A method and system for repairing waveform distortion of power grid equipment measurements

Abstract

The application is suitable for the technical field of power internet of things, and relates to a power grid equipment measurement waveform distortion repair method and system, which comprises the following steps: dividing a power grid voltage signal sequence into multiple batches according to the cycle length; constructing a nonlinear observation model based on a sine function, and representing the power grid signal of each sampling point in the batch in the form of the nonlinear observation model; constructing an extended Kalman filter based on the nonlinear observation model; obtaining the state estimation value of each sampling point by using the extended Kalman filter; performing state transmission between adjacent batches, and adding pi after the phase parameter in the state estimation value at the end of filtering of the previous batch to serve as the phase parameter in the initial state estimation value of the next batch; and merging and outputting the repaired power grid signal. The application solves the problems of low filtering precision, poor waveform distortion repair effect and the like of the traditional linear Kalman filter, and has the advantages of high filtering precision, fast convergence speed, good batch processing continuity and the like.

Description

Technical Field

[0001] This invention belongs to the field of power Internet of Things technology, and in particular relates to a method and system for repairing waveform distortion in power grid equipment measurements. Background Technology

[0002] The accuracy of measurement signals from power grid equipment is crucial for the monitoring, protection, and control of power systems. Signals acquired by measuring devices such as voltage transformers and current transformers often contain noise, harmonics, and other interference, leading to waveform distortion and affecting subsequent applications such as power quality analysis and fault diagnosis. Therefore, effectively restoring distorted waveforms is an important research direction in the field of the power Internet of Things (IoT).

[0003] Kalman filtering, as a recursive optimal estimation algorithm, is widely used in signal processing. Its basic principle is to optimally estimate the state of a noisy system by combining the system's state equations and observation equations with statistical characteristics. In power system signal processing, existing technologies have attempted to apply Kalman filtering to waveform restoration; however, these technologies suffer from problems in practical applications, including a mismatch between the linear model and the nonlinear characteristics of the power grid signal, a lack of effective initial state estimation mechanisms, a lack of state continuity guarantees between batch processing, and insufficient parameter adaptive capabilities. Patent CN103995180B provides a power system frequency estimation method considering inequality constraints. The method's steps are as follows: Step 1: Obtain the target power signal for filtering; Step 2: Determine the characteristics of the target power signal for filtering; Step 3: Establish a filtering model for the target signal, including the corresponding state equations, observation equations, state inequality constraints, and the discrete target signal; Step 4: After linearizing the filtering model, design an extended Kalman filter under equality constraints; Step 5: Select different extended Kalman filters based on whether the state at the online sampling time satisfies the inequality conditions; Step 6: Output the signal through the interface. This patent mainly corrects the estimation results by adding inequality constraints, but it fails to fundamentally solve the problem of mismatch between the linear model and the nonlinear characteristics of the sinusoidal signal, and it also fails to establish a parameterized state-space model that can directly describe the signal amplitude, phase and offset.

[0004] Therefore, how to improve filtering accuracy, achieve effective waveform distortion repair, enhance the accuracy of initial state estimation, ensure the continuity of batch processing, and enhance the adaptability to changes in power grid signal parameters are problems that urgently need to be solved by personnel in this technical field. Summary of the Invention

[0005] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method for repairing waveform distortion in power grid equipment measurements. This method solves the problems of low accuracy and poor waveform distortion repair in existing technologies due to issues such as mismatch between linear models and the nonlinear characteristics of power grid signals, lack of effective initial state estimation mechanisms, lack of state continuity guarantees between batch processing, and insufficient parameter adaptive capabilities. In addition, this invention also provides a system for repairing waveform distortion in power grid equipment measurements.

[0006] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0007] In a first aspect, the present invention provides a method for repairing waveform distortion in power grid equipment measurements, comprising the following steps:

[0008] S10. Obtain the grid voltage signal sequence to be processed, and divide the grid voltage signal sequence into multiple batches according to the period length;

[0009] S20. For each batch, construct a nonlinear observation model based on a sine function, and represent the power grid signal at each sampling point in the batch in the form of the nonlinear observation model;

[0010] S30. Based on the aforementioned nonlinear observation model, construct an extended Kalman filter;

[0011] S40. The extended Kalman filter is used to perform point-by-point filtering estimation on the sampling points in each batch to obtain the state estimate value of each sampling point;

[0012] S50. Perform state transfer between adjacent batches. Add 2π to the phase parameter in the state estimate at the end of the previous batch filtering and use it as the phase parameter in the initial state estimate for the next batch filtering.

[0013] S60. Merge the filtering results of all batches and output the repaired power grid signal.

[0014] Furthermore, in S20, the nonlinear observation model is expressed as follows:

[0015]

[0016] Where z_k is the observed value of the k-th sampling point, A is the signal amplitude, k is the sampling point index, and N is the batch length. Where C is the signal phase, C is the DC offset, and v_k is the observation noise.

[0017] Furthermore, in S30, the state vector of the extended Kalman filter is represented as follows:

[0018] .

[0019] Furthermore, the observation equation of the extended Kalman filter is expressed as follows:

[0020] .

[0021] Furthermore, the Jacobian matrix of the observation equation is expressed as follows:

[0022]

[0023] in:

[0024] .

[0025] Furthermore, S40 includes a time update for prediction: predicting the prior state estimate and prior estimate error covariance matrix at the current time based on the state estimate and covariance matrix of the previous time step;

[0026] It also includes observation updates for correction: based on the observations at the current time, the nonlinear observation model, and the Jacobian matrix, the Kalman gain is calculated, and the Kalman gain is used to correct the state prior estimate and the prior estimate error covariance matrix to obtain the state posterior estimate and the posterior estimate error covariance matrix at the current time.

[0027] Furthermore, the time update is represented as follows:

[0028]

[0029] The observation update is represented as follows:

[0030]

[0031] Where x̂_{k|k-1} is the prior estimate of the k-th state based on the first k-1 observations, x̂_{k|k} is the posterior estimate of the k-th state based on the first k observations, P_{k|k-1} is the prior estimate error covariance matrix, P_{k|k} is the posterior estimate error covariance matrix, K_k is the Kalman gain of the k-th sampling point, ŷ_k is the innovation, Q is the system noise covariance matrix, and R is the observation noise covariance.

[0032] Furthermore, before performing filtering estimation on the sampling points of each batch, the state vector and error covariance matrix are initialized based on prior information.

[0033] Furthermore, the initial state is estimated using the known sin90° position information.

[0034] Secondly, the present invention also provides a power grid equipment measurement waveform distortion repair system, comprising:

[0035] The signal segmentation module is used to acquire the power grid signal sequence to be processed and to segment the power grid signal sequence into multiple batches according to the period length;

[0036] The signal modeling module is used to build a nonlinear observation model based on a sine function for each batch;

[0037] A filter construction module is used to construct an extended Kalman filter based on the nonlinear observation model;

[0038] The filtering module is used to perform point-by-point filtering estimation on the sampling points in each batch using the extended Kalman filter to obtain the state estimate value of each sampling point.

[0039] The state transfer module is used to transfer state between adjacent batches. It adds 2π to the phase parameter in the state estimate value at the end of the previous batch of filtering and uses it as the phase parameter in the initial state estimate value of the next batch of filtering.

[0040] The signal output module is used to merge the filtering results of all batches and output the repaired power grid signal.

[0041] Compared with the prior art, the method and system for repairing waveform distortion in power grid equipment provided by this invention have at least the following advantages:

[0042] Existing technologies suffer from several drawbacks in practical applications, including a mismatch between the linear model and the nonlinear characteristics of power grid signals, a lack of effective initial state estimation mechanisms, insufficient state continuity guarantees between batch processing, and inadequate parameter adaptation capabilities. These issues lead to low accuracy and poor waveform distortion repair in linear Kalman filtering. This invention improves filtering accuracy by establishing a nonlinear observation model based on a sine function, making Kalman filtering more consistent with the physical characteristics of power grid signals. This invention utilizes a parameterized sinusoidal state model (amplitude, phase, and offset) to effectively separate signal and noise components through the Kalman filtering algorithm, achieving accurate waveform distortion repair. By leveraging known sin90° position information, this invention provides more accurate initial amplitude and phase estimates, effectively improving filter convergence speed. By transferring states between batches and accumulating 2π on the phase parameter, this invention ensures the continuity and smoothness of the processing results across the entire signal sequence. By updating the amplitude, phase, and offset parameters in the state vector in real time, this invention achieves adaptive tracking of slowly changing signal parameters. In summary, this invention solves the problems of low accuracy and poor adaptability of traditional linear Kalman filtering when processing periodic power grid signals, providing a more effective solution for signal processing in power systems. Attached Figure Description

[0043] To more clearly illustrate the solution of the present invention, a brief introduction will be given to the drawings used in the description of the embodiments below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0044] Figure 1 This is a flowchart of a method for repairing waveform distortion in power grid equipment, provided as an embodiment of the present invention. Detailed Implementation

[0045] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Preferred embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the invention.

[0046] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention.

[0047] This invention provides a method for repairing waveform distortion in power grid equipment measurements, applicable to signal processing scenarios in power systems. The method includes the following steps:

[0048] S10. Obtain the grid voltage signal sequence to be processed and divide the grid voltage signal sequence into multiple batches according to the period length; S20. For each batch, construct a nonlinear observation model based on a sine function, and represent the grid signal of each sampling point in the batch in the form of the nonlinear observation model; S30. Based on the nonlinear observation model, construct an extended Kalman filter; S40. Use the extended Kalman filter to perform point-by-point filtering estimation on the sampling points in each batch to obtain the state estimate value of each sampling point; S50. Perform state transfer between adjacent batches, add 2π to the phase parameter in the state estimate value at the end of the filtering of the previous batch, and use it as the phase parameter in the initial state estimate value of the filtering of the next batch; S60. Merge the filtering results of all batches and output the repaired grid signal.

[0049] This invention establishes a nonlinear observation model that conforms to the sinusoidal characteristics of power grid signals, and directly parameterizes the state vector into amplitude, phase, and offset. This solves the problems of low accuracy and poor waveform distortion repair effect of traditional linear Kalman filtering. It has the advantages of high filtering accuracy, fast convergence speed, and good batch processing continuity, and is suitable for real-time online signal processing scenarios in power systems.

[0050] This invention provides a method for repairing waveform distortion in power grid equipment measurements, applicable to signal processing scenarios in power systems, such as... Figure 1 As shown, in this embodiment, the method for repairing waveform distortion in power grid equipment measurements includes the following steps:

[0051] S10. Obtain the grid voltage signal sequence to be processed, and divide the grid voltage signal sequence into multiple batches according to the period length.

[0052] S20. For each batch, construct a nonlinear observation model based on a sine function, and represent the power grid signal of each sampling point in the batch in the form of a nonlinear observation model.

[0053] Specifically, in this embodiment, the nonlinear observation model is represented as follows:

[0054]

[0055] Where z_k is the observed value (voltage measurement) of the kth sampling point, A is the signal amplitude, k is the sampling point index, and N is the batch length (the number of sampling points in a complete cycle). Where is the signal phase, C is the DC offset, and v_k is the observation noise.

[0056] Furthermore, in this embodiment, the state vector of the extended Kalman filter is represented as follows:

[0057] .

[0058] The parameterized representation method described above has the following advantages compared to traditional linear models:

[0059] It is more in line with the physical characteristics of power grid signals;

[0060] Key parameters of the signal (amplitude, phase, offset) can be obtained directly.

[0061] It facilitates the analysis of signal change trends.

[0062] S30. Based on the nonlinear observation model, an extended Kalman filter is constructed.

[0063] Specifically, in this embodiment, due to the use of a nonlinear observation model, the Extended Kalman Filter (EKF) algorithm is employed. The observation equation of the Extended Kalman Filter is expressed as follows:

[0064] .

[0065] The Jacobian matrix (observation matrix H) of the observation equation is expressed as follows:

[0066]

[0067] in:

[0068] .

[0069] This design enables the observation model to accurately reflect the impact of state variables on the observed values, thus improving the filtering accuracy.

[0070] S40. Use an extended Kalman filter to perform point-by-point filtering estimation on the sampling points in each batch to obtain the state estimate value of each sampling point.

[0071] Specifically, in this embodiment, the following state-space model is adopted:

[0072] The state equations are as follows:

[0073]

[0074] Since it is assumed that the signal parameters remain basically unchanged within one period, the state transition matrix is ​​the identity matrix, the system noise w_{k-1} is zero-mean Gaussian white noise, and the covariance matrix is ​​Q.

[0075] The observation equation is as follows:

[0076]

[0077] Where h(x_k) is the aforementioned nonlinear observation function, v_k is the observation noise, which is zero-mean Gaussian white noise with covariance R.

[0078] Furthermore, in this embodiment, the prediction includes a time update: based on the state estimate and covariance matrix of the previous time step, predicting the state prior estimate and prior estimate error covariance matrix of the current time step;

[0079] It also includes observation updates for correction: based on the current observations, the nonlinear observation model, and the Jacobian matrix, the Kalman gain is calculated, and the Kalman gain is used to correct the state prior estimate and the prior estimate error covariance matrix to obtain the current state posterior estimate and the posterior estimate error covariance matrix.

[0080] Furthermore, in this embodiment, the time update is represented as follows:

[0081]

[0082] The observation update is represented as follows:

[0083]

[0084] Where x̂_{k|k-1} is the prior estimate of the k-th state based on the first k-1 observations, x̂_{k|k} is the posterior estimate of the k-th state based on the first k observations, P_{k|k-1} is the prior estimate error covariance matrix, P_{k|k} is the posterior estimate error covariance matrix, K_k is the Kalman gain of the k-th sampling point, ŷ_k is the innovation, Q is the system noise covariance matrix, and R is the observation noise covariance.

[0085] Furthermore, in this embodiment, before performing filtering estimation on the sampling points of each batch, the state vector and error covariance matrix are initialized based on prior information.

[0086] Furthermore, in this embodiment, the known sin90° position information is used for initial state estimation.

[0087] S50. Perform state transfer between adjacent batches. Add 2π to the phase parameter in the state estimate at the end of the previous batch filtering and use it as the phase parameter in the initial state estimate for the next batch filtering.

[0088] S60. Merge the filtering results of all batches and output the repaired power grid signal.

[0089] In this embodiment, by constructing the observation model of the Kalman filter using a sine function, the limitation of the traditional Kalman filter using a linear observation model is overcome, and the sinusoidal characteristics of the power grid signal are reflected more accurately, thereby significantly improving the filtering accuracy and signal restoration. This method directly parameterizes the state vector of the power grid signal into three physically meaningful parameters: amplitude, phase, and offset, i.e., x_k = [A_k; The design of _k; C_k]^T gives the state vector a clear physical meaning, allowing direct access to the key characteristic parameters of the signal, which facilitates signal analysis and subsequent processing.

[0090] In this embodiment, the initial state estimation of the Kalman filter is optimized using known sin90° position information. This fully utilizes prior signal knowledge, significantly improving the accuracy of the initial estimation and accelerating the filter's convergence speed. When processing continuous signals, this invention employs a batch processing mechanism and state continuity guarantee technology. _{out} = The phase accumulation method of _{final} + 2π solves the efficiency problem of long sequence signal processing, while ensuring the continuity of processing results, making it suitable for real-time online processing scenarios.

[0091] This invention also provides a power grid equipment measurement waveform distortion repair system. In this embodiment, the power grid equipment measurement waveform distortion repair system includes:

[0092] The signal segmentation module acquires the power grid signal sequence to be processed and divides it into multiple batches according to the period length. The signal modeling module constructs a nonlinear observation model based on a sine function for each batch. The filter construction module constructs an extended Kalman filter based on the nonlinear observation model. The filtering module uses the extended Kalman filter to perform point-by-point filtering estimation on the sampling points within each batch to obtain the state estimate value of each sampling point. The state transfer module performs state transfer between adjacent batches by adding 2π to the phase parameter in the state estimate value at the end of the previous batch's filtering and using it as the phase parameter in the initial state estimate value for the next batch's filtering. The signal output module merges the filtering results of all batches and outputs the repaired power grid signal.

[0093] The above-described method and system for repairing waveform distortion in power grid equipment measurements, compared with existing technologies, suffer from several drawbacks in practical applications. Existing technologies exhibit problems such as a mismatch between the linear model and the nonlinear characteristics of the power grid signal, a lack of effective initial state estimation mechanisms, insufficient state continuity guarantees between batch processing, and inadequate parameter adaptive capabilities. These issues lead to low accuracy in linear Kalman filtering and poor waveform distortion repair. This invention improves filtering accuracy by establishing a nonlinear observation model based on a sine function, making Kalman filtering more consistent with the physical characteristics of the power grid signal. This invention utilizes a parameterized sinusoidal state model (amplitude, phase, offset) and effectively separates signal and noise components through a Kalman filtering algorithm, achieving accurate repair of distorted waveforms. This invention provides more accurate initial amplitude and phase estimates by utilizing known sin90° position information, effectively improving the filter convergence speed. This invention ensures the continuity and smoothness of the processing results across the entire signal sequence by transferring states between batches and accumulating 2π on the phase parameter. This invention achieves adaptive tracking of slowly changing signal parameters through real-time updates of the amplitude, phase, and offset parameters in the state vector. In summary, this invention solves the problems of low accuracy and poor adaptability of traditional linear Kalman filtering when processing periodic signals from the power grid, and provides a more effective solution for signal processing in power systems.

[0094] Obviously, the embodiments described above are merely preferred embodiments of the present invention, and not all embodiments. The accompanying drawings illustrate preferred embodiments of the present invention, but do not limit the scope of the patent. The present invention can be implemented in many different forms; rather, these embodiments are provided to provide a more thorough and complete understanding of the disclosure of the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this invention.

Claims

1. A method for repairing waveform distortion in power grid equipment measurements, characterized in that, Includes the following steps: S10. Obtain the grid voltage signal sequence to be processed, and divide the grid voltage signal sequence into multiple batches according to the period length; S20. For each batch, construct a nonlinear observation model based on a sine function, and represent the power grid signal at each sampling point in the batch in the form of the nonlinear observation model; S30. Based on the aforementioned nonlinear observation model, construct an extended Kalman filter; S40. The extended Kalman filter is used to perform point-by-point filtering estimation on the sampling points in each batch to obtain the state estimate value of each sampling point; S50. Perform state transfer between adjacent batches. Add 2π to the phase parameter in the state estimate at the end of the previous batch filtering and use it as the phase parameter in the initial state estimate for the next batch filtering. S60. Merge the filtering results of all batches and output the repaired power grid signal.

2. The method for repairing waveform distortion in power grid equipment according to claim 1, characterized in that, In S20, the nonlinear observation model is expressed as follows: Where z_k is the observed value of the k-th sampling point, A is the signal amplitude, k is the sampling point index, and N is the batch length. Where C is the signal phase, C is the DC offset, and v_k is the observation noise.

3. The method for repairing waveform distortion in power grid equipment according to claim 2, characterized in that, In S30, the state vector of the extended Kalman filter is represented as follows: 。 4. The method for repairing waveform distortion in power grid equipment according to claim 3, characterized in that, The observation equation of the extended Kalman filter is expressed as follows: 。 5. The method for repairing waveform distortion in power grid equipment according to claim 4, characterized in that, The Jacobian matrix of the observation equation is expressed as follows: in: 。 6. The method for repairing waveform distortion in power grid equipment according to claim 1, characterized in that, S40 includes a time update for prediction: based on the state estimate and covariance matrix of the previous time step, predict the state prior estimate and prior estimate error covariance matrix of the current time step. It also includes observation updates for correction: based on the observations at the current time, the nonlinear observation model, and the Jacobian matrix, the Kalman gain is calculated, and the Kalman gain is used to correct the state prior estimate and the prior estimate error covariance matrix to obtain the state posterior estimate and the posterior estimate error covariance matrix at the current time.

7. The method for repairing waveform distortion in power grid equipment according to claim 6, characterized in that, The time update is represented as follows: The observation update is represented as follows: Where x̂_{k|k-1} is the prior estimate of the k-th state based on the first k-1 observations, x̂_{k|k} is the posterior estimate of the k-th state based on the first k observations, P_{k|k-1} is the prior estimate error covariance matrix, P_{k|k} is the posterior estimate error covariance matrix, K_k is the Kalman gain of the k-th sampling point, ŷ_k is the innovation, Q is the system noise covariance matrix, and R is the observation noise covariance.

8. The method for repairing waveform distortion in power grid equipment according to claim 7, characterized in that, Before performing filtered estimation on the sampling points of each batch, the state vector and error covariance matrix are initialized based on prior information.

9. A method for repairing waveform distortion in power grid equipment according to claim 8, characterized in that, Initial state estimation is performed using the known sin90° position information.

10. A system employing the method according to any one of claims 1 to 9, characterized in that, include: The signal segmentation module is used to acquire the power grid signal sequence to be processed and to segment the power grid signal sequence into multiple batches according to the period length; The signal modeling module is used to build a nonlinear observation model based on a sine function for each batch; A filter construction module is used to construct an extended Kalman filter based on the nonlinear observation model; The filtering module is used to perform point-by-point filtering estimation on the sampling points in each batch using the extended Kalman filter to obtain the state estimate value of each sampling point. The state transfer module is used to transfer state between adjacent batches. It adds 2π to the phase parameter in the state estimate value at the end of the previous batch of filtering and uses it as the phase parameter in the initial state estimate value of the next batch of filtering. The signal output module is used to merge the filtering results of all batches and output the repaired power grid signal.

Images