A method for analyzing regional distribution network status

Abstract

The present invention discloses a method for analyzing the state of a regional distribution network, comprising: forming a high-dimensional matrix from the measurement data of a synchronous measurement device of a regional distribution network according to sampling time and different nodes; normalizing the high-dimensional matrix and obtaining its covariance matrix; obtaining a state analysis index based on the measurement data according to the measurement data, and obtaining a state analysis index characterizing the operating state of the distribution network according to the state analysis index based on the measurement data; combining the state analysis index with a sliding time window, dynamically analyzing the state of the distribution network, and drawing the operation trajectory of the power grid in real time; analyzing the time complexity of the state analysis index; and using the basic data of the actual distribution network operation for simulation verification. The present invention directly analyzes the operating data of the distribution network and extracts its operating state characteristics, which can avoid a series of problems caused by inaccurate models and large amounts of calculations, and can achieve dynamic evaluation to a certain extent.

Description

Technical Field

[0001] The invention belongs to the technical field of power systems, and in particular relates to a method for analyzing the state of a regional distribution network. Background Art

[0002] With the construction and implementation of the power Internet of Things, all aspects of the power grid are showing a development trend of deep interaction and full state perception, and the power system has fully entered the era of big data. The measurement big data on the power grid side, which is mainly sourced from synchronous measurement data, has brought greater challenges to the real-time dynamic monitoring and intelligent operation and maintenance of the power grid. Therefore, from the perspective of measurement data drive, extracting the change characteristics of the operation data of large-scale and complex power grids and mapping them to the evolution characteristics of the operation status of the power grid is conducive to real-time dynamic effective early warning and evaluation in the stage of slow deterioration of the power grid status, which plays an important role in the active safety management and intelligent operation and maintenance of the power grid.

[0003] Effective power grid operation status analysis is the basis for realizing intelligent power grid operation and maintenance. At present, the research on distribution network operation status analysis can be divided into two types of methods: mechanism-driven and data-driven. The mechanism-based method is often a prevention and control mode of "simulation modeling + expected fault set". However, the mechanism-based method relies heavily on model accuracy and expert experience, and its accuracy cannot be guaranteed; at the same time, when faced with massive data generated by large-scale and complex power grids, the calculation amount is large and the real-time performance is poor. It can often only be used for offline analysis and cannot meet the requirements of dynamic monitoring and real-time defense. More classic research methods also include the analytic hierarchy process (AHP), fuzzy comprehensive evaluation method, and principal component analysis (PCA). These methods can evaluate the power grid from the perspectives of safety, economy, reliability, adaptability, and quality, but they mainly solve the problems of recent construction planning and transformation of distribution networks, and cannot evaluate the real-time status of the power grid and support intelligent operation and maintenance of distribution networks. Among the data-driven methods, there are also methods for self-learning the characteristics of the power grid status based on deep learning, which have fast calculation speed and strong robustness, but the method itself has high requirements for source data and low comprehensibility.

[0004] In addition, the linear power grid state analysis method based on the big data dimensionality reduction method can perceive events in advance and issue timely warnings when the power grid state deteriorates; the voltage stability analysis method based on singular value decomposition can depict the power grid operation trajectory in real time and has good practicality. However, when faced with large amounts of data and rich information in measurement big data, this method often uses feature extraction or dimensionality reduction methods to reduce the computational complexity, thereby losing some of the original data information and may not meet the requirements of real-time evaluation. The application of random matrix theory (RMT) can avoid information loss caused by dimensionality reduction or feature extraction. However, the evaluation of the power grid state is only from the perspective of the probability distribution of the data, and the evaluation indicators are relatively single, the mapping degree of the power grid state is insufficient, and there is a lack of analysis of the evolution direction of the power grid state. Summary of the invention

[0005] The technical problem to be solved by the present invention is to provide a method for analyzing the state of a regional distribution network, which directly analyzes the power grid operation data based on the identification of data change characteristics to avoid a series of problems caused by inaccurate models and large amounts of calculation.

[0006] In order to solve the above technical problems, the present invention provides a method for analyzing the state of a regional distribution network, comprising:

[0007] Step S1, the measurement data of the regional distribution network synchronous measurement device is formed into a high-dimensional matrix according to the sampling time and different nodes;

[0008] Step S2, normalizing the high-dimensional matrix and obtaining its covariance matrix;

[0009] Step S3, obtaining a state analysis indicator based on the measurement data according to the measurement data, and obtaining a state analysis indicator characterizing the operating state of the distribution network according to the state analysis indicator based on the measurement data;

[0010] Step S4, combining the state analysis index with the sliding time window, dynamically analyzing the state of the distribution network, and drawing the operation trajectory of the power grid in real time;

[0011] Step S5, analyzing the time complexity of the state analysis indicator;

[0012] Step S6, using basic data of actual distribution network operation to perform simulation verification, analyzing the distribution network operation status, observing the evolution direction of the distribution network status, and verifying the effectiveness of the status analysis index.

[0013] Furthermore, the specific method of forming a high-dimensional matrix in step S1 is:

[0014] Get the time vector y(t i ):

[0015]

[0016] Among them, the time vector y(t i ) is the total number of measurement points at sampling time t i Generate a state variable, l is the number of nodes equipped with synchronous measurement devices;

[0017]

[0018] Wherein, n is the number of sampling times, and the high-dimensional matrix is ​​the result of sampling n times, and is constructed by concatenating all time vectors.

[0019] Furthermore, in step S2, the high-dimensional matrix is ​​normalized and its covariance matrix is ​​obtained in the following specific manner:

[0020]

[0021] i=1,2,...l

[0022] Among them, μ(Y i ) and σ(Y i ) represent Y i The mean and variance of Y i is the i-th row of Y;

[0023]

[0024] Among them, S n represents the sample covariance matrix, Y l×n The complex conjugate transpose of .

[0025] Furthermore, in step S3, the state analysis index based on the measured data specifically includes: a measurement probability distribution index I based on the ring rate; 1 , Uniformity Index I 2 .

[0026] Furthermore, the measurement probability distribution index I based on the ring rate 1 The specific method of obtaining is:

[0027] I 1 =MSR(S n )

[0028]

[0029] Among them, I 1 It is an indicator based on the circular ring law and MSR to characterize the probability distribution of measurement data; MSR is the singular eigenvalue λ i The mean of

[0030] Uniformity index I involving distribution network changes2 The specific method of obtaining is:

[0031]

[0032]

[0033]

[0034] Among them, I 2 An index representing the uniformity of the distribution network structure and the rate of change of measurement data, i = 1, 2…l; W is S n The right eigenvector matrix, λ k For S n The kth eigenvalue of k is k The corresponding right eigenvector; L i For S i Contribution to the anomaly; node degree m of node i i is the total number of nodes connected to node i; D(·) represents the variance treatment.

[0035] Furthermore, in step S3, the specific method of obtaining the state analysis index characterizing the operating state of the distribution network according to the state analysis index based on the measurement data is:

[0036]

[0037] Furthermore, in step S4, the specific method of drawing the distribution network operation trajectory in real time is:

[0038] Assume n is the width of the time window. When the grid structure is determined, the value of n is determined by c=l / n∈(0,1). Let the step size of the sliding time window be 1. Then, among the n columns of sample values, there is 1 column of current data and n-1 columns of historical data.

[0039] Furthermore, the specific method of analyzing the time complexity in step S5 is:

[0040] Assuming there are p branches and n nodes in the power grid, the time complexity analysis and calculation steps are as follows:

[0041] Step S51, normalizing the data in the time window, with a time complexity of O(n);

[0042] Step S52, normalizing the covariance matrix, with a time complexity of O(n);

[0043] Step S53, singularizing the matrix;

[0044] Step S54, calculating the eigenvalue and the mean I1 of the eigenvalue, the time complexity is O(1);

[0045] Step S55, calculate I2, the time complexity is O(n);

[0046] Step S56, calculate I3, the time complexity is O(1).

[0047] Furthermore, the final time complexity is O(n)+O(1).

[0048] Furthermore, the specific method of simulating and verifying the basic data of the actual distribution network operation in step S6 is: based on the PSASP software, simulation verification is performed using the basic data of the actual distribution network, the distribution network operation status is analyzed under each calculation example, the evolution direction of the distribution network status is observed, and the validity of the status analysis indicators is verified.

[0049] The implementation of the present invention has the following beneficial effects: the present invention directly analyzes the distribution network operation data and extracts its operation status characteristics, which can avoid a series of problems caused by inaccurate models and large calculations, and can achieve dynamic evaluation to a certain extent. The distribution network state defined in the present invention characterizes the system's ability to withstand disturbances, reflects the dynamic performance and stability level of the power system, and can be analyzed from two aspects: real-time state evaluation and evolution direction, which is the basis for realizing intelligent operation and maintenance of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative work.

[0051] Figure 1 It is a flow chart of a method for analyzing the state of a regional distribution network according to an embodiment of the present invention.

[0052] Figure 2a -c is a schematic diagram of the distribution network state evolution trajectory results under various indicators in the embodiment of the present invention, where: Figure 2a Is the measurement probability distribution index I 1 Schematic diagram of motion trajectory, Figure 2b Is the uniformity index I 2 Schematic diagram of motion trajectory, Figure 2c It is the distribution network operation status index I 3 Schematic diagram of motion trajectory. DETAILED DESCRIPTION

[0053] The following descriptions of the embodiments refer to the accompanying drawings to illustrate specific embodiments in which the present invention may be implemented.

[0054] The data-based power grid state analysis method directly analyzes the power grid operation data and extracts its operation state characteristics, which can avoid a series of problems caused by inaccurate models and large calculations, and can achieve dynamic evaluation to a certain extent. Figure 1 As shown, an embodiment of the present invention provides a method for analyzing the state of a regional distribution network, comprising:

[0055] Step S1, the measurement data of the regional distribution network synchronous measurement device is formed into a high-dimensional matrix according to the sampling time and different nodes;

[0056] Step S2, normalizing the high-dimensional matrix and obtaining its covariance matrix;

[0057] Step S3, obtaining a state analysis indicator based on the measurement data according to the measurement data, and obtaining a state analysis indicator characterizing the operating state of the distribution network according to the state analysis indicator based on the measurement data;

[0058] Step S4, combining the state analysis index with the sliding time window, dynamically analyzing the state of the distribution network, and drawing the operation trajectory of the power grid in real time;

[0059] Step S5, analyzing the time complexity of the state analysis indicator;

[0060] Step S6, using basic data of actual distribution network operation to perform simulation verification, analyzing the distribution network operation status, observing the evolution direction of the distribution network status, and verifying the effectiveness of the status analysis index.

[0061] Step 2, normalize the high-dimensional matrix and calculate its covariance matrix;

[0062] Specifically, in this embodiment, the specific method of forming a high-dimensional matrix in step S1 is:

[0063] Get the time vector y(t i ):

[0064]

[0065] Among them, the time vector y(t i ) is the total number of measurement points at sampling time t i Generate a state variable, l is the number of nodes equipped with synchronous measurement devices;

[0066]

[0067] Where n is the number of sampling times, and the matrix is ​​the result of n samplings, which is formed by concatenating all time vectors.

[0068] In step S2, the high-dimensional matrix is ​​normalized and its covariance matrix is ​​obtained in the following specific manner:

[0069]

[0070] i=1,2,...l

[0071] Among them, μ(Y i ) and σ(Y i ) represent Y i The mean and variance of Y i is the i-th row of Y;

[0072]

[0073] Among them, S n represents the sample covariance matrix, Y l×n The complex conjugate transpose of . l×n After normalization and calculation of the covariance matrix, a series of analyses can be performed using random matrix theory.

[0074] The specific method of step S3 for obtaining the state analysis index based on the measurement data according to the measurement data is:

[0075] The state analysis indicators based on the measurement data specifically include: measurement probability distribution indicator I based on the ring rate 1 , Uniformity Index I 2 .

[0076] Measurement Probability Distribution Index Based on Ring Rate I 1 The specific method of obtaining is:

[0077] I 1 =MSR(S n )

[0078]

[0079] Among them, I 1 It is an indicator based on the circular ring law and MSR to characterize the probability distribution of measurement data; MSR is the singular eigenvalue λ i The mean of

[0080] Uniformity index I involving distribution network changes 2 The specific method of obtaining is:

[0081]

[0082] Among them, I 2 An index representing the uniformity of the distribution network structure and the rate of change of measurement data, i=1,2…l.

[0083] In step S3, the specific method of constructing the state analysis index characterizing the operation state of the distribution network by comprehensively considering the probability distribution of the measured data, the uniformity level of the change rate of the measured data and the uniformity level of the grid structure is as follows:

[0084]

[0085] The specific method of drawing the distribution network operation trajectory in real time in step S4 is:

[0086] As shown in Table 1, n is the width of the time window. When the grid structure is determined, the number of nodes l in the grid is also determined, and the value of n is determined by c = l / n∈(0,1). Let the step length of the sliding time window be 1, that is, among the n columns of sample values, there is 1 column of current data and n-1 columns of historical data.

[0087] Table 1 Sliding time window diagram

[0088]

[0089]

[0090] The specific method for analyzing the time complexity in step S5 is:

[0091] Assuming that there are p branches and n nodes in the power grid, the time complexity analysis and calculation steps are as follows:

[0092] Step S51, normalizing the data in the time window, with a time complexity of O(n);

[0093] Step S52, normalizing the covariance matrix, with a time complexity of O(n);

[0094] Step S53, singularizing the matrix;

[0095] Step S54, calculating the eigenvalue and the mean I1 of the eigenvalue, the time complexity is O(1);

[0096] Step S55, calculate I2, the time complexity is O(n);

[0097] Step S56, calculate I3, the time complexity is O(1).

[0098] Therefore, the time complexity in the embodiment of the present invention is O(n)+O(1).

[0099] The specific operation of simulating and verifying the basic data of the actual distribution network operation in step S6 is as follows:

[0100] Based on PSASP software, simulation verification is carried out using basic data of actual distribution networks. The operating status of the distribution network is analyzed in each case, the evolution direction of the distribution network status is observed, and the effectiveness of the proposed indicators is verified.

[0101] The technical solution of the present invention is further specifically described below through specific application examples.

[0102] The embodiment of the present invention is based on MATLAB software, and the operation data collected from a 110 kV distribution network in a certain province is used for simulation verification. The calculation step length Δt=0.01s and the calculation time are 5s, so there are 500 sampling points in total. The example description is shown in Table 2. The operation status of the distribution network is analyzed under each example, and the evolution direction of the distribution network status is observed to verify the effectiveness of the indicators proposed in the present invention.

[0103] Table 2 Example simulation settings

[0104]

[0105] The busbars are numbered and the node degree m of each node is calculated by the grid structure. i There are 40 110 kV buses in a distribution network of a certain province, so p=40. In this embodiment, the sliding time window width n=60 is selected, so the inner ring radius is 0.577, and the boundary between the stability and instability of the distribution network is 0.52.

[0106] The specific operation process is as follows:

[0107] Assuming the original load rate is 1, since the sliding time window width is 60, all indicators have data starting from the 60th sampling moment. In this way, it can be verified that when the load rate increases within the tolerable range of the distribution network, the proposed indicators are effective in analyzing the state of the distribution network and its evolution direction.

[0108] (1) Figure 2a As shown in the figure, at the first 100 sampling points, the load rate is low, there is only a small disturbance caused by white noise, and the voltage distribution basically conforms to the Gaussian distribution, so I 1 The value is relatively large, about 0.8; the load rate of the 101st sampling point begins to increase, the voltage begins to deviate from the per-unit value, and its probability distribution begins to deviate from the Gaussian distribution. 1 The load factor at the 151st sampling point was basically maintained at a high level of 105%, and the voltage level began to recover. 1 It began to rise and eventually maintained at around 0.8, indicating that the distribution network regained a new balance.

[0109] (2) Figure 2b As shown in the figure: For the first 100 sampling points, the load rate remains unchanged and the voltage change rate of each node is almost 0, so I 2Maintained at a relatively low level; when the load rate of the distribution network increases as a whole, the voltage level changes, and the two begin to increase; in the later stage of the simulation, I 2 It is almost the same as the value of the first 100 sampling points. This is because the growth rate of the whole network load is the same, the voltage change rate of each node is also similar, and there is no serious imbalance in the voltage change rate. Therefore, I 2 Little change; due to I 2 Considering the weighting based on the balance of the power grid structure, the change is minimal. 2 The overall fluctuation is small, the trend is stable, the consideration is more comprehensive, and it is more conducive to the analysis of the power grid status.

[0110] (3) Figure 2c Shown: the first 100 sampling points, I 3 Maintained at about 0.7, corresponding to a relatively stable power grid state; after the 101st sampling point I 3 The load rate starts to increase when the load rate starts to increase; the 155th sampling point I 3 It starts to rise, corresponding to the moment when the load rate stops increasing; after the dynamic change process, I 3 After the 400th sampling point, it basically remained at 0.7, corresponding to a relatively low level of stability.

[0111] Analysis on the evolution direction of distribution network status:

[0112] Depend on Figure 2a and Figure 2c It can be obtained that the initial stable stage of the power grid corresponds to I 3 is about 0.65, the inner diameter of the ring is about 0.577, I 3 It is approximately 1.1 times the radius of the inner ring, so the radius of the inner ring is reduced to 1.1 times, and 0.52 is taken as the stable threshold.

[0113] Index I 3 The three factors that affect the state of the power grid are considered at the same time, and the evolution process of the power grid state is characterized from a holistic perspective. Figure 2c The 125th and 200th sampling points are I 3 When the load factor drops below the stable threshold and rises above the stable threshold, the system load rate increases at the 101st sampling point, causing disturbance to the system. 3 A drop occurs, and the system runs in an unstable direction; after a period of adjustment, the load rate does not change at the 150th sampling point, I 3 It starts to rise, and the system starts to develop in a stable direction. Finally, at the 400th sampling point, the system gradually stabilizes.

[0114] The above simulation shows that I 3The obvious drop and rise correspond to the system running in the unstable and stable directions respectively. The index proposed in the present invention can characterize the operating status of the distribution network, and can evaluate the evolution direction of the distribution network state in real time, providing early warning and decision support for the operating personnel in time.

[0115] Analysis of time complexity:

[0116] The algorithm of the present invention takes about 1 / 3 of the time of the traditional method and has obvious advantages in time complexity.

[0117] From the above description, it can be seen that the beneficial effects brought by the present invention are: the present invention directly analyzes the distribution network operation data and extracts its operation status characteristics, which can avoid a series of problems caused by inaccurate models and large calculations, and can achieve dynamic evaluation to a certain extent. The distribution network state defined in the present invention characterizes the system's ability to withstand disturbances, reflects the dynamic performance and stability level of the power system, and can be analyzed from two aspects: real-time state evaluation and evolution direction, which is the basis for realizing intelligent operation and maintenance of the distribution network.

[0118] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for analyzing the state of a regional distribution network, characterized in that: include: Step S1, the measurement data of the regional distribution network synchronous measurement device is formed into a high-dimensional matrix according to the sampling time and different nodes; Step S2, normalizing the high-dimensional matrix and obtaining its covariance matrix; Step S3, obtaining a state analysis indicator based on the measurement data according to the measurement data, and obtaining a state analysis indicator characterizing the operating state of the distribution network according to the state analysis indicator based on the measurement data; Step S4, combining the state analysis index with the sliding time window, dynamically analyzing the state of the distribution network, and drawing the operation trajectory of the power grid in real time; Step S5, analyzing the time complexity of the state analysis indicator; Step S6, using basic data of actual distribution network operation to perform simulation verification, analyze the distribution network operation status, observe the evolution direction of the distribution network status, and verify the validity of the status analysis index; In the step S3, the state analysis indicators based on the measurement data specifically include: a measurement probability distribution indicator I1 based on the ring rate and a uniformity indicator I2; The specific method for obtaining the measurement probability distribution index I1 based on the ring rate is: I1=MSR(S n ) Among them, I1 is an indicator based on the circular ring law and MSR to characterize the probability distribution of measurement data; MSR is the singular eigenvalue λ i The mean of The specific method for obtaining the uniformity index I2 involving distribution network changes is as follows: Where I2 represents the index of the uniformity of the distribution network structure and the rate of change of the measured data, i = 1, 2…l; In step S4, the specific method of drawing the distribution network operation trajectory in real time is: Assume n is the width of the time window. When the grid structure is determined, the value of n is determined by c=l / n∈(0,1). Let the step size of the sliding time window be 1. Then, among the n columns of sample values, there is 1 column of current data and n-1 columns of historical data.

2. The regional distribution network status analysis method according to claim 1, characterized in that: The specific method of forming a high-dimensional matrix in step S1 is: Get the time vector y(t i ): Among them, the time vector y(t i ) is the total number of measurement points at sampling time t i Generate a state variable, l is the number of nodes equipped with synchronous measurement devices; Wherein, n is the number of sampling times, and the high-dimensional matrix is ​​the result of sampling n times, and is formed by concatenating all time vectors.

3. The method for analyzing the state of a regional distribution network according to claim 2, characterized in that: In step S2, the high-dimensional matrix is ​​normalized and its covariance matrix is ​​obtained in the following specific manner: Among them, μ(Y i ) and σ(Y i ) represent Y i The mean and variance of Y i is the i-th row of Y; Among them, S n represents the sample covariance matrix, Y l×n The complex conjugate transpose of .

4. The method for analyzing the state of a regional distribution network according to claim 1, characterized in that: In step S3, the specific method of obtaining the state analysis index characterizing the operating state of the distribution network according to the state analysis index based on the measurement data is:

5. The method for analyzing the state of a regional distribution network according to claim 4, characterized in that: The specific method of analyzing the time complexity in step S5 is: Assuming there are p branches and n nodes in the power grid, the time complexity analysis and calculation steps are as follows: Step S51, normalizing the data in the time window, with a time complexity of O(n); Step S52, normalizing the covariance matrix, with a time complexity of O(n); Step S53, singularizing the matrix; Step S54, calculating the eigenvalue and the mean I1 of the eigenvalue, the time complexity is O(1); Step S55, calculate I2, the time complexity is O(n); Step S56, calculate I3, the time complexity is O(1).

6. The method for analyzing the state of a regional distribution network according to claim 5, characterized in that: The final time complexity is O(n)+O(1).

7. The method for analyzing the state of a regional distribution network according to claim 1, characterized in that: The specific method of simulating and verifying the basic data of the actual distribution network operation in step S6 is: based on the PSASP software, using the basic data of the actual distribution network to perform simulation verification, analyze the operating status of the distribution network, observe the evolution direction of the distribution network status, and verify the validity of the status analysis indicators.

Images