A station master-slave differential protection method and system for a photovoltaic power station

Abstract

The application discloses a kind of station in and station main-slave differential protection method and system for photovoltaic power station in the technical field of relay protection, comprising: according to the manufacturer of component, capacity and topological structure, power station line is divided into multiple functional divisions, and based on the ratio of node real-time voltage and rated voltage, functional division is further divided into multiple voltage divisions;The current difference of the line end is collected by the main differential protection module as the basis for action;Slave differential protection module is based on the random forest model that is trained in advance, combined with the current collected voltage, current, line length, equipment operating life and other characteristics, the fault probability in voltage division is calculated;According to the relationship between the fault probability and the set threshold value, it is decided whether to allow the main differential protection action, delay action or lockout.The system effectively solves the problem of traditional differential protection misoperation and refusal caused by current limiting characteristics in photovoltaic power station through main-slave linkage, significantly improves the sensitivity and reliability of protection device.

Description

Technical Field

[0001] This invention relates to a master-slave differential protection method and system for photovoltaic power plants, belonging to the field of relay protection technology. Background Technology

[0002] With the rapid development of photovoltaic (PV) power generation globally, its proportion in the power system continues to increase, and PV power plants are gradually becoming one of the important power source types in the power grid. Compared with traditional thermal power and hydropower generation methods, PV power plants have the characteristics of controllable current output, distributed system structure, and a high proportion of power electronic interfaces. However, due to its unique control characteristics and the current-limiting behavior of the inverter, it exhibits short-circuit characteristics that are significantly different from those of traditional power generation systems in actual operation.

[0003] In relay protection, traditional differential protection methods are widely used in transmission lines, transformers, and other scenarios. Their basic principle is to compare the difference in current across the line; when the differential current exceeds a set threshold, it is determined to be an internal line fault and triggers the protection action. However, in photovoltaic power plant scenarios, due to their current-limiting control characteristics, when a short-circuit fault occurs, the fault current amplitude is usually low and the rise rate is slow. This means the differential current may be insufficient to meet the operating threshold of traditional differential protection, easily leading to failure to operate. Furthermore, lowering the threshold value to improve sensitivity may result in false tripping under non-fault conditions such as normal system fluctuations or load switching.

[0004] Furthermore, most existing differential protection schemes are based on static threshold judgments, lacking the comprehensive ability to identify the regional structure, electrical topology, and equipment operating status of photovoltaic power plants. Especially in practical engineering applications, due to the large scale and complex structure of photovoltaic power plants, with different component types, capacity levels, and communication topologies coexisting, traditional differential protection schemes based on uniform settings for the entire plant are difficult to adapt to the differentiated protection needs of different areas, and the protection strategies lack specificity and adaptability.

[0005] With limited fault current, traditional differential protection carries the risk of misjudgment and failure to operate.

[0006] Static threshold tuning lacks adaptability to the real-time state of the system;

[0007] There is a lack of intelligent auxiliary judgment mechanisms based on regional division and data-driven approaches;

[0008] In complex topologies, traditional protection schemes struggle to quickly and accurately locate fault areas.

[0009] Therefore, there is an urgent need to propose a new master-slave differential protection method and system that is suitable for photovoltaic power plant scenarios and has regional identification capabilities and fault probability auxiliary judgment capabilities, so as to improve the accuracy, sensitivity and reliability of protection actions. Summary of the Invention

[0010] The purpose of this invention is to overcome the shortcomings of the existing technology. Existing traditional differential protection has the risk of misjudgment and failure to operate, lacks adaptability to the real-time status of the system, and is difficult to quickly and accurately locate the fault area.

[0011] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution:

[0012] Firstly, a method for master-slave differential protection within a photovoltaic power station is provided, comprising the following steps:

[0013] Based on the component information in the photovoltaic power station, the internal power lines of the power station are divided into functional zones;

[0014] Voltage zones are divided based on the ratio of real-time voltage to rated voltage of nodes within the functional zone.

[0015] The voltage zones are numbered according to the ratio of real-time voltage to rated voltage, and the ratio of real-time voltage to rated voltage of each voltage zone is inversely proportional to its number.

[0016] Based on the pre-trained random forest model, combined with the currently collected fault feature data, the fault probability in each voltage zone and each functional zone is calculated.

[0017] The fault probability is compared with a preset threshold value to control the authority of the main differential protection action: the number size is proportional to the priority order;

[0018] When measuring the voltage zone or functional zone, a first threshold value and a second threshold value are set, wherein the first threshold value is greater than the second threshold value.

[0019] When the measured fault probability is higher than the first threshold, the differential operation protection condition is met, and the main differential protection is allowed.

[0020] When the measured fault probability is within the range of the first threshold value and the second threshold value, the differential operation protection condition is established after a delay, and the main differential protection performs the operation after the delay; a threshold range is set between the first threshold value and the second threshold value, and a 1-second delay is added within each range.

[0021] When the measured fault probability is lower than the second threshold, the differential protection condition is not met, and the main differential protection is locked out.

[0022] When measuring the functional zones, a limit threshold is set. When the measured fault probability is higher than the limit threshold, the first and second threshold values ​​of the voltage zone are both reduced by 10%.

[0023] Furthermore, the functional zoning of the power lines inside the power station includes: dividing the functional zones according to the type of photovoltaic modules, the capacity of the devices, the connection relationship of the collector wires in the communication topology, and the substation busbar wiring diagram.

[0024] Furthermore, the random forest model includes: constructing a training dataset from historical operating data of photovoltaic power plants. The training dataset includes fault samples and non-fault samples within each functional partition, and the feature parameters in each sample include at least: line length, equipment operating years, effective voltage value, negative sequence voltage ratio, zero sequence voltage mutation, total harmonic distortion rate, and peak current ratio.

[0025] Furthermore, the feature parameters are preprocessed, including:

[0026] The line length is normalized using Min-Max.

[0027] The service life of the equipment is discretized into multiple levels according to a set threshold.

[0028] The effective voltage value, negative sequence voltage ratio, zero sequence voltage abrupt change, total harmonic distortion rate, and peak current ratio were all directly extracted from historical data.

[0029] Standardize all feature parameters to the range of 0 to 1.

[0030] Furthermore, a bootstrap sampling method is used to generate multiple training subsets from the training dataset. Each training subset is used to train an unpruned decision tree, and the node splitting of the decision tree uses Gini impurity as the splitting criterion.

[0031] The final trained decision trees are used to form the random forest model.

[0032] The differential protection unit updates a portion of the decision trees in the random forest model at set intervals.

[0033] Furthermore, when the functional zone's differential operation condition is met, but the voltage zone's differential operation protection condition is met after a delay or is not met, the main differential protection will execute the operation after the delay.

[0034] Furthermore, the voltage zones are numbered according to the ratio of real-time voltage to rated voltage, including:

[0035] A node is classified as the first partition when the ratio of its real-time voltage to its rated voltage is greater than 0.95.

[0036] A node is classified as belonging to the second partition when the ratio of its real-time voltage to its rated voltage is between 0.85 and 0.95.

[0037] A node is classified as the third partition when the ratio of its real-time voltage to its rated voltage is less than 0.85.

[0038] Furthermore, the fact that the number size and priority order are proportional includes:

[0039] The fault probability calculation is performed on the voltage zones in descending order of their numbers, and the calculation stops when the condition for differential setting is met.

[0040] Secondly, an in-station master-slave differential protection system for photovoltaic power plants is provided, including:

[0041] The area division module is used to divide the lines within the photovoltaic power station into functional zones based on the equipment manufacturers, batches, capacity, and topology, and to divide the functional zones into voltage zones based on the ratio of the real-time voltage to the rated voltage of the nodes.

[0042] The main differential protection module is used to collect and compare the current at both ends of the line to determine whether the differential current exceeds the set action value, so as to determine whether there is a short circuit fault.

[0043] The fault probability calculation module is used to calculate the real-time fault probability of each voltage zone based on the random forest model trained from the historical fault data of the photovoltaic power station, combined with the real-time collected electrical and non-electrical quantity characteristics.

[0044] The differential protection module is used to authorize and control the operation of the main differential protection module based on the real-time fault probability results.

[0045] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0046] This invention solves the problem of maloperation or failure to operate caused by current-limiting characteristics in traditional differential protection in photovoltaic power plants by introducing a master-slave differential protection collaborative mechanism. Specifically, by functionally partitioning the internal lines of the power plant based on component manufacturer identification, capacity level, and topological connection relationships, clear spatial positioning of the structure can be achieved. Furthermore, voltage partitioning is performed based on the ratio of real-time voltage to nameplate rated voltage at each node, enabling the system to dynamically identify local voltage variation characteristics under different operating conditions. On this basis, a random forest model trained on historical data is used in the differential protection unit to integrate electrical and non-electrical quantities for fault probability assessment, enhancing the identifiability of regional faults and the targeted nature of the response. The master differential protection is only allowed to trip or delay when the fault probability in the corresponding area reaches a set threshold, effectively reducing the difficulty of setting differential settings, improving the sensitivity and reliability of the action, significantly suppressing maloperation and failure to operate, and improving the adaptability and safety stability of the photovoltaic power plant relay protection system in complex operating environments. Attached Figure Description

[0047] Figure 1 The diagram shown is a regional division map of a photovoltaic power station provided in Embodiment 1 of the present invention;

[0048] Figure 2 The diagram shown is a flowchart of a random forest provided in Embodiment 1 of the present invention;

[0049] Figure 3 The diagram shown is an example of a decision tree provided in Embodiment 1 of the present invention;

[0050] Figure 4 The diagram shown is a flowchart of the fault probability generation process provided in Embodiment 1 of the present invention.

[0051] Figure 5 The diagram shown is a master-slave differential flowchart provided in Embodiment 1 of the present invention. Detailed Implementation

[0052] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0053] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0054] Example 1:

[0055] This embodiment provides a master-slave differential protection method for photovoltaic power plants, such as... Figure 1The diagram illustrates a photovoltaic (PV) power station. Based on the component information, the internal power lines are functionally divided. First, identical lines are divided into n functional zones according to the topology diagram. An identical line is defined as a set of equipment of the same capacity produced in the same batch by the same manufacturer. For example, one PV line consists of 10 PV panels from manufacturer A, one inverter from manufacturer B, and a cable of almost the same length from manufacturer C. Another line also consists of the same number and capacity of PV panels and inverters, as well as the same length of cable, produced in the same batch by the above three manufacturers. These two lines can be considered identical. If one component differs from the other, the two lines are not considered identical. Furthermore, only identical lines directly connected to the same collector wire in the topology diagram can be classified into the same functional zone. In the accompanying diagram, for simplicity, each PV panel represents more than one unit, and it is assumed that the lines on each collector wire are identical. Therefore, the system can be divided into n functional zones as shown in the diagram.

[0056] The voltage zones are numbered according to the ratio of real-time voltage to rated voltage, and the ratio of real-time voltage to rated voltage of each voltage zone is inversely proportional to its number.

[0057] Specifically, the functional zones are fixed and can be entered into the computer and saved after construction. Then, the computer divides each functional zone into voltage zones based on the ratio of the real-time voltage to the rated voltage of the nodes within that zone. Specifically, it calculates the ratio of the real-time voltage to the rated voltage of each node and uses this ratio to divide the functional zone into voltage zones. During normal operation, the equipment mostly operates at its rated state; therefore, nodes with a ratio greater than 0.95 can be assigned to the same voltage zone, such as voltage zone 1.1; nodes with a ratio greater than 0.85 but not greater than 0.95 can be assigned to voltage zone 1.2; nodes with a ratio not greater than 0.85 can be assigned to voltage zone 1.3, and so on. Only areas 1.a and nb are shown in the diagram, where a and b only represent the voltage zone number. The nodes included in these areas will change in real time according to changes in node voltage. The larger the voltage zone number, the lower the node voltage, the higher the probability of it being a faulty area, and the higher its priority in fault identification. In the fault probability calculation, the voltage zones are calculated in descending order of their numbers. When the probability requirement of the differential setting is met, the main differential is allowed to operate. If the selected operating area of ​​the main differential is within this range, it can operate immediately without needing to calculate the fault probability of other voltage zones. Although the probability calculation requires some time, the current-limiting characteristics of the photovoltaic power station mean that the current during a short-circuit fault will not be very large. Fault clearing can be achieved through a certain delay. Furthermore, although the calculations required by this invention require computer processing time, the calculation speed can be increased through the use of algorithms and improvements in equipment performance, thus not affecting normal operation.

[0058] like Figure 2 and Figure 3 As shown, after region division, fault features are extracted from the historical fault dataset of photovoltaic power plants as the training set for the random forest algorithm to calculate fault probabilities. The algorithm training and usage steps are as follows:

[0059] First, the required characteristic parameters are obtained based on the historical operating data of the photovoltaic power station. These characteristic parameters include: line length, equipment operating years, effective voltage value, negative sequence voltage ratio, zero sequence voltage abrupt change, total harmonic distortion rate, and peak current ratio. These characteristic parameters include both static and dynamic characteristics.

[0060] For static feature preprocessing: Equipment usage time: Discretized according to industry standards into three levels: [0-5 years), [5-10 years), and ≥10 years, with values ​​of 0, 0.5, and 1 respectively; Line length: Min-Max normalization is used, where the minimum value is 0 and the maximum value is the maximum length of a single line within the station, which is also between 0 and 1.

[0061] For dynamic electrical measurement feature extraction: voltage features: three-phase voltage RMS value, negative sequence voltage ratio, zero sequence voltage abrupt change; current features: total harmonic distortion (THD), peak current ratio, etc. All of the above feature data are directly extracted from historical data and processed into numbers between 0 and 1. The number of extracted fault feature data can be reduced or increased according to the actual situation. For simplicity, this embodiment assumes the extraction of the above two non-electrical quantity features and five electrical quantity features.

[0062] like Figure 4 and Figure 5 As shown, the construction of the random forest model includes: Bootstrap aggregating (Bagging) sampling: Assuming there are N=10 historical dataset records, 5 fault datasets, and 5 normal datasets, each dataset contains the required 7 features. The historical datasets are sampled M times with replacement, forming K=50 subsets of training data. Each subset retains approximately 63.2% of the original data, with the remaining 36.8% used as OOB validation samples. It should be noted that the training data used in this embodiment is relatively small, but for ease of description, the values ​​are very small; the data requirements can be modified according to actual circumstances.

[0063] Set the number of candidate features when each tree splits. ; The feature number is 7 in this embodiment, meaning 2-3 features are randomly selected from the 7 features for node splitting to generate a decision tree. To improve accuracy, feature values ​​can be obtained from both electrical and non-electrical quantities. Starting from the randomly selected features, according to the pre-defined partitioning, such as... Figure 3 As shown, the splitting criteria Y and Z are self-generated random numbers between 0 and 1. Gini impurity is used as the splitting criterion, allowing the tree to grow completely until the number of leaf node samples is <5. If the above requirement is not met after two splits, Y and Z are regenerated randomly, or they can be manually adjusted until the condition is met. Fifty unpruned decision trees are independently generated for each sub-training set. In this embodiment, the samples in a node have two different categories (faulty, normal), let the proportion of samples in each category be , and the formula for calculating the Gini index is: Ultimately, there will inevitably be one type of normal and another type of faulty compartments.

[0064] One point worth mentioning is that the real-time acquired voltage / current data is processed into numbers between 0 and 1 using the same feature extraction process. The specific processing method depends on the site requirements and can be as reasonable as possible. Static parameters are discretized according to the binning rules of the training set.

[0065] Output the classification results of the dataset under each decision tree. A fault is assigned a value of 1, and a normal value is assigned a value of 0. The final fault probability value is then calculated based on the results. (Average probability of 50 trees). Input a dataset into 50 decision trees generated using the above method; each decision tree will naturally produce one result. Assuming a dataset ultimately produces 40 faulty outputs and 10 normal outputs in this random forest decision tree, the probability of the dataset corresponding to a faulty state is 80%.

[0066] This embodiment integrates 50 differentiated decision trees, which can capture the complex interaction between long-term factors such as equipment aging and instantaneous electrical characteristics, and suppress the risk of overfitting from a single tree through a probabilistic averaging mechanism. In actual deployment, it is recommended to update 10% of the trees with new fault data every quarter to maintain the model's timeliness.

[0067] Specifically, when expanding and upgrading the data from a single node to a certain area, only the addition of feature data is needed to obtain the fault probability within a specific voltage zone or functional zone under a certain state. It can be set that when the fault probability within a voltage zone is greater than 80%, the main differential protection is allowed to operate directly; when the fault probability within a voltage zone is between 60% and 80%, the main differential protection is allowed to operate after a 1-second delay, with the range set at 20%, meaning a 1-second delay is added for every 20% increase; when the fault probability within a voltage zone is less than 20%, the main differential protection is prohibited from operating within that voltage zone. However, when the fault probability within a functional zone is greater than 90%, the fault probability requirement for the aforementioned voltage zone is reduced by 10%.

[0068] It's worth mentioning that the differential protection system calculates the probability of a fault existing in different areas of the entire photovoltaic power station based on various measurement data. These areas are divided into functional zones and voltage zones. When the current fault probability in a zone exceeds a certain value, the main differential protection in that zone will activate. If the probability is within a certain range, the main differential protection in that zone can activate after a certain delay. If the probability is below a certain value, the main differential protection in that zone will be blocked. This probability threshold can be slightly lower for functional zones and slightly higher for voltage zones. If a functional zone meets the differential protection activation conditions while a voltage zone does not, a certain delay can be allowed before the main differential protection can activate. The main differential protection operates similarly to traditional differential protection, constantly monitoring the differential current and operating current. By combining main and slave differential protection, the main differential protection can reduce its operating current setting to a certain extent, improving protection sensitivity without increasing its maloperability.

[0069] Example 2:

[0070] This embodiment provides an in-station master-slave differential protection system for photovoltaic power plants, including:

[0071] The area division module is used to divide the lines within the photovoltaic power station into functional zones based on the equipment manufacturers, batches, capacity, and topology, and to divide the functional zones into voltage zones based on the ratio of the real-time voltage to the rated voltage of the nodes.

[0072] The main differential protection module is used to collect and compare the current at both ends of the line to determine whether the differential current exceeds the set action value, so as to determine whether there is a short circuit fault.

[0073] The fault probability calculation module is used to calculate the real-time fault probability of each voltage zone based on the random forest model trained from the historical fault data of the photovoltaic power station, combined with the real-time collected electrical and non-electrical quantity characteristics.

[0074] The differential protection module is used to authorize and control the operation of the main differential protection module based on the real-time fault probability results.

[0075] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for in-station master-slave differential protection of photovoltaic power plants, characterized in that: Includes the following steps: Based on the component information in the photovoltaic power station, the internal power lines of the power station are divided into functional zones; Voltage zones are divided based on the ratio of real-time voltage to rated voltage of nodes within the functional zone. The voltage zones are numbered according to the ratio of real-time voltage to rated voltage; Based on the pre-trained random forest model, combined with the currently collected fault feature data, the fault probability in each voltage zone and each functional zone is calculated. The fault probability is compared with a preset threshold value to control the authority of the main differential protection action: specifically, the number size and priority order are directly proportional; When the measured fault probability is higher than the first threshold, the differential operation protection condition is met, and the main differential protection is allowed. When the measured fault probability is within the range of the first threshold value and the second threshold value, the differential operation protection condition is established after a delay, and the main differential protection performs the operation after the delay; the first threshold value is greater than the second threshold value, and a threshold range is set between the first threshold value and the second threshold value, with an additional 1 second delay in each range. When the measured fault probability is lower than the second threshold, the differential protection condition is not met, and the main differential protection is locked out.

2. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, The functional zoning of the power lines inside the power station includes dividing the functional zones according to the type of photovoltaic modules, the capacity of the devices, the connection relationship of the collector wires in the communication topology, and the substation busbar wiring diagram.

3. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, The random forest model includes: constructing a training dataset from historical operating data of photovoltaic power plants. The training dataset includes fault samples and non-fault samples in each functional partition, and the feature parameters in each sample include at least: line length, equipment operating years, effective voltage value, negative sequence voltage ratio, zero sequence voltage mutation, total harmonic distortion rate, and peak current ratio.

4. The in-station master-slave differential protection method for photovoltaic power plants according to claim 3, characterized in that, Preprocessing the feature parameters includes: The line length is normalized using Min-Max. The service life of the equipment is discretized into multiple levels according to a set threshold. The effective voltage value, negative sequence voltage ratio, zero sequence voltage abrupt change, total harmonic distortion rate, and peak current ratio were all directly extracted from historical data. Standardize all feature parameters to the range of 0 to 1.

5. The in-station master-slave differential protection method for photovoltaic power plants according to claim 3, characterized in that, Multiple training subsets are generated from the training dataset using a bootstrap sampling method. Each training subset is used to train an unpruned decision tree, and the node splitting of the decision tree uses Gini impurity as the splitting criterion. The final trained decision trees are used to form the random forest model. The differential protection unit updates a portion of the decision trees in the random forest model at set intervals.

6. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, When the differential protection condition for the functional zone is met, but the differential protection condition for the voltage zone is met after a delay or is not met, the main differential protection will perform the action after the delay.

7. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, The voltage zones are numbered according to the ratio of real-time voltage to rated voltage, including: the ratio of real-time voltage to rated voltage of the voltage zone is inversely proportional to the number. A node is classified as the first partition when the ratio of its real-time voltage to its rated voltage is greater than 0.

95. A node is classified as belonging to the second partition when the ratio of its real-time voltage to its rated voltage is between 0.85 and 0.

95. A node is classified as the third partition when the ratio of its real-time voltage to its rated voltage is less than 0.

85.

8. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, The proportionality between the number size and priority order includes: The fault probability calculation is performed on the voltage zones in descending order of their numbers, and the calculation stops when the condition for differential setting is met.

9. The in-station master-slave differential protection method for photovoltaic power plants according to claim 1, characterized in that, When measuring the functional zones, a limit threshold is set. When the measured fault probability is higher than the limit threshold, the first and second threshold values ​​of the voltage zone are both reduced by 10%.

10. A master-slave differential protection system for a photovoltaic power station, characterized in that, The method for in-station master-slave differential protection of photovoltaic power plants as described in any one of claims 1-9 includes: The area division module is used to divide the lines within the photovoltaic power station into functional zones based on the equipment manufacturers, batches, capacity, and topology, and to divide the functional zones into voltage zones based on the ratio of the real-time voltage to the rated voltage of the nodes. The main differential protection module is used to collect and compare the current at both ends of the line to determine whether the differential current exceeds the set action value, so as to determine whether there is a short circuit fault. The fault probability calculation module is used to calculate the real-time fault probability of each voltage zone based on the random forest model trained from the historical fault data of the photovoltaic power station, combined with the real-time collected electrical and non-electrical quantity characteristics. The differential protection module is used to authorize and control the operation of the main differential protection module based on the real-time fault probability results.

Images