As shown, the sampling interval is 1 minute; visualization is achieved using the scatter function in the matplotlib library; Figure 5 and Figure 3 In comparison, the stratification phenomenon of the spatial distribution of the monitoring data series is clearer, and more outliers are separated.

[0116] Sensor performance degradation compensation comparison chart as shown below Figure 6As shown, the sampling interval is 1 minute; visualization is achieved using the plot function in the matplotlib library; the research data window is extended to years; from the perspective of ultra-long time scales, there is an obvious trend change in the boundaries of the monitoring data series, which is mainly caused by the degradation of sensor performance. In addition, the complex space environment interference is also a cause.

[0117] The relative thermal characterization energy parameter distribution diagram is as follows Figure 7 As shown, the sampling interval is 1 minute; visualization is achieved using the scatter function in the matplotlib library; as Figure 7 As shown in the figure, among the many parameter distributions, the relative thermal characterization energy parameter distribution is more stable and has a smaller variation span. It integrates the variation laws of multiple parameters and is more stable in representing the temperature characteristics of the main reflector of the intersatellite link antenna.

[0118] The temperature characteristic diagram extracted by the Euler state network model is as follows Figure 8 As shown, the sampling interval is 1 minute; visualization is achieved using the heatmap function in the seaborn library; as Figure 8 As shown in the figure, the temperature feature distribution has good texture characteristics, which shows that the Euler state network can extract generalized abnormal features from a large amount of data.

[0119] The abnormality degree curve is as follows Figure 9 As shown, the sampling interval is 1 minute; visualization is achieved using the plot function in the matplotlib library; Figure 9 As shown in the figure, the temperature anomaly measurement method of the main reflector of the intersatellite link antenna based on the characteristic space eccentricity can well characterize the degree of temperature anomaly, which verifies the effectiveness of the present invention.

[0120] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0121] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the steps in the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0122] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0123] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0124] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for detecting abnormal temperature of a main reflector of an intersatellite link antenna, characterized in that: The specific steps are as follows: 1) Obtain temperature data of the antenna main reflector through the sensor; 2) Calculate the relative thermal energy parameters under the current temperature data; 3) Analyze the relative thermal energy parameters through the Euler state network model to determine the degree of abnormality of the current temperature; 4) Monitor sensor data and go to step 1 if there is new temperature data); The specific steps for calculating the relative thermal energy parameter under the current temperature data in step 2) are as follows: 2-1) Construct a time-varying description of the external heat flow in the orbital space environment; 2-2) The acquired temperature of the antenna main reflector is formed into a temperature binary discrete sequence, and the temperature binary discrete sequence is subjected to equal time fusion to obtain a first processing result; 2-3) Compensating for performance degradation on the first processing result to obtain a second result; 2-4) Based on the time-varying description of the heat flow outside the orbital space environment and combined with the second processing results, the relative thermal characterization energy parameter, i.e., the third processing result, is calculated; The specific steps for analyzing the relative thermal energy parameters using the Euler state network model in step 3) and determining the current temperature anomaly degree are as follows: 3-1) Build the Euler state network model, initialize the Euler state network and tune the hyperparameters; 3-2) Extracting long-term time series features based on the Euler state network model according to the third processing result to obtain a fourth processing result; 3-3) Calculate the degree of temperature anomaly based on the processing results; Constructing a time-varying description of the external heat flow of the orbital space environment as described in step 2-1) includes establishing a solar radiation external heat flow characterization model, establishing a solar reflection heat flow characterization model for the Earth, and establishing a solar infrared radiation heat flow characterization model for the Earth; The solar radiation external heat flow characterization model is: ,in represents the external heat flux from solar radiation, It represents the average solar absorption ratio of the satellite in one orbital period, usually taken as 0.

3. It represents the solar radiation flux per unit area reaching the top of the Earth's atmosphere. represents the solar radiation angular coefficient, It represents the projection area of ​​the satellite's outer surface perpendicular to the direction of sunlight. ,in, Indicates time, Indicates the duration of sunlight exposure. Indicates the operating time of one orbital cycle of the satellite; The model for characterizing the heat flow reflected by the Earth's sunlight is established as follows: ,in Indicates the heat flux reflected by the Earth's sunlight. Indicates the average albedo, usually taken as 0.3, It represents the cross-sectional area of ​​the earth's solar radiation, reflecting the earth's effective receiving area in the direction of solar radiation. represents the Earth's sunlight reflection angular coefficient; The model for characterizing the earth's infrared radiation heat flow is established as follows: ,in represents the Earth's infrared radiation heat flow, represents the average infrared emissivity of the satellite surface, represents the satellite's surface area, represents the infrared radiation angular coefficient of the earth; The specific method of forming the acquired antenna main reflector temperature into a temperature binary discrete sequence and performing equal time fusion on the temperature binary discrete sequence to obtain the first processing result is as follows: Collecting the low-end temperature resistance data and the high-end temperature resistance data of the antenna main reflector, uniformly sampling the measured temperature data with a sampling period of T, and obtaining the measured temperature binary discrete sequence; pass The binary discrete series of the antenna main reflector temperature is fused in equal time, where Characterizes the temperature of the main reflector of the intersatellite link antenna at time t, The low-end temperature and high-end temperature of the main reflector of the intersatellite link antenna are measured at time t, respectively. Represent the fusion weight coefficients respectively, The time series set This is the first processing result; in, , when the temperature low-end resistance data reaches the upper limit T1, The value is 0. When the high-end temperature resistance data reaches the lower limit T2, The value is 0; The specific method of performing performance degradation compensation on the first processing result in step 2-3) to obtain the second result is as follows: pass The first processing result is subjected to performance degradation compensation, wherein The performance degradation model of the temperature measuring resistor is represented by The time series set This is the second processing result; Wherein, the temperature measuring resistor performance degradation model is: ,in Indicates the service life of the temperature measuring resistor. represents the degradation coefficient, represents the time coefficient, Indicates the initial phase of resistance performance, its value is greater than or equal to 1, represents the performance linear degradation term; The specific formula for calculating the relative thermal energy parameter, i.e., the third processing result, in step 2-4) is as follows: ,in is the relative thermal energy parameter, which is a discrete sequence about time. is the total modulation coefficient, They are respectively the external heat flux modulation coefficient of solar radiation, the heat flux modulation coefficient of the earth's sunlight reflection and the heat flux modulation coefficient of the earth's infrared radiation.

2. The method for detecting abnormal temperature of a main reflector of an intersatellite link antenna according to claim 1, wherein: The specific methods for constructing the Euler state network model, initializing the Euler state network, and tuning the hyperparameters described in step 3-1) are as follows: From the interval Initialize a matrix randomly from a uniform distribution on The value of Is a positive recursive scaling factor, and then set the reservoir matrix ; from The input weight matrix is ​​randomly initialized from a uniform distribution on The value of is a positive input scaling factor; from The bias vector is randomly initialized from a uniform distribution on The value of is a positive deviation scale coefficient.

3. The method for detecting temperature anomaly of an intersatellite link antenna main reflector according to claim 2, wherein: The specific method of extracting long-term time series features based on the Euler state network model according to the third processing result in step 3-2) to obtain the fourth processing result is as follows: Traverse each element of the third processing result set in turn, and input it into the Euler state network for recursive calculation to obtain the output state set , and the last output state As the output feature, that is, the fourth processing result.

4. The method for detecting temperature anomaly of an intersatellite link antenna main reflector according to claim 3, wherein: The specific method for calculating the degree of temperature anomaly based on the processing results in step 3-3) is as follows: Collect the output state set of the Euler state network with normal temperature , the sample set obeys the X distribution, which represents the characteristic distribution of the normal temperature state; The center of the Euler state network output feature space, that is, the expectation of the distribution, is estimated by the elements of the set. Calculate, where Represents probability, value ,in is the confidence level; pass The eccentricity of the characteristic vector output by the Euler state network is calculated, where the eccentricity is the degree of temperature anomaly.

Images