Multi-Sensor State Estimation Method and System Based on Adaptive Observation Enhancement

Abstract

The present invention discloses a multi-sensor state estimation method and system based on adaptive observation enhancement, comprising: using the stator currents of the generator's direct and quadrature axes and the generator's electrical angular velocity collected at time t as the state vector of a Kalman filter; using the measured stator currents of the direct and quadrature axes as the observation vector of the Kalman filter; if the current time is a normal sampling time for the Kalman filter, the Kalman filter performs a normal one-step estimation and measurement update; if the current time is not a normal sampling time for the Kalman filter, the Kalman filter performs only a one-step estimation and determines the number of one-step estimations; ultimately, an estimated value of the stator currents of the generator's direct and quadrature axes at the current time is obtained, and further an estimated value of the stator voltages of the generator's direct and quadrature axes is obtained. The present invention increases the number of one-step estimations between two Kalman filter samplings, thereby reducing the impact of data loss from sensors with fast sampling frequencies on the accuracy of generator voltage estimation during multi-sensor data fusion.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-sensor state prediction, and in particular to a multi-sensor state prediction method and system based on adaptive observation enhancement. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] Currently, with the continued expansion of large-scale renewable energy installed capacity, thermal power plants are shouldering an increasing burden in implementing deep peak and frequency regulation for the power grid, posing a significant challenge to the traditional power industry. While energy storage devices vary widely in type and characteristics, few are well-suited for coordinating frequency regulation with wind farms. Flywheel energy storage, with its advantages of long lifespan, clean operation, and unlimited charge and discharge cycles, can achieve frequency regulation, peak load regulation, and active and reactive power regulation. The deployment of flywheel energy storage can effectively support the transformation of new power systems and has constructive significance for promoting the low-carbon transition of the energy industry.

[0004] Generator voltage monitoring is a key technology in flywheel energy storage systems. Although multi-sensor data fusion is currently used, it still faces challenges. The generation of flywheel generator voltage is related to multiple parameters, and the sampling frequencies of the multiple sensors that collect these parameters vary. For such multi-sensor data fusion systems, the common practice is to set the sampling frequency of the data fusion filter (such as the Kalman filter) to match that of the sensor with the lower sampling frequency. However, this situation prevents some measurement information from the sensors with the faster sampling frequency from being fully integrated into the data fusion system, affecting the accuracy of the data fusion. Summary of the Invention

[0005] To solve the above problems, the present invention proposes a multi-sensor state estimation method and system based on adaptive observation enhancement. In the traditional sensor processing gap, the frequency of one-step prediction is increased, and the Mahalanobis distance is used to judge the effect of observation enhancement until the estimation accuracy meets the requirements.

[0006] In some embodiments, the following technical solutions are adopted:

[0007] A multi-sensor state estimation method based on adaptive observation enhancement, comprising:

[0008] The stator currents of the direct axis and quadrature axis of the generator and the electrical angular velocity of the generator collected at time t are used as the state vectors of the Kalman filter; the stator currents of the direct axis and quadrature axis measured are used as the observation vectors of the Kalman filter;

[0009] If the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs one-step estimation and measurement update normally;

[0010] If the current moment is not a normal sampling moment for Kalman filtering, Kalman performs only one-step estimation, and determines the number of one-step estimations by calculating the Mahalanobis distance between the one-step estimated observation vector and the observation vector obtained at the previous normal sampling moment.

[0011] Finally, the estimated values of the stator current of the direct axis and quadrature axis of the generator at the current moment are obtained, and then the estimated values of the stator voltage of the direct axis and quadrature axis of the generator are obtained.

[0012] As a further solution, the state equation of the Kalman filter is specifically:

[0013]

[0014] Among them, i d 、i q are the direct-axis and quadrature-axis stator currents at time t respectively; μ d 、μ q are the direct-axis and quadrature-axis stator voltages at time t respectively; ω e is the electrical angular velocity of the generator, R s is the stator resistance, ψ f is the permanent magnet flux, L d 、L q are the direct-axis and quadrature-axis inductances at time t, w t is the system noise at time t, and its covariance matrix is Q; is the state vector, is the derivative of the direct-axis stator current at time t, is the derivative of the quadrature-axis stator current at time t, The derivative of the generator's electrical angular velocity.

[0015] As a further solution, the observation equation of the Kalman filter is specifically:

[0016]

[0017] Among them, v t represents the measurement noise, whose covariance matrix is R; are the direct-axis and quadrature-axis stator currents measured at time t, is the measured electrical angular velocity of the generator,

[0018] As a further solution, if the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs a normal one-step estimation and measurement update. The iterative steps at time t are:

[0019] Let’s make an estimate first:

[0020]

[0021] P t - =F t P t-1 F t T +Q

[0022] Then perform measurement update:

[0023] K t =P t - H T (HP t - H T +R) -1

[0024]

[0025] P t =(IK t H)P t -

[0026] in, is the one-step prediction matrix of the state vector, P t 、P t-1 Represent the error matrix at time t and time t-1, K t is the filter gain, f(i t )’s Jacobian matrix; H is the measurement matrix; Q is the system noise w t The covariance matrix of t The covariance matrix of y t is the observation vector, P t - is the one-step estimate of the error matrix, i t is the state vector.

[0027] If the current moment is not the normal sampling moment of the Kalman filter, Kalman only performs one-step estimation. The specific process is:

[0028] Let’s make an estimate first:

[0029]

[0030] P t - =F t Pt-1 F t T +Q

[0031] Then calculate the Mahalanobis distance d, if d<d set , then stop one-step prediction and wait for measurement update; where d set is a pre-set threshold; if d>d set , then make another prediction and calculate the Mahalanobis distance until the Mahalanobis distance is less than d set Or arrive at the normal sampling time of Kalman filtering.

[0032] As a further solution, the Mahalanobis distance d is calculated as follows:

[0033]

[0034] in, is the one-step prediction matrix of the state vector, H is the observation matrix, R is the covariance of the measurement noise, y t is the observation vector.

[0035] In other embodiments, the following technical solutions are adopted:

[0036] A multi-sensor state prediction system based on adaptive observation enhancement, comprising:

[0037] The data acquisition module is used to use the stator currents of the direct axis and quadrature axis of the generator and the electrical angular velocity of the generator collected at time t as the state vector of the Kalman filter; and use the measured stator currents of the direct axis and quadrature axis as the observation vector of the Kalman filter;

[0038] The current estimation module is used to perform one-step estimation and measurement update normally when the current moment is the normal sampling moment of the Kalman filter. When the current moment is not the normal sampling moment of the Kalman filter, the Kalman filter only performs one-step estimation and determines the number of one-step estimations by calculating the Mahalanobis distance between the observation vector of the one-step estimation and the observation vector obtained at the previous normal sampling moment. Finally, the estimated stator current values of the direct and quadrature axes of the generator at the current moment are obtained.

[0039] In other embodiments, the following technical solutions are adopted:

[0040] A terminal device includes a processor and a memory, wherein the processor is used to implement various instructions; the memory is used to store multiple instructions, and the instructions are suitable for being loaded by the processor and executing the above-mentioned multi-sensor state estimation method based on adaptive observation enhancement.

[0041] In other embodiments, the following technical solutions are adopted:

[0042] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded and executed by a processor of a terminal device for the multi-sensor state estimation method based on adaptive observation enhancement.

[0043] Compared with the prior art, the present invention has the following beneficial effects:

[0044] (1) Based on the traditional Kalman filter, the method of the present invention increases the number of one-step estimation between two samplings of the Kalman filter, thereby reducing the impact of the loss of sensor data with a fast sampling frequency on the accuracy of generator voltage estimation in multi-sensor data fusion.

[0045] (2) The present invention determines whether to terminate the newly added one-step prediction by calculating the Mahalanobis distance, thereby enabling the system to adaptively add one-step prediction.

[0046] Other features and advantages of additional aspects of the present invention will be given in part in the following description and in part will become obvious from the following description or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 This is a flow chart of a multi-sensor state estimation method based on adaptive observation enhancement in an embodiment of the present invention. DETAILED DESCRIPTION

[0048] It should be noted that the following detailed description is illustrative and is intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the art to which the present application belongs.

[0049] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0050] Example 1

[0051] In one or more embodiments, a multi-sensor state estimation method based on adaptive observation enhancement is disclosed, combining Figure 1 , Figure 1 The square box is the generator angular velocity sampling frequency (faster), and the rest are the current and voltage sampling frequencies (slower).

[0052] The method of this embodiment specifically includes the following process:

[0053] (1) The stator currents of the direct and quadrature axes of the generator and the electrical angular velocity of the generator collected at time t are used as the state vectors of the Kalman filter; the stator currents of the direct and quadrature axes measured are used as the observation vectors of the Kalman filter;

[0054] In this embodiment, the state equation of the Kalman filter algorithm is specifically:

[0055]

[0056] where i d 、i q are the direct-axis and quadrature-axis stator currents at time t respectively; μ d 、μ q are the direct-axis and quadrature-axis stator voltages at time t respectively; ω e is the electrical angular velocity of the generator, R s is the stator resistance, ψ f is the permanent magnet flux, L d 、L q are the direct-axis and quadrature-axis inductances at time t, w t is the system noise at time t, and its covariance matrix is Q; is the state vector, is the derivative of the direct-axis stator current at time t, is the derivative of the quadrature-axis stator current at time t, The derivative of the generator's electrical angular velocity.

[0057] The observation equation of the Kalman filter algorithm is specifically:

[0058]

[0059] Among them, v t represents the measurement noise, whose covariance matrix is R; are the direct-axis and quadrature-axis stator currents measured at time t, is the measured electrical angular velocity of the generator.

[0060] (2) If the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs one-step estimation and measurement update normally; the specific iterative steps at time t are:

[0061] Let’s make an estimate first:

[0062]

[0063] P t - =Ft P t-1 F t T +Q

[0064] Then perform measurement update:

[0065] K t =P t - H T (HP t - H T +R) -1

[0066]

[0067] P t =(IK t H)P t -

[0068] in, is the one-step prediction matrix of the state vector, P t 、P t-1 Represent the error matrix at time t and time t-1, K t is the filter gain, f(i t )’s Jacobian matrix; H is the measurement matrix; Q is the system noise w t The covariance matrix of t The covariance matrix of y t is the observation vector, P t - is the one-step estimate of the error matrix, i t is the state vector.

[0069] (3) If the current moment is not a normal sampling moment for Kalman filtering, Kalman performs only one-step estimation, and determines the number of one-step estimations by calculating the Mahalanobis distance between the observation vector of the one-step estimation and the observation vector obtained at the previous normal sampling moment;

[0070] Specifically, if the current moment is not the normal sampling moment of the Kalman filter, Kalman only performs one-step estimation and does not perform measurement update. The specific iterative steps at time t are:

[0071] Let’s make an estimate first:

[0072]

[0073] P t - =F t Pt-1 F t T +Q

[0074] Then calculate the Mahalanobis distance:

[0075]

[0076] If d <d set , it means that the filtering accuracy is good, then stop the prediction step and wait for the measurement update; where d set is a preset threshold;

[0077] If d>d set , it means that the filtering accuracy needs to be improved, so one-step prediction is performed again, and then the Mahalanobis distance is calculated again until the Mahalanobis distance is less than the threshold or the sampling time is reached.

[0078] (4) Finally, the estimated values of the stator currents of the direct axis and quadrature axis of the generator at the current moment are obtained, and then the estimated values of the stator voltages of the direct axis and quadrature axis of the generator are obtained.

[0079] This embodiment increases the number of one-step estimations between two samplings of the Kalman filter, thereby reducing the impact of data loss of sensors with fast sampling frequencies on the accuracy of voltage estimation in multi-sensor data fusion.

[0080] Example 2

[0081] In one or more embodiments, a multi-sensor state estimation system based on adaptive observation enhancement is disclosed, comprising:

[0082] The data acquisition module is used to use the stator currents of the direct axis and quadrature axis of the generator and the electrical angular velocity of the generator collected at time t as the state vector of the Kalman filter; and use the measured stator currents of the direct axis and quadrature axis as the observation vector of the Kalman filter;

[0083] The current estimation module is used to perform one-step estimation and measurement update normally when the current moment is the normal sampling moment of the Kalman filter. When the current moment is not the normal sampling moment of the Kalman filter, the Kalman filter only performs one-step estimation and determines the number of one-step estimations by calculating the Mahalanobis distance between the observation vector of the one-step estimation and the observation vector obtained at the previous normal sampling moment. Finally, the estimated stator current values of the direct and quadrature axes of the generator at the current moment are obtained.

[0084] It should be noted that the specific implementation of each of the above modules has been described in detail in Example 1, and is the same here as in Example 1 and will not be described in detail again.

[0085] Example 3

[0086] In one or more embodiments, a terminal device is disclosed, including a server. The server includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the multi-sensor state estimation method based on adaptive observation enhancement in Example 1 is implemented. For the sake of brevity, this description is omitted here.

[0087] It should be understood that in this embodiment, the processor may be a central processing unit (CPU), or may be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), off-the-shelf field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0088] The memory may include a read-only memory and a random access memory, and provides instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.

[0089] During implementation, each step of the above method may be completed by an integrated logic circuit of hardware in a processor or by instructions in the form of software.

[0090] Example 4

[0091] In one or more embodiments, a computer-readable storage medium is disclosed, which stores a plurality of instructions suitable for being loaded by a processor of a terminal device and executing the multi-sensor state estimation method based on adaptive observation enhancement described in Example 1.

[0092] Although the above describes the specific embodiments of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without any creative work are still within the scope of protection of the present invention.

Claims

1. A multi-sensor state estimation method based on adaptive observation enhancement, characterized in that: include: The stator currents of the direct axis and quadrature axis of the generator and the electrical angular velocity of the generator collected at time t are used as the state vectors of the Kalman filter; the stator currents of the direct axis and quadrature axis measured are used as the observation vectors of the Kalman filter; If the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs one-step estimation and measurement update normally; If the current moment is not a normal sampling moment for Kalman filtering, Kalman performs only one-step estimation, and determines the number of one-step estimations by calculating the Mahalanobis distance between the one-step estimated observation vector and the observation vector obtained at the previous normal sampling moment. Finally, the estimated values of the stator current of the direct axis and quadrature axis of the generator at the current moment are obtained, and then the estimated values of the stator voltage of the direct axis and quadrature axis of the generator are obtained; The state equation of the Kalman filter is specifically: in, 、 are the direct-axis and quadrature-axis stator currents at time t respectively; 、 are the direct-axis and quadrature-axis stator voltages at time t respectively; is the electrical angular velocity of the generator, is the stator resistance, is the permanent magnet flux, 、 are the direct-axis and quadrature-axis inductances at time t, is the system noise at time t, and its covariance matrix is Q; is the state vector, is the derivative of the direct-axis stator current at time t, is the derivative of the quadrature-axis stator current at time t, The derivative of the generator's electrical angular velocity; The observation equation of the Kalman filter is specifically: in, represents the measurement noise, and its covariance matrix is ; 、 are the direct-axis and quadrature-axis stator currents measured at time t, is the measured electrical angular velocity of the generator, ; If the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs one-step estimation and measurement update normally. The iterative steps at time t are: Let’s make an estimate first: Then perform measurement update: in, is the one-step prediction matrix of the state vector, 、 Represent the error matrices at time t and time t-1 respectively, is the filter gain, for The Jacobian matrix of ; is the observation matrix; Q is the system noise The covariance matrix of To measure noise The covariance matrix of is the observation vector, is the one-step estimate of the error matrix, is the state vector.

2. The multi-sensor state estimation method based on adaptive observation enhancement according to claim 1, characterized in that: If the current moment is not the normal sampling moment of the Kalman filter, Kalman only performs one-step estimation. The specific process is: Let’s make an estimate first: Then calculate the Mahalanobis distance d, if d<d set , then stop one-step prediction and wait for measurement update; where d set is a pre-set threshold; if d>d set , then make another prediction and calculate the Mahalanobis distance until the Mahalanobis distance is less than d set Or arrive at the normal sampling time of Kalman filtering.

3. The multi-sensor state estimation method based on adaptive observation enhancement according to claim 2, characterized in that: Calculate the Mahalanobis distance d, specifically: in, is the one-step prediction matrix of the state vector, is the observation matrix, is the covariance of the measurement noise, is the observation vector.

4. A multi-sensor state prediction system based on adaptive observation enhancement, characterized in that: include: The data acquisition module is used to use the stator currents of the direct axis and quadrature axis of the generator and the electrical angular velocity of the generator collected at time t as the state vector of the Kalman filter; and use the measured stator currents of the direct axis and quadrature axis as the observation vector of the Kalman filter; The current estimation module is used to perform one-step estimation and measurement update normally when the current moment is the normal sampling moment of the Kalman filter. When the current moment is not the normal sampling moment of the Kalman filter, the Kalman filter only performs one-step estimation and determines the number of one-step estimations by calculating the Mahalanobis distance between the observation vector of the one-step estimation and the observation vector obtained at the previous normal sampling moment. Ultimately, the estimated stator current values of the direct and quadrature axes of the generator at the current moment are obtained. Among them, the state equation of the Kalman filter is specifically: in, 、 are the direct-axis and quadrature-axis stator currents at time t respectively; 、 are the direct-axis and quadrature-axis stator voltages at time t respectively; is the electrical angular velocity of the generator, is the stator resistance, is the permanent magnet flux, 、 are the direct-axis and quadrature-axis inductances at time t, is the system noise at time t, and its covariance matrix is Q; is the state vector, is the derivative of the direct-axis stator current at time t, is the derivative of the quadrature-axis stator current at time t, The derivative of the generator's electrical angular velocity; The observation equation of the Kalman filter is specifically: in, represents the measurement noise, and its covariance matrix is ; 、 are the direct-axis and quadrature-axis stator currents measured at time t, is the measured electrical angular velocity of the generator, ; If the current moment is the normal sampling moment of the Kalman filter, the Kalman filter performs one-step estimation and measurement update normally. The iterative steps at time t are: Let’s make an estimate first: Then perform measurement update: in, is the one-step prediction matrix of the state vector, 、 Represent the error matrices at time t and time t-1 respectively, is the filter gain, for The Jacobian matrix of ; is the observation matrix; Q is the system noise The covariance matrix of To measure noise The covariance matrix of is the observation vector, is the one-step estimate of the error matrix, is the state vector.

5. A terminal device comprising a processor and a memory, wherein the processor is used to implement various instructions; the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the multi-sensor state estimation method based on adaptive observation enhancement according to any one of claims 1 to 3.

6. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the multi-sensor state estimation method based on adaptive observation enhancement according to any one of claims 1 to 3.

Images