Modeling Method, Device, Terminal and Medium for High-Temperature Scenario Simulation Model of Power System

Abstract

The present application discloses a method, device, terminal and medium for modeling a high-temperature scenario simulation model of a power system, relating to the technical field of power system operation simulation. The solution provided by the present application characterizes the correlation characteristics between multi-dimensional meteorological factors through the R-vine copula function to generate a Markov multi-dimensional transition kernel; establishes a double-layer Markov chain weather model, where the upper-layer model describes the transition of typical weather states, the lower-layer model divides the 24 hours within a day into time periods, and establishes a state transition matrix within each time period, generates a non-stationary change time series of intra-day meteorological factors based on the multi-dimensional transition kernel, integrates the double-layer model to obtain a complete high-temperature scenario, realizes more accurate characterization of the non-linear dependence relationship between multiple variables in the high-temperature scenario, thereby obtaining a more accurate extreme high-temperature weather sequence, which helps to improve the accuracy of the reliability assessment of the high-temperature scenario of the power system.

Description

Technical Field

[0001] This application relates to the technical field of power system operation simulation, and particularly to a method, device, terminal and medium for modeling a power system high-temperature scenario simulation model. Background Art

[0002] Under the new power system paradigm, the supply side will be dominated by new energy power generation such as wind power / solar power with significant uncertainties; in the demand side, the proportion of responsive loads will increase significantly. Affected by factors such as user response behavior, the uncertainty of the demand side is significantly enhanced, and in extreme natural disasters, the risk of power supply interruption increases sharply. Especially the emergence of high-temperature weather is becoming more and more frequent, bringing great challenges to the safe and stable operation of the power system.

[0003] In order to improve the reliability of the new power system to cope with increasingly frequent extreme high-temperature events, the commonly used method at present is to build an operation simulation model of the power system in high-temperature scenarios, so as to evaluate the reliability of the power system in high-temperature scenarios through simulation, and optimize the design or transformation plan of the power system based on the simulation results. However, in actual scenarios, there are also complex interactions among multiple meteorological factors in high-temperature weather, and the existing power system high-temperature scenario simulation models are difficult to accurately describe the non-linear dependence relationship among multiple variables in high-temperature scenarios, resulting in the technical problem of inaccurate reliability assessment of the existing power system high-temperature scenarios. Summary of the Invention

[0004] This application provides a method, device, terminal and medium for modeling a power system high-temperature scenario simulation model, which is used to solve the technical problem of inaccurate reliability assessment of the existing power system high-temperature scenarios.

[0005] To solve the above technical problem, the first aspect of this application provides a method for modeling a power system high-temperature scenario simulation model, including:

[0006] Based on several meteorological factors corresponding to the high-temperature scenario, calculate the marginal distribution of the meteorological factors through the kernel density estimation method;

[0007] Based on the marginal distribution, construct an R-vine copula maximum spanning tree through an improved Prim's algorithm, so as to determine the multi-dimensional joint probability distribution of the meteorological factors based on the R-vine copula maximum spanning tree;

[0008] Based on historical meteorological data, by clustering the historical meteorological data, the typical weather states corresponding to the historical meteorological data are determined. Then, based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, a typical weather state transition sub-model is constructed, where the historical weather state sequence is a typical weather state sequence generated in the time order of the historical meteorological data, and the typical weather state transition sub-model is used to simulate the process of typical weather state transition;

[0009] Based on the historical meteorological data, combined with the Markov multi-dimensional transition kernel, an intraday meteorological factor change sub-model is constructed, where the Markov multi-dimensional transition kernel is obtained based on the multi-dimensional joint probability distribution, and the intraday meteorological factor change sub-model is used to simulate the non-stationary change process of the intraday meteorological factors corresponding to the typical weather state based on the typical weather state sequence output by the typical weather state transition sub-model;

[0010] Integrate the typical weather state transition sub-model and the intraday meteorological factor change sub-model to obtain a power system high-temperature scenario simulation model based on a two-layer Markov architecture.

[0011] Preferably, based on the marginal distributions, by improving Prim's algorithm, constructing an R-vine copula maximum spanning tree includes:

[0012] Based on the marginal distributions of each meteorological factor, combined with the calculation method of the Kendall correlation coefficient, calculate the correlation coefficients between each meteorological factor, and determine the edge weights between each meteorological factor according to the correlation coefficients;

[0013] Taking each meteorological factor as a node, combined with the edge weights between each meteorological factor, construct an R-vine copula maximum spanning tree by improving Prim's algorithm.

[0014] Preferably, taking each meteorological factor as a node, combined with the edge weights between each meteorological factor, constructing an R-vine copula maximum spanning tree includes:

[0015] Taking each meteorological factor as a node, selecting one from all the nodes as the starting node, associating the nodes according to the connection relationship between the nodes and the edge weights according to Prim's algorithm to determine the maximum spanning tree, and then using the Akaike information criterion to select the optimal bivariate copula function to connect the nodes of the first-layer tree to construct edges and determine the first-layer maximum spanning tree;

[0016] The loop is based on the maximum spanning tree of the previous layer, calculates the correlation coefficient between each pair of conditional variables, determines the connectivity relationship of the conditional variables, and then determines the maximum spanning tree of the current layer according to the Prim's algorithm and the Akaike information criterion, until the maximum spanning trees of all layers and the optimal bivariate copula of the edges in each layer tree are determined, and the maximum spanning tree of the R-vine copula is obtained.

[0017] Preferably, based on historical meteorological data, by clustering the historical meteorological data, the typical weather states corresponding to the historical meteorological data are determined, and then based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, a typical weather state transition sub-model is constructed, including:

[0018] Based on the preset historical meteorological data, a historical meteorological data set matrix is constructed according to the data date and the number of meteorological factors;

[0019] According to the historical meteorological data set matrix, according to the preset similarity calculation formula, the weather similarity between meteorological data on different dates is calculated, and then the historical meteorological data is clustered according to the weather similarity, the typical weather state corresponding to each date in the historical meteorological data is determined, and a historical weather state sequence is generated according to the time sequence of the historical meteorological data;

[0020] Based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, a typical weather state transition matrix and a cumulative probability transition matrix are constructed in sequence;

[0021] According to the cumulative probability transition matrix and the preset random generation algorithm, a typical weather state transition sub-model is constructed.

[0022] Preferably, based on the historical meteorological data, combined with the Markov multi-dimensional transition kernel, an intraday meteorological factor change sub-model is constructed, including:

[0023] Based on the historical meteorological data, the intraday meteorological factor sequences in the historical meteorological data with the same weather state are merged;

[0024] According to the preset intraday time period division information, the meteorological factor sequence is divided into multiple state intervals;

[0025] Through the typical weather state transition sub-model, the weather state transition matrix and the cumulative probability transition matrix of each state interval are determined respectively, and then combined with the Markov multi-dimensional transition kernel, an intraday meteorological factor change sub-model is constructed.

[0026] The second aspect of the present application provides a modeling device for a power system high-temperature scenario simulation model, including:

[0027] A meteorological factor distribution calculation unit, which is used to calculate the marginal distribution of the meteorological factors based on a number of meteorological factors corresponding to a high-temperature scenario through a kernel density estimation method;

[0028] A multi-dimensional joint probability distribution determination unit, which is used to construct a maximum spanning tree of R-vine copula based on the marginal distribution through an improved Prim's algorithm, so as to determine the multi-dimensional joint probability distribution of the meteorological factors based on the maximum spanning tree of R-vine copula;

[0029] An upper-layer sub-model construction unit, which is used to determine the typical weather states corresponding to the historical meteorological data by clustering the historical meteorological data based on the historical meteorological data, and then construct a typical weather state transition sub-model based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, wherein the historical weather state sequence is a typical weather state sequence generated according to the time sequence of the historical meteorological data, and the typical weather state transition sub-model is used to simulate the process of typical weather state transition;

[0030] A lower-layer sub-model construction unit, which is used to construct an intra-day meteorological factor change sub-model based on the historical meteorological data, combined with a Markov multi-dimensional transition kernel, wherein the Markov multi-dimensional transition kernel is obtained based on the multi-dimensional joint probability distribution, and the intra-day meteorological factor change sub-model is used to simulate the non-stationary change process of the intra-day meteorological factors corresponding to the typical weather state based on the typical weather state sequence output by the typical weather state transition sub-model;

[0031] A high-temperature scenario simulation model construction unit, which is used to integrate the typical weather state transition sub-model and the intra-day meteorological factor change sub-model to obtain a power system high-temperature scenario simulation model based on a two-layer Markov architecture.

[0032] Preferably, the multi-dimensional joint probability distribution determination unit is specifically used for:

[0033] Based on the marginal distribution of each meteorological factor, combined with the calculation method of the Kendall correlation coefficient, calculate the correlation coefficients between each meteorological factor, and determine the edge weights between each meteorological factor according to the correlation coefficients;

[0034] Taking each meteorological factor as a node, combined with the edge weights between each meteorological factor, construct a maximum spanning tree of R-vine copula through an improved Prim's algorithm.

[0035] Preferably, the upper-layer sub-model construction unit is specifically used for:

[0036] Based on the preset historical meteorological data, construct a historical meteorological data set matrix according to the data date and the number of meteorological factors;

[0037] According to the historical meteorological data set matrix, calculate the weather similarity between meteorological data of different dates according to a preset similarity calculation formula, and then cluster the historical meteorological data according to the weather similarity to determine the typical weather state corresponding to each date in the historical meteorological data, and generate a historical weather state sequence according to the time sequence of the historical meteorological data;

[0038] Based on the historical weather state sequence, combine the Markov architecture logic and the random generation algorithm to sequentially construct a typical weather state transition matrix and a cumulative probability transition matrix;

[0039] According to the cumulative probability transition matrix and a preset random generation algorithm, construct a typical weather state transition sub-model;

[0040] The lower-layer sub-model construction unit is specifically used for:

[0041] Based on the historical meteorological data, merge the intra-day meteorological factor sequences in the historical meteorological data with the same weather state;

[0042] According to the preset intra-day time period division information, divide the meteorological factor sequence into multiple state intervals;

[0043] Through the typical weather state transition sub-model, respectively determine the weather state transition matrix and the cumulative probability transition matrix of each state interval, and then combine the Markov multi-dimensional transition kernel to construct an intra-day meteorological factor change sub-model.

[0044] The third aspect of the present application provides a power system high-temperature scenario simulation model modeling terminal, including: a memory and a processor;

[0045] The memory is used to store program codes, and the program codes are used to implement the power system high-temperature scenario simulation model modeling method provided in the first aspect of the present application;

[0046] The processor is used to read and execute the program codes.

[0047] The fourth aspect of the present application provides a computer-readable storage medium, in which program codes are stored, and the program codes are used to be read and executed by a processor to implement the power system high-temperature scenario simulation model modeling method provided in the first aspect of the present application.

[0048] It can be seen from the above technical solutions that the present application has the following advantages:

[0049] The solution provided by this application characterizes the correlation characteristics among multi-dimensional meteorological factors through the R-vine copula function to generate a Markov multi-dimensional transition kernel; establishes a two-layer Markov chain weather model. The upper-layer model describes the transition of typical weather states, and the lower-layer model divides the 24 hours within a day into time periods and establishes the state transition matrix within each time period. Based on the multi-dimensional transition kernel, a non-stationary change time series of intra-day meteorological factors is generated, and the two-layer model is integrated to obtain a complete high-temperature scenario, realizing a more accurate characterization of the non-linear dependence relationship among multiple variables in the high-temperature scenario, thereby obtaining a more accurate extreme high-temperature weather sequence, which helps to improve the accuracy of the reliability assessment of the high-temperature scenario of the power system. Description of the Drawings

[0050] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following-described drawings are only some embodiments of this application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0051] Figure 1 It is a schematic flowchart of an embodiment of a method for modeling a high-temperature scenario simulation model of a power system provided by this application.

[0052] Figure 2 It is a schematic topological diagram of the first layer tree of the R-vine.

[0053] Figure 3 It is a schematic topological diagram of the second layer tree of the R-vine.

[0054] Figure 4 It is a schematic topological diagram of the complete maximum spanning tree of the R-vine copula.

[0055] Figure 5 It is a comparison diagram of the solar radiation sequence, historical data, and the output sequence of the traditional Markov model generated based on the solution of this application.

[0056] Figure 6 It is a comparison diagram of the temperature sequence, historical data, and the output sequence of the traditional Markov model generated based on the solution of this application.

[0057] Figure 7 It is a comparison diagram of the wind speed sequence, historical data, and the output sequence of the traditional Markov model generated based on the solution of this application.

[0058] Figure 8 It is a curve graph of the autocorrelation characteristics of the irradiance sequence.

[0059] Figure 9 It is a curve graph of the autocorrelation characteristics of the temperature sequence.

[0060] Figure 10 It is a curve graph of the autocorrelation characteristics of the wind speed sequence.

[0061] Figure 11 It is a schematic structural diagram of an embodiment of a modeling device for a high-temperature scenario simulation model of a power system provided by the present application.

[0062] Figure 12 It is a schematic structural diagram of an embodiment of a modeling terminal for a high-temperature scenario simulation model of a power system provided by the present application. Specific embodiments

[0063] The embodiments of the present application provide a method, device, terminal and medium for modeling a high-temperature scenario simulation model of a power system, which are used to solve the technical problem of inaccurate reliability assessment of the existing high-temperature scenario of the power system.

[0064] In order to make the invention purpose, features and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the embodiments described below are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0065] First, a detailed description of an embodiment of a method for modeling a high-temperature scenario simulation model of a power system provided by the present application is as follows:

[0066] Please refer to Figure 1 , the method for modeling a high-temperature scenario simulation model of a power system provided by the embodiments of the present application includes:

[0067] Step 101: Based on several meteorological factors corresponding to the high-temperature scenario, calculate the marginal distribution of the meteorological factors by means of kernel density estimation.

[0068] It should be noted that first, the marginal distribution of each meteorological factor (such as temperature, wind speed, solar radiation, etc.) is established by the Gaussian kernel density estimation method. Specifically: the Gaussian kernel function of non-parametric kernel density estimation is used to fit each meteorological factor. In this way, it is not necessary to assume the theoretical probability distribution of the variable, and its probability distribution can be directly obtained from historical data.

[0069] For a meteorological factor with L data samples, its PDF (Probability Density Function) can be estimated by the following kernel density function:

[0070] (1)

[0071] In the formula is the bandwidth, m l represents the l-th meteorological factor. The Gaussian kernel function can be expressed as:

[0072] (2)

[0073] Step 102: Based on the marginal distribution, construct the maximum spanning tree of the R-vine copula by improving the Prim's algorithm, and determine the multi-dimensional joint probability distribution of meteorological factors based on the maximum spanning tree of the R-vine copula;

[0074] It should be noted that the R-vine copula of the L-dimensional meteorological factors consists of trees and edges. These edges are respectively constructed by specific conditional and unconditional copulas, and each edge corresponds to two nodes. Let represent the edge set of tree , , then each edge of the R-vine can be expressed as " ", where and represents the conditional set, and the elements in the set are all nodes. It should be noted that in the first tree is an empty set. An edge in tree can be determined by two edges and in tree . Then the copula density function corresponding to the edge can be expressed as . At this time, the joint probability density function of the L-dimensional variables can be obtained from its marginal distribution and the specific R-vine structure:

[0075] (3)

[0076] If the edge shares the same nodes with , then , . The conditional CDF (Cumulative Distribution Function) in formula (3) can be calculated by the following formula:

[0077] (4)

[0078] The conditional CDF It can also be obtained by the same method. In the formula represents the conditional copula CDF of side a.

[0079] For the construction of the edges, first, the Kendall correlation coefficient is used to calculate the correlations between all meteorological factors, and these correlations are used as the weight values of each edge:

[0080] (5)

[0081] where C and D are the numbers of concordant pairs and discordant pairs, respectively.

[0082] Execute the improved Prim's algorithm. Select one node from all nodes as the starting node. Secondly, among the nodes that are connected to it but not selected, choose an edge with the largest weight, and then add a new node. Repeat this process of connecting the edge with the largest weight and adding new nodes until all nodes are selected to determine the maximum spanning tree. Then, use the Akaike information criterion (AIC) to select the optimal bivariate copula function to connect the nodes of the first-layer tree to construct the edges and determine the maximum spanning tree.

[0083] (6)

[0084] where is the number of parameters in the copula function, is the likelihood function. The selected bivariate copula families include Gaussian copula, t copula, Frank copula, Joe copula, Gumbel copula, Clayton copula. To capture the dependence relationships of variables under extreme conditions, the survival forms of Joe, Gumbel, and Clayton are also introduced.

[0085] More specifically, assume there are 5 meteorological factor variables , then the construction process of the multivariate R-vine model is as follows:

[0086] 1) Construct the first-layer tree

[0087] Input the variable data , calculate the empirical Kendall coefficient between the variables to obtain a complete connected graph. Run the adjusted Prim's algorithm to determine the first-layer maximum spanning tree. Determine the optimal bivariate copula for each edge by AIC and estimate its parameters. Then calculate the conditional variables and and through Figure 2as shown

[0088] 2) Construct the second - layer tree

[0089] Based on the constructed first - layer tree, calculate the empirical Kendall coefficient between all pairs of conditional variables to obtain the connectivity graph of conditional variables. Run the adjusted Prim's algorithm to determine the maximum spanning tree of conditional variables and the edges in the tree . Determine the optimal bivariate copula for each edge by AIC . Then, through and Equation (4), calculate the conditional variables and and . The constructed second - layer tree structure is as Figure 3 shown

[0090] 3) Construct the trees of the remaining layers

[0091] According to the construction process of the previous step, continue to calculate the correlation coefficient between each pair of conditional variables based on the maximum spanning tree of the previous layer, determine the connectivity relationship of conditional variables, and then, according to Prim's algorithm and Akaike information criterion, determine the maximum spanning tree of the current layer until the maximum spanning trees of all layers and the optimal bivariate copulas of the edges in each layer of the tree are determined , then the construction of the R - vine of multi - dimensional variables is completed. Substitute the marginal distributions of all variables and the bivariate copulas of non - conditional or conditional variables in each layer of the tree into Equation (3) to obtain the joint probability density function of multi - variable variables. The structures of the maximum spanning trees of each layer constructed are as Figure 4 shown

[0092] Step 103: Based on historical meteorological data, by clustering the historical meteorological data, determine the typical weather states corresponding to the historical meteorological data. Then, based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, construct a typical weather state transition sub - model, where the historical weather state sequence is a typical weather state sequence generated according to the time order of historical meteorological data, and the typical weather state transition sub - model is used to simulate the process of typical weather state transition;

[0093] It should be noted that the L - dimensional historical meteorological data within N days are used to construct the dataset matrix Z according to the daily average value:

[0094] (7)

[0095] Normalize the dataset to eliminate the differences between units, and then determine the typical weather states by the fuzzy clustering method. Calculate the weather similarity between any two days using the cosine - angle method:

[0096] (8)

[0097] Where r cd represents the similarity degree between the c-th day and the d-th day, , and represents the normalized values of the l-th meteorological factor on the c-th day and the d-th day. Furthermore, the fuzzy similarity matrix R = [r cd can be constructed. By the self-multiplication process of the fuzzy similarity matrix R = [r cd , the data can be clustered according to the similarity degree between any two days, and typical weather states can be obtained. The self-multiplication process of the fuzzy similarity matrix obtains the fuzzy equivalence matrix t(R), at:

[0098] (9)

[0099] (10)

[0100] Where t(R) represents the fuzzy equivalence matrix. Mapping each day in the dataset Z to the clustered typical state, the historical weather state sequence can be obtained, and K is the number of states obtained by clustering.

[0101] Next, based on the weather state data, a -order state transition matrix P is established to record the probability of each state reaching another state. Specifically, the frequency of transitioning from state m to any other possible state k is counted, and then each possible state frequency is divided by the total state frequency to obtain the transition probability corresponding to state m.

[0102] (11)

[0103] By analyzing all states in the original sequence, the state transition matrix P will finally summarize the probabilities of all states transitioning to another state.

[0104] The cumulative probability transition matrix P mentioned in this embodiment cum is calculated as follows:

[0105] (12)

[0106] The state transition matrix P is a -order matrix, and the cumulative probability transition matrix P cum is a -order matrix. The elements in the first column are all 0. Starting from the second column, the value of each element p cum,mk is the sum of the elements before the m-th row and the k-th column in matrix P.

[0107] Next, using the obtained weather state transition matrix and cumulative probability transition matrix as the Markov architecture logic, combined with a random generation algorithm, a typical weather state transition sub-model as the upper-layer model is constructed. This typical weather state transition sub-model can be used to generate a typical weather state sequence based on a preset random generation algorithm, such as the Monte Carlo method. First, set the initial state m of the typical weather, and then continuously generate uniform random numbers and compare them with the elements of the cumulative transition matrix P cum in the m-th row. If , then k is determined as the next state. Repeat this process until the typical weather state sequence within N days is completely generated .

[0108] Step 104: Based on historical meteorological data, combined with the Markov multi-dimensional transition kernel, construct an intraday meteorological factor change sub-model. The Markov multi-dimensional transition kernel is obtained based on the multi-dimensional joint probability distribution. The intraday meteorological factor change sub-model is used to simulate the non-stationary change process of intraday meteorological factors corresponding to the typical weather state sequence output by the typical weather state transition sub-model;

[0109] It should be noted that the intraday 24-hour meteorological factor sequences included in the historical days with the same state in the historical weather state sequence are merged together, and then the meteorological factor sequence is divided into time periods. The meteorological factors within the time period are divided into num state intervals state according to their magnitudes, where num is the number of states to be discretized. Each interval contains a certain range of values of the meteorological factors, and the time points whose values fall within this interval are regarded as the same state, thereby converting the original data into discrete state points. The conversion process is as follows:

[0110] (13)

[0111] In the formula , represents the maximum and minimum values of the l-th meteorological factor within the i-th time period.

[0112] Then, the upper-layer model is executed in a linked manner to establish the state transition matrix and the cumulative probability transition matrix within each time period . For a historical sequence with K typical weather states and L meteorological factors, a total of state transition matrices and cumulative probability transition matrices need to be constructed.

[0113] Subsequently, a Markov transition kernel is constructed from the joint distribution of multi-dimensional meteorological factors established by the R-vine copula function, and the cumulative joint distribution is obtained by integrating Equation (3):

[0114] (14)

[0115] Next, using this Markov multi-dimensional transition kernel, a sub-model of the intra-day meteorological factor changes as the lower-layer model is constructed. This sub-model of the intra-day meteorological factor changes can generate the intra-day meteorological factor sequence corresponding to the typical weather state k by the Markov model according to the typical weather state sequence generated by the upper-layer model.

[0116] Step 105: Integrate the typical weather state transition sub-model and the sub-model of the intra-day meteorological factor changes to obtain a power system high-temperature scenario simulation model based on a two-layer Markov architecture.

[0117] Finally, based on the integration of the typical weather state transition sub-model and the sub-model of the intra-day meteorological factor changes constructed in the previous steps, based on the typical weather state sequence output by the typical weather state transition sub-model and the intra-day meteorological factor sequence output by the sub-model of the intra-day meteorological factor changes, in the order of the typical weather state The intra-day meteorological factor sequences of each day are concatenated to obtain a complete meteorological factor sequence.

[0118] The specific process example is as follows:

[0119] 1) Establish the joint distribution of the first moment of the historical data of multi-dimensional meteorological factors in state k every day and sample to generate the values at the first moment. Compare these values with the state intervals in the first time period in state k to obtain the initial state of the meteorological factors, thereby determining the row number where the state is located in the cumulative transition matrix. Then retrieve the number of days N of state k in the typical weather state sequence generated by the upper-layer model in k .

[0120] 2) Randomly generate a uniform random number between 0 and 1 , compare it with the first column data of the cumulative joint distribution, and find the row number r where the number closest to is located. Then all the elements in row r are used as the Markov transition kernel MTK:

[0121] (15)

[0122] (16)

[0123] Compare each element in row r with the cumulative transition matrix determined by the initial state in the first time period Compare the elements in the corresponding rows to obtain the next state. Then repeat this step until N is fully generated k All time periods within a day of the meteorological factor sequence, the non-stationary change process of the intraday meteorological factors in state k is established. 3) After establishing the intraday meteorological factor sequences under K typical states, according to the positions of each day in the typical weather state sequence concatenate the intraday meteorological factor sequences to obtain a complete sequence.

[0124] In summary, the solution provided in this application uses the vine copula function to describe the correlation between multi-dimensional meteorological factors, generates a multi-dimensional joint probability distribution, and converts it into a Markov transition kernel after integration; secondly, considering the non-stationary change process of meteorological factors, a two-layer Markov model is constructed, and a meteorological factor time series is generated based on the Markov transition kernel. The generated meteorological factor time series can better approximate the actual data, making the established high-temperature scenario more accurate.

[0125] To further demonstrate the technical effects of this solution, this application also provides an experimental example implemented based on this method, specifically as follows:

[0126] In this embodiment, the vine copula function is used to characterize the correlation between meteorological factors under high-temperature weather, and a time series of meteorological factors is established based on a two-layer Markov model, which can more accurately generate wind and solar power generation sequences. Using the historical high-temperature meteorological data of a certain area from 2019 to August 2023, a meteorological factor sequence under high-temperature weather is generated. Among them, the meteorological factor time series mentioned in this example includes a solar radiation sequence, a temperature sequence, and a wind speed sequence. Compare the results with the data in August 2024 and the traditional Markov model, and the comparison results are respectively as Figures 5 to 7 shown. The lower layer model of this model divides each day into 4 time periods, and the discrete state within each time period is 10. The traditional Markov model is discretized into 10 typical states. It can be seen from the figure that the proposed method can well capture the daily cycle characteristics of meteorological factors, that is, the non-stationary change characteristics of meteorological factors. Especially for the irradiation sequence, the traditional Markov model completely fails to simulate the cycle characteristics of irradiation.

[0127] As shown in Table 1, Table 1 is a comparison of the probability distribution parameters of meteorological factors. The probability distribution parameters of the method proposed in this application are closer to the actual data in 2024 and can better capture the distribution characteristics of historical data.

[0128]

[0129] Such as Figures 8 to 10 , Figures 8 to 10They are comparative diagrams of the 24-hour autocorrelation characteristics of three meteorological factors obtained based on the solution of this application and the traditional Markov model respectively. From the autocorrelation characteristics, it can be seen that the method proposed in this application is closer to the autocorrelation curve of the historical data in 2024 and can better retain the periodic characteristics of meteorological factors.

[0130] It can be known from the experimental results that the inventive method uses the vine copula function to describe the correlation between multi-dimensional meteorological factors, generates a multi-dimensional joint probability distribution, and converts it into a Markov transition kernel after integration; secondly, considering the non-stationary change process of meteorological factors, a two-layer Markov model is constructed, and a meteorological factor time series is generated based on the Markov transition kernel. The generated meteorological factor time series can better approximate the actual data, making the established high-temperature scenario more accurate.

[0131] The above is a detailed description of an embodiment of a modeling method for a high-temperature scenario simulation model of a power system provided by this application. Next is a detailed description of an embodiment of a modeling device for a high-temperature scenario simulation model of a power system provided by this application.

[0132] Please refer to Figure 11 , a modeling device for a high-temperature scenario simulation model of a power system provided by an embodiment of this application, includes:

[0133] A meteorological factor distribution calculation unit 201, configured to calculate the marginal distribution of meteorological factors by means of kernel density estimation based on a plurality of meteorological factors corresponding to a high-temperature scenario;

[0134] A multi-dimensional joint probability distribution determination unit 202, configured to construct an R-vine copula maximum spanning tree based on the marginal distribution by improving the Prim's algorithm, so as to determine the multi-dimensional joint probability distribution of meteorological factors based on the R-vine copula maximum spanning tree;

[0135] An upper-layer sub-model construction unit 203, configured to determine the typical weather states corresponding to the historical meteorological data by clustering the historical meteorological data based on the historical meteorological data, and then construct a typical weather state transition sub-model based on the historical weather state sequence in combination with the Markov architecture logic and the random generation algorithm, where the historical weather state sequence is a typical weather state sequence generated in the time order of the historical meteorological data, and the typical weather state transition sub-model is used to simulate the process of typical weather state transition;

[0136] The lower-layer sub-model construction unit 204 is configured to construct an intraday meteorological factor change sub-model based on historical meteorological data in combination with a Markov multi-dimensional transition kernel, where the Markov multi-dimensional transition kernel is obtained based on a multi-dimensional joint probability distribution, and the intraday meteorological factor change sub-model is used to simulate the non-stationary change process of intraday meteorological factors corresponding to a typical weather state based on the typical weather state sequence output by the typical weather state transition sub-model;

[0137] The high-temperature scenario simulation model construction unit 205 is configured to integrate the typical weather state transition sub-model and the intraday meteorological factor change sub-model to obtain a power system high-temperature scenario simulation model based on a two-layer Markov architecture.

[0138] Further, the multi-dimensional joint probability distribution determination unit 202 is specifically configured to:

[0139] Based on the marginal distributions of each meteorological factor, in combination with the Kendall correlation coefficient calculation method, calculate the correlation coefficients between each meteorological factor, and determine the edge weights between each meteorological factor according to the correlation coefficients;

[0140] Taking each meteorological factor as a node, in combination with the edge weights between each meteorological factor, construct an R-vine copula maximum spanning tree through an improved Prim's algorithm.

[0141] Further, the upper-layer sub-model construction unit 203 is specifically configured to:

[0142] Based on the preset historical meteorological data, construct a historical meteorological data set matrix according to the data date and the number of meteorological factors;

[0143] According to the historical meteorological data set matrix, calculate the weather similarity between meteorological data on different dates according to a preset similarity calculation formula, and then cluster the historical meteorological data according to the weather similarity to determine the typical weather state corresponding to each date in the historical meteorological data, and generate a historical weather state sequence according to the time sequence of the historical meteorological data;

[0144] Based on the historical weather state sequence, in combination with the Markov architecture logic and the random generation algorithm, construct a typical weather state transition matrix and a cumulative probability transition matrix in sequence;

[0145] According to the cumulative probability transition matrix and a preset random generation algorithm, construct a typical weather state transition sub-model;

[0146] The lower-layer sub-model construction unit 204 is specifically configured to:

[0147] Based on the historical meteorological data, merge the intraday meteorological factor sequences in the historical meteorological data with the same weather state;

[0148] Divide the meteorological factor sequence into multiple state intervals according to the preset intra-day time period division information;

[0149] Through the typical weather state transition sub-model, determine the weather state transition matrix and the cumulative probability transition matrix of each state interval respectively, and then combine the Markov multi-dimensional transition kernel to construct an intra-day meteorological factor change sub-model.

[0150] In addition, as Figure 12 shown, a modeling terminal for a power system high-temperature scenario simulation model provided by an embodiment of the present application, the implementation types of the terminal include but are not limited to: personal computers, servers, and embedded intelligent devices. The main components of the terminal include: a memory 33 and a processor 31, and the memory 33 and the processor 31 can be connected through a communication bus 34;

[0151] The memory 33 is used to store program codes, and the program codes are used to implement the power system high-temperature scenario simulation model modeling method provided by the above embodiment;

[0152] The processor 31 is used to read and execute the program codes.

[0153] A computer-readable storage medium provided by the present application stores program codes in the computer-readable storage medium, and the program codes are used to be read and executed by a processor to implement the power system high-temperature scenario simulation model modeling method provided by the above embodiment.

[0154] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the above-described terminal, device, and unit can refer to the corresponding processes in the foregoing method embodiments, and will not be described herein again.

[0155] In several embodiments provided by the present application, it should be understood that the disclosed terminal, device, and method can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point, the displayed or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, and the indirect coupling or communication connection of the device or unit may be in an electrical, mechanical, or other form.

[0156] In the description of the present application and the above-mentioned drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of the present application described herein, for example, can be implemented in an order other than those illustrated or described herein. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device comprising a series of steps or units does not necessarily limit to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0157] It should be understood that in the present application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that there can be three relationships. For example, "A and / or B" can mean: only A exists, only B exists, and both A and B exist simultaneously. Among them, A and B can be singular or plural. The character " / " generally indicates that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or similar expressions refer to any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0158] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0159] In addition, in each embodiment of the present invention, each functional unit can be integrated in a processing unit, or each unit can exist physically alone, or two or more units can be integrated in one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0160] When the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs.

[0161] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the various embodiments of the present application.

Claims

1. A method for modeling a high-temperature scenario simulation model of a power system, characterized in that, Including: Based on several meteorological factors corresponding to high-temperature scenarios, the marginal distribution of the meteorological factors is calculated by means of kernel density estimation. Based on the marginal distribution, an R-vine copula maximum spanning tree is constructed by improving Prim's algorithm, so as to determine the multi-dimensional joint probability distribution of the meteorological factors based on the R-vine copula maximum spanning tree. Based on historical meteorological data, by clustering the historical meteorological data, the typical weather states corresponding to the historical meteorological data are determined, and then based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, a typical weather state transition sub-model is constructed, wherein the historical weather state sequence is a typical weather state sequence generated in the time order of the historical meteorological data, and the typical weather state transition sub-model is used to simulate the process of typical weather state transition. Based on the historical meteorological data, combined with the Markov multi-dimensional transition kernel, an intra-day meteorological factor change sub-model is constructed, wherein the Markov multi-dimensional transition kernel is obtained based on the multi-dimensional joint probability distribution, and the intra-day meteorological factor change sub-model is used to simulate the non-stationary change process of the intra-day meteorological factors corresponding to the typical weather state based on the typical weather state sequence output by the typical weather state transition sub-model. The typical weather state transition sub-model and the intra-day meteorological factor change sub-model are integrated to obtain a power system high-temperature scenario simulation model based on a two-layer Markov architecture.

2. The modeling method of a high-temperature scenario simulation model for a power system according to claim 1, characterized in that Based on the marginal distribution, constructing an R-vine copula maximum spanning tree by improving Prim's algorithm includes: Based on the marginal distribution of each meteorological factor, combined with the calculation method of Kendall correlation coefficient, the correlation coefficient between each meteorological factor is calculated, and the edge weights between each meteorological factor are determined according to the correlation coefficient. Taking each meteorological factor as a node, combined with the edge weights between each meteorological factor, an R-vine copula maximum spanning tree is constructed by improving Prim's algorithm.

3. A modeling method for a high-temperature scenario simulation model of a power system according to claim 2, characterized in that Taking each meteorological factor as a node, combined with the edge weights between each meteorological factor, constructing an R-vine copula maximum spanning tree by improving Prim's algorithm includes: Taking each meteorological factor as a node, selecting one from all the nodes as the starting node, associating the nodes according to the connection relationship between the nodes and the edge weights according to Prim's algorithm to determine the maximum spanning tree, and then using the Akaike information criterion to select the optimal binary copula function to connect the nodes of the first-layer tree to construct edges, thereby determining the first-layer maximum spanning tree. Based on the maximum spanning tree of the previous layer in a loop, calculate the correlation coefficient between each pair of conditional variables, determine the connectivity relationship of the conditional variables, and then according to Prim's algorithm and the Akaike information criterion, determine the maximum spanning tree of the current layer until the maximum spanning trees of all layers and the optimal binary copulas of the edges in each layer of the tree are determined to obtain the R-vine copula maximum spanning tree.

4. A modeling method for a high-temperature scenario simulation model of a power system according to claim 1, characterized in that, Based on historical meteorological data, by clustering the historical meteorological data, the typical weather states corresponding to the historical meteorological data are determined, and then based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, the construction of the typical weather state transition sub-model includes: Based on the preset historical meteorological data, a historical meteorological data set matrix is constructed according to the data date and the number of meteorological factors; According to the historical meteorological data set matrix, according to the preset similarity calculation formula, the weather similarity between meteorological data on different dates is calculated, and then the historical meteorological data is clustered according to the weather similarity, the typical weather state corresponding to each date in the historical meteorological data is determined, and a historical weather state sequence is generated according to the time sequence of the historical meteorological data; Based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, a typical weather state transition matrix and a cumulative probability transition matrix are constructed in sequence; According to the cumulative probability transition matrix and the preset random generation algorithm, a typical weather state transition sub-model is constructed.

5. A modeling method for a high-temperature scenario simulation model of a power system according to claim 1, characterized in that Based on the historical meteorological data, combined with the Markov multi-dimensional transition kernel, the construction of the intraday meteorological factor change sub-model includes: Based on the historical meteorological data, the intraday meteorological factor sequences in the historical meteorological data with the same weather state are merged; According to the preset intraday time period division information, the meteorological factor sequence is divided into multiple state intervals; Through the typical weather state transition sub-model, the weather state transition matrix and the cumulative probability transition matrix of each state interval are determined respectively, and then combined with the Markov multi-dimensional transition kernel, an intraday meteorological factor change sub-model is constructed.

6. A modeling device for a high-temperature scenario simulation model of a power system, characterized in that, Including: A meteorological factor distribution calculation unit, which is used to calculate the marginal distribution of the meteorological factors by means of kernel density estimation based on several meteorological factors corresponding to the high-temperature scenario; A multi-dimensional joint probability distribution determination unit, which is used to construct a maximum R-vine copula tree based on the marginal distribution through an improved Prim's algorithm, so as to determine the multi-dimensional joint probability distribution of the meteorological factors based on the maximum R-vine copula tree; An upper-layer sub-model construction unit, which is used to determine the typical weather states corresponding to the historical meteorological data by clustering the historical meteorological data based on the historical meteorological data, and then construct a typical weather state transition sub-model based on the historical weather state sequence, combined with the Markov architecture logic and the random generation algorithm, wherein the historical weather state sequence is a typical weather state sequence generated according to the time sequence of the historical meteorological data, and the typical weather state transition sub-model is used to simulate the typical weather state transition process; A lower-layer sub-model construction unit, which is used to construct an intraday meteorological factor change sub-model based on the historical meteorological data, combined with the Markov multi-dimensional transition kernel, wherein the Markov multi-dimensional transition kernel is obtained based on the multi-dimensional joint probability distribution, and the intraday meteorological factor change sub-model is used to simulate the non-stationary change process of the intraday meteorological factors corresponding to the typical weather state based on the typical weather state sequence output by the typical weather state transition sub-model; A high-temperature scenario simulation model construction unit is used to integrate the typical weather state transition sub-model and the intraday meteorological factor change sub-model to obtain a high-temperature scenario simulation model of the power system based on a two-layer Markov architecture.

7. The modeling device for a high-temperature scenario simulation model of a power system according to claim 6, characterized in that, The multi-dimensional joint probability distribution determination unit is specifically used for: Based on the marginal distributions of various meteorological factors, combining the Kendall correlation coefficient calculation method, calculating the correlation coefficients between various meteorological factors, and determining the edge weights between various meteorological factors according to the correlation coefficients; Taking each meteorological factor as a node, combining the edge weights between various meteorological factors, and constructing a maximum spanning tree of R-vine copula through an improved Prim's algorithm.

8. The modeling device for a high-temperature scenario simulation model of a power system according to claim 6, characterized in that The upper-layer sub-model construction unit is specifically used for: Based on the preset historical meteorological data, constructing a historical meteorological data set matrix according to the data date and the number of meteorological factors; According to the historical meteorological data set matrix, calculating the weather similarity between meteorological data on different dates according to the preset similarity calculation formula, and then clustering the historical meteorological data according to the weather similarity to determine the typical weather state corresponding to each date in the historical meteorological data, and generating a historical weather state sequence according to the time sequence of the historical meteorological data; Based on the historical weather state sequence, combining the Markov architecture logic and the random generation algorithm, successively constructing a typical weather state transition matrix and a cumulative probability transition matrix; According to the cumulative probability transition matrix and the preset random generation algorithm, constructing a typical weather state transition sub-model; The lower-layer sub-model construction unit is specifically used for: Based on the historical meteorological data, merging the intraday meteorological factor sequences in the historical meteorological data with the same weather state; According to the preset intraday time period division information, dividing the meteorological factor sequence into multiple state intervals; Through the typical weather state transition sub-model, respectively determining the weather state transition matrix and the cumulative probability transition matrix of each state interval, and then combining the Markov multi-dimensional transition kernel to construct an intraday meteorological factor change sub-model.

9. A modeling terminal for a high-temperature scenario simulation model of a power system, characterized in that, It includes: A memory and a processor; The memory is used to store program codes, and the program codes are used to implement the power system high-temperature scenario simulation model modeling method according to any one of claims 1 to 5; The processor is used to read and execute the program codes.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores program codes, and the program codes are used to be read and executed by the processor to implement the power system high-temperature scenario simulation model modeling method provided in any one of claims 1 to 5.

Images