CN116149186B - Kalman filtering method for fault estimation of satellite attitude control system - Google Patents
Kalman filtering method for fault estimation of satellite attitude control system Download PDFInfo
- Publication number
- CN116149186B CN116149186B CN202310122951.5A CN202310122951A CN116149186B CN 116149186 B CN116149186 B CN 116149186B CN 202310122951 A CN202310122951 A CN 202310122951A CN 116149186 B CN116149186 B CN 116149186B
- Authority
- CN
- China
- Prior art keywords
- estimation
- matrix
- lag
- attitude control
- control system
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 59
- 238000001914 filtration Methods 0.000 title claims abstract description 45
- 239000011159 matrix material Substances 0.000 claims abstract description 82
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 57
- 241000283153 Cetacea Species 0.000 claims abstract description 35
- 238000005457 optimization Methods 0.000 claims abstract description 27
- 239000002131 composite material Substances 0.000 claims abstract description 13
- 230000003044 adaptive effect Effects 0.000 claims description 16
- 238000005259 measurement Methods 0.000 claims description 15
- 239000002245 particle Substances 0.000 claims description 12
- 238000013461 design Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 9
- 238000002922 simulated annealing Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000002776 aggregation Effects 0.000 claims description 3
- 238000004220 aggregation Methods 0.000 claims description 3
- 230000007423 decrease Effects 0.000 claims description 3
- 238000003745 diagnosis Methods 0.000 abstract description 6
- 238000000137 annealing Methods 0.000 description 3
- 230000007547 defect Effects 0.000 description 3
- 230000005856 abnormality Effects 0.000 description 2
- 238000001514 detection method Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 241000282461 Canis lupus Species 0.000 description 1
- 241000283708 Capra aegagrus Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 150000001875 compounds Chemical class 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 238000002955 isolation Methods 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 230000007246 mechanism Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 238000012360 testing method Methods 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05B—CONTROL OR REGULATING SYSTEMS IN GENERAL; FUNCTIONAL ELEMENTS OF SUCH SYSTEMS; MONITORING OR TESTING ARRANGEMENTS FOR SUCH SYSTEMS OR ELEMENTS
- G05B13/00—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion
- G05B13/02—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric
- G05B13/04—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators
- G05B13/042—Adaptive control systems, i.e. systems automatically adjusting themselves to have a performance which is optimum according to some preassigned criterion electric involving the use of models or simulators in which a parameter or coefficient is automatically adjusted to optimise the performance
Abstract
The invention discloses a Kalman filtering method for fault estimation of a satellite attitude control system, which belongs to the technical field of spacecraft fault diagnosis and specifically comprises the following steps: considering modeling interference and state time lag, taking actuator and sensor faults as additional variables, and establishing a time lag generalized system model; designing a non-singular robust Kalman filter; expanding nonlinear terms and nonlinear time-lag terms of a time-lag generalized system based on a Taylor series, and providing a robust upper bound; optimizing a noise covariance matrix through an improved whale optimization algorithm, and replacing the original noise covariance with the optimized noise covariance matrix; and executing state estimation and modeling interference estimation of the time-lag generalized system to complete state estimation and composite fault estimation of the satellite attitude control system. By adopting the Kalman filtering method for the satellite attitude control system fault estimation, the problems of various fault estimation of the satellite attitude control system in a complex operating environment are solved, and the precision of the fault estimation is improved.
Description
Technical Field
The invention relates to the technical field of spacecraft fault diagnosis, in particular to a Kalman filtering method for satellite attitude control system fault estimation.
Background
With the rapid development of satellite in-orbit services, the number of in-orbit satellites is increasing. Accordingly, the demands for satellite stability and security are also increasing. Attitude Control Systems (ACSs) act as important sub-systems for satellites, whose normal operation is related to the efficient deployment of important tasks such as on-board payload, remote control, telemetry and orbit control. However, ACSs are also relatively prone to failure due to the numerous system devices and complex operating mechanisms. Therefore, the fault of ACSs needs to be diagnosed, and the abnormality of the system is found in time, so that the occurrence of safety accidents is avoided, and the on-orbit normal operation of satellites is ensured. Therefore, by combining with the modern fault diagnosis technology, the fault of ACSs is effectively diagnosed, and the method has important practical significance.
In the model-based fault diagnosis method, designing a residual generator is an effective method. Currently, residual generators such as observers, filters, etc. have been widely used for fault diagnosis of acs. To achieve the first two task phases of ACSs, namely: fault Detection and Isolation (FDI), unknown Input Observer (UIO) has already beenBond H _ /H ∞ The detection method is widely applied. And for the last task phase of fault diagnosis: the fault estimation is particularly important because the process not only can judge the specific form and amplitude of the fault signal, but also can provide a basis for active fault-tolerant control. In view of this, adaptive observers, sliding mode observers, interval observers, etc. have combined with robust H ∞ The method ensures the stability and the robustness of the system. However, when there is operational noise in the system, the influence of noise on the fault estimation accuracy needs to be considered. For this purpose, another residual generator, a Kalman filter, has been used for ACSs fault estimation under noise.
In view of the nonlinear characteristics of ACSs systems, extended Kalman filters based on the parity vector method, fused unscented Kalman filters, strong tracking volume Kalman filters, and other improved nonlinear filters have been used to estimate the failure of ACSs. In addition, the unknown characteristics of ACSs noise are considered, and methods such as adaptive filtering, variational Bayesian filtering and the like are combined with the nonlinear Kalman filter to better estimate system noise, so that better fault estimation is realized. However, the nonlinear Kalman filtering method of ACSs still has certain drawbacks. First, the accuracy of concurrent actuator and sensor concurrent fault (composite fault) estimation of the system still needs to be further improved. Typically, satellite actuator failures are manifested as flywheel stuck, lost motion, and ramp drift, etc.; the sensor failure is mainly represented by: the gyro sensor is locked, constant deviation, slope change and the like. In extreme cases, if compound faults occur simultaneously, the coupling of fault signals may result in poor estimation accuracy. While sensor failure may be considered a false actuator failure, deviations from the equivalent method may affect the accuracy of the estimation. Second, external disturbances of the system have a certain impact on the accuracy of the fault estimation. Although methods such as disturbance observers have been used to estimate the external disturbances of ACSs, random noise and unknown external disturbances are often present at the same time, which presents certain challenges to the design of Kalman filters. Again, the speed lag of ACSs can also affect the filtering accuracy. At present, the Kalman filter design for ACSs with time-lag links mainly focuses on ACSs with measurement time-lag, but the essence of the methods is high-dimensional transformation of a system, so that the defects of high variable dimension, larger algorithm load and the like are caused. Therefore, it is necessary to design an effective fault estimation Kalman filter to improve the composite fault estimation accuracy of acs, taking the comprehensive influence of the above factors into consideration.
In addition, for ACSs, because the ACSs operate under complex working conditions for a long time, a plurality of uncertain factors can cause the abnormality of a measuring element, so that a certain deviation exists between a measured value and a true value of noise, and the filtering precision is further reduced. Currently, population intelligent algorithms such as locust algorithm, wild goat algorithm, gray wolf algorithm and Whale Optimization Algorithm (WOA) are used for optimizing system noise, so that the influence of measurement deviation of the system noise on filtering precision is reduced. However, the existing researches and applications show that the method has the defects of low convergence speed, low optimizing precision, easiness in sinking into local optimal solution and the like to different degrees. Therefore, the method for optimizing the intelligent group by combining the effective method improves the existing intelligent group optimizing method, and has important significance for enhancing the algorithm performance, improving the noise optimizing precision and further improving the fault estimating precision of ACSs.
Considering the speed time lag of ACSs, modeling disturbance and system measurement deviation, and obtaining a time lag generalized system according to the faults of an actuator and a sensor as additional state variables of the system. Thus, the fault estimation of the ACSs translates into a state estimation of the time-lapse generalized system. In recent years, kalman filtering estimation of a generalized system becomes a hot spot for domestic and foreign research. Because of the singular matrix in the dynamic term of the generalized system, the Kalman filter with the existing singular structure is difficult to solve, singular values may appear in the covariance matrix, and the method is not easy to apply. Meanwhile, the Kalman filtering estimation method for the state time lag generalized system is still limited at present. In addition, the external interference and system measurement bias described above present additional challenges to the Kalman filter design of nonlinear generalized systems. Therefore, the invention builds a time lag generalized system according to the additional variables of ACSs as the composite faults of the executor and the sensor; a robust Kalman filter is designed based on a disturbance observer, and robust parameters are designed to solve the problem of linearization errors of a nonlinear term and a nonlinear time-lag term of a system; improved annealing simulation algorithm (ISA) and adaptive parameter improved WOA based on elite set strategy to better optimize system noise and reduce the influence of noise measurement bias on filtering accuracy. Combining robust H method, kalman
∞
The filter and disturbance estimation idea are that an improved group intelligent algorithm is utilized to design an optimized robust Kalman filter, so that ACSs composite fault estimation accuracy under complex working conditions is improved, and the method has important practical significance.
Disclosure of Invention
The invention aims to provide a Kalman filtering method for fault estimation of a satellite attitude control system, which realizes high-precision fault estimation of a spacecraft.
In order to achieve the above purpose, the invention provides a Kalman filtering method for satellite attitude control system fault estimation, which comprises the following specific steps:
step S1: taking modeling interference and state time lag into consideration, taking actuator faults and sensor faults as additional variables of a satellite attitude control system, and establishing a time lag generalized system model;
step S2: aiming at the established time lag generalized system, a non-singular robust Kalman filter is designed based on a robust filter and a disturbance observer;
the non-singular robust Kalman filter includes an interference estimation term and a nonlinear state estimation term,
an interference estimation term based on the disturbance observer is used for estimating modelable interference;
the nonlinear state estimation term is used for estimating state variables of the time lag generalized system;
step S3: based on a nonlinear term and a nonlinear time-lag term of a Taylor series expansion time-lag generalized system, a robust upper bound of a Taylor series expansion truncation error is given;
step S4: an improved simulated annealing algorithm is obtained based on an elite set strategy, a whale optimization algorithm is improved through the improved simulated annealing algorithm and the adaptive parameters, an improved whale optimization algorithm is obtained, a noise covariance matrix of a system is optimized through the improved whale optimization algorithm, and the optimized noise covariance matrix is used for replacing the original noise covariance of the system;
step S5: and executing state estimation and modeling interference estimation of the time-lag generalized system to complete state estimation and various fault estimation of the satellite attitude control system.
Preferably, in step S1,
the modelable disturbance has harmonic characteristics and is equivalent to an exogenous disturbance;
the time lag is constant time lag, and the instant time lag constant is a positive integer;
the fault distribution matrix and the modelable disturbance distribution matrix of the actuator and the sensor are both a column full order matrix.
Preferably, the model of the satellite attitude control system is:
wherein ,xk 、x k-l 、g(x k )、g d (x k-l )、u k 、y k The method comprises the following steps of: the three-axis angular velocity, angular velocity time lag, nonlinear term, nonlinear time lag term, control input and measurement output of the satellite at the kth moment; l represents a time lag constant and is a positive integer, f k and sk Representing an actuator failure and a sensor failure, respectively; A. a is that d And C is a constant matrix; the constant matrix B is the control input u k Coefficient matrix of (a); f (F) a and Fs Respectively the actuator faults f k And sensor faults s k Is a distribution matrix of (a); w (w) k and vk Respectively, process white noise and measurement white noise, d k To model interference, a constant matrix F d A distribution matrix for the same;
the equivalent formula for modeling interference is as follows:
d k ＝D 1 η k +D 2 ,η k+1 ＝Wη k
wherein ,ηk For exogenous disturbance, D 1 、D 2 And W is a constant matrix;
order they r,k ＝C r f k +v r,k ，v r,k White noise and its covariance matrix is R r,k Let f k+1 ≈f k The time lag generalized system is obtained as follows:
wherein: e=diag (I, 0), I representing a unit array of set dimensions;as a state variable of the time-lapse generalized system,for state time lag variable, ++>
White noise w of system k and vk Covariance of Q k and Rk ， and />Covariance of +.>And
preferably, the non-singular matrices G and H related to the time-lag generalized system and the designed non-singular robust Kalman filter satisfy the following formula:
the general solution of the non-singular matrix G and the matrix H is:
wherein ,y is a degree of freedom matrix>Is pseudo-inverse.
Preferably, the structure of the non-singular robust Kalman filter is as follows:
wherein , and />State variables for estimating time-lag generalized system, respectively>State time lag variable->And exogenous interference eta k ；K k and Vk For gain matrices to be designed；
Order the Define the error of variable J as +.>Obtain->The joint estimation error of η is:
wherein ,
preferably, the nonlinear error term and />Linearization is performed by adopting a Taylor series expansion mode, and after a nonlinear high-order term and a nonlinear time-lag high-order term are expressed by using an equation, the method is characterized in that:
wherein ,N i,k (i=1, 2) is +.>About->With respect toDeriving the obtained matrix; />L 1 and L2 Is a constant matrix phi 1,k and φ2,k For an unknown bounded matrix, satisfy +.>
Based on inequality quotation, getCovariance matrix robust upper bound P of (2) k+1 The method comprises the following steps:
wherein the scalar mu > 0, gamma 1 and γ2 For the robust parameters of the design,
the filter gain matrix satisfies:
and estimate the errorFor amplifying noise-> and />The robustness of (c) is guaranteed by the following inequality:
wherein ,P0 -1 Represents an initial value P of an estimated error covariance upper bound 0 Is the inverse of (2); for matrix or variable X, Z, there are
Preferably, the improved whale optimization algorithm comprises the following specific steps:
step S41a: setting an adaptive parameter p s ；
Step S42a: based on adaptive parameter p s Performing a search process of a whale optimization algorithm:
if rand () > p s Whale performs a local search for prey:
wherein p is a random number between 0 and 1, U i,k+1 Is the current particle position, U i,k * Is the position of the optimal particle, θ is a constant defining the helical shape, ω is between-1 and 1, ζ＝2ar l1 -a，ξ＝2r l2 ，r l1 and rl2 Is a random number between 0 and 1, a is a convergence factor, and decreases linearly from 2 to 0;
if rand (). Ltoreq.p s Whale performs a global search for prey:
wherein , is the location of the random particles in the current population.
Step S43a: judgment by elite set policyWhether to enter a collection, define +.>The elite aggregation policy steps are:
(1) for the fitness function ψ, ifThen->Entering a collection; otherwise, turning to the step (4);
(2) if j < Y, j=j+1,entering a collection; if j=Y-1, then +.>Enter collection, adjust>Exiting the collection;
(3) arranging elements in a collection from large to small, i.e.
(4) Ending the judgment;
step S44a: based on elite set strategy, executing simulated annealing algorithm, and accepting probability by MertopolisJudging whether to accept the new solution:
wherein ,is the position of the random particle in the elite set,/->
If it isThen->Otherwise->
If it isThen->Otherwise U i,k+1 ＝U i,k 。
Preferably, the noise covariance matrix of the system is optimized by a modified whale optimization algorithm and />The obtained optimized values are +.> and />Will-> and />Substituting into a non-singular robust Kalman filter, the specific steps are as follows:
step S41b: initializing, wherein initial values of state estimation and corresponding covariance matrixes are given;
step S42b: searching for optimal noise, optimizing by improved whale optimizing algorithm and />Optimal value +.>And
step S43b: performing filtering estimation to obtain a time-lag generalized system state variableAnd exogenous interference eta k+1 Is>The corresponding covariance robust upper bound and the filter gain matrix calculation mode are as follows:
step S44b: step S41 b-step S43b are repeated for the next set of samples.
Preferably, the joint estimation is obtainedThereafter, the following estimation values are calculated:
calculating state estimation value of time-lag generalized systemAnd estimate of exogenous interference->
From the following componentsCalculating a modelable interference estimate +.>
Calculating state estimation value of satellite attitude control systemAnd calculating the composite fault estimated value of the actuator and the sensor and />
Preferably, the obtained filtering algorithm result comprises a satellite state estimation curve, a composite fault estimation curve based on an actuator and a sensor, an improved whale optimization algorithm fitness function iteration curve and index data for evaluating the filtering algorithm, wherein the index data is root mean square error.
Therefore, the Kalman filtering method for satellite attitude control system fault estimation has the following beneficial effects:
(1) According to the technical scheme, modeling interference and state time lag are considered, faults of an actuator and a sensor are taken as additional variables of a satellite attitude control system, a time lag generalized system model is established, influences of complex factors such as ACSs (active sensor system) actuator faults, sensor faults, modeling interference, state timing time lag and system measurement deviation are fully considered, then a fault estimation Kalman filter is designed, the problem of complex fault estimation of the satellite attitude control system in a complex operation environment is solved, and the application range of the Kalman filter in the time lag generalized system is widened.
(2) According to the technical scheme, a non-singular robust Kalman filter is designed based on a robust filter and a disturbance observer aiming at an established time-lag generalized system; the non-singular robust Kalman filter has a non-singular structure, has the advantages of simple calculation, easy realization and the like, realizes multiple fault estimation and simultaneously realizes the estimation of modeling external interference; through the designed robust upper bound, linearization errors of nonlinear terms and nonlinear time-lag terms caused by Taylor series expansion are reduced, so that filtering precision is improved, and more accurate ACSs state estimation results, modeling interference estimation results, actuator and sensor fault estimation results are obtained.
(3) According to the technical scheme, aiming at noise measurement deviation of a system, an annealing simulation algorithm is improved based on an elite set strategy, and a Whale Optimization Algorithm (WOA) is improved based on the improved annealing simulation algorithm and adaptive parameters. The Improved Whale Optimization Algorithm (IWOA) effectively avoids the defects of low WOA optimization accuracy, low convergence speed, easiness in sinking into a local optimal solution and the like to a certain extent, so that system noise is better optimized, and an optimized filtering algorithm is executed based on a process noise and measurement noise covariance matrix obtained by improving whale algorithm optimization. The algorithm effectively reduces the problem of filtering precision reduction caused by system noise measurement deviation, so that the state value, the modeling interference value and various fault values of the satellite attitude control system are estimated more effectively; compared with the prior optimized filtering algorithm, the method has higher estimation accuracy.
The technical scheme of the invention is further described in detail through the drawings and the embodiments.
Drawings
FIG. 1 is a schematic diagram of a Kalman filtering method for fault estimation of a satellite attitude control system according to the present invention;
FIG. 2a is a graph of the result of estimating the X-axis state of the satellite attitude control system according to the present invention;
FIG. 2b is a diagram of the Y-axis state estimation result of the satellite attitude control system according to the present invention;
FIG. 2c is a graph of the Z-axis state estimation results of the satellite attitude control system of the present invention;
FIG. 3 is a graph of a modeled disturbance estimation result;
FIG. 4 is an actuator fault estimation result;
FIG. 5 is a sensor fault estimation result;
FIG. 6 is a graph comparing the iteration results of the adaptive functions of IWOA-ODORKF and WOA-ORDRKF.
Detailed Description
Examples
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present invention more apparent, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments of the present invention. The components of the embodiments of the present invention generally described and illustrated in the figures herein may be arranged and designed in a wide variety of different configurations.
Thus, the following detailed description of the embodiments of the invention, as presented in the figures, is not intended to limit the scope of the invention, as claimed, but is merely representative of selected embodiments of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
Embodiments of the present invention will be described in detail below with reference to the accompanying drawings.
Referring to fig. 1, a Kalman filtering method for fault estimation of a satellite attitude control system specifically includes the following steps:
step S1: and taking modeling interference and state time lag into consideration, taking the faults of the actuator and the sensor as additional variables of the satellite attitude control system, and establishing a time lag generalized system model. The fault distribution matrix and the modelable disturbance distribution matrix of the actuator and the sensor are both a column full order matrix.
The model of the satellite attitude control system of the present embodiment is:
wherein ,xk 、x k-l 、g(x k )、g d (x k-l )、u k 、y k The method comprises the following steps of: the three-axis angular velocity, angular velocity time lag, nonlinear term, nonlinear time lag term, control input and measurement output of the satellite at the kth moment; l represents a time lag constant and is a positive integer; f (f) k and sk Indicating actuator and sensor failure; A. a is that d And C is a constant matrix; the constant matrix B is the control input u k Coefficient matrix of (a); f (F) a and Fs Respectively the actuator faults f k And sensor faults s k Is a distribution matrix of (a); w (w) k and vk Respectively, process white noise and measurement white noise, d k To model interference, a constant matrix F d For its distribution matrix.
The modelable disturbance has harmonic characteristics and can be equivalently an exogenous disturbance;
the equivalent formula for modeling interference is as follows:
d k ＝D 1 η k +D 2 ,η k+1 ＝Wη k
wherein ,ηk For exogenous disturbance, D 1 、D 2 And W is a constant matrix;
order they r,k ＝C r f k +v r,k ，v r,k White noise and its covariance matrix is R r,k Let f k+1 ≈f k The time lag generalized system is obtained as follows:
wherein: e=diag (I, 0), I representing a unit array of appropriate dimensions,as a state variable of the time-lapse generalized system,is a state time lag variable-> The specific form of the other matrix is:
in addition, system white noise w k and vk Covariance of Q k and Rk ， and />Covariance of +.>Andfrom matrix-> and D1 The full detectability of the time-lapse generalized system and the feasibility of filter design are guaranteed.
The non-singular matrix G and matrix H related to the time-lag generalized system and the designed non-singular robust Kalman filter satisfy the following formula:
the general solution of the non-singular matrix G and the matrix H is:
wherein ,y is a degree of freedom matrix>Is pseudo-inverse.
Step S2: aiming at the established time lag generalized system, a non-singular robust Kalman filter is designed based on the robust filtering and the disturbance observer.
The non-singular robust Kalman filter includes an interference estimation term and a nonlinear state estimation term. The disturbance observer based disturbance estimation terms are used to estimate the modelable disturbance. The nonlinear state estimation term is used to estimate the state variable of the time-lag generalized system.
The structure of the non-singular robust Kalman filter is as follows:
wherein , and />State variables for estimating time-lag generalized system, respectively>State time lag variable->And exogenous interference eta k ；K k and Vk A gain matrix to be designed;
order the Define the error of variable J as +.>Obtain->The joint estimation error of η is:
wherein ,
step S3: nonlinear terms and nonlinear time-lag terms of a Taylor series expansion time-lag generalized system are based, and a robust upper bound of a Taylor series expansion truncation error is given.
Given the robust upper bound of the nonlinear higher order term and the nonlinear time-lag higher order term, the nonlinear error termAndlinearization is performed by adopting a Taylor series expansion mode, after a nonlinear higher-order term and a nonlinear time-lag higher-order term are expressed by using an equation, a first derivative term is reserved as a linear error term, and a truncation error is expressed as follows:the method comprises the following steps:
wherein ,N i,k (i=1, 2) is +.>About->About->Deriving the resulting matrix.
Based on inequality quotation, given positive scalar μ, design robust parameter γ 1 and γ2 ObtainingCovariance matrix robust upper bound P of (2) k+1 The method comprises the following steps:
wherein the scalar mu > 0, gamma 1 and γ2 For the robust parameters of the design,
meanwhile, the designed filter gain matrix satisfies the following conditions:
and estimate the errorFor amplifying noise-> and />The robustness of (c) is guaranteed by the following inequality:
wherein ,P0 -1 Represents an initial value P of an estimated error covariance upper bound 0 For matrix or variable X, Z, there is
Step S4: an improved simulated annealing algorithm is obtained based on an elite set strategy, a whale optimization algorithm is improved through the improved simulated annealing algorithm and the adaptive parameters, an improved whale optimization algorithm is obtained, a noise covariance matrix of a system is optimized through the improved whale optimization algorithm, and the optimized noise covariance matrix is used for replacing the original noise covariance of the system.
The improved whale optimization algorithm comprises the following specific steps:
step S41a: setting an adaptive parameter p s ；
Step S42a: based on adaptive parameter p s Performing a search process of a whale optimization algorithm:
if rand () > p s Whale performs a local search for prey:
wherein p is a random number between 0 and 1, U i,k+1 Is the current particle position, U i,k * Is the position of the optimal particle, θ is a constant defining the helical shape, ω is between-1 and 1, ζ＝2ar l1 -a，ξ＝2r l2 ，r l1 and rl2 Is a random number between 0 and 1, a is a convergence factor, and decreases linearly from 2 to 0.
If rand (). Ltoreq.p s Whale performs a global search for prey:
wherein , is the location of the random particles in the current population.
Step S43a: judgment by elite set policyWhether to enter a collection, define +.>The elite aggregation policy steps are:
(1) for the fitness function ψ, ifThen->Entering a collection; otherwise, turning to the step (4);
(2) if j < Y, j=j+1,entering a collection; if j=Y-1, then +.>Enter collection, adjust>Exiting the collection;
(3) arranging elements in a collection from large to small, i.e.
(4) Ending the judgment;
step S44a: based on elite set strategy, executing simulated annealing algorithm, and accepting probability by MertopolisJudging whether to accept the new solution:
wherein ,is the position of the random particle in the elite set,/->
If it isThen->Otherwise->
If it isThen->Otherwise U i,k+1 ＝U i,k 。
Optimizing noise covariance matrix of system by improved whale optimization algorithm and />The obtained optimized values are +.> and />Will-> and />Substituting the filtered signal into a non-singular robust Kalman filter to obtain an IWA-DORKF filtering flow, wherein the specific steps are as follows:
step S41b: initializing, wherein initial values of state estimation and corresponding covariance matrixes are given;
step S42b: searching for optimal noise, optimizing by improved whale optimizing algorithm and />Optimal value +.>And
step S43b: performing filtering estimation to obtain a time-lag generalized system state variableAnd exogenous interference eta k+1 Is>The corresponding covariance robust upper bound and the filter gain matrix calculation mode are as follows:
step S44b: step S41 b-step S43b are repeated for the next set of samples.
Step S5: state estimation and modelable stem for execution time-lag generalized systemsAnd (5) disturbing estimation, namely completing state estimation and composite fault estimation of the satellite attitude control system. Obtaining the joint estimation valueThereafter, the following estimation values are calculated: />
(1) Calculating state estimation value of time-lag generalized systemAnd estimate of exogenous interference->
(2) From the following componentsCalculating a modelable interference estimate +.>
(3) Calculating state estimation value of satellite attitude control systemAnd calculating an actuator and sensor composite fault estimation value +.> and />
The obtained filtering algorithm results comprise a satellite state estimation curve, a composite fault estimation curve based on an actuator and a sensor, an improved whale optimization algorithm fitness function iteration curve and index data for evaluating the filtering algorithm, wherein the index data is root mean square error.
To verify the performance of the present implementation, three filters are used to estimate the failure of the target acs, etc., and the three obtained filter estimation results are compared. The three filters are respectively: DORKF, WOA-DORKF, and IWOA-DORKF.
The test results are shown in fig. 2a, 2b, 2c, 3, 4 and 5, and the IWOA-DORKF filtering method based on the embodiment has the advantages of faster convergence speed, higher calculation accuracy and better capability of jumping out of the global optimal solution, and the fault estimation result of modeling interference and actuator and sensor composite to the state of the target ACSs is closest to the actual value of the corresponding variable of the target ACSs, so that the fault estimation method provided by the invention has more accurate estimation effect. At the same time, as shown in FIG. 6, the WOA-DORKF and the IWAA-DORKF adaptive function of the invention are iteratively compared. The estimation accuracy is ensured by a better adaptive function iteration mode of the IWAA in the estimation period.
Finally, it should be noted that: the above embodiments are only for illustrating the technical solution of the present invention and not for limiting it, and although the present invention has been described in detail with reference to the preferred embodiments, it will be understood by those skilled in the art that: the technical scheme of the invention can be modified or replaced by the same, and the modified technical scheme cannot deviate from the spirit and scope of the technical scheme of the invention.
Claims (10)
1. A Kalman filtering method for satellite attitude control system fault estimation is characterized by comprising the following specific steps:
step S1: taking modeling interference and state time lag into consideration, taking actuator faults and sensor faults as additional variables of a satellite attitude control system, and establishing a time lag generalized system model;
step S2: aiming at the established time lag generalized system, a non-singular robust Kalman filter is designed based on a robust filter and a disturbance observer;
the non-singular robust Kalman filter includes an interference estimation term and a nonlinear state estimation term,
an interference estimation term based on the disturbance observer is used for estimating modelable interference;
the nonlinear state estimation term is used for estimating state variables of the time lag generalized system;
step S3: based on a nonlinear term and a nonlinear time-lag term of a Taylor series expansion time-lag generalized system, a robust upper bound of a Taylor series expansion truncation error is given;
step S4: an improved simulated annealing algorithm is obtained based on an elite set strategy, a whale optimization algorithm is improved through the improved simulated annealing algorithm and the adaptive parameters, an improved whale optimization algorithm is obtained, a noise covariance matrix of a system is optimized through the improved whale optimization algorithm, and the optimized noise covariance matrix is used for replacing the original noise covariance of the system;
step S5: and executing state estimation and modeling interference estimation of the time-lag generalized system to complete state estimation and various fault estimation of the satellite attitude control system.
2. A Kalman filtering method for fault estimation of satellite attitude control system according to claim 1, wherein: in the step S1 of the process,
the modelable disturbance has harmonic characteristics and is equivalent to an exogenous disturbance;
the time lag is constant time lag, and the instant time lag constant is a positive integer;
the fault distribution matrix and the modelable disturbance distribution matrix of the actuator and the sensor are both a column full order matrix.
3. A Kalman filtering method for fault estimation of satellite attitude control system according to claim 2, wherein: the satellite attitude control system model is as follows:
wherein ,xk 、x k-l 、g(x k )、g d (x k-l )、u k 、y k The method comprises the following steps of: the three-axis angular velocity, angular velocity time lag, nonlinear term, nonlinear time lag term, control input and measurement output of the satellite at the kth moment; l represents a time lag constant and is a positive integer, f k and sk Representing an actuator failure and a sensor failure, respectively; A. a is that d And C is a constant matrix; the constant matrix B is the control input u k Coefficient matrix of (a); f (F) a and Fs Respectively the actuator faults f k And sensor faults s k Is a distribution matrix of (a); w (w) k and vk Respectively, process white noise and measurement white noise, d k To model interference, a constant matrix F d A distribution matrix for the same;
the equivalent formula for modeling interference is as follows:
d k ＝D 1 η k +D 2 ,η k+1 ＝Wη k
wherein ,ηk For exogenous disturbance, D 1 、D 2 And W is a constant matrix;
order they r,k ＝C r f k +v r,k ，v r,k White noise and its covariance matrix is R r,k Let f k+1 ≈f k The time lag generalized system is obtained as follows:
wherein: e=diag (I, 0), I representing a unit array of set dimensions; x is x k As a state variable of the time-lapse generalized system,for state time lag variable, ++>
White noise w of system k and vk Covariance of Q k and Rk ， and />Covariance of +.>And
4. a Kalman filtering method for fault estimation of satellite attitude control system according to claim 3, wherein: the non-singular matrix G and matrix H related to the time-lag generalized system and the designed non-singular robust Kalman filter satisfy the following formula:
the general solution of the non-singular matrix G and the matrix H is:
wherein ,y is a degree of freedom matrix>Is pseudo-inverse.
5. The Kalman filtering method for satellite attitude control system fault estimation according to claim 4, wherein: the structure of the non-singular robust Kalman filter is as follows:
wherein , and />Respectively used for estimating state variables x of time lag generalized system k State time lag variable->And exogenous interference eta k ；K k and Vk A gain matrix to be designed;
order the Define the error of variable J as +.>Obtain->The joint estimation error of η is:
wherein ,
6. the Kalman filtering method for satellite attitude control system fault estimation according to claim 5, wherein: nonlinear error term and />Linearization is performed by adopting a Taylor series expansion mode, and after a nonlinear high-order term and a nonlinear time-lag high-order term are expressed by using an equation, the method is characterized in that:
wherein ,N i,k (i=1, 2) is +.>About->About->Deriving the obtained matrix; />L 1 and L2 Is a constant matrix phi 1,k and φ2,k For an unknown bounded matrix, satisfy +.>
Based on inequality quotation, getCovariance matrix robust upper bound P of (2) k+1 The method comprises the following steps:
wherein the scalar mu > 0, gamma 1 and γ2 For the robust parameters of the design,
the filter gain matrix satisfies:
and estimate the errorFor amplifying noise-> and />The robustness of (c) is guaranteed by the following inequality:
wherein ,P0 -1 Representing the estimated errorCovariance upper bound initial value P 0 Is the inverse of (2); for matrix or variable X, Z, there are
7. The Kalman filtering method for fault estimation of satellite attitude control system according to claim 6, wherein: the improved whale optimization algorithm comprises the following specific steps:
step S41a: setting an adaptive parameter p s ；
Step S42a: based on adaptive parameter p s Performing a search process of a whale optimization algorithm:
if rand () > p s Whale performs a local search for prey:
wherein p is a random number between 0 and 1, U i,k+1 Is the current particle position, U i,k * Is the position of the optimal particle, θ is a constant defining the helical shape, ω is between-1 and 1, ζ＝2ar l1 -a，ξ＝2r l2 ，r l1 and rl2 Is a random number between 0 and 1, a is a convergence factor, and decreases linearly from 2 to 0;
if rand (). Ltoreq.p s Whale performs a global search for prey:
wherein ,is the location of the random particles in the current population.
Step S43a: judgment by elite set policyIf the set is entered, define that the size of the selected set is YThe elite aggregation policy steps are:
(1) for the fitness function ψ, ifThen->Entering a collection; otherwise, turning to the step (4);
(2) if j < Y, j=j+1,entering a collection; if j=Y-1, then +.>Enter collection, adjust>Exiting the collection;
(3) arranging elements in a collection from large to small, i.e.
(4) Ending the judgment;
step S44a: based on elite set strategy, executing simulated annealing algorithm, and accepting probability by MertopolisJudging whether to accept the new solution:
wherein ,is the position of the random particle in the elite set,/->
If it isThen->Otherwise->
If it isThen->Otherwise U i,k+1 ＝U i,k 。
8. The Kalman filtering method for fault estimation of satellite attitude control system according to claim 7, wherein: optimizing noise covariance matrix Q of system by improved whale optimization algorithm k Andthe obtained optimized values are respectively and />Will-> and />Substituting into a non-singular robust Kalman filter, the specific steps are as follows:
step S41b: initializing, wherein initial values of state estimation and corresponding covariance matrixes are given;
step S42b: searching for optimal noise, optimizing by improved whale optimizing algorithm and />Optimal value +.> and />
Step S43b: performing filtering estimation to obtain a time-lag generalized system state variableAnd exogenous interference eta k+1 Is>The corresponding covariance robust upper bound and the filter gain matrix calculation mode are as follows:
step S44b: step S41 b-step S43b are repeated for the next set of samples.
9. The Kalman filtering method for fault estimation of satellite attitude control system according to claim 8, wherein: obtaining the joint estimation valueThereafter, the following estimation values are calculated:
calculating state estimation value of time-lag generalized systemAnd estimate of exogenous interference->
From the following componentsCalculating a modelable interference estimate +.>
Calculating state estimation value of satellite attitude control systemAnd calculating an actuator and sensor composite fault estimation value +.> and />
10. The Kalman filtering method for fault estimation of satellite attitude control system according to claim 9, wherein: the obtained filtering algorithm results comprise a satellite state estimation curve, a composite fault estimation curve based on an actuator and a sensor, an improved whale optimization algorithm fitness function iteration curve and index data for evaluating the filtering algorithm, wherein the index data is root mean square error.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310122951.5A CN116149186B (en) | 2023-02-16 | 2023-02-16 | Kalman filtering method for fault estimation of satellite attitude control system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310122951.5A CN116149186B (en) | 2023-02-16 | 2023-02-16 | Kalman filtering method for fault estimation of satellite attitude control system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN116149186A CN116149186A (en) | 2023-05-23 |
| CN116149186B true CN116149186B (en) | 2023-08-11 |
Family
ID=86352257
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310122951.5A Active CN116149186B (en) | 2023-02-16 | 2023-02-16 | Kalman filtering method for fault estimation of satellite attitude control system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116149186B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1999010783A1 (en) * | 1997-08-22 | 1999-03-04 | Voyan Technology | A method for real-time nonlinear system state estimation and control |
| WO2014030054A2 (en) * | 2012-08-21 | 2014-02-27 | Gonzalez Juan Valdes | Method and device for determining the electrical properties of materials |
| EP2784445A2 (en) * | 2013-03-26 | 2014-10-01 | Honeywell International Inc. | Selected aspects of advanced receiver autonomous integrity monitoring application to kalman filter based navigation filter |
| CN109885075A (en) * | 2019-03-06 | 2019-06-14 | 扬州大学 | A kind of anti-interference fault tolerant control method of attitude of satellite based on T-S obscurity model building |
| CN110955231A (en) * | 2019-12-18 | 2020-04-03 | 中国科学院长春光学精密机械与物理研究所 | Satellite attitude control system tiny fault detection method based on robust observer |
| CN111999747A (en) * | 2020-08-28 | 2020-11-27 | 大连海事大学 | Robust fault detection method for inertial navigation-satellite combined navigation system |
| CN112212860A (en) * | 2020-08-28 | 2021-01-12 | 山东航天电子技术研究所 | Distributed filtering micro-nano satellite attitude determination method with fault tolerance |
| CN113253050A (en) * | 2021-05-08 | 2021-08-13 | 南昌大学 | Traveling wave fault location method based on whale optimization Kalman filtering algorithm |
-
2023
- 2023-02-16 CN CN202310122951.5A patent/CN116149186B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO1999010783A1 (en) * | 1997-08-22 | 1999-03-04 | Voyan Technology | A method for real-time nonlinear system state estimation and control |
| WO2014030054A2 (en) * | 2012-08-21 | 2014-02-27 | Gonzalez Juan Valdes | Method and device for determining the electrical properties of materials |
| EP2784445A2 (en) * | 2013-03-26 | 2014-10-01 | Honeywell International Inc. | Selected aspects of advanced receiver autonomous integrity monitoring application to kalman filter based navigation filter |
| CN109885075A (en) * | 2019-03-06 | 2019-06-14 | 扬州大学 | A kind of anti-interference fault tolerant control method of attitude of satellite based on T-S obscurity model building |
| CN110955231A (en) * | 2019-12-18 | 2020-04-03 | 中国科学院长春光学精密机械与物理研究所 | Satellite attitude control system tiny fault detection method based on robust observer |
| CN111999747A (en) * | 2020-08-28 | 2020-11-27 | 大连海事大学 | Robust fault detection method for inertial navigation-satellite combined navigation system |
| CN112212860A (en) * | 2020-08-28 | 2021-01-12 | 山东航天电子技术研究所 | Distributed filtering micro-nano satellite attitude determination method with fault tolerance |
| CN113253050A (en) * | 2021-05-08 | 2021-08-13 | 南昌大学 | Traveling wave fault location method based on whale optimization Kalman filtering algorithm |
Non-Patent Citations (1)
| Title |
|---|
| 范数有界的鲁棒卫星姿态估计方法;赵琳;邱海洋;郝勇;;华中科技大学学报(自然科学版)(04);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN116149186A (en) | 2023-05-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103472820B (en) | Method for diagnosing propulsion system failure based on partial least squares algorithm | |
| CN109799803B (en) | LFT-based aeroengine sensor and actuator fault diagnosis method | |
| Ichalal et al. | Simultaneous state and unknown inputs estimation with PI and PMI observers for Takagi Sugeno model with unmeasurable premise variables | |
| Aboutalebi et al. | A novel sensor fault detection in an unmanned quadrotor based on adaptive neural observer | |
| CN108594638B (en) | Spacecraft ACS (auto-configuration transform) on-orbit reconstruction method oriented to multitask and multi-index optimization constraints | |
| Sushchenko et al. | Processing of redundant information in airborne electronic systems by means of neural networks | |
| Rahimi et al. | Fault isolation of reaction wheels for satellite attitude control | |
| CN106249590B (en) | The method that integrated self-adaptive Nano satellite posture determines | |
| Zhao et al. | Goal representation adaptive critic design for discrete-time uncertain systems subjected to input constraints: The event-triggered case | |
| Cao et al. | A novel unscented predictive filter for relative position and attitude estimation of satellite formation | |
| CN116149186B (en) | Kalman filtering method for fault estimation of satellite attitude control system | |
| Liu et al. | A robust adaptive linear parameter-varying gain-scheduling controller for aeroengines | |
| Antoulas et al. | Model reduction of large-scale dynamical systems | |
| CN115355927A (en) | Quick initial course alignment method, system and terminal of strapdown inertial navigation system | |
| Okatan et al. | Kalman filter innovation sequence based fault detection in LEO satellite attitude determination and control system | |
| CN114567288A (en) | Distributed cooperative nonlinear system state estimation method based on variational Bayes | |
| Pham et al. | Aerodynamic sensing for hypersonics via scientific machine learning | |
| Tan et al. | A novel RPI set computation method for discrete-time LPV systems with bounded uncertainties | |
| Shu et al. | Parametric Aeroelastic Reduced-Order Modeling with Hyperparameter Optimization for Flutter Analysis | |
| Volponi et al. | A bootstrap data methodology for sequential hybrid engine model building | |
| Rahimi et al. | Fault isolation of reaction wheels onboard 3-axis controlled in-orbit satellite using ensemble machine learning techniques | |
| Sigthorsson | Control-oriented modeling and output feedback control of hypersonic air-breathing vehicles | |
| Hall et al. | A probabilistic approach for reachability set computation for efficient space situational awareness | |
| Cooper et al. | Deterministic and Probabilistic Approaches to Model and Update Dynamic Systems | |
| Schwartz et al. | Integrated modeling structural tools for the Giant Magellan Telescope design effort |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |