Method for processing synchronous phasor information of power distribution system in unbalanced and noise environment

Abstract

The invention provides a power distribution system synchronous phasor information processing method in an unbalanced and noise environment, and belongs to the technical field of synchronous phasor measurement, and the method comprises the following steps: S1, building a power distribution system AC signal dynamic state estimation model; s2, in a three-phase power distribution system in an unbalanced and noise interference environment, calculating a phasor estimated value by using a recursive filtering method; s3, establishing mutually orthogonal voltage component signals by using a second-order generalized integrator phase-locked loop; s4, performing linearization processing on the nonlinear system state equation, and designing a discrete state equation and an observation equation; s5, performing time updating by adopting recursive calculation combined with estimation and correction; and S6, filtering the estimated value of the state quantity, correcting the estimated value by using a variance matrix, and carrying out measurement updating. According to the method, the synchronous phasor information of the power distribution system can be processed in unbalanced and noise interference environments, the estimation precision is improved, the estimation error is reduced, the convergence time is shortened, and accurate data is provided for fault diagnosis.

Description

Technical Field

[0001] The present invention relates to the technical field of synchronized phasor measurement, and in particular to a synchronized phasor information processing method for a power distribution system in an unbalanced and noisy environment. Background Art

[0002] During the dynamic operation of the distribution system, the synchronized phasor information of the measurement node is easily interfered by the non-fundamental frequency signal integer harmonic signal, the lead induced electrical noise and the noise caused by the window function leakage effect, and the distribution system is in a three-phase unbalanced state. Under the conditions of low signal-to-noise ratio and harmonic distortion of the unbalanced distribution system, the synchronized phasor information is difficult to estimate accurately. Therefore, in view of the characteristics of the distribution system operation condition such as system imbalance, low signal-to-noise ratio, and rich harmonics, a synchronized phasor information processing method for the distribution system under unbalanced and noisy environment is proposed by using modern control theory, power system dynamic equivalent modeling theory and stability analysis theory. It can improve the performance of synchronized phasor information estimation under low signal-to-noise ratio and high-order harmonic conditions, and can improve the accuracy of synchronized phasor measurement in the interactive environment between users and power grid supply and demand, and provide accurate data for distribution system fault diagnosis, which has great practical significance for the development of power systems.

[0003] Consulting the published patent documents, the publication (announcement) number is: CN104423417A, which discloses a synchronous phasor accurate timing system, including: a GPS timing module, a synchronous trigger circuit, a digital signal processor, a host PC, an AD sampling circuit and a signal acquisition conditioner, wherein the output of the signal acquisition conditioner is connected to the digital signal processor through the AD sampling circuit, the digital signal processor is connected to the host PC through a high-speed interface, the signal acquisition conditioner collects the sampled voltage or current signal, and conditions and filters the signal to convert it into an analog voltage signal that is easy for the digital signal processor to process, the output end of the GPS timing module is connected to the digital signal processor, and the input end of the GPS timing module is connected to the synchronous trigger circuit; this invention only provides a synchronous phasor accurate timing system to ensure that the digital signal processor is not affected by fluctuations of factors such as temperature and realize uniform data acquisition, without considering the influence of system imbalance and noise in the actual on-site environment. Publication (Announcement) No.: CN110568309B, discloses a filter, a synchronous phasor measurement system and a method, wherein the filter described in this invention is an adaptive sliding average filter with a fixed window length; the system comprises a filter and a phase-locked loop control loop, the output end of the phase-locked loop control loop is cascaded with the input end of the filter; the method comprises obtaining a voltage signal output by a smart grid; inputting the voltage signal into the synchronous phasor measurement system, and obtaining a phasor measurement result and a frequency measurement result output by the synchronous phasor measurement system; this invention is implemented by a synchronous phasor measurement algorithm based on a phase-locked loop, using a fixed-length adaptive sliding average filter to filter the ripple output by a type 1 phase-locked loop, and providing a time scale for estimating parameters, and the final measurement results show that the method based on the synchronous phasor measurement system can accurately estimate the grid frequency and phasor under the condition of grid distortion; however, the filter gain parameter update problem and the anti-noise performance of the filter under the dynamic condition of the smart grid are not considered. Publication (Announcement) No.: CN104020352A, discloses a synchronous phasor measurement method applicable to M-type PMU units, including: combining a low-pass digital filter for phasor factors with a discrete Fourier algorithm DFT to calculate the initial measurement phasor after eliminating spectrum leakage; fitting the dynamic phasor using a second-order Taylor series to obtain the measurement error caused by the DFT averaging effect, and compensating the initial phasor based on the error to obtain an accurate dynamic measurement phasor; setting a low-order digital filter for M-type phasor measurement units PMU at the phasor upload, taking the accurate measurement phasor as input, and obtaining the final measurement phasor. By adopting the method disclosed in this invention, phasor measurement can be performed accurately and quickly when the input is a static signal or a dynamic signal; however, this invention can accurately measure low-frequency phasor signals, without considering the influence of high-frequency harmonic distortion signals and nonlinear noise in the power grid. Summary of the invention

[0004] The object of the present invention is to provide a method for processing synchronous phasor information of a power distribution system in an unbalanced and noisy environment, which can improve the estimation accuracy of phasor information during the dynamic operation of the power distribution system in an unbalanced and noisy environment.

[0005] To achieve the above object, the technical solution of the present invention is:

[0006] The present invention provides a method for processing synchronized phasor information of a power distribution system in an unbalanced and noisy environment, characterized in that the method for processing synchronized phasor information of a power distribution system in an unbalanced and noisy environment comprises the following steps:

[0007] S1. Establish a dynamic state estimation model for AC signals in a power distribution system;

[0008] S2, based on the dynamic state estimation model of the AC signal of the power distribution system, the filtered phasor estimation value is calculated by using the recursive filtering method;

[0009] S3, based on the filtered phasor estimation value, using a second-order generalized integrator phase-locked loop to establish mutually orthogonal voltage component signals;

[0010] S4. Based on the output mutually orthogonal signals, the state equation of the continuous nonlinear system is linearized to construct the linearized discrete state equation and observation equation of the extended Kalman filter of the distribution system;

[0011] S5, based on the linearized discrete state equation and observation equation of the extended Kalman filter, the time update process of the error covariance is estimated by adopting the recursive extrapolation method combining estimation and correction and the Riccati equation;

[0012] S6. Based on the time update result of the extended Kalman filter, the estimated value of the state quantity is measured and updated using the Kalman gain coefficient to obtain an optimized estimated value;

[0013] S7: Feedback the optimized estimated value to S2 for the next real-time estimation and correction.

[0014] Furthermore, the power distribution system AC signal dynamic state estimation model in step S1 specifically includes:

[0015] The AC signal x(t) of the power distribution system is:

[0016] x(t)=Xcos(2πft+φ)

[0017] Where f is the rated frequency of the distribution system; X is the synchronous phasor amplitude, t is the sampling time; φ is the angle between the first sampling point and the peak value of the input signal.

[0018] Furthermore, the recursive filtering method in step S2 specifically includes:

[0019] When calculating the phasor of the new acquisition point, there are n-1 sampling points that are exactly the same as the old data window. The old phasor is recursively calculated to obtain the new phasor, which is the recursive update phasor calculation; when the sampling point is n+r, the characteristic e of the fundamental frequency signal -j(n+r) =e -jr , the general expression of recursive filtering is:

[0020]

[0021] In the formula, x n+r is the synchronous phasor measurement signal of the power distribution system at time n+r, r is the information update time, X n+r is the amplitude of the synchronous phasor measurement signal at time n+r, X n is the amplitude of the synchronized phasor measurement signal at time n, where n is the time node.

[0022] Furthermore, the second-order generalized integrator phase-locked loop outputs mutually orthogonal signals in step S3, and a second-order infinite gain bandpass filter is used and a feedback loop is added to establish mutually orthogonal i / q voltage components X i and X q ;Optimize feedback gain coefficient k q , k i Correction X i and X q , orthogonal fundamental i / q component X i (s) and X q The transfer function expression of (s) is as follows:

[0023]

[0024] Wherein, ω is the resonant angular frequency, ω = 2πf.

[0025] Furthermore, in step S4, the continuous nonlinear system state equation is linearized to construct the linearized discrete state equation and observation equation of the extended Kalman filter of the distribution system, which specifically includes the following steps:

[0026] S4-1: Based on the coordinate transformation theory, the synchronous phasor voltage signal is orthogonally decomposed, and the actual measurement signal of the distribution system and the local reference signal generated by the second-order generalized integrator phase-locked loop are subjected to phase rotation and integral zeroing to obtain the input observation vector of the extended Kalman filter method;

[0027] The continuous nonlinear system state equation expression is as follows:

[0028]

[0029] In the formula, xn is the system state vector, u n is the measured input quantity, y n is the system observation quantity, w n-1 and v n is an uncorrelated Gaussian random noise vector with a mean of 0, f(x n-1 ,u n-1 ) is the state transfer matrix of the continuous system, h(x n ) is the continuous system measurement matrix;

[0030] S4-2: When Δx n If it is small enough, for the function f(x n-1 ,u n-1 )exist Expand the Taylor series near , and get the linearized discrete state equation and observation equation expressions as follows:

[0031]

[0032] Where Φ n,n-1 is the discrete system state transfer matrix, H n is the measurement matrix of the discrete system.

[0033] Furthermore, the time updating process of estimating the error covariance by adopting the recursive extrapolation method and the Riccati equation combined with estimation and correction in step S5 specifically includes:

[0034] S5-1: The last moment observation vector y(n-1) estimates the state vector x n Orthogonal projection on y(n-1) The calculation expression is as follows:

[0035]

[0036] In the formula, G n-1 =E(w n-1 w n-1 T ), is the expectation of the noise component of the state vector;

[0037] The recursive extrapolation method combining estimation and correction described above calculates the optimal estimate of the current state based on the latest observation data at each moment. Its recursive update equation expression is as follows:

[0038]

[0039] In the formula, K n is the filter gain matrix, P n-1 is the minimum steady-state error covariance, R n is the noise phasor covariance at time n;

[0040] S5-2: The Riccati equation is used to describe the estimation error covariance, and the prior error covariance matrix P of the constrained state estimate is n│n-1 The expression is as follows:

[0041]

[0042] Where P n│n-1 It's about x t The minimum covariance matrix of the initial condition P(t 0 )=0; when (Φ,E(w n-1 w n-1 T When )) is controllable and (Φ,H) is observable, the Riccati equation converges to a unique semi-positive definite matrix independent of the initial conditions.

[0043] Furthermore, the specific steps of measuring and updating the estimated value of the state quantity using the Kalman gain coefficient in step S6 to obtain the optimized estimated value include:

[0044] S6-1: When bad data appears in the measured value, the bad data includes

[0045] The Kalman gain coefficient K obtained by analytical method is used n Correct and update the system state to obtain the estimated value of the state quantity at time n The Kalman filter gain coefficient k is obtained from the Kalman gain matrix K i and k q , k i and k q The calculation formula is as follows:

[0046]

[0047] Where q is the expected value of process noise component w;

[0048] When the approximate conditions are met When k i and k q for:

[0049]

[0050] In the formula, p f Covariance components, p f =p 12 / r,

[0051] Using P n│n-1 , H n and R n Calculate the gain factor K nfor:

[0052]

[0053] S6-2: Based on the minimum variance of the observation error and in accordance with the principle of state unbiasedness, the prior estimate is corrected using the Kalman gain coefficient K n and the prior error covariance matrix P n-1 , calculate the optimal estimated variance matrix P n for:

[0054] P n =(IK n H n ) nn-1

[0055] S6-3: Using the Kalman gain coefficient K n , given the initial conditions and the input vector x n , the optimal estimate of the state variables is obtained by recursive calculation for:

[0056]

[0057] After collecting new observation data, based on the estimate at the previous moment and using the state transfer equation of the signal process itself, the prediction information is iteratively updated by time and measured by feedback update in the recursive calculation framework of prediction-measurement-correction, and the optimal estimate of the current state is obtained according to the recursive formula.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] (1) In a single-phase or unbalanced three-phase power distribution system, the recursive filtering method is used to reduce the processor's computational burden, reduce the method's computational complexity, and reduce the processor's computational burden.

[0060] (2) By utilizing the orthogonal property of a second-order generalized integrator phase-locked loop, a second-order infinite gain bandpass filter is used and a feedback loop is added to establish an orthogonal voltage component, thereby improving the accuracy of state estimation of the unbalanced three-phase distribution system.

[0061] (3) Taylor series is used to solve the linearization problem of optimal state estimation of nonlinear equations, establish the state space variance of noisy processes, and accurately model the field status of the distribution system.

[0062] (4) The time update equation calculates the current state and the estimated value of the error covariance matrix, and can predict and estimate the system state at the next time.

[0063] (5) The measurement update equation follows the state unbiased principle and uses the prior estimate and filter gain feedback to correct the posterior estimate at the next time.

[0064] (6) In order to adapt to the changes in the statistical characteristics of noise, the expected error covariance iteration process is modeled as the Riccati equation, and the Kalman gain coefficient is analytically solved to improve the accuracy of state estimation.

[0065] (7) In the recursive inference framework of prediction-measurement-correction, the prediction information is corrected by iterative execution of time update and measurement update feedback, without the need to store a large amount of historical data, while taking into account both computing speed and accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0067] Figure 1 It is a flow chart of a method for processing synchronized phasor information of a power distribution system in an unbalanced and noisy environment;

[0068] Figure 2 It is a structural diagram of the node topology of the IEEE 34-node distribution system;

[0069] Figure 3 It is the output diagram of the voltage phasor waveform simulation experiment collected and measured by node 836;

[0070] Figure 4 It is the simulation experiment output diagram of the recursive estimation calculation result of the A phase voltage fundamental wave signal;

[0071] Figure 5 It is the output diagram of the simulation experiment of the recursive estimation amplitude of the fundamental wave of the synchronous phasor voltage phase A;

[0072] Figure 6 It is the simulation experiment output diagram of the recursive estimation amplitude of the fifth harmonic of the synchronous phasor voltage A phase voltage;

[0073] Figure 7 It is the simulation experiment output diagram of the voltage phasor information processing result in the normal operating state;

[0074] Figure 8 It is the simulation experiment output diagram of the voltage phasor information processing result of the three-phase unbalanced state;

[0075] Fig. 9 It is the simulation experiment output diagram of the voltage phasor information processing result in a noise-free environment;

[0076] Fig.10 This is the simulation experiment output diagram of the voltage phasor information processing results in an environment with a signal-to-noise ratio of 10dB;

[0077] Fig.11 This is the output diagram of the simulation experiment of voltage estimation error based on the extended Kalman filter method;

[0078] Fig.12 This is the output diagram of the simulation experiment of voltage estimation error based on the extended Kalman filter method in a noise-free environment;

[0079] Fig.13 This is the output diagram of the simulation experiment of the voltage estimation error based on the extended Kalman filter method in an environment with a signal-to-noise ratio of 15dB;

[0080] Fig.14 It is a comparison output diagram of simulation experiments of different information processing methods under three-phase unbalanced state;

[0081] Fig.15 This is the output diagram of the simulation experiment of the voltage estimation error value based on the improved Kalman filter method in a 20dB noise environment;

[0082] Fig.16 It is the voltage phasor information estimation error based on the improved Kalman filter method under different noise environments. DETAILED DESCRIPTION

[0083] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only embodiments of a part of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention. Unless otherwise specified, the technical means used in the embodiments are conventional means well known to those skilled in the art.

[0084] The embodiment of the present invention provides a method for processing synchronized phasor information of a power distribution system in an unbalanced and noisy environment, wherein the original measurement parameters are measured by the master station of the power distribution system using a PMU device to measure the bus voltage and current information in real time, and the measurement parameters include amplitude, frequency and phase; the dispatching control center uses a cascade filter module to combine the PMU measurement values ​​of different geographical locations in a unified time coordinate system to obtain synchronized phasor information of the power distribution system The above filtering process is performed through the cascaded filtering module, thereby improving the accuracy of the measurement information.

[0085] like Figure 1 As shown, Figure 1A flowchart of a method for processing synchronized phasor information of a power distribution system in an unbalanced and noisy environment is provided in an embodiment of the present invention. The method comprises the following steps:

[0086] S1. Establish a dynamic state estimation model for AC signals in a power distribution system;

[0087] S2, based on the dynamic state estimation model of the AC signal of the power distribution system, the filtered phasor estimation value is calculated by using the recursive filtering method;

[0088] S3, based on the filtered phasor estimation value, using a second-order generalized integrator phase-locked loop to establish mutually orthogonal voltage component signals;

[0089] S4. Based on the output mutually orthogonal signals, the state equation of the continuous nonlinear system is linearized to construct the linearized discrete state equation and observation equation of the extended Kalman filter of the distribution system;

[0090] S5, based on the linearized discrete state equation and observation equation of the extended Kalman filter, the time update process of the error covariance is estimated by adopting the recursive extrapolation method combining estimation and correction and the Riccati equation;

[0091] S6. Based on the time update result of the extended Kalman filter, the estimated value of the state quantity is measured and updated using the Kalman gain coefficient to obtain an optimized estimated value;

[0092] S7: Feedback the optimized estimated value to S2 for the next real-time estimation and correction.

[0093] Specifically, the power distribution system AC signal dynamic state estimation model in step S1 specifically includes:

[0094] The AC signal x(t) of the power distribution system is:

[0095] x(t)=Xcos(2πft+φ)

[0096] Where f is the rated frequency of the distribution system; X is the synchronous phasor amplitude, t is the sampling time; φ is the angle between the first sampling point and the peak value of the input signal.

[0097] Specifically, the recursive filtering method in step S2 specifically includes:

[0098] When calculating the phasor of the new acquisition point, there are n-1 sampling points that are exactly the same as the old data window. The old phasor is recursively calculated to obtain the new phasor, which is the recursive update phasor calculation; when the sampling point is n+r, the characteristic e of the fundamental frequency signal -j(n+r) =e -jr , the general expression of recursive filtering is:

[0099]

[0100] In the formula, x n+r is the synchronous phasor measurement signal of the power distribution system at time n+r, r is the information update time, X n+r is the amplitude of the synchronous phasor measurement signal at time n+r, X n is the amplitude of the synchronized phasor measurement signal at time n, where n is the time node.

[0101] Specifically, the second-order generalized integrator phase-locked loop outputs mutually orthogonal signals in step S3, and a second-order infinite gain bandpass filter is used and a feedback loop is added to establish mutually orthogonal i / q voltage components X i and X q ; Effectively reduce the feedback gain coefficient k of the initial offset q , k i Correction X i and X q , orthogonal fundamental i / q component X i (s) and X q The transfer function expression of (s) is as follows:

[0102]

[0103] Wherein, ω is the resonant angular frequency, ω = 2πf.

[0104] Specifically, in step S4, the linearization process of the continuous nonlinear system state equation and the construction of the linearized discrete state equation and observation equation of the extended Kalman filter of the power distribution system specifically include the following steps:

[0105] S4-1: Based on the coordinate transformation theory, the synchronous phasor voltage signal is orthogonally decomposed, and the actual measurement signal of the distribution system and the local reference signal generated by the second-order generalized integrator phase-locked loop are subjected to phase rotation and integral zeroing to obtain the input observation vector of the extended Kalman filter method;

[0106] The continuous nonlinear system state equation expression is as follows:

[0107]

[0108] In the formula, x n is the system state vector, u n is the measured input quantity, y n is the system observation quantity, w n-1 and v n is an uncorrelated Gaussian random noise vector with a mean of 0, f(x n-1 ,u n-1 ) is the state transfer matrix of the continuous system, h(x n ) is the continuous system measurement matrix;

[0109] S4-2: When Δx n If it is small enough, for the function f(x n-1 ,u n-1 )exist Expand the Taylor series near , and get the linearized discrete state equation and observation equation expressions as follows:

[0110]

[0111] Where Φ n,n-1 is the discrete system state transfer matrix, H n is the measurement matrix of the discrete system.

[0112] Specifically, the time updating process of estimating the error covariance by adopting the recursive extrapolation method combining estimation and correction and the Riccati equation in step S5 specifically includes:

[0113] S5-1: The last moment observation vector y(n-1) estimates the state vector x n Orthogonal projection on y(n-1) The calculation expression is as follows:

[0114]

[0115] In the formula, G n-1 =E(w n-1 w n-1 T ), is the expectation of the noise component of the state vector;

[0116] The recursive extrapolation method combining estimation and correction described above calculates the optimal estimate of the current state based on the latest observation data at each moment. Its recursive update equation expression is as follows:

[0117]

[0118] In the formula, K n is the filter gain matrix, P n-1 is the minimum steady-state error covariance, R n is the noise phasor covariance at time n;

[0119] S5-2: The Riccati equation is used to describe the estimation error covariance, and the prior error covariance matrix P of the constrained state estimate is n│n-1 The expression is as follows:

[0120]

[0121] Where P n│n-1 It's about x tThe minimum covariance matrix of the initial condition P(t 0 )=0; when (Φ,E(w n-1 w n-1 T When )) is controllable and (Φ,H) is observable, the Riccati equation converges to a unique semi-positive definite matrix independent of the initial conditions.

[0122] Specifically, the specific steps of measuring and updating the estimated value of the state quantity using the Kalman gain coefficient in step S6 to obtain the optimized estimated value include:

[0123] S6-1: When bad data appears in the measurement value, the Kalman gain coefficient K is obtained by analytical method. n Correct and update the system state, effectively suppress the influence of noise, and obtain the estimated value of the state quantity at time n The Kalman filter gain coefficient k is obtained from the Kalman gain matrix K i and k q , k i and k q The calculation formula is as follows:

[0124]

[0125] Where q is the expected value of process noise component w;

[0126] When the approximate conditions are met When k i and k q for:

[0127]

[0128] In the formula, p f Covariance components, p f =p 12 / r,

[0129] Using P n│n-1 , H n and R n Calculate the gain factor K n for:

[0130]

[0131] S6-2: Based on the minimum variance of the observation error and in accordance with the principle of state unbiasedness, the prior estimate is corrected using the Kalman gain coefficient K n and the prior error covariance matrix P n-1 , calculate the optimal estimated variance matrix P n for:

[0132] Pn =(IK n H n ) nn-1

[0133] S6-3: Using the Kalman gain coefficient K n , given the initial conditions and the input vector x n , the optimal estimate of the state variables is obtained by recursive calculation for:

[0134]

[0135] After collecting new observation data, based on the estimate at the previous moment and using the state transfer equation of the signal process itself, the prediction information is iteratively updated by time and measured by feedback update in the recursive calculation framework of prediction-measurement-correction, and the optimal estimate of the current state is obtained according to the recursive formula.

[0136] Example

[0137] like Figure 2 As shown, the structure diagram of the node topology of the IEEE 34-node distribution system is shown; in order to analyze the effectiveness of the information processing method based on the extended Kalman filter, the IEEE 34-node system is used as a test platform to perform information processing analysis on the synchronous voltage phasor information of the IEEE 34-node distribution system, where the sampling frequency is 12.8kHz and the fundamental frequency is 50Hz. In the IEEE34-node distribution system, the voltage phasor information of node 836 is observed based on the extended Kalman filter method, and the update process is used to observe the distribution system state model, where, in the case of high-order harmonic distortion, the state variables include fundamental, third harmonic, fifth harmonic and seventh harmonic components.

[0138] The main advantage of the extended Kalman filter information processing method lies in the response under distortion conditions, because it can be easily extended with additional states to solve the state estimation problem of the fundamental component and higher harmonic components. The extended Kalman filter method can not only process the fundamental component, but also be extended to observe the harmonic components of the voltage phasor signal, which improves the estimation performance of the method. The computational cost of the harmonic component includes two additional state quantities for each harmonic and one additional state quantity for each DC component.

[0139] In order to analyze the effectiveness of the synchronous phasor information processing method of the distribution system under unbalanced and noise interference environment, the information processing analysis of the synchronous voltage phasor information of the IEEE34-node distribution system is carried out, which includes a generator in Arizona, USA connected to the distribution system; two distributed power sources, photovoltaic power generation and wind power generation, respectively, and the initial output state of each distributed energy source is at the stable operation point [i]; 10 loads L1-L10, with a power consumption of 0.5MW+0.1Mvar; 1 transformer and 33 AC lines. The sampling frequency is 12.8kHz and the fundamental frequency is 50Hz.

[0140] Since the PMU device collects distribution system information from a global perspective, the measurement process and short-circuit fault process bring measurement noise interference and harmonic oscillation respectively, forming a nonlinear distribution system with strong transients, low signal-to-noise ratio and high-order harmonic distortion environment. Aiming at the accuracy problem of synchronous phasor measurement under low signal-to-noise ratio and high-order harmonic distortion conditions, a quasi-optimal estimation information processing method based on extended Kalman filtering is used to achieve accurate estimation of voltage phasor information under fault conditions.

[0141] When the distribution system fails or is disturbed, the electromagnetic quantities of each power component and the mechanical quantities such as the speed of the rotating motor will change. At this time, the distribution system will experience complex electromechanical transient processes, including phase fluctuations, noise interference, and a large number of high-order harmonics. In order to accurately estimate the synchronized phasor information in a low signal-to-noise ratio and high-order harmonic distortion environment, the state equation and observation equation of the nonlinear discrete-time system are linearized, and synchronized phasor information processing based on the extended Kalman filter is performed under the prediction-measurement-correction recursive framework.

[0142] In the IEEE 34-node distribution system, the voltage phasor information of node 836 is observed based on the improved Kalman filter method, and the time update and measurement update equations are used to observe the distribution system state. In the case of high-order harmonic distortion, the state variables include the fundamental, third harmonic, fifth harmonic and seventh harmonic components.

[0143] In the prediction-measurement-correction recursive framework, the state to be estimated is recursively deduced by feedback. First, the Kalman filter method is extended to estimate the system state at a certain moment, and then feedback is obtained by considering the noisy measurement variables.

[0144] The extended Kalman filter method consists of two parts: the time update equation and the measurement update equation. The time update equation part mainly plays a predictive estimation role. It is responsible for calculating the estimated value of the current state and the error covariance matrix, and proposing a predictive estimate for the next time state. The measurement update equation part mainly plays a correction role. It is responsible for feedback, combining the prior estimate with the new measurement value, and proposing a corrected posterior estimate for the next time state.

[0145] like Figure 3 As shown in the simulation experiment, the voltage phasor signal measurement value U based on the extended Kalman filter method is a ,U b ,U c , shows the voltage phasor waveform collected and measured in node 836.

[0146] like Figure 4 As shown in the simulation experiment, the voltage i / q component estimation value U based on the extended Kalman filter method is i / Uq shows the recursive estimation calculation results of the A-phase voltage fundamental signal based on the extended Kalman filter information processing method.

[0147] like Figure 5 As shown, the harmonic analysis of the synchronous phasor voltage waveform is performed, and the amplitude U of the fundamental wave of the phase A voltage is recursively estimated. i .

[0148] like Figure 6 As shown, the harmonic analysis of the synchronous phasor voltage waveform is performed, and the amplitude U of the fifth harmonic of the A phase voltage is recursively estimated. i .

[0149] from Figure 3 , Figure 4 , Figure 5 and Figure 6 It can be seen from the simulation result output diagram that under complex working conditions, the information processing method based on the extended Kalman filter can still accurately extract the steady-state component of the PMU measurement value, capture the response of the distribution system on different time scales, and accurately feed back the steady-state component to the diagnostic application.

[0150] In order to verify the effectiveness of the information processing method based on extended Kalman filtering under complex working conditions of the distribution system, the results of synchronized phasor information processing based on the extended Kalman filtering method in node 888 are analyzed.

[0151] like Figure 7 As shown in the figure, the simulation experiment output diagram of the voltage phasor information processing result in normal operation; the node voltage phasor diagram in the normal operation distribution system is composed of phasors with a phase difference of 120°, and its amplitude is the phase voltage amplitude of 480V. The extended Kalman filter method is used to estimate the node voltage i / q components as U ai =479.985V, U aq =477.569V. The motion trajectories of the three-phase voltage components of the node voltage on the i / q plane are all circles, and their radii are R a =0.994, R b =0.994, R c=0.994, and moves counterclockwise over time.

[0152] like Figure 8 As shown in the figure, the simulation experiment output diagram of the voltage phasor information processing results under the three-phase unbalanced state; in order to verify the effectiveness of the information processing method based on the extended Kalman filter under the condition of three-phase unbalanced distribution system, the voltage of phase A suddenly changes to 80% of the original value at 1s, and the voltages of phases B and C remain unchanged. At this time, the node voltage is seriously unbalanced. The measured voltage amplitudes of phase A and phase B are U a =261.279V, U b =363.678V. The estimated values ​​based on the extended Kalman filter method are U a-ekf =259.932V and U b-ekf =361.853V. At this time, the i / q components of the voltage phasor of node 888 observed based on the extended Kalman filter method are U ai =259.901V, U aq =59.964V; U bi =361.992V, U bq =361.715V. The three-phase voltage components of the node voltage move in circles of different radii on the i / q plane, and their radii are R a =0.487, R b =0.995, R c =0.995, and moves counterclockwise over time.

[0153] like Fig. 9 As shown in the figure, the simulation experiment output diagram of the voltage phasor information processing results in a noise-free environment is the estimation result of the state estimation of the voltage phasor information of node 888 based on the synchronous phasor information processing method of the extended Kalman filter. The voltage measurement amplitudes of phase A and phase B measured by the PMU device are U a =367.652V. The estimated values ​​based on the extended Kalman filter method are U a-ekf =367.942V. At this time, the i / q components of the voltage phasor of node 888 observed based on the extended Kalman filter method are U ai =366.917V, U aq =368.964V. The three-phase voltage components of the node voltage move in circles of different radii on the i / q plane, with radii Ra=0.999, and move counterclockwise over time. As the signal-to-noise ratio decreases, it can be seen that the estimation error based on the extended Kalman filter method increases. When SNR=10dB, the estimation error is the largest, as shown in Fig.10 As shown, the signal-to-noise ratio is 10dB. At this time, the estimated values ​​based on the extended Kalman filter method are U a-ekf=368.965V. At this time, the i / q components of the voltage phasor of node 888 observed based on the extended Kalman filter method are U ai =367.263V, U aq =370.659V; estimated error e ai =0.001%, e aq =0.008%.

[0154] The simulation results show that the use of the extended Kalman filter method to update the value of the measurement noise covariance matrix can filter out dynamic and bad data from the original PMU data. In the case of the corrected extended Kalman filter method, it will be updated according to the measurement results to filter out bad data and extract the steady-state component, which has a certain compensation effect on the phase angle measurement error of the PMU, and the extracted fundamental component has a higher accuracy.

[0155] like Fig.11 As shown in the figure, the simulation experiment output diagram of the voltage estimation error based on the extended Kalman filter method, under different noise environments, according to the voltage phasor information at the exit of the 888 node, the extended Kalman filter method is used to analyze and process the voltage phasor information of the voltage source, and the estimation error is as follows Fig.11 As shown. The voltage phasor information estimation error based on the extended Kalman filter method under different noise environments is shown, where the center line of each box represents the median error, the top and bottom of the box are the 75% quantile and the 2% quantile, and the cross symbol is the outlier. The average error e of the extended Kalman filter phasor estimation method under different noise environments of 40dB, 30dB, 20dB, 15dB, and 10dB is shown. EKF They are 0.010%, 0.032%, 0.123%, 0.182% and 0.324% respectively. The relative error of the extended Kalman filter phasor estimation method varies in the range of [0, 0.91%], and the relative error is small.

[0156] In order to verify the accuracy of the synchronized phasor information processing method based on extended Kalman filter, the extended Kalman filter method and the second-order generalized integrator (SOGI) method were used to estimate the voltage phasor signal in different noise environments, and their estimation accuracies were compared.

[0157] like Fig.12As shown in the figure, the simulation experiment output diagram of the voltage estimation error based on the extended Kalman filter method in a noise-free environment; the A-phase short circuit fault occurred in the distribution branch 888-890 at 2s. The information processing method based on the extended Kalman filter method can quickly and accurately estimate the steady-state system after the fault. The estimated state curve is in good agreement with the actual measurement curve, with an estimated error of 0.0283% and a convergence time of 0.017s; the estimated curve based on the SOGI method deviates significantly from the actual measurement curve before and after the fault, with an estimated error of 0.0907% and a convergence time of 0.031s.

[0158] like Fig.13 As shown, the simulation experiment output diagram of voltage estimation error based on the extended Kalman filter method in an environment with a signal-to-noise ratio of 15dB;

[0159] like Fig.14 As shown in the figure, the simulation experiment comparison output diagram of different information processing methods shows that under different signal-to-noise ratio conditions, the synchronous phasor information processing method based on the extended Kalman filter can quickly and accurately estimate the voltage phasor signal, and the estimated curve obtained by the SOGI method is closer to the actual measured waveform signal. The simulation results show that the estimation effect of the information processing method based on the extended Kalman filter is significantly better than that of the SOGI method.

[0160] like Fig.15 As shown in the figure, the simulation experiment output diagram of the voltage estimation error value based on the improved Kalman filter method in a 20dB noise environment is analyzed; the state estimation results of the voltage phasor information of node 888 based on the synchronous phasor information processing method of the improved Kalman filter in a 20dB noise environment are analyzed. The estimated value of the voltage amplitude measured by the PMU device is 367.942V, and the estimated error range is 0.001-0.008%. The simulation results show that the dynamic bad data can be filtered out from the original PMU data by updating the value of the measurement noise covariance matrix using the improved Kalman filter method. The time update equation calculates the estimated value of the current state and the error covariance matrix, and can predict and estimate the system state at the next time. The measurement update equation follows the state unbiased principle and uses the prior estimate and the filter gain feedback to correct the posterior estimate of the next time. In the case of correcting the improved Kalman filter method, it will be updated according to the measurement results to filter out the bad data and extract the steady-state component, taking into account both the calculation speed and accuracy, and the extracted fundamental component has a high accuracy.

[0161] like Fig.16As shown in the figure, under different noise environments, the voltage phasor information estimation error based on the improved Kalman filter method has a higher state estimation accuracy in a low noise environment, and the average voltage amplitude estimation error range is 0.010-0.123%; in a low noise environment, the state estimation error increases, and the average estimation error range is 0.182-0.324%, and the overall estimation accuracy is high. This method models the expected error covariance iteration process as the Riccati equation, and analytically solves the Kalman gain coefficient, which adapts to the changes in the statistical characteristics of noise and improves the state estimation accuracy.

[0162] In the recursive calculation framework of prediction-measurement-correction, this method iteratively executes time update and measurement update feedback to correct the prediction information, uses coordinate transformation theory to orthogonally decompose the input signal to achieve phase locking, and then uses a cascade filtering module to eliminate the influence of harmonic components on the phase-locked result. There is no need to store a large amount of historical data, and both calculation speed and accuracy are taken into account to achieve online signal processing of synchronous phasors in voltage unbalanced and noisy environments.

[0163] According to the simulation experiment results and the comparative experiment results, it can be seen that the synchronized phasor information processing method based on the extended Kalman filter processes the synchronized phasor information in the low signal-to-noise ratio and high-order harmonic distortion environment and eliminates the bad data. The coordinate transformation theory is used to orthogonally decompose the input signal to achieve phase locking, and then the cascade filtering module is used to eliminate the influence of the harmonic component on the phase locking result. The average estimation error is 0.0283%, and the convergence time is 0.017s. The fundamental voltage signal is estimated quickly and accurately under fault conditions, and the accuracy problem of synchronized phasor measurement under the conditions of user-grid supply and demand interaction and voltage imbalance and harmonics in the distribution system is solved.

[0164] The above is a detailed description of an embodiment of the present invention, but the content is only a preferred embodiment of the present invention and cannot be considered to limit the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the present invention should still fall within the scope of the patent coverage of the present invention.

Claims

1. A method for processing synchronous phasor information of a power distribution system in an unbalanced and noisy environment, characterized in that: The synchronized phasor information processing method comprises the following steps: S1. Establish a dynamic state estimation model for AC signals in a power distribution system; S2, based on the dynamic state estimation model of the AC signal of the power distribution system, the filtered phasor estimation value is calculated by using the recursive filtering method; S3, based on the filtered phasor estimate, using the mutually orthogonal signals output by the second-order generalized integrator phase-locked loop; S4. Based on the output mutually orthogonal signals, the state equation of the continuous nonlinear system is linearized to construct the linearized discrete state equation and observation equation of the extended Kalman filter of the distribution system; S5, based on the linearized discrete state equation and observation equation of the extended Kalman filter, the time update process of the error covariance is estimated by adopting the recursive extrapolation method combining estimation and correction and the Riccati equation; S6. Based on the time update result of the extended Kalman filter, the estimated value of the state quantity is measured and updated using the Kalman gain coefficient to obtain an optimized estimated value; S7: Feedback the optimized estimated value to S2 for the next real-time estimation and correction.

2. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The dynamic state estimation model of the AC signal of the power distribution system in step S1 is specifically in the following form: The AC signal x(t) of the power distribution system is: x(t)=Xcos(2πft+φ) Where f is the rated frequency of the distribution system, X is the synchronous phasor amplitude, t is the sampling time, and φ is the angle between the first sampling point and the peak value of the input signal.

3. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The recursive filtering method in step S2 specifically includes: When calculating the phasor of the new acquisition point, there are n-1 sampling points that are exactly the same as the old data window. The old phasor is recursively calculated to obtain the new phasor, which is the recursive update phasor calculation; when the sampling point is n+r, the characteristic e of the fundamental frequency signal -j(n+r) =e -jr , the general expression of recursive filtering is: In the formula, x n+r is the synchronous phasor measurement signal of the power distribution system at time n+r, r is the information update time, X n+r is the amplitude of the synchronous phasor measurement signal at time n+r, X n is the amplitude of the synchronized phasor measurement signal at time n, where n is the time node.

4. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The second-order generalized integrator phase-locked loop in step S3 outputs mutually orthogonal signals, uses a second-order infinite gain bandpass filter and adds a feedback loop to establish mutually orthogonal i / q voltage components X i and X q ;Optimize feedback gain coefficient k q , k i Correction X i and X q , orthogonal fundamental i / q component X i (s) and X q The transfer function expression of (s) is as follows: Wherein, ω is the resonant angular frequency, ω = 2πf.

5. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The linearization of the continuous nonlinear system state equation in step S4 to construct the linearized discrete state equation and observation equation of the extended Kalman filter of the power distribution system specifically includes the following steps: S4-1: Based on the coordinate transformation theory, the synchronous phasor voltage signal is orthogonally decomposed, and the actual measurement signal of the distribution system and the local reference signal generated by the second-order generalized integrator phase-locked loop are subjected to phase rotation and integral zeroing to obtain the input observation vector of the extended Kalman filter method; The state equation of the continuous nonlinear system is expressed as follows: In the formula, x n is the system state vector, u n is the measured input quantity, y n is the system observation quantity, w n-1 and v n is an uncorrelated Gaussian random noise vector with a mean of 0, f(x n-1 ,u n-1 ) is the state transfer matrix of the continuous system, h(x n ) is the continuous system measurement matrix; S4-2: When Δx n If it is small enough, for the function f(x n-1 ,u n-1 )exist Expand the Taylor series near , and get the linearized discrete state equation and observation equation expressions as follows: Where Φ n,n-1 is the discrete system state transfer matrix, H n is the measurement matrix of the discrete system.

6. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The specific steps of the time updating process of estimating the error covariance by adopting the recursive extrapolation method combining estimation and correction and the Riccati equation in step S5 include: S5-1: Estimate the state vector x based on the observation vector y(n-1) at the previous moment n Orthogonal projection on y(n-1) The calculation expression is as follows: In the formula, G n-1 =E(w n-1 w n-1 T ), is the expectation of the noise component of the state vector; The recursive calculation method combining estimation and correction calculates the optimal estimate of the current state based on the latest observation data at each moment. Its recursive update equation expression is as follows: In the formula, K n is the filter gain matrix, P n-1 is the minimum steady-state error covariance, R n is the noise phasor covariance at time n; S5-2: The Riccati equation is used to describe the estimation error covariance, and the prior error covariance matrix P of the constrained state estimate is n│n-1 The expression is as follows: Where P n│n-1 It's about x t The minimum covariance matrix of the initial condition P(t0) = 0; when (Φ, E(w n-1 w n-1 T )) is controllable and (Φ,H) is observable, the Riccati equation converges to a unique semi-positive definite matrix independent of the initial conditions.

7. The method for processing synchronous phasor information of a power distribution system in an unbalanced and noise interference environment according to claim 1, characterized in that: The specific steps of measuring and updating the estimated value of the state quantity by using the Kalman gain coefficient to obtain the optimized estimated value in step S6 include: S6-1: When bad data appears in the measured value, the Kalman gain coefficient K n Correct and update the system state to obtain the estimated value of the state quantity at time n The Kalman filter gain coefficient k is obtained from the Kalman gain matrix K i and k q , k i and k q The calculation formula is as follows: Where q is the expected value of process noise component w; When the approximate conditions are met When k i and k q for: In the formula, p f Covariance components, p f =p 12 / r, Using P n│n-1 , H n and R n Calculate the gain factor K n for: S6-2: Based on the minimum variance of the observation error and in accordance with the principle of state unbiasedness, the prior estimate is corrected using the Kalman gain coefficient K n and the prior error covariance matrix P n-1 , calculate the optimal estimated variance matrix P n for: P n =(I-K n H n )P nn-1 S6-3: Using the Kalman gain coefficient K n , given the initial conditions and the input vector x n , the optimal estimate of the state variables is obtained by recursive calculation for: After collecting new observation data, based on the estimate at the previous moment and using the state transfer equation of the signal process itself, the prediction information is iteratively updated by time and measured by feedback update in the recursive calculation framework of prediction-measurement-correction, and the optimal estimate of the current state is obtained according to the recursive formula.

Images