A protection principle setting calculation method, system and device for multi-type converters and a medium

Abstract

This invention relates to the field of converter protection technology. It discloses a method, system, equipment, and medium for calculating protection principle settings for multiple types of converters. The method includes: acquiring multi-dimensional electrical signals from multiple types of converters; identifying fault characteristics and classifying levels using a multi-level converter fault hierarchy model; determining the boundary conditions of a valve group short-circuit current attenuation prediction model; obtaining dynamic characteristic parameters of short-circuit current attenuation through the prediction model; performing distributed collaborative optimization of protection parameters using a distributed intelligent collaborative protection algorithm; conducting multi-scenario simulation tests using a converter fault simulation calculation platform; and correcting setting parameters based on test data to form a protection setting value system adapted to multiple types of converters. This invention integrates hierarchical modeling, dynamic prediction, collaborative optimization, and simulation verification technologies, adapting to the topological differences and operating conditions of multiple types of converters, and improving the universality, accuracy, and reliability of protection setting calculations.

Description

Technical Field

[0001] This invention relates to the field of converter protection technology, and in particular to a protection principle setting calculation method, system, equipment and medium for multiple types of converters. Background Technology

[0002] With the rapid development of power electronics technology, various types of converters are increasingly widely used in new energy power generation, flexible DC transmission, and other fields. The diversity of their topologies and operating conditions places stringent requirements on the adaptability of protection systems. Traditional protection setting calculation methods are mostly designed for single-type converters, making it difficult to take into account the differences in fault characteristics of converters with different topologies. Furthermore, the short-circuit current of the converter valve group exhibits complex attenuation characteristics when a fault occurs, and existing methods lack accurate characterization of this dynamic process. This results in insufficient adaptability of protection setting values ​​to actual fault scenarios, affecting the reliability and accuracy of protection actions. At the same time, in a distributed grid architecture, converter protection needs to achieve multi-node coordinated response. Traditional centralized setting calculation modes cannot meet the requirements of real-time performance and coordination. Therefore, it is urgent to construct a new protection principle setting calculation system adapted to multiple types of converters. Summary of the Invention

[0003] In view of the above-mentioned existing problems, the present invention provides a protection principle setting calculation method, system, equipment and medium for multiple types of converters.

[0004] Therefore, the technical problem solved by the present invention is that the existing technology relies only on static electrical parameters for protection setting, ignores the dynamic characteristics and hierarchical features of faults, and lacks distributed intelligent collaboration and simulation verification closed loop, which limits the universality and reliability of setting parameters.

[0005] To solve the above technical problems, the present invention provides the following technical solution: a protection principle setting calculation method for multiple types of converters, comprising: acquiring multi-dimensional electrical quantity signals of multiple types of converters, identifying fault characteristics and dividing levels using a multi-level converter fault hierarchical model, and determining the boundary conditions of the valve group short-circuit current attenuation prediction model. The short-circuit current attenuation dynamic characteristic parameters are obtained through the valve group short-circuit current attenuation prediction model, and the distributed collaborative optimization of the protection parameters is completed by the distributed intelligent collaborative protection algorithm. Multi-scenario simulation tests were conducted using a converter fault simulation calculation platform. Based on the test data, the setting parameters were corrected to form a protection setting value system that is compatible with multiple types of converters.

[0006] As a preferred embodiment of the protection principle setting calculation method for multi-type converters described in this invention, the multi-level converter fault hierarchical model includes: using valve group switch status signals, bridge arm current signals, and DC side voltage signals as basic input datasets; analyzing the second derivative of each bridge arm current, the nonlinear change of the DC side voltage, and the coupling relationship between bridge arms; and outputting characteristic values ​​representing the fault level to classify the levels corresponding to different fault types.

[0007] As a preferred embodiment of the protection principle setting calculation method for multi-type converters described in this invention, the valve group short-circuit current attenuation prediction model includes, based on equivalent inductance, equivalent resistance, equivalent capacitance and initial short-circuit current amplitude, combined with energy loss and circuit topology changes during fault development, dynamically predicting the attenuation curve of short-circuit current over time through the superposition of fundamental attenuation component and multi-order harmonic components, and obtaining dynamic characteristic parameters including attenuation rate and oscillation frequency.

[0008] As a preferred embodiment of the protection principle setting calculation method for multi-type converters described in this invention, the distributed collaborative optimization of protection parameters by the distributed intelligent collaborative protection algorithm includes, with distributed protection nodes as units, each node, based on local protection setting parameters, reference values, and inter-node communication delays, iteratively optimizing to obtain a local optimal solution by minimizing an objective function that includes parameter deviations, protection action sensitivity, and inter-node parameter differences, and outputting globally unified protection setting parameters through inter-node collaborative verification and fusion.

[0009] As a preferred embodiment of the protection principle setting calculation method for multiple types of converters described in this invention, the multi-scenario simulation test includes: constructing a simulation environment consistent with the actual operating scenario by importing the topology model parameters, protection configuration parameters, and optimized protection setting parameters of multiple types of converters; conducting simulation tests according to a preset fault scenario sequence; collecting electrical quantities and protection action data in real time during the simulation process; and performing time-series alignment and feature extraction on the simulation data.

[0010] As a preferred embodiment of the protection principle setting calculation method for multi-type converters described in this invention, the step of correcting the setting parameters based on test data includes classifying and correcting the protection setting parameters according to the deviation between simulation data and actual requirements. During the correction process, the coupling relationship between the basic setting value, fault characteristic parameters and model adaptation coefficients is considered, and the protection action deviation is made to meet the requirements through iterative adjustment.

[0011] As a preferred embodiment of the protection principle setting calculation method for multi-type converters described in this invention, the step of correcting the setting parameters based on test data further includes feeding the corrected parameters back to the distributed intelligent cooperative protection algorithm, and re-optimizing and simulating the protection parameters until the protection action deviation meets the preset threshold.

[0012] This invention provides a protection principle setting calculation system for multiple types of converters.

[0013] As a preferred embodiment of the protection principle setting calculation system for multi-type converters described in this invention, it includes a multi-level converter fault feature acquisition and preprocessing unit and a hierarchical fault model construction and analysis unit connected in one direction. The former acquires multi-dimensional electrical quantity signals during the operation of the converter and transmits them to the latter after filtering and processing. The hierarchical fault model construction and analysis unit is bidirectionally connected to the valve group short-circuit current decay dynamic prediction unit and the distributed intelligent collaborative protection parameter optimization unit, respectively, and interactively transmits the fault hierarchy classification results and characteristic parameters. The output of the valve group short-circuit current decay dynamic prediction unit is unidirectionally connected to the input of the distributed intelligent collaborative protection parameter optimization unit, providing it with short-circuit current decay characteristic parameters. The distributed intelligent collaborative protection parameter optimization unit is bidirectionally connected to the converter fault simulation test and data acquisition unit, transmitting the optimized setting parameters to the latter while receiving the simulation test data fed back by the latter. The output of the converter fault simulation test and data acquisition unit is unidirectionally connected to the input of the multi-type converter protection setting parameter correction and output unit, transmitting simulation data for parameter correction. The multi-type converter protection setting parameter correction and output unit integrates all data, completes the calculation of the final protection setting value, and outputs it.

[0014] The present invention provides a computer device, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of a protection principle setting calculation method for multiple types of converters.

[0015] The present invention provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, it implements the steps of a protection principle setting calculation method for multiple types of converters.

[0016] The beneficial effects of this invention are: by using a multilevel converter fault hierarchical model to accurately identify the fault level and characteristic patterns of different topology converters, and by combining a valve group short-circuit current attenuation prediction model to dynamically capture the current attenuation characteristics during the fault development process, the invention solves the problem that traditional methods rely on static parameters and cannot adapt to dynamic changes in faults, thus enabling the setting parameters to be accurately matched with the fault scenario. By introducing a distributed intelligent collaborative protection algorithm, it adapts to the distributed deployment requirements of various types of converters, realizes multi-node collaborative optimization of protection parameters, and breaks through the real-time and collaborative limitations of the traditional centralized computing mode. Multi-scenario simulation tests are conducted through a converter fault simulation calculation platform. A closed-loop optimization process is formed by combining the comprehensive correction mechanism of the tuning parameters. This makes up for the lack of dedicated simulation verification and dynamic parameter correction in existing technologies, and greatly improves the versatility and reliability of the tuning results. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic flowchart of a protection principle setting calculation method for multiple types of converters, provided as an embodiment of the present invention.

[0019] Figure 2 This is a schematic diagram of the working module of a protection principle setting calculation system for multiple types of converters, provided as an embodiment of the present invention. Detailed Implementation

[0020] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0021] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a protection principle setting calculation method for multiple types of converters, including: S1: Collect multi-dimensional electrical signals from various types of converters, use a multi-level converter fault hierarchical model to identify fault characteristics and classify levels, and determine the boundary conditions of the valve group short-circuit current attenuation prediction model.

[0022] S2: Obtain the dynamic characteristic parameters of short-circuit current attenuation through the valve group short-circuit current attenuation prediction model, and complete the distributed collaborative optimization of protection parameters through the distributed intelligent collaborative protection algorithm.

[0023] S3: Utilize the converter fault simulation calculation platform to conduct multi-scenario simulation tests, and correct the setting parameters based on the test data to form a protection setting value system that is compatible with multiple types of converters.

[0024] It should be noted that the deep integration of multiple technologies enables precise adaptation to various types of converters. It relies on a fault hierarchy model to accurately capture the fault level characteristics of different topology converters, and combines a short-circuit current decay prediction model to dynamically depict the changes in electrical quantities during fault development, breaking the limitations of traditional methods that rely on static parameters. Furthermore, it introduces a distributed intelligent collaborative algorithm to adapt to the multi-node protection requirements of distributed power grid architectures, and utilizes a dedicated fault simulation calculation platform to conduct multi-scenario tests, forming a dynamic correction mechanism for setting parameters, comprehensively improving the versatility and accuracy of setting calculations.

[0025] Example 2 is an embodiment of the present invention. Based on the above embodiment, a protection principle setting calculation method for multiple types of converters is provided.

[0026] Furthermore, in this embodiment, step S1 collects multi-dimensional electrical quantity signals from multiple types of converters, identifies fault characteristics and classifies levels using a multi-level converter fault hierarchical model, and determines the boundary conditions of the valve group short-circuit current attenuation prediction model. Specific steps include S101-S102: S101: Using valve group switching status signal, bridge arm current signal and DC side voltage signal as basic input dataset, by analyzing the second derivative of each bridge arm current, the nonlinear change of DC side voltage and the coupling relationship between bridge arms, it outputs characteristic values ​​that characterize the fault level, which are used to classify the levels corresponding to different fault types.

[0027] Based on the differences in topology and operating conditions of various types of converters, the switching status signals of converter valve groups, the arm current signals, and the DC side voltage signals are extracted to construct the basic input dataset for a multilevel converter fault hierarchical model. To address the topological differences between different types of converters, such as voltage source and current source converters, key parameters such as the number of bridge arms, valve group switching device models, and DC-side energy storage element configurations for each type of converter are defined. The number of bridge arms can include common configurations such as 3 or 5 arms. IGBT or IGCT series valve group switching devices are selected, and the DC-side voltage level is set to range from several hundred to several thousand volts. Subsequently, high-precision sensing equipment is used to collect valve group switching status signals, bridge arm current signals, and DC-side voltage signals during converter operation. The sampling frequency of the sensing equipment is set to 10kHz to 100kHz to ensure the capture of high-frequency dynamic changes. The valve group switching status signal acquisition resolution reaches the microsecond level, the bridge arm current signal acquisition accuracy is controlled within ±0.5%, and the DC-side voltage signal acquisition error does not exceed ±1%.

[0028] During the data acquisition process, the operating parameters of the converter need to be recorded simultaneously, including the output power range of 100kW to 100MW, the switching frequency of 500Hz to 20kHz, and the modulation method. The acquired multi-dimensional signals are time-aligned to remove abnormal data caused by sensor interference. Finally, a basic input dataset is formed, including switching state sequence, current and voltage waveform data, and operating parameters. The sample duration of this dataset is set to 0.5 seconds before the fault and 2 seconds after the fault, with no less than 100,000 data points. This provides comprehensive and accurate raw data support for the subsequent fault stratification model, ensuring that the model can fully explore the differences in fault characteristics of different types of converters.

[0029] S102: Based on the equivalent inductance, equivalent resistance, equivalent capacitance, and initial short-circuit current amplitude, combined with the energy loss and circuit topology changes during the fault development process, the decay curve of the short-circuit current over time is dynamically predicted by superimposing the fundamental attenuation component and multiple harmonic components, and dynamic characteristic parameters including attenuation rate and oscillation frequency are obtained.

[0030] The fault feature identification and hierarchical division of the input dataset are performed by using a multilevel converter fault hierarchical model, the electrical quantity change law corresponding to different fault types is separated, and the initial boundary conditions of the valve group short-circuit current attenuation prediction model are determined. Fault feature identification and hierarchical classification are achieved through a multilevel converter fault hierarchical model. During implementation, the basic input dataset is first classified according to converter type and operating conditions. The classification dimensions include topology type, output power range, and switching frequency level. The output power range is divided into three ranges: 100kW-1MW, 1MW-10MW, and 10MW-100MW. The switching frequency level is divided into three levels: 500Hz-5kHz, 5kHz-10kHz, and 10kHz-20kHz.

[0031] Subsequently, the fault hierarchical model of the multilevel converter was launched. The model first performs logical analysis on the input valve group switching status signal to identify whether the turn-on and turn-off timing of the switching devices conforms to the normal operation law. Then, through time domain analysis, it extracts the peak value, rise rate, oscillation frequency of the bridge arm current signal and the drop amplitude and recovery time of the DC side voltage signal. Among them, the current peak extraction accuracy reaches the ampere level, and the voltage drop amplitude identification error is controlled within ±2%.

[0032] Based on the aforementioned characteristic parameters, the model is divided into three levels according to the scope and severity of the fault impact: the first level represents a single switching device fault, the second level represents a single bridge arm fault, and the third level represents multiple bridge arms or DC-side faults. Simultaneously, the electrical quantity change patterns corresponding to different fault types are separated; for example, a short-circuit fault corresponds to a sudden increase in bridge arm current and a sharp drop in voltage, while an open-circuit fault corresponds to current interruption and voltage fluctuation. Finally, based on the fault level and change patterns, the initial boundary conditions of the valve group short-circuit current attenuation prediction model are determined, including the fault initiation time, initial current amplitude, and initial values ​​of equivalent circuit parameters. The fault initiation time positioning accuracy reaches 1 microsecond, and the initial current amplitude error does not exceed ±3%, laying the foundation for the accurate operation of the subsequent prediction model.

[0033] Furthermore, in this embodiment, step S2 obtains the dynamic characteristic parameters of short-circuit current attenuation through the valve group short-circuit current attenuation prediction model, and completes the distributed collaborative optimization of protection parameters through a distributed intelligent collaborative protection algorithm. The specific steps include: S201: Based on the initial boundary conditions and fault characteristic parameters, a valve group short-circuit current decay prediction model is built. By analyzing the energy loss of current-carrying elements and the changes in circuit topology during the fault development process, the dynamic characteristic parameters of short-circuit current decay are obtained. Based on the determined initial boundary conditions, including parameters such as the fault initiation time, initial current amplitude, equivalent inductance, and initial equivalent resistance, and combined with data such as the current rise rate and voltage drop amplitude in the fault characteristic parameters, the core input variables of the model are identified.

[0034] The model framework is built based on circuit theory and fault dynamics principles. The model includes an energy loss calculation module and a topology change analysis module. The energy loss calculation module focuses on the conduction loss, turn-off loss and line loss of switching devices. The conduction loss is calculated with a coefficient of 1%-5% of the rated current of the device, and the turn-off loss is dynamically adjusted in combination with the switching frequency. The topology change analysis module tracks changes in the connection status of bridge arms and the operating status of switching devices in real time during fault development. It sets a topology change trigger threshold; a topology abrupt change is identified when the current change rate exceeds 100A / μs or the voltage change rate exceeds 1kV / μs. During model runtime, initial boundary conditions and fault characteristic parameters are input. The energy loss calculation module iteratively calculates the energy loss value every millisecond. Simultaneously, the topology change analysis module captures the abrupt change time and type of change in the circuit topology, dynamically adjusting the equivalent circuit parameters in the model. The equivalent inductance is adjusted from 1mH to 100mH, the equivalent resistance from 0.1Ω to 10Ω, and the equivalent capacitance from 1μF to 100μF.

[0035] Through continuous iterative calculations, dynamic characteristics of short-circuit current decay within 0-2 seconds after a fault occurs are obtained, including decay rate, oscillation frequency, and harmonic component amplitude. The decay rate calculation accuracy reaches 0.1A / ms, and the oscillation frequency identification error does not exceed ±5Hz, providing accurate dynamic current data support for subsequent protection parameter optimization.

[0036] S202: Based on distributed protection nodes, each node iteratively optimizes a local optimal solution by minimizing an objective function that includes parameter deviation, protection action sensitivity, and parameter differences between nodes, based on local protection setting parameters, reference values, and inter-node communication delays. The solution is then verified and fused between nodes to output globally unified protection setting parameters.

[0037] The distributed intelligent collaborative protection algorithm is used to achieve distributed collaborative optimization of protection parameters. During implementation, the protection configuration requirements of various types of converters are first clarified, including the action threshold range and action delay requirements of core protection functions such as overcurrent protection, overvoltage protection, and valve group fault protection. The action threshold for overcurrent protection is set to 1.2-2.5 times the rated current, the action threshold for overvoltage protection is set to 1.1-1.5 times the rated voltage, and the action delay is required to be controlled within 10-100 microseconds.

[0038] The acquired dynamic characteristics of short-circuit current attenuation, including attenuation rate, oscillation frequency, and harmonic component amplitude, are fused with protection setting parameters such as action threshold adaptation coefficient and delay coordination coefficient to form an algorithm input parameter matrix. The dimensions of the parameter matrix are set according to the number of distributed nodes, with each node corresponding to 10-20 input parameters. Next, a distributed intelligent collaborative protection algorithm framework is built, dividing the system into 3-10 protection collaborative nodes based on the distributed deployment of the converter. Each node is configured with an independent computing unit and communication module, with the communication module having a transmission rate of no less than 1Gbps to ensure real-time parameter interaction.

[0039] During algorithm execution, each node first performs local optimization calculations based on its local input parameters, adjusting the protection setting values ​​using an iterative optimization method. The number of iterations is set to 50-100, and the parameter adjustment step size for each iteration is dynamically set according to the current error. Local optimization stops when the error is less than 1%. Subsequently, each node shares the local optimal solution through the communication module, and a weighted fusion strategy is used to integrate the local optimal solutions from different nodes. The weight coefficient is set to 0.1-0.5 according to the importance of the node's location. Abnormal solutions are eliminated through mutual verification, and finally, globally unified protection setting parameters are output, ensuring that the protection parameters are adapted to the protection requirements of various types of converters and meet the requirements of coordinated operation of distributed nodes.

[0040] Furthermore, in this embodiment, step S3 utilizes a converter fault simulation calculation platform to conduct multi-scenario simulation tests, and corrects the setting parameters based on the test data to form a protection setting value system adapted to multiple types of converters. Specific steps include S301-S303: S301: By importing the topology model parameters, protection configuration parameters, and optimized protection setting parameters of multiple types of converters, a simulation environment consistent with the actual operating scenario is constructed. Simulation tests are performed according to the preset fault scenario sequence. Electrical quantities and protection action data during the simulation process are collected in real time, and the simulation data is time-series aligned and feature extracted.

[0041] Multi-scenario simulation tests and response data were conducted using a converter fault simulation calculation platform. During implementation, detailed topology model parameters of various types of converters were first imported into the simulation platform, including the number of bridge arms, switching device parameters, equivalent inductance, equivalent resistance, and equivalent capacitance. The switching device parameters included on-state voltage drop, switching time, and maximum withstand current. The equivalent circuit parameters were set separately according to different converter types, with the error controlled within ±5%.

[0042] Import the optimized protection setting parameters, including the action threshold and action delay of each protection function, and configure the simulation model of the protection device to clarify the logical judgment process and output signal type of the protection action. Then, build a simulation environment consistent with the actual operation scenario, and set basic parameters such as grid voltage level, load type and operation mode. The grid voltage level includes 10kV-500kV, and the load types include resistive, inductive, capacitive loads and mixed loads.

[0043] Triggering conditions for various fault scenarios are defined, including valve group short circuit fault, bridge arm open circuit fault, DC side short circuit fault, etc. Each fault scenario is set with different fault occurrence location, fault start time and fault severity. The fault start time covers different stages of converter operation, and the fault severity is set according to the short circuit resistance value from 0.01Ω to 1Ω.

[0044] The simulation process is initiated sequentially according to a preset fault scenario sequence. The simulation duration is set from 0.5 seconds before the fault to 3 seconds after the fault, and the simulation step size is set to 1 microsecond. Key data such as bridge arm current, valve group voltage, protection action signals, and fault clearing time are collected in real time during the simulation. The sampling frequency of the collected data is consistent with S1 to ensure data timing synchronization. Finally, the collected simulation data is processed for timing alignment and outlier removal. Information such as electrical quantity values ​​and action delay time at the moment of protection action is extracted to form a dataset including protection response data under each fault scenario. The dataset has no less than 50 samples, providing comprehensive scenario coverage and data support for subsequent parameter setting correction.

[0045] S302: Based on the deviation between simulation data and actual requirements, the protection setting parameters are classified and corrected. During the correction process, the coupling relationship between the basic setting value, fault characteristic parameters and model adaptation coefficients is considered. The protection action deviation is adjusted iteratively to meet the requirements.

[0046] Based on the simulation response data, key parameters are corrected and a protection setting value system is formed. During implementation, the collected simulation test response data is first analyzed in depth and classified and statistically analyzed according to fault type, converter type and fault severity. The accuracy of protection action and the deviation of action delay from theoretical requirements in each group of data are calculated. The accuracy is statistically analyzed into three categories: correct action, false action and failure to act. The deviation value calculation accuracy reaches 1 microsecond.

[0047] For different types of converters, the focus is on analyzing the adaptability of protection setting parameters to fault scenarios. When the protection action deviation of a certain type of converter exceeds 5% or false tripping or failure to trip occurs under a specific fault scenario, the corresponding setting parameters are identified as correction targets. These correction targets include key parameters such as action threshold, action delay, and cooperative weighting coefficient. Subsequently, a parameter correction model is established. Based on the magnitude of the deviation value and the variation law of fault characteristic parameters in the simulation data, the correction direction and correction magnitude of each parameter are determined. The correction magnitude is dynamically adjusted according to 1%-10% of the deviation value, with larger deviation values ​​resulting in larger correction magnitudes. At the same time, it is ensured that the corrected parameters do not exceed the rated tolerance range of the equipment, such as the action threshold not exceeding 80% of the maximum withstand current and voltage of the device.

[0048] S303: Feed the corrected parameters back to the distributed intelligent cooperative protection algorithm, and re-optimize and simulate the protection parameters until the protection action deviation meets the preset threshold.

[0049] Through multiple iterative corrections, the protection action deviation of various converters under all fault scenarios was controlled within 5%, with no false trips or failures to trip. Finally, all corrected protection setting parameters were integrated and categorized by converter type and protection function type to form a protection setting value system adaptable to multiple converter types. This system clearly defines key parameters such as protection action thresholds, action delays, and coordination logic for different converter types under various fault scenarios. Parameter lookup tables and adjustment specifications were also developed to facilitate invocation and maintenance in practical applications. The entire process, through closed-loop correction of simulation data and setting parameters, ensures that the protection setting value system possesses good universality, accuracy, and reliability, meeting the protection needs of multiple converter types under complex fault scenarios and completing the entire process of setting calculations for the new protection principle.

[0050] Example 3 is the third embodiment of the present invention, which differs from the previous two embodiments in that: This embodiment also provides a protection principle setting calculation system for multiple types of converters, including: The multilevel converter fault feature acquisition and preprocessing unit and the hierarchical fault model construction and analysis unit are unidirectionally connected. The former acquires multi-dimensional electrical signals during the operation of the converter and transmits them to the latter after filtering and processing. The hierarchical fault model construction and analysis unit is bidirectionally connected to the valve group short-circuit current decay dynamic prediction unit and the distributed intelligent collaborative protection parameter optimization unit, respectively, and interactively transmits the fault hierarchy classification results and characteristic parameters. The output of the valve group short-circuit current decay dynamic prediction unit is unidirectionally connected to the input of the distributed intelligent collaborative protection parameter optimization unit, providing it with short-circuit current decay characteristic parameters. The distributed intelligent collaborative protection parameter optimization unit is bidirectionally connected to the converter fault simulation test and data acquisition unit, transmitting the optimized setting parameters to the latter while receiving the simulation test data fed back by the latter. The output of the converter fault simulation test and data acquisition unit is unidirectionally connected to the input of the multi-type converter protection setting parameter correction and output unit, transmitting simulation data for parameter correction. The multi-type converter protection setting parameter correction and output unit integrates all data, completes the calculation of the final protection setting value, and outputs it.

[0051] This embodiment also provides an electronic device, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the protection principle setting calculation method for multi-type converters proposed in the above embodiment.

[0052] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements a protection principle setting calculation method for multiple types of converters as proposed in the above embodiments.

[0053] The storage medium proposed in this embodiment and the method for calculating the protection principle setting of multiple types of converters proposed in the above embodiments belong to the same inventive concept. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0054] Based on the above description of the implementation methods, those skilled in the art can clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0055] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0056] Example 4 is an embodiment of the present invention. Based on the above embodiments, a protection principle setting calculation method for multiple types of converters is provided.

[0057] The preferred expression for the fault hierarchy model of a multilevel converter is: in, The fault level feature values ​​output by the fault hierarchy model for multilevel converters. Duration of the fault The bridge arm current vector, The DC-side voltage vector The number of converter bridge arms, The weighting coefficient for the second derivative of the current in the k-th bridge arm is... This is the DC-side voltage correction factor corresponding to the k-th bridge arm. Let be the fault development rate coefficient of the k-th bridge arm. The time when the fault begins. Let be the coupling coefficient between the k-th bridge arm and other bridge arms. is the fault level weight coefficient for the k-th bridge arm.

[0058] Specifically, the multilevel converter fault hierarchy model achieves accurate fault level division through comprehensive calculation of multi-dimensional electrical signals. During implementation, the value range and physical meaning of various parameters involved in the model are first clarified. The number of converter arms is set to 3 to 5, depending on the type of converter. The weight coefficient of the second derivative of the current corresponding to each arm is allocated from 0.1 to 0.8 according to the importance of the arm. The DC side voltage correction coefficient is set from 0.5 to 1.2 according to the voltage level. The fault development rate coefficient is dynamically adjusted from 0.3 to 1.5 according to the converter switching frequency. The fault initiation time is located at the microsecond level using a high-precision timing device. The coupling coefficient between arms is set from 0.05 to 0.4 according to the topology connection relationship. The fault level weight coefficient is set from 0.2 to 0.9 according to the fault influence range.

[0059] Furthermore, during implementation, the model first receives input data such as fault duration, current of each bridge arm, and DC side voltage. It captures the characteristics of current mutation through second derivative calculation, corrects the nonlinear changes of voltage signal using exponential function, quantifies the mutual influence between different bridge arms through product calculation, and then combines the calculation results with weighted summation of each weight coefficient to finally output fault level feature value.

[0060] The value range of this feature is set to 0 to 10, where 0 to 3 corresponds to the first level of a single switching device fault, 3 to 7 corresponds to the second level of a single bridge arm fault, and 7 to 10 corresponds to the third level of a multi-bridge arm or DC side fault. Through this quantitative division, the electrical quantity change patterns of different fault types are accurately separated, providing accurate initial boundary conditions for the subsequent valve group short-circuit current attenuation prediction model. This ensures that the model can adapt to the fault characteristic differences of multiple types of converters, and improves the accuracy of fault identification and the rationality of the level division.

[0061] Preferably, the expression for the valve group short-circuit current attenuation prediction model is: in, The predicted valve group short-circuit current decay value, The time after the short circuit occurs. The equivalent inductance of the converter, This is the equivalent resistance of the converter. This is the equivalent capacitance of the converter. This represents the initial short-circuit current amplitude. The initial phase angle of the current. To attenuate the order of harmonic components, The amplitude coefficient of the p-th harmonic component is... Let p be the time constant of the p-th harmonic component. The angular frequency of the p-th harmonic component is... Let be the phase angle of the p-th harmonic component.

[0062] Specifically, the valve group short-circuit current attenuation prediction model dynamically captures the attenuation law of short-circuit current after a fault occurs. During implementation, it is necessary to first determine the standard values ​​of various parameters of the model. Among them, the equivalent inductance of the converter is calculated to be 1 millihenry to 100 millihenry based on the topology and component parameters; the equivalent resistance is set to 0.1 ohms to 10 ohms based on line and device losses; the equivalent capacitance is set to 1 microfarad to 100 microfarads combined with DC side energy storage configuration; the amplitude of the initial short-circuit current is determined by the current sampling value at the time of fault initiation; the initial phase angle of the current is set to 0 to 360 degrees based on the grid phase at the time of fault occurrence; the order of the attenuated harmonic components is set to 3 to 8th order based on the current oscillation characteristics; the amplitude coefficient of each harmonic component is set to 0.05 to 0.3 according to the harmonic content ratio; the time constant is set to 0.1 milliseconds to 10 milliseconds combined with the energy loss rate; the angular frequency is adjusted to 50 Hz to 500 Hz according to the grid frequency and fault characteristics; and the phase angle is set to 0 to 180 degrees based on the harmonic generation mechanism.

[0063] Furthermore, during model execution, the fundamental frequency attenuation component is first calculated based on the transient analysis theory of RLC circuits. The attenuation trend of the current is simulated using an exponential function, and the oscillation characteristics of the current are characterized by a cosine function. Then, each harmonic component is superimposed, and the harmonic amplitude is adjusted over time using a power function. The oscillation law of the harmonics is described using a sine function, ultimately outputting the short-circuit current attenuation curve within 0 to 2 seconds after the fault occurs. This model, by comprehensively considering equivalent circuit parameters, initial conditions, and harmonic components, accurately recreates the dynamic process of the short-circuit current from its inception to stability. The obtained parameters, such as attenuation rate and oscillation frequency, provide dynamic data support for protection setting calculations, solving the problem that traditional static parameters cannot adapt to the dynamic development of faults and improving the adaptability of protection setting values ​​to actual fault scenarios.

[0064] Preferably, the expression for the distributed intelligent collaborative protection algorithm is: in, This is the optimal protection setting vector output by the distributed intelligent cooperative protection algorithm. To protect the tuning parameter vector, For the parameter weight vector, This is the distributed node communication delay vector. The number of distributed protection nodes. Here are the reference values ​​for the protection parameters of the i-th node. Let be the upper limit value of the protection parameters for the i-th node. Let be the lower limit value of the protection parameter for the i-th node. Let be the objective function for the protection action of the i-th node. Let be the collaborative communication weight between the i-th node and the j-th node. Let L be the L2 norm of the parameter vectors of the i-th node and the j-th node.

[0065] Specifically, the distributed intelligent collaborative protection algorithm achieves global optimization of protection parameters through multi-node collaborative optimization. During implementation, the configuration standards of relevant algorithm parameters are first clarified. The number of distributed protection nodes is set to 3 to 10 based on the scale of converter deployment. The reference value of protection parameters for each node is determined based on the rated parameters and fault tolerance capability of the equipment. The upper limit of the parameter is set at 80% of the maximum tolerance capability of the equipment, and the lower limit is set at 120% of the normal operating parameters. The parameter weight vector is allocated from 0.1 to 0.6 according to the priority of protection functions. The distributed node communication delay vector is determined to be 1 microsecond to 10 microseconds through actual communication tests. The collaborative communication weight between nodes is set from 0.2 to 0.8 according to the communication link quality.

[0066] During algorithm implementation, an optimization objective function for the protection setting parameters is first constructed. This function comprehensively considers the deviation of parameters from reference values, the sensitivity of parameters to protection actions, and the coordination and consistency of parameters among nodes. Normalization eliminates dimensional differences between different parameters, partial derivatives are used to calculate the direction of parameter adjustment, and the second norm quantifies the degree of difference between parameters among nodes. Each distributed node performs 50 to 100 iterations based on locally collected fault data and parameters. In each iteration, the parameter step size is dynamically adjusted according to the current error. Local optimization stops when the error is less than 1%. Subsequently, the local optimal solution is shared through a high-speed communication module. A weighted fusion strategy is used to integrate the results from each node, eliminating outliers that deviate from the overall trend, and finally outputting globally unified protection setting parameters.

[0067] Preferably, the simulation accuracy calibration expression of the converter fault simulation calculation platform is: in, For the calibration coefficients of the simulation platform, The simulation output parameter vector, This is the actual running parameter vector. This is the error vector of the simulation model. For parameter dimensions, This is the error attenuation coefficient. To calibrate the reference coefficient, This is the simulation parameter matrix.

[0068] Specifically, the simulation accuracy calibration expression of the converter fault simulation calculation platform is used to improve the consistency between simulation data and actual operating data. When implementing it, the value range and calibration standard of each parameter need to be clearly defined. The parameter dimension is set to 10 to 20 dimensions according to the comprehensive requirements of simulation testing, including key electrical quantities such as current, voltage, and power. The error attenuation coefficient is set to 0.5 to 2.0 according to the accuracy requirements of the simulation model. The calibration reference coefficient is calibrated to 0.8 to 1.2 through actual equipment test data. The simulation parameter matrix includes multiple parameters such as converter topology, fault type, and protection configuration.

[0069] During calibration, simulation output parameters and actual operating parameters are first collected and time-aligned at the same time points. The squared difference between corresponding parameters in each group is calculated, and the influence of large errors on the calibration results is weakened by using an exponential function. Then, the sum of the squares of the two groups of parameters is calculated as a normalization factor, and the nonlinear relationship between parameters is corrected by second-order partial derivative calculations, finally obtaining the calibration coefficient of the simulation platform. The value range of this coefficient is set from 0.8 to 1.2. When the coefficient is between 0.95 and 1.05, it indicates that the simulation accuracy meets the requirements. When the coefficient exceeds this range, the equivalent circuit parameters and fault triggering conditions in the simulation model are adjusted according to the calibration results until the calibration coefficient meets the requirements.

[0070] Preferably, the comprehensive correction expression for the protection new principle setting calculation parameters of the multi-type converter is: in, To set the comprehensive correction value for the calculation parameters, To protect the fundamental vector of the tuning value, For the fault characteristic parameter vector, To adapt the coefficient vector to the model, For the number of parameter types, For the correction index of the type I parameter, The correction coefficient for the type I parameter, Let be the partial derivatives of the type I fundamental parameters with respect to the characteristic parameters. Let be the partial derivative of the type I feature parameter with respect to the fitting coefficient.

[0071] Specifically, during implementation, the configuration standards for each parameter should be clearly defined first. The number of parameter types should be set to 8 to 15 categories based on the comprehensiveness of the protection function, including key parameters such as action threshold, action delay, and coordination weight. The correction index should be set to 0.5 to 2.0 based on the sensitivity of the parameters, and the correction coefficient should be set to 0.8 to 1.2 based on the deviation of the simulation data.

[0072] During the correction process, the basic vector of protection setting value, the vector of fault characteristic parameters, and the vector of model adaptation coefficients are first obtained. The coupling relationship between the basic value and the characteristic parameters and adaptation coefficients is calculated one by one according to the parameter type. The mutual influence between parameters is quantified by fractional operation. The correction intensity is adjusted by exponential function. Then, the influence of characteristic parameters on basic parameters and the influence of adaptation coefficients on characteristic parameters are calculated by partial derivatives. The product of the two is multiplied by the correction coefficient to obtain the incremental correction value of the parameter.

[0073] The coupling correction value and the incremental correction value are added together to obtain the comprehensive correction value of each parameter. The corrected parameters must ensure that they do not exceed the rated tolerance range of the equipment and the reasonableness requirements of the protection action.

[0074] Example 5 is an embodiment of the present invention. Based on the above embodiments, a protection principle setting calculation method for multiple types of converters is provided.

[0075] Specifically, in step S201, the short-circuit current attenuation prediction model of the valve group and the dynamic characteristic parameters are obtained through sub-steps S2011 to S2014: S2011: Collect time-domain data of valve group short-circuit current under different fault levels of various types of converters. The sampling frequency is set from 10kHz to 100kHz to ensure the capture of high-frequency dynamic changes. The time-domain data is segmented by a sliding window with a length of 1ms to 10ms. The peak current, rise rate and fluctuation frequency information in each window are extracted. The current peak extraction accuracy reaches the ampere level, the rise rate calculation accuracy is 0.1A / ms, and the fluctuation frequency identification error does not exceed ±5Hz. Finally, a raw feature dataset of short-circuit current including more than 100,000 samples is constructed.

[0076] S2012: Based on the fault level results output by the multilevel converter fault stratification model, the original feature dataset is classified and filtered according to the first to third levels. An abnormal data removal threshold is set, and data that deviates from the feature distribution of the level by 3 times the standard deviation are judged as outliers and removed. The proportion of valid feature samples is not less than 95%.

[0077] S2013: Perform trend analysis on effective feature samples to clarify the type of short-circuit current decay trend over time. Combine core parameters such as equivalent inductance, equivalent resistance, and equivalent capacitance of the converter to establish the correlation between decay trend and circuit parameters. Build the basic framework of the valve group short-circuit current decay prediction model, which includes an energy loss calculation module and a topology change analysis module.

[0078] S2014: The gradient descent iterative optimization method is used to adjust the model parameters. The number of iterations is set to 50 to 100. The parameter adjustment step size of each iteration is dynamically set according to the current error. The optimization stops when the error is less than 1%. The adjusted parameters include the attenuation coefficient, phase parameter and harmonic component coefficient. Finally, the characteristic difference between the short-circuit current attenuation curve output by the model and the actual collected data is controlled within 5%, ensuring that the obtained short-circuit current attenuation dynamic characteristic parameters are accurate and reliable, and providing data support for subsequent protection parameter optimization.

[0079] Step S202 involves optimizing the operation and protection parameters of the distributed intelligent cooperative protection algorithm through steps S2021 to S2023: S2021: The dynamic characteristic parameters output by the valve group short-circuit current attenuation prediction model, such as attenuation rate, oscillation frequency, and harmonic component amplitude, are classified according to functional types such as overcurrent protection, overvoltage protection, and valve group fault protection. The parameter input range corresponding to each type of protection is clarified. The input range of the associated parameters for overcurrent protection is set to the interval corresponding to 1.2 to 2.5 times the rated current. The input range of the associated parameters for overvoltage protection is set to the interval corresponding to 1.1 to 1.5 times the rated voltage. The input range of the associated parameters for valve group fault protection is set according to the device's tolerance capability. Based on the classification results, an input parameter matrix with a dimension of 10 to 20 times the number of distributed nodes is constructed.

[0080] S2022: Based on the distributed deployment architecture of multiple types of converters, divide into 3 to 10 protection and coordination nodes. Each node is configured with an independent computing unit and communication module. The communication module transmission rate is not less than 1Gbps. Clarify the processing authority of each node for different types of parameters, formulate information interaction protocols between nodes, and ensure that the parameter transmission delay is controlled within 1 microsecond to 10 microseconds to ensure the real-time performance and integrity of parameter transmission.

[0081] S2023: The distributed intelligent collaborative protection algorithm is activated. Each node performs local optimization calculations based on local input parameters and adjusts the protection setting value using an iterative optimization method. The number of iterations is 50 to 100. Each iteration dynamically adjusts the step size according to the current error. When the error is less than 1%, a local optimum is output. Subsequently, each node shares the local optimum through the communication module. A weighted fusion strategy based on node importance is adopted, with the weight coefficient set to 0.1 to 0.5 according to the importance of the node's location. The local optimum is integrated and cross-verified, and abnormal solutions that deviate from the overall trend are eliminated. The threshold for judging abnormal solutions is set to deviate from the mean by 3 times the standard deviation. Finally, globally unified protection setting parameters are output to ensure that the parameters are adapted to the protection requirements of multiple types of converters and meet the collaborative action requirements of distributed nodes.

[0082] Step S301 involves sub-steps S3011 to S3013 to achieve converter fault simulation testing and response data acquisition: S3011: Import detailed topology model parameters of various types of converters into the converter fault simulation calculation platform, including the number of bridge arms, switching device parameters, equivalent circuit parameters, etc., with parameter errors controlled within ±5%. At the same time, import optimized protection setting parameters and protection device simulation models, clarify the logical judgment process and output signal type of protection actions, build a simulation environment consistent with the actual operation scenario, set the grid voltage level to 10kV to 500kV, and the load types include resistive, inductive, capacitive and mixed loads. Define various fault scenarios such as valve group short circuit, bridge arm open circuit, DC side short circuit, etc. Each fault scenario sets different fault occurrence location, start time and severity. The fault start time covers different stages of converter operation such as startup, stabilization, and load change. The fault severity is graded according to the short circuit resistance value from 0.01Ω to 1Ω.

[0083] S3012: The simulation process is started sequentially according to the preset fault scenario sequence. The simulation duration is set from 0.5 seconds before the fault to 3 seconds after the fault, and the simulation step size is 1 microsecond. The platform's data acquisition module collects key data such as bridge arm current, valve group voltage, protection action signal and fault clearing time in real time. The acquisition frequency is consistent with S1, which is 10kHz to 100kHz to ensure data timing synchronization.

[0084] S3013: Perform time-series alignment processing on the collected simulation data, use interpolation to fill in missing data, with the missing data filling ratio not exceeding 1%, remove outliers using the 3σ criterion, extract key information such as electrical quantity values ​​and action delay time at the moment of protection action, and form a dataset including protection response data under various fault scenarios. The dataset has no less than 50 samples, providing comprehensive scenario coverage and accurate data support for subsequent setting parameter correction.

Claims

1. A protection principle setting calculation method for multiple types of converters, characterized in that: include, Multi-dimensional electrical signals from various types of converters are collected, and fault characteristics are identified and levels are divided using a multi-level converter fault hierarchical model. Boundary conditions for the valve group short-circuit current attenuation prediction model are determined. The short-circuit current attenuation dynamic characteristic parameters are obtained through the valve group short-circuit current attenuation prediction model, and the distributed collaborative optimization of the protection parameters is completed by the distributed intelligent collaborative protection algorithm. Multi-scenario simulation tests were conducted using a converter fault simulation calculation platform. Based on the test data, the setting parameters were corrected to form a protection setting value system that is compatible with multiple types of converters.

2. The protection principle setting calculation method for multiple types of converters as described in claim 1, characterized in that: The multilevel converter fault hierarchy model includes using valve group switching status signals, arm current signals, and DC side voltage signals as basic input datasets. By analyzing the second derivative of each arm current, the nonlinear change of the DC side voltage, and the coupling relationship between the arms, it outputs feature values ​​that characterize the fault hierarchy, which are used to classify the hierarchy corresponding to different fault types.

3. The protection principle setting calculation method for multiple types of converters as described in claim 2, characterized in that: The valve group short-circuit current decay prediction model includes, based on equivalent inductance, equivalent resistance, equivalent capacitance and initial short-circuit current amplitude, combined with energy loss and circuit topology changes during fault development, dynamically predicting the decay curve of short-circuit current over time by superimposing fundamental attenuation components and multi-order harmonic components, and obtaining dynamic characteristic parameters including decay rate and oscillation frequency.

4. The protection principle setting calculation method for multiple types of converters as described in claim 3, characterized in that: The distributed collaborative optimization of protection parameters by the distributed intelligent collaborative protection algorithm includes the following steps: taking distributed protection nodes as units, each node, based on local protection setting parameters, reference values, and inter-node communication delays, iteratively optimizes to obtain a local optimal solution by minimizing an objective function that includes parameter deviations, protection action sensitivity, and inter-node parameter differences, and outputs globally unified protection setting parameters through inter-node collaborative verification and fusion.

5. The protection principle setting calculation method for multiple types of converters as described in claim 4, characterized in that: The multi-scenario simulation test includes constructing a simulation environment consistent with the actual operating scenario by importing topology model parameters, protection configuration parameters and optimized protection setting parameters of multiple types of converters, conducting simulation tests according to a preset fault scenario sequence, collecting electrical quantities and protection action data in real time during the simulation process, and performing time-series alignment and feature extraction on the simulation data.

6. The protection principle setting calculation method for multiple types of converters as described in claim 5, characterized in that: The method of correcting the setting parameters based on test data includes classifying and correcting the protection setting parameters according to the deviation between the simulation data and the actual requirements. During the correction process, the coupling relationship between the basic setting value, fault characteristic parameters and model adaptation coefficients is considered, and the protection action deviation is adjusted iteratively to meet the requirements.

7. The protection principle setting calculation method for multiple types of converters as described in claim 6, characterized in that: The step of correcting the tuning parameters based on test data also includes feeding the corrected parameters back to the distributed intelligent collaborative protection algorithm, and re-optimizing and simulating the protection parameters until the protection action deviation meets the preset threshold.

8. A protection principle setting calculation system for multiple types of converters, employing the protection principle setting calculation method for multiple types of converters as described in any one of claims 1 to 7, characterized in that, include: The multilevel converter fault feature acquisition and preprocessing unit and the hierarchical fault model construction and analysis unit are unidirectionally connected. The former acquires multi-dimensional electrical signals during the operation of the converter and transmits them to the latter after filtering and processing. The hierarchical fault model construction and analysis unit is bidirectionally connected to the valve group short-circuit current decay dynamic prediction unit and the distributed intelligent collaborative protection parameter optimization unit, respectively, and interactively transmits the fault hierarchy classification results and characteristic parameters. The output of the valve group short-circuit current decay dynamic prediction unit is unidirectionally connected to the input of the distributed intelligent collaborative protection parameter optimization unit, providing it with short-circuit current decay characteristic parameters. The distributed intelligent collaborative protection parameter optimization unit is bidirectionally connected to the converter fault simulation test and data acquisition unit, transmitting the optimized setting parameters to the latter while receiving the simulation test data fed back by the latter. The output of the converter fault simulation test and data acquisition unit is unidirectionally connected to the input of the multi-type converter protection setting parameter correction and output unit, transmitting simulation data for parameter correction. The multi-type converter protection setting parameter correction and output unit integrates all data, completes the calculation of the final protection setting value, and outputs it.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the protection principle setting calculation method for multi-type converters as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the protection principle setting calculation method for multi-type converters as described in any one of claims 1 to 7.

Images