CN-121980267-A - Method and device for generating abnormal event of offshore wind power receiving end power grid, terminal equipment and storage medium

CN121980267ACN 121980267 ACN121980267 ACN 121980267ACN-121980267-A

Abstract

The invention discloses a method and a device for generating an abnormal event of an offshore wind power receiving end power grid, terminal equipment and a storage medium, and belongs to the field of intelligent power grids. The method comprises the steps of respectively extracting time sequence characteristics from meteorological data, offshore wind power output data, load data of a receiving end power grid and tide operation data to obtain meteorological characteristics, wind power characteristics, load characteristics and tide characteristics, constructing a multidimensional characteristic matrix, inputting the multidimensional characteristic matrix into a preset depth generation model to perform characteristic mapping to obtain potential vectors, performing abnormal mode learning according to the potential vectors under the constraint of the meteorological characteristics to obtain a coupling distribution rule, and performing tide simulation according to an abnormal event disturbance path generated by the coupling distribution rule, the meteorological characteristics and the tide operation data under the constraint of operation boundary and tide consistency constructed by the load data and the tide operation data to generate abnormal event time sequence data.

Inventors

DAI YUE
ZHAO RUIFENG
GUO WENXIN
LAN TIAN
WU YUEZHOU
LI YUMIN

Assignees

广东电网有限责任公司电力调度控制中心

Dates

Publication Date: 20260505
Application Date: 20260122

Claims (10)

1. The method for generating the abnormal event of the offshore wind power receiving end power grid is characterized by comprising the following steps of: Acquiring meteorological data, offshore wind power output data, load data of a receiving end power grid and tide operation data; respectively extracting time sequence characteristics from the meteorological data, the offshore wind power output data, the load data of the receiving-end power grid and the power flow operation data to obtain meteorological characteristics, wind power characteristics, load characteristics and power flow characteristics, and constructing a multidimensional characteristic matrix according to the meteorological characteristics, the wind power characteristics, the load characteristics and the power flow characteristics; inputting the multidimensional feature matrix into a preset depth generation model for feature mapping to obtain potential vectors used for representing abnormal mode statistics rules, and carrying out abnormal mode learning according to the potential vectors under the constraint of meteorological features to obtain a coupling distribution rule used for representing meteorological disturbance and abnormal states of a power grid; Performing time sequence simulation according to the coupling distribution rule, the meteorological features and the wind power features to generate an abnormal event disturbance path; Constructing an operation boundary constraint and a power flow consistency constraint according to the load data and the power flow operation data, and performing power flow simulation according to the abnormal event disturbance path under the operation boundary constraint and the power flow consistency constraint to generate abnormal event time sequence data; and storing the abnormal event time sequence data into a preset event sample library so as to enable the power grid security assessment to be carried out according to the abnormal event time sequence data in the event sample library.
2. The method for generating the abnormal event of the offshore wind power receiving end power grid according to claim 1, wherein the preset depth generation model comprises a variation encoder; Inputting the multidimensional feature matrix into a preset depth generation model for feature mapping to obtain potential vectors used for representing abnormal mode statistical rules, wherein the method comprises the following steps: inputting the multi-dimensional feature matrix into a variation encoder of a preset depth generation model to perform dimension reduction and feature compression processing on the multi-dimensional feature matrix, and generating a mean vector and a variance vector corresponding to potential space; and performing reparameterization sampling operation according to the mean value vector and the variance vector to obtain a potential vector used for representing an abnormal mode statistical rule.
3. The method for generating the abnormal event of the offshore wind power receiving end power grid according to claim 2, wherein the preset depth generation model further comprises the steps of generating an countermeasure network and a diffusion submodel; The abnormal mode learning is carried out according to the potential vector under the constraint of meteorological features to obtain a coupling distribution rule for representing meteorological disturbance and an abnormal state of a power grid, and the method comprises the following steps: performing dimension splicing and fusion on the meteorological features serving as condition vectors and the potential vectors to obtain a joint input vector used for representing the weather constraint; Inputting the combined input vector into the generation countermeasure network to generate a candidate abnormal mode sample for representing the coupling relation of the meteorological power grid, wherein the candidate abnormal mode sample comprises time sequence characteristics corresponding to a wind power output fluctuation mode and a power grid state response mode; Comparing the candidate abnormal pattern sample with a preset real abnormal pattern sample to obtain distributed difference feedback information; adjusting and generating coding parameters of the countermeasure network and the variation encoder according to the distributed difference feedback information, and regenerating candidate abnormal mode samples according to the adjusted countermeasure network and the variation encoder; And inputting the regenerated candidate abnormal mode sample into the diffusion submodel for reverse denoising, correcting the time sequence evolution logic of the abnormal mode, extracting the mapping relation between the meteorological disturbance parameters and the power grid abnormal state parameters, and taking the mapping relation as a coupling distribution rule for representing the meteorological disturbance and the power grid abnormal state.
4. The method for generating an abnormal event of an offshore wind power receiving end power grid according to claim 3, wherein the generating an abnormal event disturbance path by performing time sequence simulation according to the coupling distribution rule, the meteorological features and the wind power features comprises: extracting an association mapping relation between the meteorological features and the wind power features according to the coupling distribution rule, and taking the meteorological features and the wind power features at a starting time point as initial disturbance states under the preset starting time point; constructing a full-period time sequence simulation frame of an abnormal event by taking the initial disturbance state as a starting point, injecting the association mapping relation as dynamic constraint into a simulation process, and generating a time sequence track of wind power side disturbance quantity according to the full-period time sequence simulation frame based on the power fluctuation rate and the set rotating speed in the wind power characteristics; Simulating according to a coupling distribution rule and a time sequence track of wind power side disturbance quantity to generate a time sequence track of receiving end power grid side disturbance quantity, wherein the receiving end power grid side disturbance quantity comprises bus voltage disturbance quantity, line power flow disturbance quantity, frequency disturbance quantity and energy storage output disturbance quantity; And generating an abnormal event disturbance path according to the time sequence track of the wind power side disturbance quantity and the time sequence track of the receiving end power grid side disturbance quantity.
5. The method for generating the abnormal event of the offshore wind power receiving end power grid according to claim 4, wherein the load data comprises load node active demand, load node reactive demand, load fluctuation index, load ramp rate, maximum allowable frequency deviation and system frequency, and the power flow operation data comprises bus voltage amplitude, line power flow, line maximum power, energy storage output and node phase angle difference; Constructing an operation boundary constraint and a power flow consistency constraint according to the load data and the power flow operation data, wherein the method comprises the following steps: determining the maximum allowable fluctuation range, the maximum climbing rate and the stable operation power range of the load power according to the load node active demand, the load node reactive demand, the load fluctuation index and the load climbing rate, and constructing the operation boundary constraint according to the maximum allowable fluctuation range, the maximum climbing rate and the stable operation power range; Constructing voltage phase angle association constraint according to the bus voltage amplitude and the node phase angle difference, constructing power balance constraint according to energy storage output, load node active demand and load node reactive demand, and constructing line power flow transmission constraint according to line power flow and the node phase angle difference; And taking the voltage phase angle association constraint, the power balance constraint and the line power flow transmission constraint as the power flow consistency constraint.
6. The method for generating an abnormal event of a wind power receiving end grid at sea according to claim 5, wherein under the operation boundary constraint and the power flow consistency constraint, performing power flow simulation according to the abnormal event disturbance path to generate abnormal event time sequence data, comprising: Inputting the disturbance path of the abnormal event into a preset power grid dynamic power flow simulation model, initializing simulation parameters, carrying out time sequence power flow simulation according to preset time steps, and outputting power grid state data of each time step, wherein the simulation parameters comprise a convergence threshold, the maximum iteration times and a node voltage initial value; Determining injection power variation, power loss variation and power flow variation according to the operation boundary conditions and the output power grid state data of each time step, and calculating to obtain an active power association error according to the injection power variation, the power loss variation, the power flow variation and a preset active power reference value; calculating to obtain a voltage correlation error according to a bus voltage amplitude value in the power flow consistency constraint and a preset node voltage reference value, and calculating to obtain a line overload correlation error according to a line power flow and a line maximum power in the power flow consistency constraint; Calculating to obtain a load flow consistency score according to the active power correlation error, the voltage correlation error and the line overload correlation error; According to the operation boundary constraint and the output power grid state data of each time step, calculating to obtain the deviation of the actual power fluctuation range and the maximum allowable fluctuation range of the load and the deviation of the actual climbing rate and the maximum climbing rate, and screening the load data in the stable operation power range based on the deviation; calculating the sum of node injection active power, the sum of node outflow active power and power loss according to the screened load data, and calculating to obtain an active power balance error according to the sum of node injection active power, the sum of node outflow active power and the power loss; Calculating to obtain a frequency deviation error according to a preset rated frequency, a system frequency in the screened load data and a maximum allowable frequency deviation; Calculating to obtain a voltage out-of-limit error according to the busbar voltage amplitude in the power flow consistency constraint, and calculating to obtain a physical constraint satisfaction degree score according to the active power balance error, the frequency deviation error and the voltage out-of-limit error; calculating to obtain a model confidence score according to the discrimination confidence coefficient corresponding to the generated countermeasure network, the reconstruction confidence coefficient corresponding to the variation encoder and the noise prediction confidence coefficient corresponding to the diffusion submodel; And calculating to obtain a comprehensive score according to the model confidence score, the physical constraint satisfaction score and the tide consistency score, and integrating the power grid state data with the comprehensive score being greater than or equal to a preset score threshold value to generate abnormal event time sequence data.
7. The method for generating an abnormal event in an offshore wind power receiving grid according to claim 1, further comprising, before extracting the timing characteristics: acquiring space positioning information of meteorological monitoring points, offshore wind farms and receiving end grid nodes; converting the space positioning information into the same preset geographic coordinate system to obtain first meteorological data, first offshore wind power output data, first load data and first tide operation data which are used for representing coordinate unification; carrying out spatial correlation according to the spatial positioning information of the meteorological monitoring points and the spatial positioning information of the offshore wind farm to obtain a first spatial relationship used for representing the mapping relationship between the first meteorological data and the first offshore wind power output data; performing spatial correlation according to the spatial positioning information of the offshore wind farm and the spatial positioning information of the receiving end power grid node to obtain a second spatial relationship used for representing the mapping relationship between the first offshore wind power output data and the first load data; According to the first spatial relationship and the second spatial relationship, carrying out spatial coordinate matching, and eliminating corresponding first meteorological data, first offshore wind power output data, first load data and first tide operation data under the condition of space positioning information deficiency and association conflict to obtain second meteorological data, second offshore wind power output data, second load data and second tide operation data after elimination; And supplementing and aligning the second meteorological data, the second offshore wind power output data, the second load data and the second tide operation data according to a linear interpolation method to obtain final meteorological data, offshore wind power output data, load data of a receiving end power grid and tide operation data.
8. An abnormal event generating device of an offshore wind power receiving end power grid is characterized by comprising: the data acquisition module is used for acquiring meteorological data, offshore wind power output data, load data of a receiving end power grid and tide operation data; The characteristic matrix construction module is used for respectively carrying out time sequence characteristic extraction on the meteorological data, the offshore wind power output data, the load data of the receiving end power grid and the tide operation data to obtain meteorological characteristics, wind power characteristics, load characteristics and tide characteristics, and constructing a multidimensional characteristic matrix according to the meteorological characteristics, the wind power characteristics, the load characteristics and the tide characteristics; The disturbance abnormal coupling distribution rule learning module is used for inputting the multidimensional feature matrix into a preset depth generation model for feature mapping to obtain potential vectors used for representing abnormal mode statistics rules, and carrying out abnormal mode learning according to the potential vectors under the constraint of meteorological features to obtain coupling distribution rules used for representing meteorological disturbance and abnormal states of a power grid; the abnormal event disturbance path generation module is used for carrying out time sequence simulation according to the coupling distribution rule, the meteorological features and the wind power features to generate an abnormal event disturbance path; the abnormal event time sequence data generation module is used for constructing an operation boundary constraint and a power flow consistency constraint according to the load data and the power flow operation data, and carrying out power flow simulation according to the abnormal event disturbance path under the operation boundary constraint and the power flow consistency constraint to generate abnormal event time sequence data; The abnormal event time sequence data storage module is used for storing the abnormal event time sequence data into a preset event sample library so as to enable the power grid security assessment to be carried out according to the abnormal event time sequence data in the event sample library.
9. 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 a method for generating an abnormal event of an offshore wind power receiving grid according to any one of claims 1 to 7 when the computer program is executed.
10. 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 execute an abnormal event generating method of an offshore wind power receiving grid according to any one of claims 1 to 7.

Description

Method and device for generating abnormal event of offshore wind power receiving end power grid, terminal equipment and storage medium Technical Field The invention relates to the technical field of smart grids, in particular to a method and a device for generating an abnormal event of an offshore wind power receiving end grid, terminal equipment and a storage medium. Background In recent years, the global energy transformation process is accelerated, large-scale offshore wind power gradually becomes an important component of a power system, the high-proportion access of the offshore wind power and the clean energy duty ratio are continuously improved, and the operation mode of a receiving end power grid is evolved from centralized control to distributed interaction. The operation state of the receiving end power grid is extremely sensitive to meteorological conditions under the influence of offshore environment, particularly, typhoon, cold tide, extreme storm and other extreme meteorological events frequently occur, the output of a fan is suddenly reduced due to the triggering of a protection mechanism or the stall control action, and further, the chain reactions such as line overload, bus voltage drop and abnormal charging and discharging of an energy storage device are caused, so that the safety and the stability of the power grid are threatened. At present, the research of the abnormal event of the power grid is mainly focused on detection and classification based on a historical sample, for example, a disturbance scene is manually set, single disturbance parameters such as load mutation amplitude and line fault type are set, then simulation analysis is carried out by combining a power grid simulation platform, or manual statistical modeling is carried out based on historical fault data, so that the classification of the known abnormal event is realized. The historical fault samples which can be collected in the mode are rare, the combined scenes of various meteorological conditions and wind power output states are difficult to cover, the multivariable coupling effect under meteorological driving cannot be reflected, and the generated scene event lacks of reality. Disclosure of Invention The embodiment of the invention provides a method, a device, a terminal device and a storage medium for generating an abnormal event of a marine wind power receiving end power grid, which can effectively solve the problems that a disturbance scene is set by manually setting a single disturbance parameter, a historical fault sample is rare and a combined scene of various meteorological conditions and wind power output states is difficult to cover in the prior art. The embodiment of the invention provides a method for generating an abnormal event of a marine wind power receiving end power grid, which comprises the following steps: Acquiring meteorological data, offshore wind power output data, load data of a receiving end power grid and tide operation data; respectively extracting time sequence characteristics from the meteorological data, the offshore wind power output data, the load data of the receiving-end power grid and the power flow operation data to obtain meteorological characteristics, wind power characteristics, load characteristics and power flow characteristics, and constructing a multidimensional characteristic matrix according to the meteorological characteristics, the wind power characteristics, the load characteristics and the power flow characteristics; inputting the multidimensional feature matrix into a preset depth generation model for feature mapping to obtain potential vectors used for representing abnormal mode statistics rules, and carrying out abnormal mode learning according to the potential vectors under the constraint of meteorological features to obtain a coupling distribution rule used for representing meteorological disturbance and abnormal states of a power grid; Performing time sequence simulation according to the coupling distribution rule, the meteorological features and the wind power features to generate an abnormal event disturbance path; Constructing an operation boundary constraint and a power flow consistency constraint according to the load data and the power flow operation data, and performing power flow simulation according to the abnormal event disturbance path under the operation boundary constraint and the power flow consistency constraint to generate abnormal event time sequence data; and storing the abnormal event time sequence data into a preset event sample library so as to enable the power grid security assessment to be carried out according to the abnormal event time sequence data in the event sample library. Further, the preset depth generation model includes a variation encoder; Inputting the multidimensional feature matrix into a preset depth generation model for feature mapping to obtain potential vectors used for representing abnormal mode statistical rules, where