Power system station area protection method and device, terminal and storage medium

Abstract

The application provides a substation area protection method and device of a power system, a terminal and a storage medium. A substation area protection device is arranged in a substation, a hierarchical protection architecture is formed with a local protection device, and a time sequence relationship of the substation area protection and the local protection is configured. A substation main protection discriminates faults of elements in the substation according to shared electrical quantities, and protects the fault elements. When the substation main protection does not effectively act or the fault is not effectively removed, a substation backup protection checks and locates the fault, and performs corresponding cooperation control. The substation area protection acts before the local protection, and when the substation area protection does not effectively act, the local protection realizes protection. Thus, the fault removal is accelerated, the cascading failure is prevented, and the overall protection performance of the substation is improved.

Description

Technical Field

[0001] This invention relates to the field of power system relay protection technology, specifically to a station-area protection method, device, terminal, and storage medium for power systems dominated by new energy sources. Background Technology

[0002] With the large-scale integration of new energy power generation, the power system is shifting from being dominated by synchronous generators to being primarily based on new energy. When new energy is connected to the grid via inverters, fault current is limited and phase is controlled, reducing the sensitivity of traditional power frequency phasor protection; high harmonic content and large frequency offset decrease the accuracy of Fourier algorithm phasor extraction; system inertia and short-circuit capacity decrease, and the fault propagation range expands; after a single fault is cleared, power flow transfer can easily cause severe overload of non-faulty equipment, triggering maloperation of local backup protection, resulting in cascading trips and large-scale disconnection of new energy from the grid.

[0003] Existing protection schemes mainly adopt a localized configuration mode of "main protection + backup protection". Each bay protection device makes independent decisions based on the electrical quantities of its own bay, lacking information sharing and coordination among multiple bays at the station level. Although some schemes propose the concept of station-area protection, they do not establish a clear functional division and timing coordination relationship between station-area protection and local protection. Moreover, the station-area backup protection function is singular and lacks the ability to coordinate the handling of conditions such as power flow transfer overload and high and low voltage ride-through of new energy sources.

[0004] Therefore, a station-area protection scheme for power systems is needed. While retaining the existing local protection configuration, a hierarchical protection architecture of station-area and local protection should be constructed to realize ultra-fast station-area main protection, centralized station-area backup protection with integrated multi-decision, and coordinated with the high and low voltage ride-through strategy of new energy sources. Summary of the Invention

[0005] This invention aims to provide a substation protection method, device, terminal, and storage medium for power systems. While retaining the existing local protection configuration, it constructs a hierarchical protection architecture for substation-local protection, realizes ultra-fast substation main protection and centralized substation backup protection with integrated multi-decision, and coordinates with the high and low voltage ride-through strategy of new energy sources; thereby improving the fault defense capability in scenarios with a high proportion of new energy access.

[0006] Firstly, a method for station-area protection in a power system is provided, including:

[0007] A substation-level protection device is installed, consisting of a main substation protection system and a backup substation protection system. The device employs two data acquisition strategies: when a broadband measurement device is available in the substation, its high-frequency sampling data is used to control the action time of the main substation protection system within 10ms; when broadband measurement is not available, existing conventional sampling data is used to complete the judgment, ensuring the action time of the main substation protection system is controlled within 12ms. The backup substation protection system serves as backup protection in cases where the main substation protection fails to respond correctly or the fault cannot be completely isolated, and it also suppresses cascading faults in the power system. The station-area protection device and the local protection device are configured independently, forming a hierarchical protection architecture of the power system from station-area to local level. The local protection device consists of local main protection and local backup protection. The timing relationship between the station-area main protection and the local main protection and the local backup protection is as follows: the station-area main protection operates within 12 milliseconds, the local main protection operates between 15 and 40 milliseconds, the station-area backup protection operates between 100 and 200 milliseconds, and the local backup protection operates between 500 and 2000 milliseconds. After a fault occurs, the station-area main protection / station-area backup protection will perform fault diagnosis and send a trip signal before the local main protection / local backup protection. When the station-area main protection / station-area backup protection fails to operate effectively, the local main protection / local backup protection will perform fault diagnosis and send a trip signal. The station protection device is deployed in the bay layer within the substation's safety zone I, and its communication interface establishes a data link with at least one of the following types of devices: merging unit, data acquisition and execution unit, broadband measurement device, and local protection device. The protection scope of the main protection system covers the substation's lines, busbars, and transformers. A component fault discrimination model is configured for each independent protected object among the busbars, lines, and transformers. This component fault discrimination model takes a uniformly collected time-domain electrical quantity sampling sequence within the substation as input, synchronously separating the instantaneous voltage, instantaneous current, and fault components at the ports of each protected device to complete fault discrimination. The output of the component fault discrimination model is one of the following 25 scenarios: Normal; Overload; Inrush Current; High Voltage; Low Voltage; 3 types of single-phase ground fault within the zone; 3 types of two-phase ground fault within the zone; 3 types of two-phase short-circuit fault within the zone; 1 type of three-phase short-circuit fault within the zone; 3 types of single-phase ground fault outside the zone; 3 types of two-phase ground fault outside the zone; 3 types of two-phase short-circuit fault outside the zone; 1 type of three-phase short-circuit fault outside the zone. The station-area backup protection reuses the component fault discrimination model of the station-area main protection and is configured with a comprehensive verification model. After a 100ms delay from the start of the station-area protection device, the station-area backup protection retrieves the secondary discrimination conclusion of the component fault discrimination model, sends the conclusion to the comprehensive verification model to complete the comparison and verification between multiple devices and fault location, and issues a trip signal accordingly to assume the backup protection function in the event that the station-area main protection fails to operate reliably or the fault cannot be completely isolated. At the same time, based on the discrimination conclusions continuously output by the component fault discrimination model, the station-area backup protection identifies states such as power flow transfer, equipment overload, high voltage, and low voltage, and takes corresponding coordinated measures to prevent the occurrence of cascading faults in the power system. The station backup protection performs corresponding coordination control, specifically including the following two types of operating conditions: (1) When it is determined that the system has a power flow shift and any line within the protection range of the station protection device is in a state of severe overload operation, a blocking signal is applied to the local backup protection corresponding to the device to prevent it from tripping due to severe overload; (2) When it is determined that the new energy grid connection point has high voltage or low voltage and the voltage condition is sufficient to trigger the new energy high voltage ride-through or low voltage ride-through response, an alarm signal is sent to the reactive power compensation device to suppress the continuous rise or fall of the grid connection point voltage, and a coordination signal is sent to the new energy high voltage ride-through control unit or low voltage ride-through control unit at the same time; The substation protection device adopts the following data access schemes for substations at different construction stages: (1) For incremental smart substations, the broadband measurement device provides sampling data to the substation protection device through the communication network, with a sampling rate of not less than 10kHz; (2) For existing smart substations, the merging unit or acquisition execution unit provides sampling data to the substation protection device through the communication network, with a sampling rate of not less than 4kHz; (3) For existing conventional substations, the current transformer and voltage transformer directly provide analog signals to the substation protection device through the electrical circuit, and the substation protection device completes the analog-to-digital conversion itself, with a conversion rate of not less than 4.8kHz.

[0008] The power system substation protection method provided in this application is a targeted improvement on the hierarchical coordination, data acquisition methods, component fault identification, backup cross-verification, and cascading fault coordination control between substation protection and local protection in scenarios where substation protection devices are simultaneously configured with substation main protection and substation backup protection. By configuring the action sequence between substation main protection, substation backup protection, local main protection, and local backup protection, the substation main protection can quickly identify and clear faults. When the substation main protection fails to operate effectively or the fault is not effectively cleared, the substation backup protection performs re-judgment, cross-verification, and fault location, thereby forming a multi-level protection system in which substation main protection and substation backup protection cooperate with each other.

[0009] By configuring component fault discrimination models for busbars, lines, and transformers respectively, and acquiring instantaneous voltage, instantaneous current, and fault component information at the boundaries of each protected component, it is possible to distinguish between different operating states such as intra-zone faults, inter-zone faults, inrush current, overload, and high / low voltage. This improves the accuracy of the main protection system in identifying complex faults and non-fault transients, and reduces the probability of protection maloperation and failure to operate. By cross-validating the output results of different component fault discrimination models through a comprehensive verification model, fault location can be completed using information from multiple components within the substation. Furthermore, it improves the fault tolerance and operational reliability of the substation backup protection system when the output of a single sampling channel or a single fault discrimination model is abnormal.

[0010] In addition, by identifying power flow shifts, severe line overloads, and high and low voltage conditions at renewable energy grid connection points, and sending blocking, regulating, or coordinating signals to the corresponding local backup protection, reactive power compensation devices, and renewable energy high and low voltage cross-through control units, it is possible to block cascading trips caused by severe overloads, suppress further increases or decreases in voltage at renewable energy grid connection points, reduce the risk of unnecessary grid disconnection of renewable energy units and grid disconnection of entire areas, and achieve proactive prevention and control of overload-type cascading faults and renewable energy grid disconnection-type cascading faults.

[0011] Meanwhile, by configuring different data access methods for incremental smart substations, existing smart substations, and existing conventional substations, it is possible to meet the data requirements of both primary and backup protection in the substation area while ensuring compatibility with existing data acquisition equipment and communication conditions at different construction stages. This improves the engineering adaptability and application potential of the substation protection method. In summary, the embodiments of this application can improve the speed, selectivity, reliability, fault tolerance, and cascading fault prevention capabilities of the new power system substation protection.

[0012] Secondly, a station-area protection device for a power system is provided, comprising: Data acquisition interface, component fault diagnosis module, comprehensive verification module, and signal output interface; The data acquisition interface is used to receive electrical quantities and circuit breaker status quantities of three types of components in the substation: lines, busbars, and transformers. The output end of the data acquisition interface is connected to the data input end of the component fault diagnosis module; the diagnosis result output end of the component fault diagnosis module is connected to the data input end of the comprehensive verification module; the component fault diagnosis module and the comprehensive verification module each have independent signal output ends, and their respective signal output ends are connected to the signal output interface. The component fault discrimination module is configured with the component fault discrimination model, which is used to discriminate faults in the received component electrical quantities. The fault types it can identify include at least overload, inrush current, high voltage, low voltage, 10 types of short circuit faults within the zone, and 10 types of short circuit faults outside the zone. The component fault discrimination module also executes the station domain main protection function corresponding to the discrimination result based on the discrimination result of the component fault discrimination model. The integrated verification module is configured with the integrated verification model, which is used to perform cross-verification between the judgment results output by the component fault judgment module for different components, so as to determine the location of the fault; the integrated verification module also executes the station backup protection function corresponding to the result based on the verification and location results of the integrated verification model. The signal output interface is connected to the component fault diagnosis module and the integrated verification module, and is used to output one or more of the following signals based on the diagnosis results output by the component fault diagnosis module and the integrated verification module: outputting a trip signal to the circuit breaker of the corresponding faulty component, outputting a blocking signal to the local backup protection device of the corresponding component, outputting an alarm signal to the reactive power compensation device, and outputting a coordination signal to the new energy high and low voltage ride-through control unit.

[0013] Thirdly, a station-area protection terminal for a power system is provided, comprising a memory and a processor, wherein the memory is used to store program code, and the processor executes the program code to implement the station-area protection method for the power system as described in any one of the first aspects.

[0014] Fourthly, a storage medium is provided, wherein the storage medium stores program code that implements the station-area protection function of the power system as described in any one of the first aspects. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0016] Figure 1 A schematic flowchart of a power system station area protection method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram illustrating the configuration of the station domain-local hierarchical protection architecture and station domain protection device provided in an embodiment of the present invention. Figure 3 A schematic diagram illustrating the timing coordination relationship between site protection and local protection provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the integrated verification area division for station protection provided in an embodiment of the present invention; Figure 5 A schematic diagram illustrating the identification accuracy of the component fault discrimination model provided in this embodiment of the invention on a test dataset; Figure 6 A schematic diagram illustrating the recognition accuracy under inrush current conditions provided in an embodiment of the present invention; Figure 7 A schematic diagram illustrating the fault identification accuracy of a current transformer under saturation conditions, provided in an embodiment of the present invention. Figure 8 A schematic diagram illustrating the identification accuracy under single-phase high-resistance grounding fault conditions provided in an embodiment of the present invention; Figure 9 This is a schematic diagram comparing the accuracy of fault identification at different sampling frequencies provided in an embodiment of the present invention. Figure 10 This is a schematic diagram comparing the voltage response of the new energy grid connection point when the backup protection of the substation is in operation and when it is not, as provided in an embodiment of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0018] The following detailed description of some embodiments of this application is provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0019] Please see Figure 1 This invention provides a substation protection method for a power system, comprising steps S110 to S150. The implementation methods for each step are described below with reference to the accompanying drawings.

[0020] In some embodiments, the protection architecture of the power system is a hierarchical protection architecture of substation-local protection. This hierarchical architecture includes independent substation protection devices and local protection devices. The local protection devices are installed within the electrical compartments of the substation's busbars, lines, and transformers to protect each component and facility. The substation protection devices are installed within the substation to protect all components and facilities, including busbars, lines, and transformers. The substation protection devices are configured to implement substation primary protection and substation backup protection, and the local protection devices are configured to implement local primary protection and local backup protection. The substation protection method includes: The timing relationship of the station domain primary protection, the station domain backup protection, the local primary protection, and the local backup protection shall meet the following conditions: The operating time of the station-area primary protection is within 12 milliseconds, the operating time of the local primary protection is between 15 and 40 milliseconds, the operating time of the station-area backup protection is between 100 and 200 milliseconds, and the operating time of the local backup protection is between 500 and 2000 milliseconds. The station backup protection serves as backup protection when the station main protection fails to operate effectively or the fault is not effectively cleared.

[0021] Step S110: Install a substation-area protection device in the substation, which, together with the existing local protection device, forms a hierarchical protection architecture of substation-local protection; the substation-area protection device consists of substation-area main protection and substation-area backup protection. Please refer to [link / reference]. Figure 2 The station-local hierarchical protection architecture provided in this embodiment is composed of station-area protection devices and local protection devices. The station-area protection devices are deployed in the safety zone I bay layer of the substation and are connected to one or more of the merging unit, acquisition and execution unit, broadband measurement device, and local protection device. They can also work in coordination with devices such as measurement and control, bus protection, safety and stability control, and intelligent waveform recording, and obtain shared information for the entire station by relying on the redundant dual network of the substation process layer and station control layer.

[0022] In some embodiments, the substation protection device and the local protection device independently complete fault identification, action judgment, and trip output; in the event of a power system fault, if either the substation protection device or the local protection device fails to operate effectively, the other device can independently perform the protection function. After a fault occurs, the station-area main protection / station-area backup protection performs fault diagnosis and sends a trip signal before the local main protection / local backup protection; when the station-area main protection / station-area backup protection fails to operate effectively, the local main protection / local backup protection performs fault diagnosis and sends a trip signal. The data acquisition method for substation protection devices is determined according to the type of substation, as shown in Table 1. Among them, the broadband measurement device is compatible with the fundamental phasor measurement function and can uniformly measure the fundamental, harmonic, interharmonic and oscillation characteristic quantities in the 0-2500Hz frequency band of the power grid. It is preferentially deployed in substations with new energy grid connection.

[0023] In some embodiments, the substation protection device is configured in the safety zone I bay layer of the substation and is connected to one or more of the substation's merging unit, acquisition and execution unit, broadband measurement device, and local protection device. The substation protection device collects electrical data of all components of the substation's busbars, lines, and transformers through a broadband measurement device, so that the operating time of the substation's main protection is within 10ms. When the substation does not deploy broadband measurement devices, the station protection device collects conventional time-domain data through merging units, acquisition and execution units, or current and voltage transformers to ensure that the operating time of the station main protection is within 12ms.

[0024]

[0025] Table 1 In some embodiments, the station protection device employs different data access methods for different types of substations, including: In incremental smart substations, the station protection device is communicatively connected to the broadband measurement device, and the sampling frequency is not less than 10kHz. In existing smart substations, the station protection device is communicatively connected to the merging unit or the acquisition and execution unit, and the sampling frequency is not less than 4kHz. In existing conventional substations, the station protection device is electrically connected to the current and voltage transformers, and the sampling frequency is not less than 4.8kHz.

[0026] In this embodiment, the substation area is the protection zone defined with the substation as the center, including the busbars and transformers within the substation, as well as the lines that are electrically connected to the substation and included in the protection range of the substation area protection device.

[0027] Step S120: For the three types of components in the substation—lines, busbars, and transformers—configure a component fault discrimination model for each independent object and complete model training using simulation samples.

[0028] First, the model input is constructed. The component fault discrimination model uses the shared time-domain electrical quantities within the substation as the data source, and synchronously acquires the instantaneous values ​​of voltage and current and fault component information according to the protection boundary of the protected component. The electrical quantities required for various components are shown in Table 2.

[0029]

[0030] Table 2 For any protected component, the instantaneous values ​​of the three-phase voltage and three-phase current directly related to its protection boundary are selected as the basic inputs, and fault component information reflecting the changes in electrical quantities before and after the fault is extracted simultaneously. The original instantaneous value sequence and the fault component sequence are concatenated by channel to form the multi-channel time-domain input of the model. Then, each input channel is scaled and normalized according to the statistics of the training set to form the complete input vector of the model. Among them, the original quantities retain the absolute operating state and amplitude information of the system, while the fault components highlight the abrupt changes caused by the fault. The two complement each other, which is conducive to the model capturing weak fault features in the early stage of the fault.

[0031] The component fault discrimination model is constructed using a deep neural network with causal temporal modeling capabilities. This network consists of an input mapping layer, several temporal feature extraction layers, and an output layer connected sequentially. The model performs inference only using current and historical sampling information at each sampling time, without relying on future sampling points, thus meeting the real-time and causal requirements of the protection device's online operation. The output layer has two outputs: one is the main discrimination head, outputting the probability of the fault category; the other is a state auxiliary head, outputting the probabilities of operating states such as overload, high voltage, and low voltage. Both outputs share the same temporal feature extraction backbone, but each uses independent output parameters. This invention does not limit the specific network form of the temporal feature extraction layer; any structure with temporal feature extraction capabilities, such as recurrent neural networks, temporal convolutional networks, or attention networks, can be applied.

[0032] The main discriminator outputs the probability of fault categories, covering normal operation, 10 types of short-circuit faults within the zone, and 10 types of short-circuit faults outside the zone. These 10 short-circuit faults consist of 3 types of single-phase grounding faults, 3 types of two-phase grounding faults, 3 types of two-phase short-circuit faults, and 1 type of three-phase short-circuit fault. When the model is used for transformer objects, an inrush current category is added in addition to the aforementioned categories. The network backbone remains unchanged; only the output dimension of the main discriminator is expanded. The status auxiliary head outputs the probabilities of overload, high voltage, and low voltage operating states. In summary, the discrimination result of the component fault discrimination model is one of the following 25 types: normal, overload, inrush current, high voltage, low voltage, A-phase ground fault within the zone, B-phase ground fault within the zone, C-phase ground fault within the zone, AB two-phase ground fault within the zone, AC two-phase ground fault within the zone, BC two-phase ground fault within the zone, AB two-phase short circuit fault within the zone, AC two-phase short circuit fault within the zone, BC two-phase short circuit fault within the zone, three-phase short circuit fault within the zone, A-phase ground fault outside the zone, B-phase ground fault outside the zone, C-phase ground fault outside the zone, AB two-phase ground fault outside the zone, AC two-phase ground fault outside the zone, BC two-phase ground fault outside the zone, AB two-phase short circuit fault outside the zone, AC two-phase short circuit fault outside the zone, BC two-phase short circuit fault outside the zone, and three-phase short circuit fault outside the zone.

[0033] In some embodiments, the station domain main protection configures a component fault discrimination model for each component of the substation's busbars, lines, and transformers. The component fault identification model uses shared time-domain electrical quantities within the substation as the data source to synchronously acquire the instantaneous voltage values, instantaneous current values, and fault component information at the boundaries of each component within the substation's corresponding domain, thereby achieving fault identification. The discrimination results of the component fault discrimination model include at least one of the following: normal, overload, inrush current, high voltage, low voltage, 10 types of short circuit faults within the zone, and 10 types of short circuit faults outside the zone. The 10 types of short-circuit faults include 3 single-phase ground faults, 3 two-phase ground faults, 3 two-phase short-circuit faults, and 1 three-phase short-circuit fault.

[0034] During model training, the main discrimination task uses multi-class cross-entropy loss, while the state discrimination task uses binary cross-entropy loss. The weighted sum of these two losses is then used for parameter updates. Current transformer saturation is treated as a distortion condition in the measured signal rather than an independent category. Samples containing saturation are labeled according to their actual fault status (in-zone, out-of-zone, or no-fault) to enhance the model's discriminative adaptability under transient transformer saturation conditions. The entire discrimination process is data-driven, requiring no manual threshold setting and independent of pre-defined grid topology and operating modes.

[0035] Step S130: The station's main protection implements online monitoring based on the point-by-point rolling inference of the component fault discrimination model, and sends a trip signal to the circuit breaker of the corresponding faulty component when the action criteria are met.

[0036] Specifically, when the model output changes from a normal state to an intra-zone fault state, the station's main protection does not immediately trip, but instead enters continuous consistency verification: a verification window of n consecutive sampling points is set, and tripping only occurs when all n sampling points within the window are confirmed to be of the same type of intra-zone fault. In this embodiment, n=10. If the output within the window returns to the normal state, or if the verification time reaches the preset upper limit and the condition is still not met, it is considered a transient disturbance or an occasional misjudgment, and the protection automatically resets and resumes monitoring. At 10kHz sampling, 10 verification points occupy only 1ms. After superimposing the model judgment and recovery time, the station's main protection's action time is within 10ms when data is collected by a broadband measurement device; and within 12ms when data is collected by conventional methods.

[0037] Please see Figure 3The protection systems at each level form four independent lines of defense, arranged from fastest to slowest. Their operating times are as follows: station-level main protection within 12ms (within 10ms when data is acquired via broadband measurement devices); local main protection between 15 and 40ms; station-level backup protection between 100 and 200ms; and local backup protection operating within 500 to 2000ms according to tiered time limits. After a fault occurs, the station-level main protection activates first and sends a trip signal. If the station-level main protection fails to operate effectively, the local main protection then determines the fault and sends an output. These two systems constitute a coordinated relationship where the station-level protection takes precedence over the local protection in the main protection hierarchy. When the aforementioned main protection fails to effectively clear the fault, the station backup protection starts after a 100ms delay, completes cross-checking and fault location, and sends a trip signal. Simultaneously, it sends a blocking signal to the corresponding local backup protection. If the station backup protection also fails to operate effectively, the local backup protection operates last according to the tiered time limit. These two constitute a coordination relationship where the station protection precedes the local protection in the backup protection level. Therefore, station protection operates before local protection in both the main and backup protection levels, and each level is independently redundant; if any level fails to operate effectively, the next level can still continue to function.

[0038] Step S140: Configure a fault discrimination model for the backup protection component of the station area and a comprehensive verification model as backup protection when the main protection of the station area fails to operate effectively or the fault is not effectively cleared. This includes sub-steps S141 and S142.

[0039] Step S141: Main protection action confirmation. The moment t1 when the output of the component fault discrimination model changes from normal to fault state is taken as the start time of the station protection device. The station backup protection waits for 100ms confirmation time limit from t1: If an action message of the station main protection or local main protection is received during this period, and the corresponding circuit breaker position status indicates that the circuit breaker has been tripped, and the output of the component fault discrimination model returns to normal state, then it is determined that the main protection has been effectively operated and the fault has been effectively cleared, and the system returns to online monitoring; if a main protection action message has been received, but the circuit breaker has not been tripped, the model continues to output a fault state, or the faulty component has not been isolated, then it is determined that the fault has not been effectively cleared, and the system proceeds to step S142; in addition, if the model output continuously returns to normal during the waiting period, the system exits the time delay in advance to prevent false start caused by data fluctuations.

[0040] Step S142: Cross-validation and fault location. See also Figure 4The comprehensive verification area is pre-divided according to the electrical topology connection relationship within the station. LB1~LB3 is the line-bus comprehensive verification area, and TB1~TB3 is the transformer-bus comprehensive verification area. Each area can accommodate at least two component fault discrimination models. After a 100ms delay, the station backup protection acquires the re-discrimination results of each component fault discrimination model and sends them to the comprehensive verification model: the model of the faulty component itself should be judged as an in-zone fault, and the models of adjacent components directly electrically connected to it should be judged as out-of-zone faults simultaneously; the comprehensive verification model performs consistency verification on the output of each model within the area. When the model results remain stable within 10 consecutive sampling points and the fault types match each other, the faulty component and fault type are confirmed, and a trip signal is sent to the circuit breaker of that component.

[0041] Furthermore, when the model of the faulty component itself cannot provide effective output due to extreme conditions such as sampling channel disconnection or complete data loss, the comprehensive verification model can still complete fault-tolerant positioning: after the adjacent component model detects the fault characteristics outside the zone, it triggers the start timer. After confirming that none of the main protections have operated, it traverses the output of the other models in the comprehensive verification area. If the models of related components all output the same type of fault outside the zone, but the model of a certain component has no effective output, it is determined that the fault is located in that component and the output trips. The station area information redundancy is used to overcome the risk of failure to operate caused by the failure of a single channel.

[0042] In some embodiments, the station backup protection and the station main protection share a component fault discrimination model and are configured with a comprehensive verification model; After the station area protection device is activated and a 100-millisecond delay has elapsed, the station area backup protection re-identifies the busbar, line, and transformer based on the component fault discrimination model. After a 100ms delay following the activation of the station area protection device, the station area backup protection performs another judgment based on the component fault discrimination model; the result of the component fault discrimination model is input into the comprehensive verification model, the comprehensive verification model performs cross-verification and fault location on the results of each component fault discrimination model, and sends a trip signal; Meanwhile, the station backup protection identifies power flow shifts, equipment overloads, high voltage and low voltage characteristics based on the continuous output of the component fault discrimination model, and performs corresponding coordination control to prevent cascading failures in the new power system.

[0043] Step S150: Based on the continuous output of the component fault discrimination model, the station backup protection identifies characteristics such as power flow transfer, equipment overload, high / low voltage and performs coordinated control to prevent cascading failures in the power system.

[0044] Specifically, the station backup protection determines the changes in the operating status of each line based on the continuous output of the component fault discrimination model and in combination with the voltage, current and circuit breaker position status of each line; when one or more lines are taken out of operation and the active power or current of the remaining lines continues to increase and exceeds the corresponding operating limit, it is determined that the power system has experienced a power flow shift.

[0045] For overload-type cascading faults: When the station backup protection identifies a power flow shift in the power system and any line within the protection range of the station backup protection device is in a state of severe overload operation, the station backup protection sends a blocking signal to the local backup protection of that line to cut off the path of overload current triggering maloperation of the local backup protection, and prevents cascading trips from spreading step by step.

[0046] The blocking signal has a preset effective time limit; within the preset effective time limit, when the overload condition of the line disappears or the power flow returns to the allowable range, the blocking signal is released; when the station backup protection fails to operate effectively, the corresponding circuit breaker fails to trip, or the fault persists, the blocking signal is released so that the local backup protection performs protection actions according to its preset operating time limit.

[0047] For cascading faults caused by new energy grid disconnection: When high / low voltage is detected at the grid connection point of a new energy source, which will trigger high / low voltage ride-through, the station backup protection will correlate the voltage status with the fault identification results of the entire station, and implement differentiated control after distinguishing the causes of voltage anomalies. If there is a clear fault within the station, it is determined to be a fault-type low voltage. While quickly clearing the fault, an alarm signal is sent to the reactive power compensation device. This signal carries the adjustment demand for injected reactive power to suppress the voltage at the grid connection point from continuing to drop and assist the unit in passing the low voltage ride-through. If the low voltage is caused by fault propagation from outside the area, a coordination signal is sent to the new energy high / low voltage ride-through control unit to adjust the ride-through holding time and grid disconnection criteria, and the logic for cumulative grid disconnection caused by external disturbances is blocked or delayed. If there are no fault characteristics in the entire station, it is determined to be a non-fault cause such as power flow transfer or large load switching, and only the reactive power compensation device is used for smooth voltage adjustment. Symmetrical handling of high-voltage situations: After a fault is cleared, transient overvoltages are detected by an alarm signal that drives the reactive power compensation device to absorb excess reactive power, suppressing further increases in grid connection voltage and preventing the unit from disconnecting from the grid at high voltage. The above prevention and control and comprehensive verification processes are executed in parallel, enabling proactive intervention in the early stages of a cascading fault.

[0048] In some embodiments, the mating control includes: When a power flow shift is detected in the power system and any component of the station protection device is in a state of severe overload operation, a blocking signal is sent to the local backup protection of the component. When a voltage over-limit event is detected at the grid connection point of the new energy source connected to the substation, causing a high / low voltage ride-through event in the new energy unit, an alarm signal is sent to the reactive power compensation device of the substation, and a coordination signal is sent to the new energy high / low voltage ride-through control unit. This embodiment of the invention also provides a station-area protection device for a power system, which consists of a data acquisition interface, a component fault discrimination module, a comprehensive verification module, and a signal output interface. The data acquisition interface is used to receive electrical quantities and circuit breaker status quantities from three types of components within the substation: lines, busbars, and transformers. The data source is determined according to the substation type shown in Table 1: sampling data from broadband measurement devices is received in incremental smart substations; sampling data from merging units or acquisition execution units is received in existing smart substations; and analog quantities from current and voltage transformers are directly received in existing conventional substations. The data acquisition interface is connected to the data input terminal of the component fault diagnosis module. The output terminal of the component fault diagnosis module is connected to the data input terminal of the comprehensive verification module. The signal output terminals of both the component fault diagnosis module and the comprehensive verification module are connected to the signal output interface.

[0049] The component fault identification module is pre-set with the component fault identification model described in step S120. The electrical quantities accessed via the data acquisition interface are analyzed and identified through this model. The fault types that can be identified include at least overload, inrush current, high voltage, low voltage, 10 types of short circuit faults within the zone, and 10 types of short circuit faults outside the zone. When the identification result output by the model meets the preset action conditions, the component fault identification module executes the corresponding station main protection function based on the result.

[0050] The integrated verification module is pre-set with the integrated verification model described in step S142. It receives multiple discrimination results corresponding to different components from the component fault discrimination module through the model, performs cross-verification on them to achieve fault location, and executes the corresponding station backup protection function according to the verification results.

[0051] The signal output interface is connected to the component fault diagnosis module and the comprehensive verification module. Based on the diagnosis results of the two modules, it outputs one or more of the following signals: a trip signal to the circuit breaker of the corresponding faulty component, a blocking signal to the local backup protection device of the corresponding component, an alarm signal to the reactive power compensation device, and a coordination signal to the new energy high and low voltage ride-through control unit.

[0052] In some embodiments, the station protection device includes a data acquisition interface, a component fault detection module, a comprehensive verification module, and a signal output interface; The data acquisition interface is used to receive electrical quantities and circuit breaker status quantities of all components of the substation's busbars, lines, and transformers. The component fault discrimination module has the component fault discrimination model built in. The fault types identified include at least overload, inrush current, high voltage, low voltage, 10 types of short circuit faults within the zone and 10 types of short circuit faults outside the zone. The module executes the corresponding station domain main protection function according to the discrimination result of the component fault discrimination model. The integrated verification module has the integrated verification model built in, which is used to perform cross-verification and fault location on the judgment results of different components output by the component fault judgment module, and to execute the corresponding station area backup protection function according to the verification result of the integrated verification model. The signal output interface is used to output a trip signal to the circuit breaker of the corresponding faulty component, output a blocking signal to the local backup protection device of the corresponding component, send an alarm signal to the reactive power compensation device, and send a coordination signal to the new energy high / low voltage ride-through control unit, based on the judgment results of the component fault judgment module and the integrated verification module.

[0053] For example, to verify the actual effectiveness of the above method, a simulation system of a 220kV substation with wind power and photovoltaic access was built in PSCAD / EMTDC, with its wiring topology being the same as... Figure 4 Consistency: The wind farm uses an equivalent model of a doubly-fed induction generator (DFIG), with a rated active power of 2MW and a rated capacity of 2.5MVA per unit. 100 equivalent units total approximately 200MW, connected to bus bus1 via line L1. The photovoltaic power station consists of an equivalent model of a photovoltaic array, DC-DC converter, grid-connected inverter, and step-up transformer, with a capacity of 100MVA, connected to bus bus3 via line L3. The sampling frequency is 10kHz, collecting data for one cycle before and two cycles after the fault. The actual state is labeled point-by-point according to the sampling time. The fault parameter settings for the training and test sets are shown in Table 3.

[0054]

[0055] Table 3 exist Figure 5 The results show the identification accuracy of three types of models—transformer, busbar, and line—on the test dataset: the identification accuracy of fault-free sections is 100%; there are short-term fluctuations in the output during the initial stage of a fault, and the line, transformer, and busbar models recover 100% accuracy and remain stable at the 17th, 22nd, and 51st sampling points, respectively, indicating that the time-series electrical quantities at multiple measurement points in the station area contain effective features for distinguishing the faults of the three types of components.

[0056] Please see Figure 6 , Figure 6This diagram illustrates the identification accuracy of a transformer component fault discrimination model under inrush current conditions. This test verifies the operational safety of the station's main protection system under transformer no-load closing conditions. Specifically, under renewable energy access conditions, various inrush current scenarios generated during transformer operation are simulated by setting different closing angles and core residual magnetism combinations. Figure 6 As can be seen, under all test conditions, the model consistently classified the inrush current as a non-fault transient category, without misclassifying it as an internal fault, and the station's main protection did not send a trip signal. It should be noted that the amplitude and waveform distortion characteristics of the inrush current are easily confused with internal transformer faults, and traditional protection systems usually require additional inrush current blocking criteria. In this embodiment, the difference in timing characteristics between inrush current and internal faults is learned autonomously by the model during training, eliminating the need for manual setting of the inrush current blocking criteria, thus preventing protection maloperation caused by no-load closing.

[0057] Figure 7 The accuracy of the bus component fault discrimination model under current transformer saturation conditions is shown. This test superimposed severe current transformer saturation conditions on fault samples both inside and outside the bus zone. It is understandable that with numerous bus connection branches, the fault current in the fault branch increases sharply during an external fault, easily leading to transient saturation of the current transformer in that branch, resulting in distortion of the secondary waveform and affecting the accuracy of the discrimination. During the test, the model output experienced short-term fluctuations due to distorted data in the early stages of the fault; as time-series information accumulated, the discrimination results were corrected point by point, restoring 100% accuracy at the 65th sampling point and remaining stable. After consistency verification of 10 consecutive sampling points as described in step S130, the operating time of the station's main protection under this condition was approximately 7.5 ms, meeting the time limit requirement of less than 10 ms. Furthermore, the model did not misclassify external faults as internal faults during the saturation period, ensuring the selectivity of the protection.

[0058] Single-phase high-resistance ground faults are one of the operating conditions that traditional local protection systems struggle to detect reliably: the fault current is weak, the voltage drop is small, and the fault characteristics are further weakened by the weak feed characteristics of new energy power sources, leading to insufficient sensitivity of protection systems based on impedance or current criteria. To address this condition, this test, based on the test set parameters shown in Table 3, further increased the transition resistance to 400Ω and 500Ω, and simulated a single-phase ground fault on the line. The identification accuracy of the line component fault discrimination model was recorded. Figure 8 Test results show that the model maintains high recognition accuracy under both transition resistance conditions, recovers to 100% accuracy at the 29th sampling point and remains stable; after including the continuous consistency verification window, the operating time of the station's main protection is approximately 3.9ms. Compared with traditional protection methods that rely on long delays to clear high-resistance faults from backup sections, this invention achieves rapid clearing of weak-feature faults while ensuring sensitivity.

[0059] To examine the impact of sampling frequency on model discrimination performance and to provide a basis for setting the two-stage operating time limits of the main protection system in the substation area, this test, targeting the same transformer protection object, constructed two sets of comparative models with sampling frequencies of 10kHz and 4kHz respectively under the same training strategy. The comparison of their recognition accuracy is shown in [reference needed]. Figure 9 The test results show that the model with a 10kHz sampling frequency recovered accurately after 2.2ms after the fault occurred, while the model with a 4kHz sampling frequency required 5.75ms. This indicates that the higher the sampling frequency, the higher the temporal resolution of transient information, and the shorter the initial ambiguity phase of fault identification. Based on this, it is estimated that when data is collected using a broadband measurement device, the model recovery time of 2.2ms plus a verification window of approximately 1ms allows the main protection action time of the substation to be controlled within 10ms. When data is collected using conventional methods, the model recovery time of 5.75ms plus a verification window of approximately 2.5ms, totaling approximately 8.25ms, allows the action time to be controlled within 12ms. These results verify the rationality of the two time limit settings in this invention and also indicate that existing substations can meet the action time requirements based on existing data collection conditions.

[0060] Furthermore, multiple short-term fault disturbances with adjacent time intervals were set in the external area of ​​the wind power access line L1 to test the cascading fault prevention capability. The disturbance points were Bus2, L3, and L4 in sequence. Voltage drops were conducted to the wind power grid connection point through the transformer, bus, and line, causing the unit to repeatedly enter the low voltage ride-through state. The station backup protection locked the voltage anomaly location based on the voltage, current, and circuit breaker status at multiple points within the station area. Combined with the output of the fault discrimination model of each component, the above disturbances were judged one by one as external disturbances of L1, and a coordination signal was sent to the new energy power station. The logic for cumulative disconnection of low voltage ride-through times caused by external disturbances was blocked or the threshold was adjusted. The discrimination results and corresponding measures are shown in Table 4.

[0061]

[0062] Table 4 Figure 10 The effective voltage response of the wind turbine grid connection point was compared under two operating conditions: when the backup protection of the station area was not in operation and when it was in operation. When the backup protection was not in operation, the unit disconnected from the grid because the number of low voltage rides reached the threshold in a short period of time, and a significant transient overvoltage appeared at the grid connection point at the moment of disconnection. After the backup protection was in operation, although the voltage at the grid connection point dropped multiple times during the disturbance, the unit did not disconnect from the grid unnecessarily. After the disturbance ended, the voltage gradually returned to the stable operating range, which verified the active prevention and control effect of the present invention on the grid disconnection type cascading fault of new energy.

[0063] This invention also provides a power system station domain protection terminal, which includes at least a memory and a processor coupled to each other. The memory stores program code that can be loaded and executed by the processor. When the processor calls and executes the program code, it completes each step of the power system station domain protection method described in any of the above embodiments.

[0064] This invention also provides a storage medium containing program code that implements the station domain protection function of the power system described in any of the above embodiments. When the program code is read and run by a processor, it executes all or part of the steps of the aforementioned method.

[0065] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention; any equivalent substitutions, partial modifications or improvements made based on the concept and principles of the present invention shall be deemed to fall within the protection scope defined by the claims of the present invention.

Claims

1. A method for station-area protection in a power system, characterized in that, The power system protection architecture is a hierarchical protection architecture of substation-local protection. This hierarchical architecture includes independent substation protection devices and local protection devices. The local protection devices are installed within the electrical compartments of the substation's busbars, lines, and transformers to protect each component and facility. The substation protection devices are installed within the substation to protect all components and facilities, including busbars, lines, and transformers. The substation protection devices are configured to implement substation primary protection and substation backup protection, and the local protection devices are configured to implement local primary protection and local backup protection. The site domain protection method includes: The timing relationship of the station domain primary protection, the station domain backup protection, the local primary protection, and the local backup protection shall meet the following conditions: The operating time of the station-area primary protection is within 12 milliseconds, the operating time of the local primary protection is between 15 and 40 milliseconds, the operating time of the station-area backup protection is between 100 and 200 milliseconds, and the operating time of the local backup protection is between 500 and 2000 milliseconds. The station backup protection serves as backup protection when the station main protection fails to operate effectively or the fault is not effectively cleared.

2. The substation protection method for a power system according to claim 1, characterized in that, The station-level protection device and the local protection device each independently complete fault identification, action judgment, and trip output; in the event of a power system fault, if either the station-level protection device or the local protection device fails to operate effectively, the other device can independently perform the protection function. After a fault occurs, the station-area main protection / station-area backup protection performs fault identification and sends trip signals before the local main protection / local backup protection; When the station-domain primary protection / station-domain backup protection fails to operate effectively, the local primary protection / local backup protection performs fault diagnosis and sends a trip signal.

3. The substation protection method for a power system according to claim 1, characterized in that, The substation protection device is configured in the safety zone I bay layer of the substation and is connected to one or more of the substation's merging unit, acquisition and execution unit, broadband measurement device, and local protection device. The substation protection device collects electrical data of all components of the substation's busbars, lines, and transformers using a broadband measurement device, so that the operating time of the substation's main protection is within 10ms.

4. The substation protection method for a power system according to claim 1, characterized in that, include: The main protection system of the substation is configured with a component fault discrimination model for each component of the busbar, line, and transformer of the substation. The component fault identification model uses shared time-domain electrical quantities within the substation as the data source to synchronously acquire the instantaneous values ​​of boundary voltage and current of each component within the corresponding substation area, as well as fault component information, to achieve fault identification. The results of the component fault discrimination model include at least one or more of the following: normal, overload, inrush current, high voltage, low voltage, short circuit fault within the zone, and short circuit fault outside the zone.

5. The substation protection method for a power system according to claim 4, characterized in that, include: The station backup protection and the station main protection share the component fault discrimination model and are configured with a comprehensive verification model; After a 100ms delay following the activation of the station domain protection device, the station domain backup protection performs another judgment based on the component fault discrimination model. The results of the component fault discrimination model are input into the comprehensive verification model. The comprehensive verification model performs cross-verification and fault location on the results of each component fault discrimination model and sends a trip signal. The station backup protection identifies power flow shifts, equipment overloads, and high / low voltage characteristics based on the continuous output of the component fault discrimination model, and performs corresponding coordinated control.

6. The substation protection method for a power system according to claim 5, characterized in that, The execution of the corresponding coordination control includes: When a power flow shift is detected in the power system and any component of the station protection device is in a state of severe overload operation, a blocking signal is sent to the local backup protection of the component. When a voltage over-limit event is detected at the grid connection point of the new energy source connected to the substation, causing a high / low voltage ride-through event in the new energy unit, an alarm signal is sent to the reactive power compensation device of the substation, and a coordination signal is sent to the high / low voltage ride-through control unit of the new energy source.

7. The substation protection method for a power system according to claim 3, characterized in that, The substation protection device is configured in the safety zone I bay layer of the substation, including: In incremental smart substations, the station protection device is communicatively connected to the broadband measurement device, and the sampling frequency is not less than 10kHz. In existing smart substations, the station protection device is communicatively connected to the merging unit or the acquisition and execution unit, and the sampling frequency is not less than 4kHz. In existing conventional substations, the station protection device is electrically connected to the current and voltage transformers, and the sampling frequency is not less than 4.8kHz.

8. A station-area protection device for a power system, characterized in that, The station protection device is applied to the power system of claim 1, and includes: a data acquisition interface, a component fault detection module, a comprehensive verification module, and a signal output interface; The data acquisition interface is used to receive electrical quantities and circuit breaker status quantities of all components of the substation's busbars, lines, and transformers. The component fault identification module has the component fault identification model built in, and identifies at least the following fault types: overload, inrush current, high voltage, low voltage, 10 types of short circuit faults within the zone and 10 types of short circuit faults outside the zone. Based on the results of the component fault identification model, it executes the corresponding station domain main protection function. The integrated verification module has the integrated verification model built in, which is used to perform cross-verification and fault location on the judgment results of different components output by the component fault judgment module, and execute the corresponding station area backup protection function according to the result of the integrated verification model. The signal output interface is used to output a trip signal to the circuit breaker of the corresponding faulty component, output a blocking signal to the local backup protection device of the corresponding component, send an alarm signal to the reactive power compensation device, and send a coordination signal to the new energy high and low voltage ride-through control unit, based on the judgment results of the component fault judgment module and the integrated verification module.

9. A station-area protection terminal for a power system, characterized in that, The system includes a memory and a processor, the memory being used to store program code, and the processor executing the program code to implement the power system station protection method according to any one of claims 1 to 7.

10. A storage medium, characterized in that, The storage medium contains program code that implements the station area protection function of the power system according to any one of claims 1 to 7.

Images