Method and system for calculating injection transfer distribution factor of power system
By performing discrete wavelet transform denoising reconstruction of PMU measured data, the problem of inaccurate calculation of power system injection transfer distribution factor in the existing technology is solved, and accurate distribution factor calculation that quickly adapts to the changes in the power grid under the characteristics of new energy is achieved, ensuring that The safe operation of the power grid.
Patent Information
- Application Number
- CN202411033273.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-15
AI Technical Summary
The existing mathematical simulation model calculates the injection transfer distribution factor of the power system that cannot adapt to network topology and load changes in time, resulting in incorrect calculation results or inability to use. Especially under the characteristics of intermittent and random output of new energy, it is difficult to accurately predict the failure after a failure. Changes in trends affect power grid safety.
Discrete wavelet transform (DWT) is used to denoise and reconstruct the measured PMU data. The low-frequency variation of load and high-frequency variation of noise are used to maximize the load increment of low-frequency variation, thereby calculating a more accurate injection transfer distribution factor.
Through the data calculation after DWT denoising and reconstruction, it can quickly and accurately adapt to the changes in grid parameters, improve the accuracy and real-time nature of the distribution factor, and ensure the safe prediction and operation of the grid when a fault occurs.
Smart Images

Figure CN118970912A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of power system dispatching and operation, and in particular relates to a method and system for calculating a power system injection transfer distribution factor. Background Art
[0002] In order to reduce the computation time of solving nonlinear power flow models or linearized power flow models, the currently widely used safety analysis and assessment method for power systems is to solve the linear sensitivity distribution factor (DFs) in order to predict the changes in the power flow of each line after a certain power generation equipment or a certain line fails, and to predict whether the line is overloaded or exceeds the limit. The existing mathematical simulation model-based distribution factor solution relies on an accurate and timely updated network topology parameter. Therefore, these model-based distribution factors are unable to adapt to the rapid disturbances and changes from the network topology, generators or loads. These disturbances and changes will significantly change some topologies and parameters, making the distribution factors calculated by the model without timely data update errors or even unusable. Data errors and the inability to grasp the current status of the transmission line in time, and the inability to deal with the fault in time and effectively when it occurs, are the causes of many serious power outages. Therefore, the existing technology for monitoring and protecting power systems based on system models no longer fully meets the recent development needs of the power industry, especially with the rapid growth of new energy, and the intermittent and random characteristics of its output, which exacerbates the impact of such characteristics on the power system. In view of this, a method and system that is independent of power system parameters and can calculate the power system sensitivity distribution factor more quickly and accurately appears to be more practical and more urgent.
[0003] The widespread use of synchronous phase measurement devices (PMU) makes the above requirements possible. PMU measures voltage, current and power synchronously at different locations for different devices at a very high speed (generally 50 times per second). It is a power system monitoring tool that can closely follow the system operating point status and topology changes.
[0004] The method of calculating the power system injection transfer distribution factor based on PMU measured data is to use the least squares method to remove noise from the PMU measured data. However, given that:
[0005] 1) The noise model of PMU measured data is mostly impulse noise with an oscillating tail. The least squares method is not suitable for this type of noise.
[0006] 2) Because the change of power system load is low frequency, the empirical value is no higher than 2 Hz, and the main noise frequency is higher than 10 Hz. The sampling frequency of PMU measurement data is 50 Hz, and the incremental data for calculating the injection transfer distribution factor is ΔP(k): = ΔP(kΔt)-ΔP((k-1)Δt). This incremental data method just removes the slowly changing load increments and strengthens the noise increments that change in the high-frequency band, so that the noise of the incremental data obtained for calculating the injection transfer distribution factor overwhelms the normal load increments. The least squares method cannot eliminate this enhanced noise.
[0007] It is a recognized fact that the least squares method is not very effective in removing impulse noise, and is even more powerless in removing noise when the noise overwhelms the normal signal. Therefore, the injection transfer distribution factor calculated by the existing method will be seriously distorted and difficult to put into practical application. Summary of the invention
[0008] In order to solve the deficiencies of the prior art, the present invention aims to provide a method for calculating the injection transfer distribution factor of the power system, which utilizes the characteristics of low-frequency changes in load and high-frequency changes in noise, uses discrete wavelet transform (DWT) to denoise and reconstruct the PMU measured data, and uses the data denoised and reconstructed by discrete wavelet transform (DWT) as a reference to obtain incremental data for calculating the injection transfer distribution factor, and retains the load increment of low-frequency changes as much as possible. Therefore, the denoising effect is significant and the calculated distribution factor is more accurate.
[0009] In order to achieve the above object, the technical solution adopted by the present invention is:
[0010] A method for calculating a power system injection transfer distribution factor comprises the following steps:
[0011] 1) Obtaining power system grid topology data;
[0012] 2) Obtain real-time measurement data of PMU on the power grid at each moment in a certain time period;
[0013] 3) Performing DWT discrete wavelet transform on the PMU real-time measurement data to remove noise and reconstruct;
[0014] 4) Based on the PMU measured data reconstructed by discrete wavelet transform denoising, the bus static injected active power and branch active power flow increment matrix are established;
[0015] 5) Calculate the power system injection transfer distribution factor based on the bus active power and branch active power flow increment matrix.
[0016] Among them, in step 1), it is assumed that the power system consists of a group of N+1 buses, wherein the N+1th bus is a reference bus; and V={0, 1, ..., N} is used to represent the N+1 bus set, and E={1, ..., L} is used to represent the L branch set; then the power system grid topology data includes N+1 buses, L branches and the corresponding grid topology association matrix.
[0017] In step 2), at the time 0 and every time kΔt thereafter in a certain time period, the real-time measurement data of the PMU includes the bus net injected active power and branch active power flow data, where k = 0, 1, ..., M, Δt = 1 / 50 second, that is, the sampling frequency is 50 Hz, M is an integer multiple of 2 / Δt = 100 and M>N;
[0018] In the PMU measured data obtained, at time t = 0, the static injected active power of bus i∈V and the active power flow of branch (m,n)∈E are marked as P i (0) and P (m,n) (0); the bus static injected active power and branch active power flow at time kΔt are marked as P i (kΔt) and P (m,n) (kΔt); use P i represents the internal time sampling sequence signal of the bus i∈V, which is P i (kΔt); P (m,n) represents the time sampling sequence signal of the (m,n)∈E branch, which is P (m,n) (kΔt).
[0019] Step 3) specifically includes the following steps:
[0020] 3-1) P i and P (m,n) Perform discrete wavelet transform respectively to obtain the corresponding discrete wavelet transform component combination f(P i ) and f(P (m,n) ); each component combination contains s components;
[0021] 3-2) Perform discrete wavelet transform denoising and reconstruction; perform discrete wavelet transform denoising and reconstruction on the PMU measured data with 2 Hz as the threshold; calculate the frequency of each component of the discrete wavelet transform, and perform frequency threshold denoising on the component containing the target frequency based on the threshold, that is, set the discrete wavelet transform components with frequencies greater than the threshold to zero, keep the discrete wavelet transform components with frequencies less than or equal to the threshold unchanged, and then reconstruct the processed components as the denoised signal; the discrete wavelet transform component combination f(P i ) and f(P (m,n) ) The discrete signals after denoising and reconstruction are respectively identified as and For k = 0, 1, ..., M, and The corresponding sampling values are and
[0022] The increment of the net active power injected by the denoised and reconstructed bus corresponding to the time kΔt in step 4) is:
[0023]
[0024] Logo is an M-dimensional vector, whose kth component is Logo Reconstruct the bus active power increment matrix for denoising; is an M×N matrix, whose i-th column vector is
[0025] The increment of the active power flow of the branch corresponding to the denoising and reconstruction at time kΔt is:
[0026]
[0027] Logo is the branch power flow increment vector, is an L-dimensional vector, whose kth component is
[0028] Step 5) specifically includes:
[0029] definition Identify the injection transfer distribution factor of branch (m, n) with respect to bus i, that is, the sensitivity of the active power flow on branch (m, n) to the injected active power of bus i;
[0030] Ψ (m,n) is an N-dimensional vector whose i-th component is Then we can calculate
[0031]
[0032] For each branch (m, n), calculate N injection transfer distribution factors A total of LN injection transfer distribution factors are obtained.
[0033] Correspondingly, the present invention also proposes a calculation system for power system injection transfer distribution factor, the calculation system includes a data acquisition module, a discrete wavelet transform denoising and reconstruction module, a matrix generation module and a calculation module;
[0034] The data acquisition module includes a first data acquisition module and a second data acquisition module, wherein the first data acquisition module is used to obtain the power system grid topology data, and the second data acquisition module is used to obtain the PMU measured data on the power grid at each moment during a certain period of time;
[0035] The discrete wavelet transform denoising and reconstruction module performs discrete wavelet transform on the acquired PMU measured data, performs frequency threshold denoising and reconstructs the data;
[0036] The matrix generation module is used to establish a bus net injected active power increment matrix and a branch active power flow increment matrix based on discrete wavelet transform denoising and reconstruction data;
[0037] The calculation module is used to calculate the power system injection transfer distribution factor.
[0038] The present invention has the following beneficial effects:
[0039] 1. The present invention utilizes the characteristics of low-frequency changes in load and high-frequency changes in noise, and uses discrete wavelet transform (DWT) to denoise and reconstruct the PMU measured data. At the same time, the data after denoising and reconstructing by discrete wavelet transform (DWT) is used as a reference to obtain incremental data for calculating the injection transfer distribution factor: Keep the load increments with low frequency changes as large as possible.
[0040] 2. The present invention calculates the power system injection transfer distribution factor based on the real-time measurement data of the PMU, and is capable of adapting to the changes in grid parameters caused by rapid disturbances and changes in the network topology, generator sets or loads, and quickly and accurately calculates the power system injection transfer distribution factor, solving the problem of inaccurate distribution factor calculation caused by the intermittent and random characteristics of new energy output. Based on the distribution factor, the present invention accurately and quickly predicts the changes in the flow of each line after a certain power generation equipment or a certain line fails, predicts whether the line is overloaded and exceeds the limit, and ensures the safe operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS
[0041] Figure 1 It is a system flow chart of the present invention. DETAILED DESCRIPTION
[0042] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not intended to limit the scope of use of the present invention.
[0043] The present invention proposes a method for calculating a power system injection transfer distribution factor, which specifically comprises the following steps:
[0044] 1) Obtain power system grid topology data.
[0045] Assume that the power system consists of a set of N+1 buses, where the N+1th bus is a reference bus; and V={0, 1, ..., N} represents the N+1 bus set, and E={1, ..., L} represents the L branch set; then the power system grid topology data includes N+1 buses, L branches and the corresponding grid topology association matrix.
[0046] 2) Obtain real-time measurement data of PMU on the power grid at each moment in a certain time period;
[0047] In a certain time period, starting from time 0 and at each time kΔt thereafter, the real-time measurement data of the PMU includes the bus net injected active power and branch active power flow data, where k=0,1,…,M, Δt=1 / 50 second, that is, the sampling frequency is 50 Hz, M is an integer multiple of 2 / Δt=100 and M>N.
[0048] In the PMU measured data obtained, at time t = 0, the static injected active power of bus i∈V and the active power flow of branch (m,n)∈E are marked as P i (0) and P (m,n) (0); the bus static injected active power and branch active power flow at time kΔt are marked as P i (kΔt) and P (m,n) (kΔt); use P i represents the internal time sampling sequence signal of the bus i∈V, which is P i (kΔt)(k=0,1,…,M);P (m,n) represents the time sampling sequence signal of the (m,n)∈E branch, which is P (m,n) (kΔt)(k=0,1,…,M).
[0049] 3) Perform DWT discrete wavelet transform on the PMU real-time measurement data to remove noise and reconstruct. Specifically including:
[0050] 3-1) P i and P (m,n) Perform discrete wavelet transform (DWT) respectively to obtain the corresponding discrete wavelet transform (DWT) component combination f(P i )(i∈V) and f(P (m,n) )((m,n)∈E); each component combination contains s components.
[0051] 3-2) Perform discrete wavelet transform (DWT) denoising and reconstruction.
[0052] The discrete wavelet transform (DWT) is subjected to discrete wavelet transform (DWT) denoising and reconstruction. The load change frequency of the power system is low, concentrated below 2 Hz; while the noise frequency is high, concentrated above 10 Hz. The present invention uses 2 Hz as a threshold to perform discrete wavelet transform (DWT) denoising and reconstruction on the PMU measured data; calculates the frequency of each component of the discrete wavelet transform (DWT), and performs frequency threshold denoising on the component containing the target frequency based on the threshold. That is, the discrete wavelet transform (DWT) components with a frequency greater than the threshold are set to zero, the discrete wavelet transform (DWT) components with a frequency less than or equal to the threshold are kept unchanged, and then the processed components are reconstructed as the denoised signal; the discrete wavelet transform (DWT) component combination f(P i )(i∈V) and f(P (m,n) The discrete signals after denoising and reconstruction are respectively identified as and For k = 0, 1, ..., M, and The corresponding sampling values are and
[0053] 4) Based on the PMU measured data reconstructed by discrete wavelet transform (DWT) denoising, the bus static injected active power and branch active power flow increment matrix are established.
[0054] The net active power increment of the denoised and reconstructed bus corresponding to the time kΔt is:
[0055]
[0056] Logo is an M-dimensional vector, whose kth component is Logo Reconstruct the bus active power increment matrix for denoising; is an M×N matrix, whose i-th column vector is
[0057] The increment of the active power flow of the branch corresponding to the denoising and reconstruction at time kΔt is:
[0058]
[0059] Logo is the branch power flow increment vector, is an L-dimensional vector, whose kth component is
[0060] 5) Calculate the power system injection transfer distribution factor based on the bus active power and branch active power flow increment matrix. Specifically include:
[0061] definition Identify the injection transfer distribution factor of branch (m, n) with respect to bus i, that is, the sensitivity of the active power flow on branch (m, n) to the injected active power of bus i;
[0062] Ψ (m,n) is an N-dimensional vector whose i-th component is Then we can calculate:
[0063]
[0064] For each branch (m, n), calculate N injection transfer distribution factors A total of LN injection transfer distribution factors are obtained.
[0065] like Figure 1 As shown, the corresponding present invention also proposes a calculation system based on the power system injection transfer distribution factor, the calculation system includes a data acquisition module, a discrete wavelet transform denoising and reconstruction module, a matrix generation module and a calculation module;
[0066] The data acquisition module includes a first data acquisition module and a second data acquisition module, wherein the first data acquisition module is used to acquire the power system grid topology data, and the second data acquisition module is used to acquire the real-time measurement data of the PMU on the grid at each moment during a certain period of time;
[0067] The discrete wavelet transform (DWT) denoising and reconstruction module is used to perform discrete wavelet transform (DWT) on the bus static injected active power and branch active power flow for a certain period of time (selected as 30 seconds) obtained by the second data acquisition module, and then filter out the components of the discrete wavelet transform (DWT) corresponding to more than 2 Hz, and generate denoised and reconstructed data for the bus static injected active power and branch active power flow.
[0068] The incremental matrix generation module establishes the incremental matrix of bus static injected active power and branch active power flow based on the discrete wavelet transform (DWT) denoising and reconstructed bus static injected active power and branch active power flow time series data.
[0069] The calculation module is used to calculate the power system injection transfer distribution factor according to the bus static injection active power and branch active power flow increment matrix obtained by the matrix generation module.
[0070] The traditional technology for monitoring and protecting the power system based on power system parameters and mathematical models cannot be applied when the power system is disturbed due to changes in system parameters, and therefore no longer fully meets the recent development needs of the power industry, especially the rapid growth of new energy sources, whose intermittent and random characteristics of output have exacerbated the impact of such characteristics on the power system. Different from the traditional method and system that completely relies on power system parameters to solve the power flow model to calculate the power system injection transfer distribution factor, the present invention calculates the power system injection transfer distribution factor by reconstructing the PMU real-time measurement data through discrete wavelet transform denoising, so that the distribution factor can be quickly and accurately solved when the power system parameters change, so as to quickly and accurately predict the changes in the power flow of each line after a certain power generation equipment or a certain line fails, predict whether the line is overloaded and exceeds the limit, and ensure the safe operation of the power grid.
[0071] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.
Claims
1. A method for calculating a power system injection transfer distribution factor, characterized in that: The specific steps include: 1) Obtaining power system grid topology data; 2) Obtain real-time measurement data of PMU on the power grid at each moment in a certain time period; 3) Performing DWT discrete wavelet transform on the PMU real-time measurement data to remove noise and reconstruct; 4) Based on the PMU measured data reconstructed by discrete wavelet transform denoising, the bus static injected active power and branch active power flow increment matrix are established; 5) Calculate the power system injection transfer distribution factor based on the bus active power and branch active power flow increment matrix.
2. The method for calculating the power system injection transfer distribution factor according to claim 1, characterized in that: In step 1), it is assumed that the power system consists of a group of N+1 buses, wherein the N+1th bus is a reference bus; and V={0, 1, ..., N} is used to represent the N+1 bus set, and E={1, ..., L} is used to represent the L branch set; then the power system grid topology data includes N+1 buses, L branches and the corresponding grid topology association matrix.
3. The method for calculating the power system injection transfer distribution factor according to claim 2, characterized in that: In step 2), at the time 0 and every time kΔt thereafter in a certain time period, the real-time measurement data of the PMU includes the bus net injected active power and branch active power flow data, where k = 0, 1, ..., M, Δt = 1 / 50 second, that is, the sampling frequency is 50 Hz, M is an integer multiple of 2 / Δt = 100 and M>N; In the PMU measured data obtained, at time t = 0, the static injected active power of bus i∈V and the active power flow of branch (m,n)∈E are marked as P i (0) and P (m,n) (0); the bus static injected active power and branch active power flow at time kΔt are marked as P i (kΔt) and P (m,n) (kΔt); use P i represents the internal time sampling sequence signal of the bus i∈V, which is P i (kΔt); P (m,n) represents the time sampling sequence signal of the (m,n)∈E branch, which is P (m,n) (kΔt).
4. The method for calculating the power system injection transfer distribution factor according to claim 3, characterized in that: Step 3) specifically includes the following steps: 3-1) P i and P (m,n) Perform discrete wavelet transform respectively to obtain the corresponding discrete wavelet transform component combination f(P i ) and f(P (m,n) ); each component combination contains s components; 3-2) Perform discrete wavelet transform denoising and reconstruction; perform discrete wavelet transform denoising and reconstruction on the PMU measured data with 2 Hz as the threshold; calculate the frequency of each component of the discrete wavelet transform, and perform frequency threshold denoising on the component containing the target frequency based on the threshold, that is, set the discrete wavelet transform components with frequencies greater than the threshold to zero, keep the discrete wavelet transform components with frequencies less than or equal to the threshold unchanged, and then reconstruct the processed components as the denoised signal; the discrete wavelet transform component combination f(P i ) and f(P (m,n) ) The discrete signals after denoising and reconstruction are respectively identified as and For k = 0, 1, ..., M, and The corresponding sampling values are and 5. The method for calculating the power system injection transfer distribution factor according to claim 4, characterized in that: The increment of the net active power injected by the denoised and reconstructed bus corresponding to the time kΔt in step 4) is: Logo is an M-dimensional vector, whose kth component is Logo Reconstruct bus active power increment matrix for denoising; is an M×N matrix, whose i-th column vector is The increment of the active power flow of the branch corresponding to the denoising and reconstruction at time kΔt is: Logo is the branch power flow increment vector, is an L-dimensional vector, whose kth component is 6. The method for calculating the power system injection transfer distribution factor according to claim 5, characterized in that: Step 5) specifically includes: definition Identify the injection transfer distribution factor of branch (m, n) with respect to bus i, that is, the sensitivity of the active power flow on branch (m, n) to the injected active power of bus i; Ψ (m,n) is an N-dimensional vector whose i-th component is Then we can calculate For each branch (m, n), calculate N injection transfer distribution factors A total of LN injection transfer distribution factors are obtained.
7. The method for calculating the power system injection transfer distribution factor according to claim 6 is implemented based on a calculation system for the power system injection transfer distribution factor, and is characterized in that: The computing system includes a data acquisition module, a discrete wavelet transform denoising and reconstruction module, a matrix generation module and a computing module; The data acquisition module includes a first data acquisition module and a second data acquisition module, wherein the first data acquisition module is used to obtain the power system grid topology data, and the second data acquisition module is used to obtain the PMU measured data on the power grid at each moment during a certain period of time; The discrete wavelet transform denoising and reconstruction module performs discrete wavelet transform on the acquired PMU measured data, performs frequency threshold denoising and reconstructs the data; The matrix generation module is used to establish a bus net injected active power increment matrix and a branch active power flow increment matrix based on discrete wavelet transform denoising and reconstruction data; The calculation module is used to calculate the power system injection transfer distribution factor.
