Differential protection maloperation suppression method and system for power distribution network under distributed photovoltaic access

Abstract

The invention relates to the technical field of power system power distribution network protection, and discloses a power distribution network differential protection maloperation suppression method under distributed photovoltaic access, which comprises the following steps of: 1, constructing a steady-state equivalent model and a fault equivalent model of a photovoltaic power supply; 2, performing photovoltaic island division based on a breadth-first search algorithm; 3, establishing a protection stationing optimization model; step 4, based on the voltage drop degree of the photovoltaic grid-connected point; and 5, constructing a self-adaptive differential protection criterion. According to the power distribution network differential protection maloperation suppression method and system under distributed photovoltaic access, whether photovoltaic access exists in a protection area or not is judged to dynamically adjust the criterion type, positive and negative sequence fault component auxiliary criteria are matched to accurately recognize the fault type, a voltage locking criterion and an inverse time limit criterion are introduced for the self-starting working condition of a motor, and the fault protection maloperation suppression method and system under distributed photovoltaic access are achieved. And non-linear change of short-circuit current after photovoltaic access is adapted, false triggering under special working conditions is avoided, and the protection misoperation rate is effectively reduced.

Description

Technical Field

[0001] This invention relates to the field of power system distribution network protection technology, specifically to a method and system for suppressing maloperation of differential protection in distribution networks with distributed photovoltaic access. Background Technology

[0002] As the global energy transition deepens, the penetration rate of distributed photovoltaic power in distribution networks continues to rise. Traditional distribution networks have gradually evolved from a single-source radial structure to a complex network with multiple sources, fundamentally changing their operating characteristics and fault features. Against this backdrop, relay protection systems, especially differential protection as the main protection, face severe challenges.

[0003] Existing differential protection schemes for distribution networks are mainly divided into three categories: first, traditional differential protection based on current amplitude comparison, which uses fixed braking coefficient and starting threshold value; second, adaptive differential protection scheme, which adjusts the setting value according to the system operation mode; and third, differential protection based on generalized synchronization, which focuses on solving the data synchronization problem.

[0004] However, existing technologies have two major flaws in adapting to the differential protection of distribution networks after distributed photovoltaic (PV) integration, which seriously threaten the safe and stable operation of the distribution network: Insufficient adaptability of criteria: Traditional differential protection uses fixed current amplitude and phase criteria, which cannot adapt to the nonlinear changes in short-circuit current after photovoltaic access. The short-circuit current provided by photovoltaic will dynamically adjust with the grid connection point voltage. It also does not have a special mechanism designed for special operating conditions such as motor self-starting and asymmetrical faults. When the motor self-starts, the current amplitude is close to the fault current, which is easy to trigger the protection action falsely. The negative sequence current component generated by asymmetrical faults will interfere with the criteria for identifying the fault type, ultimately leading to a significant increase in the protection false trip rate. Short-circuit current calculation is inefficient: Traditional short-circuit current calculation uses a global iterative algorithm, which does not take into account the voltage drop characteristics of the photovoltaic grid connection point to process the distribution network in a partitioned manner. The calculation efficiency is low and it is difficult to meet the millisecond-level setting requirements of adaptive protection. At the same time, the traditional calculation method does not accurately incorporate the photovoltaic fault output model, and cannot accurately reflect the current output characteristics during photovoltaic faults. This results in large errors in the short-circuit current calculation results, which cannot provide a reliable basis for the dynamic adjustment of protection criteria. Therefore, a method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access were proposed. Summary of the Invention

[0005] (a) Technical problems to be solved To address the shortcomings of existing technologies, this invention provides a method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic (PV) access. It features efficient and accurate short-circuit current calculation through a partitioned iterative algorithm that combines a PV steady-state and fault equivalent model with voltage dip characteristics at the grid connection point. Furthermore, it effectively suppresses maloperation by using a differential criterion that dynamically adjusts based on whether the protected area contains PV, an auxiliary criterion for positive and negative sequence fault components, and a specific criterion for motor self-starting. Finally, it achieves a balance between the economic efficiency and power supply reliability of the distribution network through protection point optimization using a breadth-first search algorithm for PV islanding and distribution estimation. This invention solves the problems of high maloperation rates and inefficient short-circuit current calculations in existing distribution network differential protection schemes, which suffer from insufficient criterion adaptability and difficulty in supporting setting.

[0006] (II) Technical Solution To achieve the above-mentioned efficient and accurate short-circuit current calculation by constructing a photovoltaic steady-state and fault equivalent model and combining iterative algorithms based on grid connection point voltage drop characteristics; to effectively suppress protection maloperation by using differential protection criteria based on whether the protection zone contains photovoltaics, auxiliary criteria for positive and negative sequence fault components, and specific criteria for motor self-starting; and to achieve a balance between distribution network economy and power supply reliability by optimizing protection point deployment through a breadth-first search photovoltaic islanding partitioning and distribution estimation algorithm, this invention provides the following technical solution: a method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access, comprising the following steps: Step 1: Construct a steady-state equivalent model and a fault equivalent model for the photovoltaic power source. The steady-state equivalent model corrects the output current by incorporating compensation coefficients for ambient temperature and light intensity. The fault equivalent model constructs a voltage-controlled current source model based on low voltage ride-through requirements. Step 2: Perform photovoltaic islanding based on breadth-first search algorithm. Input distribution network topology parameters, photovoltaic access parameters and fault node information. Starting with each photovoltaic node, search according to preset principles to form candidate islands, and select the candidate island with the highest fitness. Step 3: Establish a protection deployment optimization model and solve it using a distribution estimation algorithm. The model aims to minimize economic cost and is constrained by ensuring that the average power supply reliability of the distribution network is not lower than a preset threshold. Step 4: Based on the voltage drop at the photovoltaic grid connection point, the distribution network is divided into upstream and downstream fault regions, and the fault point is split into two split nodes. The short-circuit current is calculated using a partitioned iterative algorithm. Step 5: Construct adaptive differential protection criteria. Depending on whether there is photovoltaic access in the protection zone, adopt criteria with braking coefficient or setting value criteria based on short-circuit current calculation results. Introduce positive and negative sequence fault components to assist in identifying fault types. Introduce special criteria for motor self-starting conditions.

[0007] Preferably, the compensation coefficients of the steady-state equivalent model of the photovoltaic power source in step 1 include a temperature compensation coefficient a, a light compensation coefficient b, and c. The value of a ranges from 0.002 to 0.003, the value of b ranges from 0.0025 to 0.0035, and the value of c ranges from 0.0004 to 0.0006. The temperature correction is the difference between the actual ambient temperature and the rated ambient temperature, and the light correction is the difference between the actual light intensity and the rated light intensity. The corrected photovoltaic short-circuit current is obtained by multiplying the compensation coefficients by the rated short-circuit current, and the corrected photovoltaic open-circuit voltage is obtained by multiplying the compensation coefficients by the rated open-circuit voltage. The fault equivalent model satisfies: When the voltage at the photovoltaic grid connection point At that time, the reference value of photovoltaic output active current reactive current reference value ; when hour, , , ; when hour, , ; in The rated voltage of the photovoltaic grid connection point. This is the rated output current of the photovoltaic system. This is a reference value for the active power output of the photovoltaic system. This represents the reactive current coefficient, with a value between 1.4 and 1.6. This is the maximum allowable coefficient for output current, and its value ranges from 1.1 to 1.3.

[0008] Preferably, the preset principles of the breadth-first search algorithm in step 2 include the active power balance principle, the maximum load principle, the principle of prioritizing the restoration of power supply to important loads, the connectivity principle, and the principle of non-faulty island areas. The fitness calculation formula for candidate islands is as follows: , Let i be the active power of load i within the island. Let be the island affiliation state variable for node i, where 1 indicates that the node belongs to an island, and 0 indicates that the node does not belong to an island; the specific algorithm flow is as follows: Step 2.1: Mark the starting photovoltaic node as visited and add it to the queue; Step 2.2: Take a node from the queue, visit all the neighboring nodes of the node, and determine whether the neighboring nodes are faulty nodes and whether the active power balance principle is satisfied; Step 2.3: Mark the adjacent nodes that meet the conditions as visited and add them to the queue; Step 2.4: Repeat steps S2-S3 until the queue is empty, forming candidate islands; Step 2.5: Calculate the fitness of all candidate islands and select the candidate island with the highest fitness as the final island region.

[0009] Preferably, the economic cost of the protection deployment optimization model in step 3 includes protection investment cost, protection operation and maintenance cost, and system power outage loss cost; The cost of protecting investment is calculated using the equivalent annual value method, and the formula is as follows: M represents the number of protection devices installed. Let be the present value of the investment cost of a single protection device, i be the discount rate, and p be the design service life of the protection device; the protection operation and maintenance cost is the product of the protection investment cost and a preset proportional coefficient, the preset proportional coefficient ranging from 4% to 6%; the system power outage loss cost is the sum of the products of the unit power outage loss at each load point, the power outage time at each load point, and the active power at each load point; the formula for calculating the average power supply reliability rate of the distribution network is... , The annual rated power supply hours are taken as 8760 hours. Let i be the number of users at load point i. The average annual power outage time for load point i is set to a preset threshold of 99.98%.

[0010] Preferably, the zoning thresholds for the degree of voltage drop at the photovoltaic grid connection point in step 4 include a first threshold and a second threshold, where the first threshold is 0.2U_N and the second threshold is 0.9U_N. When the photovoltaic grid connection point voltage is higher than the second threshold or between the first and second thresholds, the photovoltaic unit belongs to the upstream region of the fault; when the photovoltaic grid connection point voltage is lower than the first threshold, the photovoltaic unit belongs to the downstream region of the fault. The specific process of the zoning iterative algorithm is as follows: Step 4.1: Iteration in the downstream region of the fault: Taking one of the split nodes as the equilibrium node, set the initial iteration parameters of the network in the region, calculate the initial output power of the photovoltaic according to the photovoltaic equivalent model, perform normal component network power flow calculation to obtain the normal component of the voltage of each node in the region, solve the initial short-circuit current, calculate the fault component of the voltage of each node using the node impedance matrix, superimpose the normal component and the fault component to obtain the node voltage after the fault, update the photovoltaic state, and repeat the iteration until the convergence condition is met. Step 4.2: Data transmission: Convert the photovoltaic output power in the downstream area into a constant impedance model, calculate the power transmitted from the balancing node to the area and convert it into a constant impedance model, and then superimpose the two constant impedance models and apply them to another split node. Step 4.3: Iteration of upstream region of fault: Using the split node after adding constant impedance model as the balancing node, repeat the iteration process of step S1 to obtain the final short-circuit current.

[0011] Preferably, the braking coefficient of the adaptive differential protection criterion in step 5 is adjusted according to the branch type in the protection zone: 0.2-0.45 when there are no undetected branches, 0.2-0.35 when there are load branches, and 0.22-0.34 when there are inverter-type distributed power supply branches. The criterion with braking coefficient is in the form of: , , These are the current phasors measured by the protection at both ends of the line, and K is the braking coefficient. This is the minimum starting current, and its value ranges from 0.1 to 0.2. ; The setting value criterion based on the short-circuit current calculation results is as follows: , These are the setting values ​​for the protection on both sides of line mn, determined by the short-circuit current calculation results; In the auxiliary criteria for positive and negative sequence fault components, the negative sequence current amplitude criterion and the negative sequence current phase criterion are used for asymmetrical faults, while the positive sequence fault current amplitude criterion is used for symmetrical faults.

[0012] Preferably, the special criteria introduced in step 5 for the self-starting condition of the motor include voltage blocking criteria and inverse time criterion; Voltage blocking criterion is , The line voltage difference between any two phases in the distribution network. The voltage setting value is 0.65-0.75U_N; The formula for calculating the inverse time limit criterion is as follows: , , These are the positive-sequence fault current phasors at both ends of the line. This is the threshold value for the inverse time protection starting current. This is the curve shape coefficient, and its value ranges from 0.8 to 1.2. This is the horizontal translation coefficient for the curve, and its value ranges from 0.9 to 1.1. It is the inverse time limit protection time constant with a value of 0.2-0.3.

[0013] Preferably, the convergence condition of the partitioning iterative algorithm in step 4 is: , Let be the voltage at the photovoltaic grid connection point at node i during the t-th iteration. Let be the photovoltaic grid-connected voltage at node i in the (t+1)th iteration, and ε be the convergence threshold with a value of 0.0008-0.0012pu. The node impedance matrix is ​​obtained by inverting the node admittance matrix constructed from the distribution network topology parameters. In the node admittance matrix, when i=j... Let k be the neighboring nodes of node i. Let i be the line impedance between node i and node k; when i ≠ j , Let be the line impedance between node i and node j.

[0014] The system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic (PV) access is used to implement the method for suppressing maloperation of differential protection in distribution networks under distributed PV access. It includes a PV equivalent modeling module, an islanding module, a protection point optimization module, a short-circuit current calculation module, a differential protection setting module, a data storage module, and a communication module. Each module interacts with the other through the communication module, which has a transmission rate of 100 Mbps and a transmission delay of no more than 10 ms. The photovoltaic equivalent modeling module outputs photovoltaic output current characteristic parameters and voltage characteristic parameters; The island division module outputs the final island region and the island affiliation status of each node; The protection point optimization module outputs the optimal protection point location; The short-circuit current calculation module outputs the short-circuit current calculation result; The differential protection setting module outputs the differential protection setting value; The data storage module is used to store distribution network topology parameters, photovoltaic access parameters, fault data, calculation results, and differential protection criterion parameters.

[0015] Preferably, the protection deployment optimization module includes a population generation unit, a fitness evaluation unit, a probability model construction unit, and a sampling iteration unit; The population generation unit generates an initial population with binary codes. The length of each individual is equal to the number of distribution network lines. The code value 0 indicates that the corresponding line does not have a protection device installed, and the code value 1 indicates that the corresponding line has a protection device installed. The fitness evaluation unit calculates the economic cost of each individual in the initial population and the average power supply reliability of the distribution network, and selects dominant individuals that meet the constraints of the average power supply reliability of the distribution network. The probability model construction unit calculates the initial probability model. and updated to , Let N be the encoded value of the l-th bit of the k-th individual in the dominant population, where N is the number of individuals in the dominant population. For learning efficiency, the value is 0.25-0.35; The sampling iteration unit generates a new population based on the updated probability model, iterates until the population fitness converges, and outputs the optimal protection point location.

[0016] (III) Beneficial Effects Compared with the prior art, the present invention provides a method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access, which has the following beneficial effects: 1. The method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access constructs an adaptive differential protection criterion through a differential protection setting module. It first determines whether there is photovoltaic access within the protection zone to dynamically adjust the criterion type. It is combined with positive and negative sequence fault components to assist in accurate identification of fault types. For motor self-starting conditions, voltage blocking criterion and inverse time criterion are introduced to adapt to the nonlinear changes of short-circuit current after photovoltaic access, avoid false triggering under special conditions, and effectively reduce the protection maloperation rate.

[0017] 2. The method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access utilizes a photovoltaic equivalent modeling module to construct steady-state and fault equivalent models. Combined with the partitioned iterative algorithm adopted by the short-circuit current calculation module, it divides the region into iterative calculations according to the voltage drop characteristics of the photovoltaic grid connection point. The logic of constructing the node impedance matrix is ​​integrated to accurately reflect the characteristics of photovoltaic fault current, replacing the traditional global iteration, improving calculation efficiency and accuracy, and providing a reliable basis for dynamic adjustment of protection criteria. Attached Figure Description

[0018] Figure 1 This is a flowchart of the method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access, as described in this invention. Figure 2 This is a diagram showing the system module connection structure of the present invention; Figure 3 This is a flowchart illustrating the photovoltaic islanding logic of the present invention. Figure 4 This is the logic diagram for calculating the partitioned iterative short-circuit current of the present invention. Detailed Implementation

[0019] The technical solutions of the present invention will be clearly and completely described below with reference to the embodiments and accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] Please see Figure 1-4 A method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic (PV) grid integration includes the following steps: Step 1: Construct a steady-state equivalent model and a fault equivalent model for the photovoltaic power source. The steady-state equivalent model corrects the output current by incorporating compensation coefficients for ambient temperature and light intensity. The fault equivalent model constructs a voltage-controlled current source model based on low voltage ride-through requirements. Step 2: Perform photovoltaic islanding based on breadth-first search algorithm. Input distribution network topology parameters, photovoltaic access parameters and fault node information. Starting with each photovoltaic node, search according to preset principles to form candidate islands, and select the candidate island with the highest fitness. Step 3: Establish a protection deployment optimization model and solve it using a distribution estimation algorithm. The model aims to minimize economic cost and is constrained by ensuring that the average power supply reliability of the distribution network is not lower than a preset threshold. Step 4: Based on the voltage drop at the photovoltaic grid connection point, the distribution network is divided into upstream and downstream fault regions, and the fault point is split into two split nodes. The short-circuit current is calculated using a partitioned iterative algorithm. Step 5: Construct adaptive differential protection criteria. Depending on whether there is photovoltaic access in the protection zone, adopt criteria with braking coefficient or setting value criteria based on short-circuit current calculation results. Introduce positive and negative sequence fault components to assist in identifying fault types. Introduce special criteria for motor self-starting conditions.

[0021] The system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic (PV) access is used to implement the method for suppressing maloperation of differential protection in distribution networks under distributed PV access. It includes a PV equivalent modeling module, an islanding module, a protection point optimization module, a short-circuit current calculation module, a differential protection setting module, a data storage module, and a communication module. Each module interacts with the other through the communication module, which has a transmission rate of 100 Mbps and a transmission delay of no more than 10 ms. The photovoltaic equivalent modeling module outputs photovoltaic output current characteristic parameters and voltage characteristic parameters; The island division module outputs the final island region and the island affiliation status of each node; The protection point optimization module outputs the optimal protection point location; The short-circuit current calculation module outputs the short-circuit current calculation result; The differential protection setting module outputs the differential protection setting value; The data storage module is used to store distribution network topology parameters, photovoltaic access parameters, fault data, calculation results, and differential protection criterion parameters.

[0022] Example 1: This embodiment details the construction process of the steady-state equivalent model and the fault equivalent model of the photovoltaic power source. The specific operations are as follows: First, collect the rated parameters of the photovoltaic power source, including the rated short-circuit current. Rated open circuit voltage Rated output current and active power reference value ; This is the maximum output current of the photovoltaic system during normal operation. This is the preset active power output target value for photovoltaic systems; Simultaneously, ambient temperature (T) and light intensity (S) data during actual operation are collected using environmental monitoring equipment, with 25℃ typically selected as the rated ambient temperature. 1000W / m² is the rated light intensity Environmental monitoring equipment can use the SHT30 temperature and humidity sensor and the BH1750 light sensor to ensure data acquisition accuracy.

[0023] 2. When constructing the steady-state equivalent model, a compensation coefficient is set to correct the impact of environmental factors on photovoltaic output; The temperature compensation coefficient 'a' is set to 0.0025, which falls within the range of 0.002-0.003 and is close to the commonly used correction value in the industry; the illumination compensation coefficient 'b' is set to 0.003, which falls within the range of 0.0025-0.0035; and the illumination compensation coefficient 'c' is set to 0.0005, which falls within the range of 0.0004-0.0006. Calculate temperature correction , This represents the difference between the actual ambient temperature and the rated ambient temperature; it is used to calculate the illumination correction. , This represents the difference between the actual light intensity and the rated light intensity. Through formula Calculate the short-circuit current correction factor The open-circuit voltage correction factor is calculated using the following formula. : The corrected photovoltaic short-circuit current is , The corrected photovoltaic short-circuit current is; the corrected photovoltaic open-circuit voltage is... , The corrected photovoltaic open-circuit voltage is used to complete the correction of the steady-state equivalent model.

[0024] Third, when constructing the fault equivalent model, strictly follow the low voltage ride-through requirements to construct the voltage-controlled current source model; First, set the rated voltage of the photovoltaic grid connection point. Determine the reactive current coefficient This value falls within the range of 1.4-1.6 and meets the national standard requirements for reactive power support during photovoltaic low-voltage ride-through; the maximum allowable output current factor... The current should be kept within the range of 1.1-1.3 to avoid damage to photovoltaic power electronic devices due to overcurrent. When the voltage at the photovoltaic grid connection point is detected At that time, control the reference value of the photovoltaic output active current. , Reference values ​​for active current and reactive current output from photovoltaic systems. , This is a reference value for the reactive current output of the photovoltaic system; when At that time, according to the formula Calculate the reference value of reactive current according to Calculate the reference value for active current. This represents the theoretical value of the photovoltaic output active current under normal operating conditions. This represents the maximum active current under photovoltaic overcurrent limiting; when At that time, set , To maximize reactive power support.

[0025] IV. Model Validation: The constructed equivalent model is compared with the actual photovoltaic operation data. Photovoltaic output current and voltage data are collected under different ambient temperatures and light intensities. The error between the corrected model output and the actual output is calculated to ensure that the error is controlled within a reasonable range.

[0026] Traditional photovoltaic equivalent models do not consider dynamic environmental changes, resulting in large output errors. This embodiment corrects for environmental influences through compensation coefficients, reducing model errors and providing an accurate basis for subsequent calculations. The fault model conforms to low voltage ride-through requirements, dynamically adjusts current output, avoids protection adaptation problems caused by photovoltaic fault characteristics, and ensures the accuracy of fault simulation.

[0027] Example 2: This embodiment implements a photovoltaic islanding method, and the specific operation is as follows: 1. Preprocess distribution network data, inputting distribution network topology parameters, photovoltaic access parameters, and fault node information; Distribution network topology parameters include line impedance, node connection relationships, and number of nodes; photovoltaic access parameters include photovoltaic access location, capacity, and output power characteristics; fault node information includes fault node number and fault type; The pre-defined principles for islanding are as follows: the active power balance principle requires that the total active power output of photovoltaics within the island is not less than the total active power demand of the load; the maximum load principle requires that more load nodes be prioritized for islanding to reduce the scope of power outages; the principle of prioritizing the restoration of power to important loads requires that important load nodes such as hospitals and transportation hubs be marked and prioritized for islanding; the connectivity principle requires that nodes assigned to islands must form a seamless network; and the principle of non-faulty island areas requires that faulty nodes and directly related faulty line nodes be excluded to prevent the spread of faults.

[0028] 2. Initialize the breadth-first search algorithm, using a first-in-first-out queue to manage nodes to be visited; each photovoltaic node is sequentially used as the starting node, marked as visited and added to the queue. The visited states are stored in an array, with 1 marked as visited and 0 marked as unvisited; simultaneously, initialize the node ownership state variable. , Let i be the island affiliation state variable for node i. The initial values ​​are all 0, where 1 indicates that the node belongs to an island and 0 indicates that it does not belong to an island.

[0029] 3. Execute the search process: Take the first node from the queue, traverse all adjacent nodes of the node, and determine whether the adjacent nodes are faulty nodes. If the adjacent node number is the same as the faulty node number or the adjacent nodes are connected through the faulty line, then it is determined to be a faulty node. Simultaneously calculate the sum of the total active power of the nodes already included in the island and the active power of the loads of adjacent nodes to determine whether the active power balance principle is satisfied. If the adjacent node is not faulty and meets the active power balance requirement, then mark it as visited and add it to the queue, and update the status accordingly. .

[0030] Fourth, repeat the above search process until the queue is empty, forming a candidate island; generate all candidate islands in sequence, starting with each photovoltaic unit.

[0031] V. Calculate the fitness of candidate islands according to the formula. calculate, Let i be the active power of load i within the island; select the candidate island with the highest fitness as the final island region, and output the island affiliation state variable of each node. .

[0032] Traditional island partitioning often relies on a single principle, making it difficult to balance power supply feasibility and efficiency. This embodiment uses multiple constraints to first ensure power supply feasibility and then maximize power supply efficiency. Breadth-first search ensures island connectivity and fault-free operation, while fitness calculation selects the optimal solution, reducing power outage losses and improving power supply recovery capabilities after faults.

[0033] Example 3: This embodiment implements the construction and solution of the protection point optimization model. The specific operations are as follows: I. Define the objective function and constraints of the protection point optimization model: 1. The objective function is to minimize economic cost, and its expression is: Among them, protecting investment costs Calculated using the equal annual value method, the formula is as follows: M represents the number of protection devices installed, with an initial range from 1 to the number of distribution network lines; denoted as the present value of the investment cost of a single protection device; i is the discount rate, which is taken as 15% with reference to the economic evaluation standard for power engineering; p is the design service life of the protection device, which is taken as 30 years with reference to the industry average lifespan. Protection operation and maintenance costs , The preset proportional coefficient is set to 5%, which falls within the range of 4%-6%, referencing the industry average annual maintenance cost of protection devices as a percentage of investment cost. System power outage losses It is the sum of the products of the unit power outage loss at each load point, the power outage time at the load point, and the active power at the load point.

[0034] 2. The constraint condition is that the average power supply reliability rate (ASAI) of the distribution network is ≥ 99.98%, calculated according to the following formula: The annual rated power supply hours are taken as 8760 hours. Let i be the number of users at load point i. Let be the average annual power outage time at load point i.

[0035] II. Solving the model based on the distribution estimation algorithm: 1. The population generation unit generates an initial population with binary encoding. The length of each individual is equal to the number of distribution network lines. The encoding value 0 indicates that no protection device is installed on the corresponding line, and 1 indicates that a protection device is installed. The population size is set to 50 to balance computational efficiency and optimization accuracy. 2. The fitness assessment unit calculates the economic cost F and the average power supply reliability ASAI of each individual, and selects individuals that meet ASAI≥99.98% to be included in the dominant population; 3. The probability model construction unit calculates the initial probability model. , is the encoding value of the l-th bit of the k-th individual in the dominant population, and N is the number of individuals in the dominant population; According to the formula Update the probability model. For learning efficiency, a value of 0.3 is chosen, placing it within the range of 0.25-0.35, thus balancing model stability and update sensitivity. 4. The sampling iteration unit generates a new population by sampling based on the updated probability model, repeats the fitness evaluation, probability model update, and sampling steps, and iterates until the population fitness converges. The economic cost fluctuation is ≤0.5% for 10 consecutive generations, and the optimal protection placement location is output.

[0036] III. Model Validation: The optimal protection deployment scheme is compared with other deployment schemes to verify the advantages of the optimal scheme in terms of economy and reliability.

[0037] Traditional protection deployment models have a single objective, and optimization algorithms are prone to getting trapped in local optima. This embodiment integrates multiple cost factors to balance economy and reliability. The distribution estimation algorithm iterates at the population level, converges quickly, and optimizes the global distribution to output the optimal deployment scheme, reducing costs while meeting reliability requirements.

[0038] Example 4: This embodiment implements a short-circuit current calculation method, and the specific operation is as follows: 1. Set partition thresholds, with the first threshold being 0.2. The second threshold is 0.9. , The rated voltage of the photovoltaic grid connection point; Voltage at each photovoltaic grid-connected point is detected using a voltage sensor. ,Will The photovoltaic system belongs to the upstream area of ​​the fault, including and Two scenarios; The photovoltaic system belongs to the downstream area of ​​the fault; Simultaneously, the fault point is split into two split nodes. and , For the upstream side, On the downstream side, the split nodes are connected by virtual lines to simulate the current transmission on both sides of the fault point.

[0039] II. Construct the node impedance matrix and the node admittance matrix Y based on the distribution network topology parameters; When i=j Let adj(i) be the set of neighboring nodes of node i. Let be the line impedance between node i and node k; when i ≠ j, if node i and node j are adjacent, then , Let be the line impedance between node i and node j. If they are not adjacent, then... =0; Inverting the nodal admittance matrix Y yields the nodal impedance matrix. , This represents the voltage response of node i when a unit current is injected into node j.

[0040] III. Iteration of the downstream region of the fault: Split nodes As a balancing node, the balancing node voltage is set to the rated voltage. Set the initial iteration parameters of the network within the set region, and set the initial value of the node voltage to . The initial value of photovoltaic output power is set to 50% of the rated power; The initial photovoltaic output power was calculated based on the photovoltaic equivalent model constructed in Example 1, and the normal voltage components of each node in the region were obtained by solving the power flow equations using the Newton-Raphson method. t is the number of iterations; Calculate the initial short-circuit current , This represents the normal component of the fault point voltage during the t-th iteration. The self-impedance of the fault point Grounding impedance; Calculate the voltage fault components of each node using the node impedance matrix , Let i be the mutual impedance between node i and the fault point; The post-fault node voltage is obtained by superimposing the normal component and the fault component. ,according to Adjusting photovoltaic and Repeat the iteration until the convergence condition is met. , Let be the voltage at the photovoltaic grid connection point at node i during the t-th iteration. Let be the voltage at the photovoltaic grid connection point at node i during the (t+1)th iteration. It falls within the range of 0.0008-0.0012 pu.

[0041] IV. Data Transmission: The photovoltaic output power in the downstream area is calculated according to the formula. Transform into a constant impedance model , Let be the grid connection voltage of the j-th photovoltaic cell in the downstream area. and Let j be the active and reactive power output of the j-th photovoltaic cell in the downstream region, superimposed with all... Obtain the total constant impedance ; Power flow calculations are performed on the downstream normal component network to obtain the balanced nodes. Power transmitted into the area Transformed into a constant impedance model , For balancing nodes The normal voltage component, Let S be the conjugate complex number; Will and After superposition, it is added to the split node. .

[0042] V. Iteration of the upstream region of the fault: Split nodes after adding a constant impedance model For the equilibrium node, repeat the iterative process of step 3 until the convergence condition is met, and obtain the final short-circuit current calculation result.

[0043] Traditional global iteration is computationally intensive, inefficient, and prone to errors. This embodiment divides the system into voltage drop zones and focuses on iterating regions with similar characteristics, thereby improving computational efficiency. It incorporates photovoltaic models and node impedance matrices to accurately reflect fault current characteristics, reduce computational errors, and provide reliable data for protection settings.

[0044] Example 5: This embodiment implements adaptive differential protection criteria, and the specific operation is as follows: I. Classify the types of protected areas and set parameters, detect whether each protected area contains photovoltaic access, and clarify the branch types, including branches without undetected branches, branches with loads, and branches with inverter-type distributed power sources; Adjust the range of values ​​for the braking coefficient K: when there are no undetected branches, take 0.3 (range 0.2-0.45); when there are load branches, take 0.25 (range 0.2-0.35); when there are inverter-type distributed power supply branches, take 0.28 (range 0.22-0.34). Set minimum starting current Between 0.1 and 0.2 scope, This is the rated current of the line.

[0045] II. Application of Differential Protection Criteria: 1. When there is no photovoltaic grid connection within the protected area, a criterion with a braking coefficient is adopted, the expression of which is: , and The current phasors measured by the protection at both ends of the line include amplitude and phase information; If the condition is met, the fault is determined to be within the zone, and the protection system is activated; otherwise, the fault is determined to be outside the zone, and the protection system does not activate. 2. When photovoltaic power is connected within the protected area, the setting value criterion based on the short-circuit current calculation result is adopted, and the expression is: , The setting value for the protection on both sides of line mn is taken as the maximum differential current when a three-phase metallic short-circuit fault occurs at the outlet outside the protection zone; if this formula is satisfied, it is determined to be an internal fault and the protection will operate.

[0046] III. Introducing Positive and Negative Sequence Fault Components as Auxiliary Criteria: Asymmetrical faults employ negative sequence current amplitude and negative sequence current phase criteria. The negative sequence current amplitude criterion requires that the negative sequence current amplitudes at both ends of the line be equal. The negative sequence current phase criterion requires that the phase difference between the negative sequence currents at both ends of the line be ≤30°; for symmetrical faults, the positive sequence fault current amplitude criterion is used, requiring that the amplitudes of the positive sequence fault currents at both ends of the line be equal. .

[0047] IV. Special criteria for motor self-starting operation: Introducing voltage blocking criteria , The line voltage difference between any two phases in the distribution network. Take 0.7 Between 0.65 and 0.75 scope; Introducing the inverse time limit criterion , and These represent the amplitudes of the positive-sequence fault current phasors at both ends of the line. Take 1.5 times the rated current of the motor. A value of 1.0 falls within the range of 0.8-1.2. A value of 1.0 falls within the range of 0.9-1.1. A value of 0.25 falls within the range of 0.2-0.3.

[0048] Traditional differential protection criteria are fixed and cannot adapt to current changes after photovoltaic access, making it prone to false tripping or failure to trip. This embodiment dynamically adjusts the criteria according to the type of protection zone, combines positive and negative sequence components to help identify faults, and designs special criteria for motor self-starting, effectively reducing the protection false tripping rate and improving protection reliability.

[0049] Example 6: I. System Hardware Configuration and Module Design: Photovoltaic equivalent modeling module: It adopts an industrial control computer with an Intel Core i7-12700 CPU, 32GB of memory, and a 1TB SSD. It is equipped with a MATLAB / Simulink 2022b simulation platform, outputs photovoltaic output current and voltage characteristic parameters, and the calculation cycle is 100ms. Islanding module: Employs an embedded processor, a quad-core ARM Cortex-A57 with a clock speed of 1.43GHz, and implements the search logic using Python. It takes as input distribution network topology parameters and fault node information, and outputs islanded areas and node affiliation status variables. Processing delay ≤ 50ms; The protection deployment optimization module includes a population generation unit, a fitness evaluation unit, a probability model construction unit, and a sampling iteration unit. The population generation unit uses an FPGA chip and generates binary codes using Verilog HDL. The fitness evaluation unit uses a DSP processor with 32-bit floating-point precision. The probability model construction unit uses an MCU microcontroller. The sampling iteration unit works in conjunction with the FPGA chip to output the optimal protection deployment location, with an overall computational latency of ≤200ms. Short-circuit current calculation module: It adopts a dedicated digital signal processor, implements matrix inversion and iterative logic in C language, and outputs short-circuit current calculation results with a calculation delay of ≤10ms; Differential protection setting module: It adopts a programmable logic controller, stores protection criterion parameters, performs criterion logic operations, outputs differential protection setting values, and has a response time of ≤5ms; Data storage module: adopts industrial-grade solid-state drive with a capacity of 2TB, stores various parameters and results, supports data read and write speed ≥500MB / s, and data retention period ≥10 years; Communication module: An Ethernet communication module is adopted, which supports a transmission rate of 100Mbps and a transmission delay of no more than 10ms. The communication protocol adopts Modbus TCP to realize data interaction between modules.

[0050] II. System Operation Flow: Initialization phase: The data storage module pre-imports various parameters, and each module synchronizes the parameters through the communication module; Real-time operation phase: The photovoltaic equivalent modeling module collects environmental data and calculates photovoltaic parameters, and sends them to relevant modules; the islanding module receives fault information and performs a search; the protection layout optimization module outputs the optimal layout; the short-circuit current calculation module performs partition iteration; the differential protection setting module outputs setting values ​​and sends them to the protection execution device. Fault handling phase: When a fault occurs, the short-circuit current module outputs the result first, the protection setting module quickly calculates and outputs the setting value, the protection actuator operates, and the data storage module records the fault information.

[0051] III. System Verification: Build a simulation test platform to simulate various working conditions, verify the system output accuracy, response time, and false alarm rate, and ensure that the requirements of power distribution network protection are met.

[0052] Traditional protection systems are decentralized, have slow data interaction, and a high false alarm rate. This system integrates multiple module functions, achieves real-time data interaction through high-speed communication, and selects hardware to adapt to the operational needs of each module, ensuring fast system response, high accuracy, and reduced false alarm rate, thus providing reliable protection support for photovoltaic grid connection.

[0053] In summary, the method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access constructs an adaptive differential protection criterion through a differential protection setting module. It first determines whether there is photovoltaic access within the protection zone to dynamically adjust the criterion type. Combined with positive and negative sequence fault components, it assists in accurately identifying fault types. For motor self-starting conditions, it introduces voltage blocking and inverse time criterion to adapt to the nonlinear changes in short-circuit current after photovoltaic access, avoids false triggering under special conditions, and effectively reduces the protection maloperation rate.

[0054] Furthermore, the method and system for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access utilizes a photovoltaic equivalent modeling module to construct steady-state and fault equivalent models. Combined with the partitioned iterative algorithm adopted by the short-circuit current calculation module, it divides the region into iterative calculations according to the voltage drop characteristics of the photovoltaic grid connection point. It incorporates the logic of constructing a node impedance matrix to accurately reflect the characteristics of photovoltaic fault current, replacing the traditional global iteration, improving calculation efficiency and accuracy, and providing a reliable basis for dynamic adjustment of protection criteria. This solves the problems of insufficient adaptability of existing distribution network differential protection schemes, which leads to high protection maloperation rate and inefficient short-circuit current calculation that makes it difficult to support setting.

[0055] The relevant modules involved in this system are all hardware system modules or functional modules that combine computer software programs or protocols with hardware in the prior art. The computer software programs or protocols involved in these functional modules are technologies known to those skilled in the art and are not improvements to this system. The improvement of this system lies in the interaction or connection between the modules, that is, in improving the overall structure of the system to solve the corresponding technical problems that this system aims to address.

[0056] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic (PV) grid integration, characterized in that, Includes the following steps: Step 1: Construct a steady-state equivalent model and a fault equivalent model for the photovoltaic power source. The steady-state equivalent model corrects the output current by incorporating compensation coefficients for ambient temperature and light intensity. The fault equivalent model constructs a voltage-controlled current source model based on low voltage ride-through requirements. Step 2: Perform photovoltaic islanding based on breadth-first search algorithm. Input distribution network topology parameters, photovoltaic access parameters and fault node information. Starting with each photovoltaic node, search according to preset principles to form candidate islands, and select the candidate island with the highest fitness. Step 3: Establish a protection deployment optimization model and solve it using a distribution estimation algorithm. The model aims to minimize economic cost and is constrained by ensuring that the average power supply reliability of the distribution network is not lower than a preset threshold. Step 4: Based on the voltage drop at the photovoltaic grid connection point, the distribution network is divided into upstream and downstream fault regions, and the fault point is split into two split nodes. The short-circuit current is calculated using a partitioned iterative algorithm. Step 5: Construct adaptive differential protection criteria. Depending on whether there is photovoltaic access in the protection zone, adopt criteria with braking coefficient or setting value criteria based on short-circuit current calculation results. Introduce positive and negative sequence fault components to assist in identifying fault types. Introduce special criteria for motor self-starting conditions.

2. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The compensation coefficients of the steady-state equivalent model of the photovoltaic power source mentioned in step 1 include a temperature compensation coefficient a, a light compensation coefficient b, and c. The value of a ranges from 0.002 to 0.003, the value of b ranges from 0.0025 to 0.0035, and the value of c ranges from 0.0004 to 0.0006. The temperature correction is the difference between the actual ambient temperature and the rated ambient temperature, and the light correction is the difference between the actual light intensity and the rated light intensity. The corrected photovoltaic short-circuit current is obtained by multiplying the compensation coefficients by the rated short-circuit current, and the corrected photovoltaic open-circuit voltage is obtained by multiplying the compensation coefficients by the rated open-circuit voltage. The fault equivalent model satisfies: When the voltage at the photovoltaic grid connection point At that time, the reference value of photovoltaic output active current reactive current reference value ; when hour, , , ; when hour, , ; in The rated voltage of the photovoltaic grid connection point. This is the rated output current of the photovoltaic system. This is a reference value for the active power output of the photovoltaic system. This represents the reactive current coefficient, with a value between 1.4 and 1.

6. This is the maximum allowable coefficient for output current, and its value ranges from 1.1 to 1.

3.

3. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The preset principles of the breadth-first search algorithm described in step 2 include the active power balance principle, the maximum load principle, the principle of prioritizing the restoration of power supply to important loads, the connectivity principle, and the principle of non-faulty island areas. The fitness calculation formula for candidate islands is as follows: , Let i be the active power of load i within the island. Let be the island affiliation state variable for node i, where 1 indicates that the node belongs to an island, and 0 indicates that the node does not belong to an island; the specific algorithm flow is as follows: Step 2.1: Mark the starting photovoltaic node as visited and add it to the queue; Step 2.2: Take a node from the queue, visit all the neighboring nodes of the node, and determine whether the neighboring nodes are faulty nodes and whether the active power balance principle is satisfied; Step 2.3: Mark the adjacent nodes that meet the conditions as visited and add them to the queue; Step 2.4: Repeat steps S2-S3 until the queue is empty, forming candidate islands; Step 2.5: Calculate the fitness of all candidate islands and select the candidate island with the highest fitness as the final island region.

4. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The economic cost of the protection deployment optimization model described in step 3 includes protection investment cost, protection operation and maintenance cost, and system power outage loss cost; The cost of protecting investment is calculated using the equivalent annual value method, and the formula is as follows: M represents the number of protection devices installed. Let i be the present value of the investment cost of a single protection device, i be the discount rate, and p be the design service life of the protection device. The protection operation and maintenance cost is the product of the protection investment cost and the preset proportional coefficient, which ranges from 4% to 6%; the system power outage loss cost is the sum of the products of the unit power outage loss at each load point, the power outage time at the load point, and the active power at the load point. The formula for calculating the average power supply reliability of a distribution network is as follows: , The annual rated power supply hours are taken as 8760 hours. Let i be the number of users at load point i. The average annual power outage time for load point i is set to a preset threshold of 99.98%.

5. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The partitioning thresholds for the voltage drop at the photovoltaic grid connection point mentioned in step 4 include a first threshold and a second threshold. The first threshold is 0.2U_N, and the second threshold is 0.9U_N. When the photovoltaic grid connection point voltage is higher than the second threshold or between the first and second thresholds, the photovoltaic unit belongs to the upstream region of the fault; when the photovoltaic grid connection point voltage is lower than the first threshold, the photovoltaic unit belongs to the downstream region of the fault. The specific process of the partitioning iterative algorithm is as follows: Step 4.1: Iteration in the downstream region of the fault: Taking one of the split nodes as the equilibrium node, set the initial iteration parameters of the network in the region, calculate the initial output power of the photovoltaic according to the photovoltaic equivalent model, perform normal component network power flow calculation to obtain the normal component of the voltage of each node in the region, solve the initial short-circuit current, calculate the fault component of the voltage of each node using the node impedance matrix, superimpose the normal component and the fault component to obtain the node voltage after the fault, update the photovoltaic state, and repeat the iteration until the convergence condition is met. Step 4.2: Data transmission: Convert the photovoltaic output power in the downstream area into a constant impedance model, calculate the power transmitted from the balancing node to the area and convert it into a constant impedance model, and then superimpose the two constant impedance models and apply them to another split node. Step 4.3: Iteration of upstream region of fault: Using the split node after adding constant impedance model as the balancing node, repeat the iteration process of step S1 to obtain the final short-circuit current.

6. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The braking coefficient value range of the adaptive differential protection criterion described in step 5 is adjusted according to the branch type in the protection zone: 0.2-0.45 when there are no undetected branches, 0.2-0.35 when there are load branches, and 0.22-0.34 when there are inverter-type distributed power supply branches. The criterion with braking coefficient is in the form of: , , These are the current phasors measured by the protection at both ends of the line, and K is the braking coefficient. This is the minimum starting current, and its value ranges from 0.1 to 0.

2. ; The setting value criterion based on the short-circuit current calculation results is as follows: , These are the setting values ​​for the protection on both sides of line mn, determined by the short-circuit current calculation results; In the auxiliary criteria for positive and negative sequence fault components, the negative sequence current amplitude criterion and the negative sequence current phase criterion are used for asymmetrical faults, while the positive sequence fault current amplitude criterion is used for symmetrical faults.

7. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 1, characterized in that, The special criteria introduced in step 5 for the self-starting condition of the motor include voltage blocking criteria and inverse time criterion; Voltage blocking criterion is , The line voltage difference between any two phases in the distribution network. The voltage setting value is 0.65-0.75U_N; The formula for calculating the inverse time limit criterion is as follows: , , These are the positive-sequence fault current phasors at both ends of the line. This is the threshold value for the inverse time protection starting current. This is the curve shape coefficient, and its value ranges from 0.8 to 1.

2. This is the horizontal translation coefficient for the curve, and its value ranges from 0.9 to 1.

1. It is the inverse time limit protection time constant with a value of 0.2-0.

3.

8. The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access according to claim 5, characterized in that, The convergence condition of the partitioning iterative algorithm described in step 4 is: , Let be the voltage at the photovoltaic grid connection point at node i during the t-th iteration. Let be the photovoltaic grid-connected voltage at node i in the (t+1)th iteration, and ε be the convergence threshold with a value of 0.0008-0.0012pu. The node impedance matrix is ​​obtained by inverting the node admittance matrix constructed from the distribution network topology parameters. In the node admittance matrix, when i=j... Let k be the neighboring nodes of node i. Let be the line impedance between node i and node k; When i≠j , Let be the line impedance between node i and node j.

9. A differential protection maloperation suppression system for distribution networks under distributed photovoltaic access, characterized in that, The method for suppressing maloperation of differential protection in distribution networks under distributed photovoltaic access as described in claims 1-8 includes a photovoltaic equivalent modeling module, an islanding module, a protection point optimization module, a short-circuit current calculation module, a differential protection setting module, a data storage module, and a communication module; each module interacts with data through the communication module, which has a transmission rate of 100Mbps and a transmission delay of no more than 10ms. The photovoltaic equivalent modeling module outputs photovoltaic output current characteristic parameters and voltage characteristic parameters; The island division module outputs the final island region and the island affiliation status of each node; The protection point optimization module outputs the optimal protection point location; The short-circuit current calculation module outputs the short-circuit current calculation result; The differential protection setting module outputs the differential protection setting value; The data storage module is used to store distribution network topology parameters, photovoltaic access parameters, fault data, calculation results, and differential protection criterion parameters.

10. The differential protection maloperation suppression system for distribution networks under distributed photovoltaic access according to claim 9, characterized in that, The protection deployment optimization module includes a population generation unit, a fitness evaluation unit, a probability model construction unit, and a sampling iteration unit. The population generation unit generates an initial population with binary codes. The length of each individual is equal to the number of distribution network lines. The code value 0 indicates that the corresponding line does not have a protection device installed, and the code value 1 indicates that the corresponding line has a protection device installed. The fitness evaluation unit calculates the economic cost of each individual in the initial population and the average power supply reliability of the distribution network, and selects dominant individuals that meet the constraints of the average power supply reliability of the distribution network. The probability model construction unit calculates the initial probability model. and updated to , Let N be the encoded value of the l-th bit of the k-th individual in the dominant population, where N is the number of individuals in the dominant population. For learning efficiency, the value is 0.25-0.35; The sampling iteration unit generates a new population based on the updated probability model, iterates until the population fitness converges, and outputs the optimal protection point location.

Images