CN-122026281-A - Micro-grid overcurrent protection method, device, medium and equipment

CN122026281ACN 122026281 ACN122026281 ACN 122026281ACN-122026281-A

Abstract

The application discloses a method, a device, a medium and equipment for protecting micro-grid overcurrent. The application belongs to the field of overcurrent protection, and relates to a method for generating a two-dimensional time-frequency characteristic map by collecting voltage and current transient signals of preset nodes of a micro-grid, carrying out time-frequency conversion, inputting a deep convolutional neural network model based on grid connection, island and multi-type fault data training, and outputting a real-time operation mode and a fault type. And inputting the information serving as dynamic constraint conditions into a multi-target genetic algorithm model, and optimally calculating the optimal action time limit and the current setting value of each level of protection device. And finally, transmitting the optimal fixed value to a distributed protection device, updating the local fixed value, and driving the circuit breaker to trip when the fault characteristics meet the action criteria so as to realize the self-adaptive overcurrent protection of the micro-grid. The application effectively solves the problem that the overcurrent of the micro-grid cannot be accurately and efficiently protected.

Inventors

SHENG QINGYI
LIU QINMING
Rui Qingtao
CHEN JINHUA
WANG XINGNAN
LIU FUJUN
TANG XIAOBO

Assignees

广东电网有限责任公司广州供电局

Dates

Publication Date: 20260512
Application Date: 20251205

Claims (10)

1. A microgrid overcurrent protection method, characterized by comprising: Acquiring a voltage transient signal and a current transient signal of a preset node of the micro-grid; Performing time-frequency conversion processing on the voltage transient signal and the current transient signal to generate a two-dimensional time-frequency characteristic spectrum containing fault transient energy distribution information, and inputting the two-dimensional time-frequency characteristic spectrum into a preset deep convolutional neural network model to output a real-time running mode and a fault type of the micro-grid; the deep convolutional neural network model is obtained by training grid-connected mode data, island mode data and multi-type fault data based on a micro-grid; Inputting the real-time operation mode and the fault type as dynamic constraint conditions into a preset multi-target genetic algorithm model, so that the multi-target genetic algorithm model calculates the optimal action time limit and the optimal current setting value of each level of protection devices of the micro-grid by taking the minimized total-grid fault removal time and the minimized protection coordination level difference deviation as optimization targets on the premise of meeting protection selectivity constraint; Issuing the optimal action time limit and the optimal current setting value to each distributed protection device so that the distributed protection devices update the local protection setting value based on the optimal action time limit and the optimal current setting value to obtain an updated optimal current setting value; and driving the circuit breaker to trip when the fault characteristics of the micro-grid are monitored to meet the preset action criteria according to the updated optimal current setting value, and protecting the micro-grid from overcurrent.
2. The method for protecting a micro-grid from overcurrent according to claim 1, wherein the step of obtaining a voltage transient signal and a current transient signal of a preset node of the micro-grid is specifically: acquiring voltage transient analog signals of a preset node in real time through a preset voltage sensor, and acquiring current transient analog signals of the preset node in real time through a preset current sensor; Transmitting the acquired voltage transient analog signals and current transient analog signals to a preset data acquisition unit, so that the data acquisition unit converts the analog signals into discrete digital signals through a high-speed analog-to-digital converter to obtain voltage transient signals and current transient signals of preset nodes of the micro-grid; the data acquisition unit is provided with a synchronous clock source so as to ensure that sampling phases of the voltage transient signal and the current transient signal are synchronous.
3. The method for protecting a micro-grid from overcurrent according to claim 1, wherein the performing time-frequency transformation on the voltage transient signal and the current transient signal generates a two-dimensional time-frequency characteristic map containing fault transient energy distribution information, specifically: extracting a transient current signal from the voltage transient signal and the current transient signal; Performing time-frequency decomposition operation on the transient current signal based on a wavelet transformation technology or a short-time Fourier transformation technology to obtain time-frequency domain data representing time-frequency characteristics of the transient current signal; Performing preset arithmetic operation on the time-frequency domain data of the time-frequency characteristics of the obtained representation transient current signals to obtain a two-dimensional data matrix; Wherein, the row index of the two-dimensional data matrix corresponds to the time sequence, the column index corresponds to the frequency sequence, and the matrix element value corresponds to the energy density value; And mapping the row index of the two-dimensional data matrix into a map horizontal axis and mapping the column index into a map vertical axis, and distributing pixel gray scale or color depth according to the energy density value of the elements of the two-dimensional data matrix to obtain a two-dimensional time-frequency characteristic map containing transient state energy distribution information of faults.
4. The method for protecting a micro-grid from overcurrent according to claim 1, wherein the inputting the two-dimensional time-frequency characteristic spectrum into a preset deep convolutional neural network model to output a real-time operation mode and a fault type of the micro-grid is specifically as follows: Inputting the two-dimensional time-frequency characteristic spectrum into a deep convolutional neural network model, so that each convolutional layer of the deep convolutional neural network model carries out local characteristic extraction on the two-dimensional time-frequency characteristic spectrum, each pooling layer carries out characteristic dimension reduction, each full-connection layer carries out high-dimensional characteristic fusion, and finally, a probability vector of the current real-time running mode of the micro-grid and a probability vector of a fault type are output through a Softmax activation function; The probability vector of the real-time operation mode comprises grid-connected mode probability and island mode probability, and the probability vector of the fault type comprises single-phase grounding fault corresponding probability and inter-phase short circuit fault corresponding probability.
5. The microgrid overcurrent protection method according to claim 1, wherein the real-time operation mode and the fault type are used as dynamic constraint conditions and are input into a preset multi-objective genetic algorithm model, specifically: selecting an operation mode corresponding to the probability maximum value from probability vectors of the real-time operation mode as a current real-time operation mode of the micro-grid, and selecting a fault type corresponding to the probability maximum value from probability vectors of the fault type as a current fault type; If the real-time operation mode is an island mode, generating a first constraint instruction by combining the current fault type, and inputting the first constraint instruction into the multi-target genetic algorithm model, wherein the first constraint instruction is used for reducing a minimum protection matching level difference threshold value and reducing a current setting value searching lower limit; And if the real-time operation mode is a grid-connected mode, generating a second constraint instruction by combining the current fault type, and inputting the second constraint instruction into the multi-target genetic algorithm model, wherein the second constraint instruction is a maximum protection cooperation level difference threshold.
6. The method for protecting a micro-grid from overcurrent according to claim 5, wherein the calculating the optimal action time limit and the optimal current setting value of each stage of protection device of the micro-grid by using the minimized total-grid fault removal time and the minimized protection coordination level difference deviation as optimization targets on the premise that the multi-target genetic algorithm model meets the protection selectivity constraint is specifically as follows: Determining a protection selective constraint condition based on the first constraint instruction or the second constraint instruction, wherein the constraint condition comprises that the main protection action time is smaller than the backup protection action time and the action time limit difference value of the upper and lower protection devices is not smaller than the minimum protection coordination level difference threshold value set in the first constraint instruction or the second constraint instruction; the multi-objective genetic algorithm model takes action time limit and current setting value of a protection device as optimization variables to generate a plurality of groups of initial populations; For each group of initial populations, respectively calculating a first target value for minimizing the total network fault removal time and a second target value for minimizing the protection coordination level difference deviation, namely eliminating invalid populations which do not meet the protection selectivity constraint, and executing crossover and mutation on the valid populations which meet the protection selectivity constraint to generate a plurality of groups of child populations until iteration reaches a preset convergence condition; The object of the minimized total network fault removal time is that the arithmetic average value of the main protection action time of all fault points is minimum, and the object of the minimized protection coordination level difference deviation is that the sum of the deviation of the actual protection coordination level difference between all main protection and backup protection pairs and the preset standard protection coordination level difference is minimum.
7. The method for protecting the micro-grid from overcurrent according to claim 1, wherein the optimal action time limit and the optimal current setting value are issued to each distributed protection device, so that the distributed protection devices update the local protection setting value based on the optimal action time limit and the optimal current setting value to obtain an updated optimal current setting value, and the circuit breaker is driven to trip when the fault characteristics of the micro-grid are monitored to meet the preset action criteria according to the updated optimal current setting value, specifically: packaging the calculated optimal action time limit and optimal current setting value as control instructions, and issuing the control instructions to each distributed protection device through a high-speed industrial Ethernet or an optical fiber communication network; the method comprises the steps that each distributed protection device receives and analyzes the optimal action time limit parameter and the optimal current setting value parameter in a control instruction, replaces the action time limit parameter in a local protection setting value with the analyzed optimal action time limit based on a built-in setting value storage module, and replaces the current setting value parameter in the local protection setting value with the analyzed optimal current setting value; And according to the updated optimal current setting value, after the fault characteristics of the micro-grid are monitored to meet a preset action criterion and the transient active power and the transient reactive power at the moment of the fault are calculated to confirm that the fault point is in a protection forward region, the circuit breaker is driven to trip.
8. A microgrid overcurrent protection device, comprising: the acquisition module is used for acquiring voltage transient signals and current transient signals of a preset node of the micro-grid; the first processing module is used for performing time-frequency conversion processing on the voltage transient signals and the current transient signals, generating a two-dimensional time-frequency characteristic spectrum containing fault transient energy distribution information, and inputting the two-dimensional time-frequency characteristic spectrum into a preset deep convolutional neural network model so as to output a real-time running mode and a fault type of the micro-grid; the deep convolutional neural network model is obtained by training grid-connected mode data, island mode data and multi-type fault data based on a micro-grid; The second processing module is used for inputting the real-time operation mode and the fault type into a preset multi-target genetic algorithm model as dynamic constraint conditions, so that the multi-target genetic algorithm model calculates the optimal action time limit and the optimal current setting value of each level of protection devices of the micro-grid by taking the minimized total-grid fault removal time and the minimized protection coordination level difference deviation as optimization targets on the premise of meeting protection selectivity constraint; The protection module is used for transmitting the optimal action time limit and the optimal current setting value to each distributed protection device so that the distributed protection devices update the local protection setting value based on the optimal action time limit and the optimal current setting value to obtain an updated optimal current setting value, and driving the circuit breaker to trip when the fault characteristics of the micro-grid are monitored to meet the preset action criteria according to the updated optimal current setting value to protect the micro-grid from overcurrent.
9. A computer readable storage medium, characterized in that the computer readable storage medium comprises a stored computer program, wherein the computer program, when run, controls a device in which the computer readable storage medium is located to perform the microgrid overcurrent protection method according to any one of claims 1 to 7.
10. A terminal device comprising a processor, a memory and a computer program stored in the memory and configured to be executed by the processor, the processor implementing the microgrid overcurrent protection method according to any one of claims 1 to 7 when executing the computer program.

Description

Micro-grid overcurrent protection method, device, medium and equipment Technical Field The present invention relates to the field of overcurrent protection, and in particular, to a method, an apparatus, a medium, and a device for protecting a micro-grid from overcurrent. Background Along with the transformation of the global energy structure to clean and distributed, the micro-grid system containing high-proportion new energy sources such as photovoltaic and wind power is widely applied, so that the new energy sources can be consumed in situ, and the two modes of grid connection and island can be flexibly switched, and the regional power supply reliability is improved. However, the existence of a large number of power electronic inverter interface devices (such as new energy converters and energy storage converters) in the micro-grid causes significant differences between the fault characteristics of the micro-grid and the traditional synchronous grid, and brings serious challenges to the relay protection system. In the grid-connected mode, the micro-grid can provide sufficient short-circuit current depending on a large grid, the current amplitude is obviously higher than the normal load current during faults, the traditional overcurrent protection based on the power frequency current magnitude can realize reliable action through a fixed setting value, but in the island mode, the inverter is limited by a rapid current limiting control strategy (usually only can provide short-circuit current with 1.2-1.5 times of rated current), the fault current amplitude is greatly reduced, even the fault current is close to the normal load current, so that the traditional overcurrent protection is difficult to distinguish normal overload from fault state, and the problems of refusal action (such as current is not up to the setting value during high-impedance ground faults) or misoperation (such as misjudgment as faults during load fluctuation) are easily caused. In order to cope with the contradiction, a part of self-adaptive protection scheme is provided in the prior art, but the scheme still has the obvious defects that one scheme depends on a plurality of groups of constant value tables calculated offline in advance, constant value table lookup switching is carried out according to a preset mode switching signal (such as a grid-connected switch state), transient characteristics of micro-grid source load fluctuation (such as sudden change of new energy source output and random adjustment of load) or complex faults (such as alternate short circuit and alternate single-phase ground fault occurrence) cannot be adapted in real time, protection constant values are adjusted according to a simple logic rule (such as a current change rate threshold), excavation of high-dimensional characteristics of fault signals is lacked, selective matching of protection devices of each level is difficult to ensure under the scenes of high-impedance ground faults, mode switching transition processes and the like, accidents such as time limit crossing, override trip and the like of upper and lower protection actions frequently occur, and safety and power supply continuity of micro-grid equipment are seriously threatened. These deficiencies result in the prior art being unable to accurately and efficiently protect against the overcurrent of the microgrid. Disclosure of Invention The invention provides a method, a device, a medium and equipment for protecting a micro-grid from overcurrent, which are used for solving the problem that the prior art cannot accurately and efficiently protect the micro-grid from overcurrent. In a first aspect, the present application provides a microgrid overcurrent protection method, including: Acquiring a voltage transient signal and a current transient signal of a preset node of the micro-grid; Performing time-frequency conversion processing on the voltage transient signal and the current transient signal to generate a two-dimensional time-frequency characteristic spectrum containing fault transient energy distribution information, and inputting the two-dimensional time-frequency characteristic spectrum into a preset deep convolutional neural network model to output a real-time running mode and a fault type of the micro-grid; the deep convolutional neural network model is obtained by training grid-connected mode data, island mode data and multi-type fault data based on a micro-grid; Inputting the real-time operation mode and the fault type as dynamic constraint conditions into a preset multi-target genetic algorithm model, so that the multi-target genetic algorithm model calculates the optimal action time limit and the optimal current setting value of each level of protection devices of the micro-grid by taking the minimized total-grid fault removal time and the minimized protection coordination level difference deviation as optimization targets on the premise of meeting protection selectivity constraint; Issuing