Spacecraft fault diagnosis method based on probability invariant space

Abstract

A spacecraft fault diagnosis method based on a probability invariant space comprises the following steps: firstly, establishing a discrete state space model of a spacecraft system; then, injecting a fault mode to obtain estimated observation data; eliminating the state quantity of unknown probability distribution according to the equivalent space model, obtaining a residual matrix of the estimated observed quantity, and calculating an estimated mean value and a covariance matrix in each fault mode; then, when the spacecraft operates in orbit, obtaining a mean value and a covariance matrix of an in-orbit observed quantity residual matrix according to the equivalent space model; then, according to the estimated mean value, the covariance matrix, the on-orbit mean value and the covariance matrix, the K-L divergence of the probability distribution of the on-orbit operation data and the estimated data of each fault mode is obtained; and finally, according to the K-L divergence, identifying the fault mode with the closest probability density during in-orbit operation of the spacecraft to complete fault diagnosis. When the method is applied to on-orbit fault diagnosis of the spacecraft, the complexity of a fault diagnosis algorithm is reduced only by comparing probability distribution differences of observation data.

Description

Technical Field

[0001] The invention relates to a spacecraft fault diagnosis method based on probability invariant space, belonging to the technical field of spacecraft. Background Art

[0002] Spacecraft system engineering is rapidly advancing, and its structure is becoming increasingly complex. Due to the special operating environment, the system requires extremely high reliability and stability. However, the spacecraft system has the characteristics of long-term on-orbit operation and harsh on-orbit environment, which greatly increases its probability of failure during on-orbit operation. In addition, the manufacturing cost of spacecraft is high. If the fault diagnosis cannot be completed in time and correctly when the fault occurs, it will cause huge losses. Research on fault diagnosis algorithms can effectively improve its risk resistance.

[0003] Traditional fault diagnosis methods rely on manual monitoring on the ground. With the increasing types and frequency of space missions carried out in my country, the number of spacecraft in orbit is increasing. Fault diagnosis methods that rely on the ground have high latency and consume manpower and material resources. Especially in deep space exploration missions, reducing dependence on ground monitoring systems and improving the ability to diagnose faults on spacecraft in orbit have great engineering significance. For example, the Chinese patent CN113867307A, a spacecraft intelligent fault diagnosis method based on a deep neural network, improves the accuracy of spacecraft fault diagnosis based on a deep neural network. However, while improving the accuracy of the fault diagnosis algorithm for spacecraft, it also increases the complexity of the algorithm. It is difficult to operate effectively on orbit when computing resources are scarce. Differences in the structural design of spacecraft systems also limit the applicability of the diagnosis algorithm. Summary of the invention

[0004] The technical problem to be solved by the present invention is: to overcome the shortcomings of the prior art, to provide a spacecraft fault diagnosis method based on probability invariant space, to reduce the complexity of the spacecraft diagnosis method, to complete the fault diagnosis method using on-orbit observation data, to increase the scope of application of the diagnosis method, and to realize on-orbit fault diagnosis of spacecraft.

[0005] The purpose of the present invention is achieved through the following technical solutions:

[0006] A method for diagnosing spacecraft faults based on probability invariant space comprises: establishing a discrete state space model of a spacecraft system; injecting a fault mode to obtain estimated observation data; eliminating the state quantity of unknown probability distribution according to an equivalent space model to obtain a residual matrix of the estimated observation quantity, and calculating the estimated mean and covariance matrix under each fault mode; obtaining the mean and covariance matrix of the residual matrix of the on-orbit observation quantity according to the equivalent space model; obtaining the KL divergence of the probability distribution of on-orbit operation data and estimated data of each fault mode according to the estimated mean and covariance matrix and the on-orbit mean and covariance matrix; and identifying the fault mode with the closest probability density when the spacecraft is running on-orbit according to the KL divergence to complete the fault diagnosis.

[0007] More specific:

[0008] (1) Establish a discrete state space model based on the spacecraft system:

[0009]

[0010] Among them, x[t+1] is the state of the system at time t+1; x[t] is the state of the system at time t; u[t] is the control input of the system at time t; y[t] is the output of the system at time t; f[t] is the fault of the system at time t; v[t] is the process noise of the system at time t; ε[t] is the measurement noise of the system at time t; A is the system parameter matrix; B u is the control parameter matrix of the system; B f is the system control fault parameter matrix; B v is the system control noise parameter matrix; C is the system output parameter matrix; D u Transfer parameter matrix to the system; D f Transfer fault parameter matrix to the system; D ε is the system origin noise parameter matrix.

[0011] (2) Based on the discrete state space model in (1), different fault modes are injected, namely B f , D f and f[t], and obtain the observed data under different fault modes, that is, the control vector u s,i and the output vector y s,i , where i is the failure mode number.

[0012] (3) Select the output, control input, state, fault, process noise and measurement noise vectors of the past n moments: y[t], u[t], x[t], f[t], v[t] and ε[t] to form the window vectors z, x, f and e:

[0013] z=(y[t-n+1] T ,...,y[t] T,u[t-n+1] T ,...,u[t] T ) T

[0014] x=(x[t-n+1] T ,...,x[t] T ,x[t+1] T ) T

[0015] f = (f[t-n+1] T ,...,f[t] T ) T

[0016] e=(v[t-n+1] T ,...,v[t] T ,ε[t-n+1] T ,...,ε[t] T ) T .

[0017] (4) Based on the discrete state space model and the window vectors z, x, f and e, the equivalent space model is obtained as follows:

[0018] Lz=Hx+Ff+Ne

[0019] in

[0020]

[0021] Where L is the parameter matrix related to output and control input; H is the parameter matrix related to state; F is the parameter matrix related to state; N is the parameter matrix related to noise.

[0022] (5) In step (4), by N H H=0, then we get N H , where N H is the left null space N of the parameter matrix H H , and then update the equivalent space model to N H Lz=N H Ff+N H Ne.

[0023] (6) According to the control vector u in step (2) s,i and the output vector y s,i , and get the residual matrix N H Lz s,i , and calculate the estimated mean μ of the residual matrix under each fault mode s,i and the covariance matrix ∑ s,i :

[0024]

[0025] Where w is the number of observation data samples.

[0026] (7) According to the control vector u during on-orbit operation o and the output vector y o Repeat steps (3) to (6) to obtain the mean μ of the residual matrix for each fault mode during on-orbit operation. o and the covariance matrix ∑ o .

[0027] (8) According to step (6) and step (7), the KL divergence of the probability distribution of the on-orbit operation data and the estimated data of each fault mode is obtained by the following formula:

[0028]

[0029] Here, n is the dimension of the covariance matrix.

[0030] (9) According to step (8), the fault diagnosis result is KL divergence D i The fault mode corresponding to the minimum value is the fault diagnosis result.

[0031] An electronic device, comprising:

[0032] Processor; and

[0033] a memory for storing computer program instructions;

[0034] When the computer program instructions are loaded and executed by the processor, the processor executes the above-mentioned spacecraft fault diagnosis method based on probability invariant space.

[0035] A computer-readable storage medium stores computer program instructions, which, when loaded and executed by a processor, cause the processor to execute the above-mentioned spacecraft fault diagnosis method based on probability invariant space.

[0036] A computer program product stored on a non-transitory computer-readable medium, the computer program product comprising program codes for implementing the above-mentioned spacecraft fault diagnosis method based on probability invariant space.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] (1) The method of the present invention can reduce the complexity of on-orbit fault diagnosis of spacecraft and is suitable for on-orbit spacecraft with high fault diagnosis accuracy requirements but very limited computing resources.

[0039] (2) The method of the present invention performs fault diagnosis by comparing the KL divergence of on-orbit data and fault mode estimation data. It only needs to observe the on-orbit operation data of the spacecraft for calculation. It is not limited by the design structure of the spacecraft and has strong adaptability.

[0040] (3) The present invention designs a fault diagnosis algorithm for spacecraft on-orbit based on probability invariant space. When performing fault diagnosis, it only needs to collect observation data of the spacecraft on-orbit. By calculating the KL divergence of the observation data and the estimated data under each fault mode and comparing the probability density function of the fault mode, fault diagnosis is achieved. This algorithm has the advantages of low computational complexity and no need for special design based on differences in spacecraft structure.

[0041] (4) The present invention eliminates the state quantity in the discrete state space model through the equivalent space model, wherein the probability distribution of the state quantity is unknown. All the measurement data retained by the equivalent space model are of known probability distribution, thereby realizing a fault diagnosis method based on probability invariant space.

[0042] (5) The present invention performs fault diagnosis by calculating the KL divergence of the residual matrix, realizes probability-based similarity comparison of multiple groups of random variables, performs fault diagnosis on spacecraft data in the form of probability space, and avoids complex algorithm design. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Flow chart of the method of the present invention. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions and advantages of the present invention more clear, the embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0045] A method for spacecraft fault diagnosis based on probability invariant space, such as Figure 1 As shown, including:

[0046] (1) Based on the spacecraft system model, the actuator and sensor faults of the spacecraft are diagnosed. Taking the satellite system with flywheel actuator, star sensor and gyro sensor as an example, its discrete state space model is:

[0047]

[0048] C=I6,D=0 3×3 ,

[0049]

[0050]

[0051] (2) Based on the state space model in (1), different failure modes are injected, considering the flywheel failure: Flywheel friction torque increase fault: Star sensitivity constant value failure: a1 is a constant; gyro constant fault: a2 is a constant, and the observed data u under different fault modes are obtained. s,i = {u s,1 ,u s,2 ,u s,3 ,u s,4},y s,i ={y s,1 ,y s,2 ,y s,3 ,y s,4}.

[0052] (3) According to the equivalent space model Lz = Hx + Ff + Ne, through N H H=0 solves the unknown probability distribution state quantity x parameter matrix H left null space N H , get the residual matrix N H Lz s,i , through N H Lz s,i Calculate the estimated mean μ of the residual matrix under four fault modes s,i ={μ s,1 ,μ s,2 ,μ s,3 ,μ s,4} and the covariance matrix ∑ s,i ={∑ s,1 ,∑ s,2 ,∑ s,3 ,∑ s,4}, where the number of observed samples w = 50.

[0053] (4) Recording the control vector u during orbit operation o and the output vector y o Repeat step (3) to obtain the on-orbit residual matrix N H Lz o , and the mean μ of the residual matrix under each fault mode during on-orbit operation o and the covariance matrix ∑ o .

[0054] (5) According to step (3) and step (4), the KL divergence of the probability distribution of the on-orbit operation data and the estimated data of each fault mode is obtained by the following formula:

[0055]

[0056] Here, n is the dimension of the covariance matrix.

[0057] (6) According to step (5), the fault diagnosis result is KL divergence D i = the minimum value among [D1, D2, D3, D4], and the fault mode corresponding to the minimum value is the fault diagnosis result.

[0058] In summary, the above embodiments verify the feasibility and effectiveness of the spacecraft fault diagnosis algorithm based on probability invariant space proposed in the present invention.

[0059] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

[0060] The contents not described in detail in the specification of the present invention belong to the common knowledge of those skilled in the art.

[0061] Although the present invention has been disclosed as above in the form of a preferred embodiment, it is not intended to limit the present invention. Any person skilled in the art may make possible changes and modifications to the technical solution of the present invention by using the methods and technical contents disclosed above without departing from the spirit and scope of the present invention. Therefore, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall fall within the protection scope of the technical solution of the present invention.

Claims

1. A method for spacecraft fault diagnosis based on probability invariant space, characterized in that: include: S1. Establish a discrete state space model of the spacecraft system; S2, injecting fault mode to obtain estimated observation data; S3, determining an equivalent space model based on the discrete state space model, and obtaining an updated equivalent space model after eliminating the state quantity of the unknown probability distribution of the position; S4. According to the estimated observation data and the updated equivalent space model, the residual matrix of the estimated observation quantity is obtained, and the estimated value μ of the residual matrix under each fault mode is calculated. s,i and the covariance matrix Σ s,i ; S5. Repeat S3 to S4 according to the control vector and output vector during on-orbit operation to obtain the estimated mean μ of the residual matrix under each fault mode during on-orbit operation. o and the covariance matrix Σ o ; S6, according to μ s,i ,Σ s,i , μ o ,Σ o , obtain the KL divergence of the probability distribution of on-orbit operation data and each fault mode estimation data; S7. Complete fault diagnosis by identifying the fault mode with the closest probability density when the spacecraft is in orbit based on the KL divergence.

2. The method for diagnosing a fault in a spacecraft according to claim 1, characterized in that: The discrete state space model of the spacecraft system is: Among them, x[t+1] is the state of the system at time t+1; x[t] is the state of the system at time t; u[t] is the control input of the system at time t; y[t] is the output of the system at time t; f[t] is the fault of the system at time t; v[t] is the process noise of the system at time t; ε[t] is the measurement noise of the system at time t; A is the system parameter matrix; B u is the control parameter matrix of the system; B f is the system control fault parameter matrix; B v is the system control noise parameter matrix; C is the system output parameter matrix; D u Transfer parameter matrix to the system; D f Transfer fault parameter matrix to the system; D ε is the system origin noise parameter matrix.

3. The method for diagnosing a fault in a spacecraft according to claim 1, characterized in that: In S3, the output, control input, state, fault, process noise and measurement noise vectors of the past n moments are selected: y[t], u[t], x[t], f[t], v[t] and ε[t] to form window vectors z, x, f and e: z=(y[t-n+1] T ,...,y[t] T ,u[t-n+1] T ,...,u[t] T ) T x=(x[t-n+1] T ,...,x[t] T ,x[t+1] T ) T f=(f[t-n+1] T ,...,f[t] T ) T e=(v[t-n+1] T ,...,v[t] T ,ε[t-n+1] T ,...,ε[t] T ) T Based on the discrete state space model and the window vectors z, x, f and e, the equivalent space model is obtained according to the following formula: Lz=Hx+Ff+Ne in Where L is the parameter matrix related to output and control input; H is the parameter matrix related to state; F is the parameter matrix related to state; N is the parameter matrix related to noise.

4. The method for diagnosing a fault in a spacecraft according to claim 3, characterized in that: By N H H=0, then we get N H , where N H is the left null space N of the parameter matrix H H , and then update the equivalent space model to N H Lz=N H Ff+N H Ne.

5. The method for diagnosing a fault in a spacecraft according to claim 1, characterized in that: The failure mode is estimated by observation data including, that is, the control vector u s,i and the output vector y s,i , where i is the fault mode number; in S4, the estimated mean μ of the residual matrix under each fault mode s,i and the covariance matrix ∑ s,i for: Where w is the number of observation data samples.

6. The method for diagnosing a fault in a spacecraft according to claim 1, characterized in that: In S6, the KL divergence of the data probability distribution is: Here, n is the dimension of the covariance matrix.

7. The method for diagnosing a fault in a spacecraft according to claim 1, characterized in that: In S7, the fault diagnosis result is KL divergence D i The fault mode corresponding to the minimum value is the fault diagnosis result.

8. An electronic device comprising: processor; as well as a memory for storing computer program instructions; When the computer program instructions are loaded and executed by the processor, the processor performs the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having computer program instructions stored thereon, wherein when the computer program instructions are loaded and executed by a processor, the processor is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product stored on a non-transitory computer readable medium, the computer program product comprising a program code for carrying out the method according to any one of claims 1 to 7.

Images