An adaptive current protection method and system based on support vector machine algorithm

Abstract

The present invention relates to an adaptive current protection method and system based on a support vector machine algorithm. First, the voltage and current of each line of a distributed new energy distribution network are collected in real time; a judgment threshold δ and a positive-negative sequence current ratio β are introduced to determine the fault location and type; secondly, the collected electrical data is processed using a support vector machine algorithm to achieve online identification of photovoltaic equivalent impedance and calculation of current protection setting values under fault conditions; finally, a current adaptive protection scheme for distribution lines is proposed based on the algorithm: in the event of a fault, if the measured value is greater than the setting value, the protection will reliably operate; otherwise, the protection will not operate reliably. The present invention enhances the reliability of the operation of a distributed photovoltaic power distribution network and has practical value for optimizing the operation of the distribution network.

Description

Technical Field

[0001] The present invention belongs to the field of distributed power distribution network protection and control, and relates to an adaptive current protection method and system based on a support vector machine algorithm. Background Art

[0002] Amidst the increasingly severe energy crisis, photovoltaic power generation has seen significant growth. With the increasing penetration of photovoltaics in the power grid, reliable grid operation places higher demands on the sensitivity and selectivity of relay protection. Fault detection and protection technology for photovoltaic power distribution networks is a crucial research topic and a key guarantee for reliable distribution network operation. The integration of photovoltaic power sources can affect the voltage and current flow distribution in the grid. In particular, during a fault, the addition of photovoltaic power sources can increase or drain the fault current, making the original current protection setting unable to meet protection requirements, leading to false or failed protection device operation.

[0003] To address the challenges posed by the grid connection of distributed generation (DG) to traditional power grid protection, domestic and foreign scholars have proposed corresponding solutions: 1) Determine the capacity and location of DG installation based on the size of the distribution network, load changes, line parameters, etc.; 2) Connect a current limiter to the distributed generation grid connection point to suppress the short-circuit current provided by the DG to the fault point; 3) Retain the original line current protection device and use machine learning algorithms to optimize the current protection setting; 4) Add intelligent electronic devices and communication equipment to the original protection to achieve adaptive protection setting based on real-time current information. Method 1) Adjusting the DG connection location and capacity based on the distribution network scale is highly economical and does not require changes to existing current protection devices. However, for more specialized distribution networks, the DG connection location and capacity may be limited. Method 2) Modifying existing current protection devices is less economical and practical, and the installation of current limiters limits the capacity of distributed generation installations to a certain extent. Method 3) Relying on machine learning algorithms to optimize the setting of current protection can shorten fault duration and improve protection responsiveness. However, this method over-relies on machine learning, and the quality of the algorithm directly affects the setting effect of current protection, and the robustness of the algorithm has a significant impact on the optimization of current protection. Method 4) Relying on communication technology to monitor information at each point in real time, fundamentally eliminating the impact of DG grid connection on current protection, but this method has high requirements for communication technology. If the communication system fails, it can cause the distribution network current protection to malfunction, affecting the safe operation of the power system. In summary, current protection technology for high-penetration new energy distribution networks urgently needs breakthroughs and upgrades.

[0004] Through the above analysis, the problems and defects of the existing technology are as follows:

[0005] (1) The current protection technology of distribution network at the current stage will be affected by the grid connection of new energy, resulting in an increased possibility of false operation and refusal to operate. The current protection technology that can be applied to large-scale new energy distribution network is not yet mature and cannot adapt to the development needs of new power system with full access of power electronic equipment. There is a lack of current protection technology suitable for large-scale distributed power distribution network.

[0006] (2) The access of distributed power sources changes the traditional distribution network from single-end power supply to multi-end power supply. The flow direction and current size of the distribution network will be affected by the access of distributed power sources. The original current protection setting value cannot meet the requirements of distributed power distribution network protection, which may cause the distribution network protection to operate incorrectly and reduce the stability and safety of the distribution network operation. Summary of the Invention

[0007] The purpose of the present invention is to provide an adaptive current protection method and system based on a support vector machine algorithm, aiming to reformulate the current protection setting value of the distributed power distribution network to complete the current protection of the distribution line and solve the problem of false operation and refusal of current protection when new energy is connected to the distribution network.

[0008] To achieve the above-mentioned purpose, the technical solution of the present invention is: an adaptive current protection method based on a support vector machine algorithm includes: first, real-time collection of voltage and current quantities of each line of a distributed new energy distribution network; introduction of a judgment threshold δ and a positive-negative sequence current ratio β to determine the fault location and type; second, using a support vector machine algorithm to process the collected electrical data, to achieve online identification of photovoltaic equivalent impedance under fault conditions and calculation of current protection setting values; finally, based on the algorithm, a current adaptive protection scheme for distribution lines is proposed: in the event of a fault, if the measured value is greater than the setting value, the protection will reliably operate; otherwise, the protection will not operate reliably. The present invention enhances the reliability of the operation of a distributed photovoltaic power distribution network and has practical value for optimizing the operation of the distribution network; specifically includes the following steps:

[0009] Step S1: Real-time collection of voltage and current at the beginning and end of each line of the distributed power distribution network;

[0010] Step S2: Using the Kalman filter algorithm to filter the sampled voltage and current to eliminate various noise effects;

[0011] Step S3: introducing the fault determination value δ and the fault type determination value β, calculating the current difference ΔI between the beginning and end of each line and the ratio β of the positive and negative sequence currents, and determining the fault type;

[0012] Step S4: predict the equivalent impedance Z of the photovoltaic system based on the support vector machine algorithm according to the voltage and current information of the photovoltaic power grid connection point PV , according to Z PVRe-adjust the photovoltaic distribution network current protection setting value I' set ;

[0013] Step S5: According to the photovoltaic system current protection setting value I' set The current protection criterion design is carried out based on the relationship between the magnitude of the actual current I flowing through the line.

[0014] Another object of the present invention is to provide an adaptive current protection method based on a support vector machine algorithm and an adaptive current protection system based on a support vector machine algorithm, comprising:

[0015] The phasor acquisition module is used to collect the voltage and current at the beginning and end of each line in the distributed new energy distribution network in real time;

[0016] The filtering module is used to filter the collected data using the Kalman filtering algorithm to remove the noise influence in the power system;

[0017] The symmetrical component module is used to decompose the fault current into sequence components during a fault and determine the fault type based on the ratio of positive and negative sequence currents.

[0018] The current protection module is used to compare the photovoltaic system current protection setting value I' set The current protection criterion design is carried out based on the relationship between the magnitude of the actual current I flowing through the line.

[0019] Another object of the present invention is to provide a computer device, which includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the adaptive current protection method based on the support vector machine algorithm.

[0020] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to perform the steps of the adaptive current protection method based on the support vector machine algorithm.

[0021] Another object of the present invention is to provide an information data processing terminal, which is used to implement the adaptive current protection system based on the support vector machine algorithm.

[0022] Compared with the prior art, the present invention has the following beneficial effects:

[0023] First, in view of the technical problems existing in the above-mentioned prior art and the difficulty of solving these problems, this paper closely combines the technical solutions to be protected by the present invention and the results and data during the research and development process, and analyzes in detail and in depth how the technical solutions of the present invention solve the technical problems and some creative technical effects brought about by solving the problems. The specific description is as follows:

[0024] The present invention aims to utilize line electrical quantity signals to implement current protection for distributed renewable energy distribution networks, solving the problem of current protection misoperation under centralized access to distributed renewable energy. First, the present invention collects the voltage and current of each line in the distributed renewable energy distribution network in real time; introduces a fault determination value δ and a positive-negative sequence current ratio β to determine the fault location and type; secondly, utilizes a support vector machine algorithm to process the collected electrical data, enabling online identification of photovoltaic equivalent impedance under fault conditions and real-time calculation of current protection setting values; finally, based on the algorithm, a current adaptive protection scheme for distribution lines is proposed: in the event of a fault, if the measured value is greater than the setting value, the protection reliably operates; otherwise, the protection reliably does not operate. The current protection method based on electrical quantity signals proposed in the present invention solves the problem that traditional relay protection is not applicable to line current protection in distributed renewable energy distribution networks. This scheme can effectively improve the difficulty in setting the setting value of current protection in distributed energy distribution networks, provide feasible working conditions for current protection, reduce the risk of misoperation or refusal to operate in distributed energy distribution networks, and significantly improve the safety of distributed energy distribution network operations.

[0025] Second, considering the technical solution as a whole or from the perspective of the product, the technical effects and advantages of the technical solution to be protected by the present invention are described in detail as follows:

[0026] (1) The present invention proposes a current protection method based on electrical quantity signals, which solves the problem of current protection threshold setting in distributed new energy distribution networks, enables current protection elements to have the working conditions for correct operation, greatly reduces the risk of protection misoperation, makes current protection threshold setting applicable to distributed new energy distribution network distribution lines, and improves the safety and reliability of distributed energy distribution network operation.

[0027] (2) The present invention proposes a method for calculating the equivalent impedance of a photovoltaic power source, so that the internal equivalent impedance value of the new energy source is not affected by the control strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 This is a flow chart of an adaptive current protection method based on a support vector machine algorithm provided by an embodiment of the present invention;

[0029] Figure 2 This is a schematic diagram of an adaptive current protection method based on a support vector machine algorithm provided by an embodiment of the present invention;

[0030] Figure 3 This is a schematic diagram of a simulation model of a distributed renewable energy distribution line provided by an embodiment of the present invention.

[0031] Figure 4 This is a waveform diagram of the voltage and current collected by Line 3 of the distributed new energy distribution system provided by the embodiment of the present invention after Kalman filtering.

[0032] Figure 5 Schematic diagram of the correlation between the equivalent impedance of a photovoltaic system estimated based on a vector machine algorithm and the true value provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0033] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments of the present invention. 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 any creative work.

[0034] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0035] In view of the problems existing in the prior art, the present invention provides an adaptive current protection method and system based on a support vector machine algorithm. The present invention is described in detail below with reference to the accompanying drawings.

[0036] 1. Explanatory Examples In order to enable those skilled in the art to fully understand how to implement the present invention, this section provides an illustrative example that expands upon the technical solutions of the claims.

[0037] like Figure 1 As shown, the adaptive current protection method based on the support vector machine algorithm provided by the embodiment of the present invention includes the following steps:

[0038] S101, real-time collection of voltage and current at the beginning and end of each line of the distributed power distribution network;

[0039] S102, filtering the sampled voltage and current information using a Kalman filter algorithm to eliminate various noise effects;

[0040] S103, introducing the section determination threshold δ and the positive-to-negative sequence current ratio β, calculating the current difference ΔI at the beginning and end of each line and the positive-to-negative sequence current ratio β, to determine the fault type;

[0041] S104, predicting the equivalent impedance Z of the photovoltaic system based on the support vector machine algorithm according to the voltage and current information of the photovoltaic power grid connection point PV , according to Z PV Re-adjust the photovoltaic distribution network current protection setting value I' set ;

[0042] S105, according to the photovoltaic system current protection setting value I' set The current protection criterion design is carried out based on the relationship between the magnitude of the actual current I flowing through the line.

[0043] Based on the above technical solution, the embodiment of the present invention can also be improved as follows:

[0044] In step S101 provided by the embodiment of the present invention, the sampling frequency is set to 10kHz, and the voltage and current at the beginning and end of each line of the distributed power distribution network are collected in real time using the voltage transformer and the current transformer at both ends of the line to form a voltage sequence and a current sequence. The data window length is saved as N, and N=200, U={u(1),u(2),...,u(n)}, I={i(1),i(2),...,i(n)}, where u(n)={u(1),u(2),...,i(n)} n1 ,u n2 ,...,u nk}, i(n)={i n1 ,i n2 ,...,i nk}, where n is the line number, k is the counting symbol of the number of sampling points, k = 1, 2, ..., N. Since the voltage and current at the beginning and end of the line are collected in real time, in order to effectively distinguish the electrical quantities at the beginning and end of the line, U (1) ={u (1) (1),u (1) (2),...,u (1) (n)}, I (1) ={i (1) (1),i (1) (2),...,i (1) (n)} represents the voltage and current information at the line head end, U (2) ={u (2) (1),u (2) (2),...,u (2) (n)}, I (2) ={i (2) (1),i (2) (2),...,i (2) (n)} represents the voltage and current information at the end of the line.

[0045] When the real-time current variation exceeds the threshold value of the startup judgment, it is determined that the line has a fault, that is, the fault is started. The judgment criteria for startup can be described as:

[0046] |i nk -i n(k-1) |>0.56I N

[0047] Where i nk 、i n(k-1) are two adjacent sampling points in the current sequence of node n; I N Refers to the current rating of the distribution line.

[0048] In step S102 provided in the embodiment of the present invention, the voltage and current signals collected in the above steps are subjected to Kalman filtering to eliminate the influence of various noises. The filtering algorithm used is as follows:

[0049]

[0050] Where Y n It can be expressed as a voltage sequence u(n)={u n1 ,u n2 ,...,u nk} or current sequence i(n)={i n1 ,i n2 ,...,i nk}; b is the Kalman filter gain coefficient, which is generally between 0.1 and 1.0; M is the highest harmonic order of the voltage and current signals; ε(o) is the error in the filter; η(0) is the noise interference with a mean value of zero; ω, is the angular frequency and phase of the highest harmonic; ΔT is the sampling interval at a frequency of 10 kHz.

[0051] According to the above formula, the voltage signal and the current signal are Kalman filtered, and the data window length N is taken as 200. The voltage and current signals after filtering are u′(n)={u′ n1 ,u′ n2 ,...,u′ nk}、i′(n)={i′ n1 ,i′ n2 ,...,i′ nk}, where n is still the line number, k is still the sampling point counting symbol, k = 1, 2, ..., N.

[0052] The voltage and current collected in step S103 during the fault are used to determine the fault type and location by using the ratio of the absolute values of the positive and negative sequence currents β and the difference ΔI between the start and end currents. The specific steps are as follows:

[0053] The single-phase section of the distributed power distribution network can use the current difference between the beginning and end of the section to determine whether a short circuit fault has occurred in the section. Using the full-cycle Fourier algorithm, the effective value of the current at the beginning and end of each line is calculated and recorded as The following criterion can be set: when the absolute value of the current difference |ΔI| at the beginning and end of a line section satisfies |ΔI|≥δ at m consecutive sampling points, it is determined that a fault has occurred in the section; otherwise, the section is determined to be operating normally.

[0054]

[0055] If the above formula is satisfied, it is judged to be an internal fault of the line; otherwise, it is judged to be an external fault of the line. The setting formula of the judgment value δ is as follows:

[0056]

[0057] The setting formula of β is as follows:

[0058]

[0059] Where, They represent the effective current values at the beginning and end of the line respectively, the subscript n represents the nth line node number, and k represents the sampling counting point; Respectively represent the positive sequence current and negative sequence current at the end of the line; γ s represents the safety factor for threshold setting, generally ranging from 0.5 to 1.0; ε represents the error introduced during current measurement, generally ranging from 0.02 to 0.05. When m consecutive sampling points satisfy the condition |ΔI| ≥ δ, a fault can be determined within the line section. Since positive sequence current exists in all types of faults and negative sequence current exists in asymmetric faults, the fault type can be determined based on the β value as follows:

[0060]

[0061] In step S104 provided by the embodiment of the present invention, the equivalent impedance Z of the photovoltaic system is predicted based on the support vector machine algorithm according to the fault type and location determined in step S103. PV , according to Z PV Re-adjust the photovoltaic distribution network current protection setting value I' set , the specific steps are as follows:

[0062]

[0063]

[0064] Where, Γ 3 is the correlation between the predicted value and the actual value, when Γ 3The closer the value is to 1, the higher the correlation between the value predicted by the algorithm and the actual value, and the closer the predicted value is to the true value; is the predicted value of the voltage at node i by the algorithm, y i The voltage filtering value u′(n) of the node i is u′(n)={u′ n1 ,u′ n2 ,...,u′ nk}, is the predicted value of the current at node i by the algorithm, x i The current filtering value of the node i is i′(n)={i′ n1 ,i′ n2 ,...,i′ nk}, which is the voltage and current sequence after filtering in the above step 2, k is the sampling counting point; E PV is the effective value of the output voltage of the distributed photovoltaic power supply, I PV is the effective value of the output current of the distributed photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by this algorithm.

[0065] After the above algorithm predicts the equivalent impedance of the photovoltaic power source, the line current protection setting value is recalculated based on its value. The setting calculation formula is as follows:

[0066]

[0067] Where μ is the fault type coefficient, which is 1 for symmetrical faults and 0 for asymmetrical faults. Es is the electromotive force of the system power supply, Zs is the equivalent impedance of the system power supply, E PV is the equivalent electromotive force of the photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by the support vector machine algorithm, Z1 is the total impedance of the line from the system power outlet bus to the protection installation location, and Z2 is the total impedance of the line from the photovoltaic system grid connection point to the protection installation location. When there is a fault inside or outside the line area, different setting value calculation methods are used.

[0068] Furthermore, in step S105, according to the photovoltaic system current protection setting value I' set The effective value of the current at the beginning and end of the actual line The size relationship is used as the criterion for current protection. The specific steps are as follows:

[0069] When a fault occurs, the current flowing through the protection installation or The reset current protection value I' set Perform real-time comparison, and the protection action criteria are:

[0070]

[0071] When the short-circuit current meets the above criteria, the current protection will be activated, where: are the effective values of the current at the beginning and end of the line respectively; r is the braking coefficient, and its value range is between 1.1 and 1.3.

[0072] The specific current protection steps are as follows:

[0073] When detected |i nk -i n(k-1) |>0.56I N When the current protection starts:

[0074] (1) When |ΔI|≥δ is detected at m consecutive sampling points, if β>0.15, or When , it is determined that an asymmetric fault has occurred on the line in the area, and the current protection is activated;

[0075] (2) When |ΔI|≥δ is detected at m consecutive sampling points, if β<0.1, or When , it is determined that a symmetrical fault has occurred on the line in the area, and the current protection is activated;

[0076] When detected |i nk -i n(k-1) |>0.56I N When , but it is detected that each line does not satisfy |ΔI|≥δ at m consecutive sampling points, then no fault occurs on each line and the system is judged to be operating normally;

[0077] When each line locally satisfies |i nk -i n(k-1) |>0.56I N When , there is no fault in each line and the system is judged to be operating normally.

[0078] The adaptive current protection system based on the support vector machine algorithm provided in the embodiment of the present invention includes:

[0079] The phasor acquisition module is used to collect the voltage and current at the beginning and end of each line in the distributed new energy distribution network in real time;

[0080] The filtering module is used to filter the collected data using the Kalman filtering algorithm to remove the noise influence in the power system;

[0081] The symmetrical component module is used to decompose the fault current into sequence components during a fault and determine the fault type based on the ratio of positive and negative sequence currents.

[0082] The current protection module is used to compare the photovoltaic system current protection setting value I' setThe current protection criterion design is carried out based on the relationship between the magnitude of the actual current I flowing through the line.

[0083] 2. Application Examples: In order to demonstrate the creativity and technical value of the technical solution of the present invention, this section provides application examples of the claimed technical solution on specific products or related technologies.

[0084] 3. Evidence of the effects of the embodiments: The embodiments of the present invention have achieved some positive effects during the development or use process, and indeed have great advantages over the existing technology. The following content describes them with reference to the data, charts, etc. of the experimental process.

[0085] Figure 3 This is a schematic diagram of a photovoltaic power distribution network built based on the actual topology and parameters of a domestic new energy station. The proposed protection principle is simulated and analyzed. The system base capacity is 100MVA, the base voltage is 35kV, and the system equivalent impedance is Z S =j0.18Ω, line parameters R = 0.128Ω / km, X = 0.75Ω / km. The capacities of photovoltaic power sources DG1, DG2, and DG3 are 5MVA, 8MVA, and 10MVA respectively. The capacities of load 1, load 2, and load 3 are 4+j0.2MVA, 5+j0.1MVA, and 8+j0.3MVA respectively. The equivalent impedance Z PV The lengths of lines Line1, Line2, Line3, and Line4 are 3.8km, 3.1km, 1.0km, and 2.4km respectively.

[0086] The embodiment of the present invention verifies the operating performance of the proposed current protection mode through the voltage information and current information collected at the circuit breakers QF3 and QF33, and reliably determines the line fault type and fault location.

[0087] according to Figure 2 The flowchart shown implements an adaptive current protection method based on a support vector machine algorithm.

[0088] S1, Figure 3 Taking the schematic diagram of photovoltaic power distribution network as an example, the voltage and current at the beginning and end of each line of distributed power distribution network are collected in real time. The sampling frequency is set to 10kHz. The voltage and current at the beginning and end of each line of distributed power distribution network are collected in real time using the voltage transformer and current transformer at both ends of the line to form a voltage sequence and a current sequence. The data window length is N, and N=200, U={u(1),u(2),...,u(n)}, I={i(1),i(2),...,i(n)}, where u(n)={u n1 ,u n2 ,...,u nk}, i(n)={i n1 ,i n2 ,...,i nk}, where n is the line node number, k is the counting symbol of the sampling point number, k = 1, 2, ..., N. Since the voltage and current at the beginning and end of the line are collected in real time, in order to effectively distinguish the electrical quantities at the beginning and end, U (1) ={u (1) (1),u (1) (2),...,u (1) (n)}, I (1) ={i (1) (1),i (1) (2),...,i (1) (n)} represents the voltage and current information at the head end of the line, U (2) ={u (2) (1),u (2) (2),...,u (2) (n)}, I (2) ={i (2) (1),i (2) (2),...,i (2) (n)} represents the voltage and current information at the end of the line.

[0089] S2, such as Figure 4 As shown, the voltage and current collected in S1 are Kalman filtered to eliminate system noise and provide accurate data for subsequent steps. The filtering algorithm is as follows:

[0090]

[0091] Where Y n It can be expressed as a voltage sequence or a current sequence. In this embodiment, the voltage and current at node 3 are collected as u(3)={u 31 ,u 32 ,...,u 3k},i(3)={i 31 ,i 32 ,...,i 3k b is the Kalman filter gain coefficient, and in this embodiment, b=0.15; M is the highest harmonic order of the voltage and current signals, and in this embodiment, M=7; ε(o) is the error in the filter; η(0) is the noise interference with a mean value equal to zero; ω, is the unknown angular frequency and phase of the nth harmonic; ΔT is the sampling interval at a frequency of 10 kHz, and in this embodiment, ΔT=0.1 ms.

[0092] According to the above formula, the voltage signal and the current signal are Kalman filtered, and the data window length N=200 is taken. The voltage and current signals after filtering are u′(3)={u′ 31,u′ 32 ,...,u′ 3k}、i′(3)={i′ 31 ,i′ 32 ,...,i′ 3k}, k is the counting symbol of the number of sampling points, k = 1, 2,…, N.

[0093] Depend on Figure 4 It can be seen that the voltage and current after Kalman filtering effectively filter out high-order harmonic interference, the generated data curve is smoother, and the impact of noise on data analysis is reduced.

[0094] S3. Calculate the current difference ΔI between the start and end points of each line at the time of the fault, as well as the positive-to-negative sequence current ratio β, using the filtered data from S2. Use the relationship between δ and ΔI as a criterion to determine the fault location, and use the β value to determine the fault type. The specific steps are as follows:

[0095] In this embodiment, the sampling current of line Line3 is taken as an example i(3)={i 31 ,i 32 ,...,i 3k}, the effective value of the current obtained after full-cycle Fourier filtering is When the absolute value of the current difference |ΔI| at the beginning and end of line section 3 satisfies |ΔI|≥δ at m consecutive sampling points, it is determined that a fault has occurred in the section; otherwise, the section is determined to be operating normally.

[0096]

[0097] The setting formula of threshold δ is as follows:

[0098]

[0099] The setting formula of β is as follows:

[0100]

[0101] Where, They represent the effective current values at the beginning and end of Line 3 respectively, and k represents the sampling counting point; Respectively represent the positive sequence current and negative sequence current at the end of Line 3; γ srepresents the safety factor for threshold setting, generally ranging from 0.5 to 1.0, and in this embodiment, 0.8 is used. ε represents the error introduced during current measurement, generally ranging from 0.02 to 0.05, and in this embodiment, 0.025 is used. When m consecutive sampling points satisfy |ΔI| ≥ δ, a fault can be determined within the line section. Since positive sequence current exists in all types of faults and negative sequence current exists in asymmetric faults, the fault type can be determined based on the β value as follows:

[0102]

[0103] In this embodiment, a short-circuit fault occurs within Line 3 for 1.0s. For a single-phase short-circuit fault, |ΔI| = 0.32kA; for a two-phase short-circuit fault, |ΔI| = 0.41kA; for a three-phase short-circuit fault, |ΔI| = 1.36kA. For a short-circuit fault outside Line 3 for 1.0s, |ΔI| approaches 0 for single-phase, two-phase, and three-phase short-circuit faults. Therefore, a δ value of 0.3kA is sufficient for effective line fault detection. For a short-circuit fault within the three-phase zone, β = 0.06; for a short-circuit fault outside the three-phase zone, β = 0.05; for a short-circuit fault within the single-phase zone, β = 0.29; for a short-circuit fault outside the single-phase zone, β = 0.28; for a short-circuit fault within the two-phase zone, β = 0.32; for a short-circuit fault outside the two-phase zone, β = 0.35.

[0104] S4. The filtered voltage and current U′={u′(1),u′(2),...,u′(n)}, I′={i′(1),i′(2),...,i′(n)} are used to predict the equivalent impedance Z of the photovoltaic system based on the support vector machine algorithm. PV , according to Z PV Re-adjust the photovoltaic distribution network current protection setting value I' set , the specific steps are as follows:

[0105]

[0106] Where, Γ 3 is the correlation between the predicted value and the actual value, when Γ 3 The closer the value is to 1, the higher the correlation between the value predicted by the algorithm and the actual value, and the closer the predicted value is to the true value; is the predicted value of the voltage at node i by the algorithm, y i The voltage filtering value u′(n) of the node i is u′(n)={u′ n1 ,u′ n2 ,...,u′ nk}, is the predicted value of the current at node i by the algorithm, x i The current filtering value of the node i is i′(n)={i′ n1 ,i′n2 ,...,i′ nk}, which is the voltage and current sequence after filtering in the above step 2, k is the sampling counting point; E PV The voltage for photovoltaic power supply is I PV Distribute current for photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by this algorithm.

[0107] Depend on Figure 5 It can be seen that the fitting degree of the equivalent impedance of each photovoltaic system (DG1, DG2, DG3) estimated by support vector machine and the true value are The equivalent impedance of the photovoltaic power source can be estimated more accurately.

[0108] After the above algorithm predicts the equivalent impedance of the photovoltaic power source, the line current protection setting value is recalculated based on its value. The setting calculation formula is as follows:

[0109]

[0110] Where μ is the fault type coefficient, which is 1 for symmetrical faults and 0 for asymmetrical faults. Es is the electromotive force of the system power supply, Zs is the equivalent impedance of the system power supply, E PV is the equivalent electromotive force of the photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by the support vector machine algorithm, Z1 is the total impedance of the line from the system power outlet bus to the protection installation location, and Z2 is the total impedance of the line from the photovoltaic system grid connection point to the protection installation location. When there is a fault inside or outside the line area, different setting value calculation methods are used.

[0111] S5. According to Figure 2 As shown in the flow chart, faults are set at the midpoint of line 3 and outside the line 3 section, and the fault data obtained by simulation are imported into the program to verify the feasibility and accuracy of the method proposed in this paper. In the embodiment of the present invention, the error ε = 0.025, the section judgment threshold δ = 0.2, and the safety factor γ for threshold setting s =0.8, ΔT=0.1ms, M=7.

[0112] Set a three-phase short circuit grounding fault, taking line 3 as an example:

[0113] (1) If the fault occurs outside the Line 3 section, on Line 1, the measuring device measures ΔI = 0, β = 0.07, μ = 1, and I′ in Line 3. set =4.1537kA, the fault current is 4.0245kA, and the current protection in Line3 section does not operate; while in Line1, ΔI=1.8523kA, β=0.06, μ=1, I′set =3.9853kA, the fault current is 3.9987kA, and the current protection element in Line 3 section is working normally;

[0114] (2) If the fault occurs in section 3 of the line, the measuring device measures ΔI = 1.8632 kA, β = 0.06, μ = 1, I′ set =3.9901kA, the fault current is 3.9965kA, and the current protection components in the section are working normally;

[0115] Set a single-phase short circuit grounding fault to occur, taking line 3 as an example:

[0116] (1) If the fault occurs outside the Line 3 section, on Line 1, the measuring device measures ΔI = 0, β = 0.37 in Line 3. I′ set =2.9857kA, the fault current is 3.0024kA, and the current protection in Line3 section does not operate; while in Line1, ΔI=0.7541kA, β=0.35, I′ set =3.1694kA, the fault current is 3.1567kA, and the current protection element in Line 3 section is working normally;

[0117] (2) If the fault occurs in section 3 of the line, the measuring device measures ΔI = 0.893 kA, β = 0.29, I′ set =2.3544kA, the fault current is 2.361kA, and the current protection components in the section are working normally;

[0118] Suppose a two-phase short circuit grounding fault occurs, taking line 3 as an example:

[0119] (1) If the fault occurs outside the Line 3 section, on Line 1, the measuring device measures ΔI = 0, β = 0.41 in Line 3. I′ set =3.76049kA, the fault current is 1.992kA, and the current protection in Line3 section does not operate; while in Line1, ΔI=1.923kA, β=0.45, I′ set =3.96087kA, the fault current is 3.97603kA, and the current protection element in Line 3 section of the line is working normally;

[0120] (2) If the fault occurs in Line 3, the measuring device measures ΔI = 1.183 kA, β = 0.40, I′set =3.76904kA, the fault current is 3.77144kA, and the current protection element in the Line 3 section of the line is working normally;

[0121] Test results derived from a large amount of simulation data have proven that in distributed renewable energy distribution network scenarios, the use of adaptive current setting can improve the phenomenon of current protection components failing to operate correctly.

[0122] In summary, this embodiment verifies the correctness and feasibility of the present invention. The adaptive current protection method based on the support vector machine algorithm for high-penetration photovoltaic power distribution networks provided by the embodiment of the present invention can operate normally in new energy distribution networks, laying the foundation for the adaptability of current protection in distributed new energy distribution networks.

[0123] The above are preferred embodiments of the present invention. Any changes made according to the technical solution of the present invention, as long as the resulting functions and effects do not exceed the scope of the technical solution of the present invention, shall fall within the scope of protection of the present invention.

Claims

1. An adaptive current protection method based on support vector machine algorithm, characterized in that: The steps include: Step S1: Real-time collection of voltage and current at the beginning and end of each line of the distributed power distribution network; Step S2: Using the Kalman filter algorithm to filter the sampled voltage and current to eliminate various noise effects; Step S3: introduce the fault type starting value δ and the fault type judgment value β, calculate the current difference ΔI at the beginning and end of each line and the fault type judgment value β, and judge the fault type; the specific implementation is as follows: Determine whether a fault has occurred in the line section: Use the full-cycle Fourier algorithm to calculate the effective value of the current at both ends of the line, which is recorded as On this basis, the fault type starting value δ is introduced. If the absolute value of the current difference |ΔI| between the beginning and end of a line section satisfies |ΔI|≥δ at m consecutive sampling points, the fault type judgment is initiated: If the above formula is satisfied, it is judged to be a fault within the line area; otherwise, it is judged to be a fault outside the line area. The setting formula of the fault type starting value δ is as follows: The setting formula of the fault type judgment value β is as follows: Where n is the line number, k represents the sampling counting point; Respectively represent the positive sequence current and negative sequence current at the end of the line; γ s represents the safety factor of the threshold setting; ε represents the error caused by the current measurement process; when m consecutive sampling points all satisfy |ΔI|≥δ, it is judged that a fault has occurred in the line section. Since positive sequence current exists in all types of faults and negative sequence current exists in asymmetric faults, the fault type is judged as follows based on the β value: Step S4: predict the equivalent impedance Z of the photovoltaic system based on the support vector machine algorithm according to the voltage and current information of the photovoltaic power grid connection point PV , according to Z PV Re-adjust the photovoltaic system current protection setting value I' set ; Step S5: According to the photovoltaic system current protection setting value I' set The current protection criterion design is carried out based on the relationship between the magnitude of the actual line current I and the actual line current. The specific current protection steps are as follows: When detected |i nk -i n(k-1) |>0.56I N When the current protection starts, i nk 、i n(k-1) are two adjacent sampling points in the current sequence of node n; I N Refers to the current rating of the distribution line: (1) When |ΔI|≥δ is detected at m consecutive sampling points, if β>0.15, or When , it is determined that an asymmetric fault occurs in the line within the area, and the current protection action is executed, and r is the braking coefficient; (2) When |ΔI|≥δ is detected at m consecutive sampling points, if β<0.1, or When , it is determined that a symmetrical fault has occurred on the line in the area, and the current protection action is executed; When detected |i nk -i n(k-1) |>0.56I N When , but it is detected that each line does not satisfy |ΔI|≥δ at m consecutive sampling points, then no fault occurs on each line and the system is judged to be operating normally; When each line locally satisfies |i nk -i n(k-1) |>0.56I N When , there is no fault in each line and the system is judged to be operating normally.

2. The adaptive current protection method based on support vector machine algorithm according to claim 1, characterized in that: In the step S1, the sampling frequency is set to 10kHz, and the voltage and current at the beginning and end of each line of the distributed power distribution network are collected in real time by using the voltage transformer and the current transformer at both ends of the line to form a voltage sequence and a current sequence. The data window length is saved as N, and N=200, U={u(1),u(2),...,u(n)}, I={i(1),i(2),...,i(n)}, where u(n)={u n1 ,u n2 ,...,u nk }, i(n)={i n1 ,i n2 ,...,i nk }, where n is the line number, k is the sampling point number symbol, k = 1, 2, ..., N; due to the real-time acquisition of the voltage and current at the beginning and end of the line, in order to effectively distinguish the electrical quantities at the beginning and end of the line, U (1) ={u (1) (1),u (1) (2),...,u (1) (n)}, I (1) ={i (1) (1),i (1) (2),...,i (1) (n)} represents the voltage and current information at the head end of the line, U (2) ={u (2) (1),u (2) (2),...,u (2) (n)}, I (2) ={i (2) (1),i (2) (2),...,i (2) (n)} represents the voltage and current information at the end of the line.

3. The adaptive current protection method based on support vector machine algorithm according to claim 2, characterized in that: In step S2, the voltage and current values sampled are filtered using a Kalman filter algorithm to eliminate various noise effects. The filtering algorithm is as follows: Where Y n Expressed as voltage sequence u(n)={u n1 ,u n2 ,...,u nk } or current sequence i(n)={i n1 ,i n2 ,...,i nk }; b is the Kalman filter gain coefficient; M is the highest harmonic order of the voltage and current signals; ε(o) is the error in the filter; η(0) is the noise interference with a mean value equal to zero; ω, are the angular frequency and phase of the highest harmonic respectively; ΔT is the sampling interval at a frequency of 10kHz; According to the above formula, the voltage and current signals are Kalman filtered, and the data window length N is taken as 200. The voltage and current sequences after filtering are u′(n)={u′ n1 ,u′ n2 ,...,u′ nk }、i′(n)={i′ n1 ,i′ n2 ,...,i′ nk }.

4. The adaptive current protection method based on support vector machine algorithm according to claim 1, characterized in that: In step S4, the filtered voltage and current sequence u′(n)={u′ n1 ,u′ n2 ,...,u′ nk }, i′(n)={i′ n1 ,i′ n2 ,...,i′ nk }, predict the equivalent impedance Z of the photovoltaic system based on the support vector machine algorithm PV , according to Z PV Re-adjust the photovoltaic system current protection setting value I' set , as follows: Where, Γ 3 is the correlation between the predicted value and the actual value, when Γ 3 The closer the value is to 1, the higher the correlation between the value predicted by the support vector machine algorithm and the actual value, and the closer the predicted value is to the true value; is the predicted value of node i voltage by support vector machine algorithm, y i The voltage filtering value u′(n) of the node i is u′(n)={u′ n1 ,u′ n2 ,...,u′ nk }, is the predicted value of node i current by support vector machine algorithm, x i The current filtering value of the node i is i′(n)={i′ n1 ,i′ n2 ,...,i′ nk }, which is the voltage and current sequence after filtering in step S2, k is the sampling count point, i=1,2,...,n; E PV is the effective value of the output voltage of the distributed photovoltaic power supply, I PV is the effective value of the output current of the distributed photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by the algorithm; After the support vector machine algorithm predicts the equivalent impedance of the photovoltaic system, the line current protection setting value is recalculated based on its value. The setting calculation formula is as follows: Where μ is the fault type coefficient, which is 1 for symmetrical faults and 0 for asymmetrical faults. Es is the electromotive force of the system power supply, Zs is the equivalent impedance of the system power supply, E PV is the equivalent electromotive force of the photovoltaic power source, Z PV is the equivalent impedance of the photovoltaic system calculated by the support vector machine algorithm, Z1 is the total impedance of the line from the system power outlet bus to the protection installation location, and Z2 is the total impedance of the line from the photovoltaic system grid connection point to the protection installation location. When there is a fault inside or outside the line area, different setting value calculation methods are used.

5. An adaptive current protection system based on a support vector machine algorithm using the adaptive current protection method based on a support vector machine algorithm as claimed in any one of claims 1 to 4, characterized in that: include: The phasor acquisition module is used to collect the voltage and current at the beginning and end of each line in the distributed power distribution network in real time; The filtering module is used to filter the collected data using the Kalman filtering algorithm to remove the noise influence in the power system; The symmetrical component module is used to decompose the fault current into sequence components during a fault and make a judgment on the fault type through the fault type judgment value; The current protection module is used to compare the photovoltaic system current protection setting value I' set The current protection criterion design is carried out based on the relationship between the magnitude of the actual current I flowing through the line.

6. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the adaptive current protection method based on the support vector machine algorithm as claimed in any one of claims 1 to 4.

Images