Joint scene generation method, system and device considering wind-light correlation and medium

Abstract

The invention discloses a combined scene generation method, system, equipment and medium considering wind and light correlation, and relates to the field of new energy power generation dispatching, and the technical scheme is characterized in that the method comprises the steps: obtaining wind and light output data, carrying out marginal probability modeling through multi-distribution and kernel density estimation, and screening an optimal probability model of each variable; based on the optimal marginal distribution, constructing a wind and light output joint probability model by adopting a Copula function, evaluating and screening an optimal Copula model by utilizing a maximum likelihood estimation and information criterion (AIC / BIC), and generating a joint output scene considering correlation; an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed, and a typical daily joint output scene set is generated; a density peak fast search clustering (CFSFDP) algorithm is combined with a DTW distance to generate a clustering decision diagram to identify an extreme scene. According to the invention, the problem of insufficient scene generation typicality caused by the fact that dynamic time similarity and initial center sensitivity are not considered when the wind-solar combined scene is generated in the prior art is solved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of new energy power generation scheduling, in particular, it relates to a joint scene generation method, system, device and medium considering wind-solar correlation. BACKGROUND

[0002] With large-scale renewable energy such as hydropower, wind power and photovoltaic power being connected to the grid as a normal state, the operation of the power system is facing increasingly significant uncertainty challenges. Resources such as wind power and photovoltaic power are affected by the natural environment and show dramatic and uncontrollable volatility.

[0003] At present, there are many studies on the generation of wind-solar scenes, and different clustering methods are used to generate typical daily curves of wind-solar, but none of them considers the correlation, spatial density characteristics and time sequence form diversity of wind-solar output, resulting in insufficient typicality of the generated scenes, which is difficult to fully reflect the complex spatio-temporal correlation characteristics of wind-solar joint output in actual operation. SUMMARY

[0004] The purpose of the present application is to provide a joint scene generation method, system, device and medium considering wind-solar correlation. The present application solves the problem of insufficient typicality of the generated scenes caused by not considering the correlation, spatial density characteristics and time sequence form diversity of wind-solar output in the related art when generating wind-solar scenes.

[0005] In a first aspect, the present application provides a joint scene generation method considering wind-solar correlation, the method comprising:

[0006] S101: Obtain wind-solar output data, model the probability distribution by using multi-candidate distribution and kernel density estimation, and select the optimal probability model of each variable through goodness-of-fit test and precision evaluation index;

[0007] S102: Based on the optimal edge distribution fitting result, construct a wind-solar output joint probability model by using multiple Copula functions, and select the optimal dependent structure model through comprehensive evaluation of maximum likelihood estimate value, AIC and BIC, to generate a joint output scene considering wind-solar correlation;

[0008] S103: Based on the wind-solar output joint scene, an improved density-variance dynamic time warping clustering (DV-DTWk-Medoids) method is proposed, the density peak value and variance fusion strategy are used to optimize the selection of initial center, dynamic time warping (DTW) is introduced as a similarity measure, and a typical daily joint output scene set is generated through the Medoids center update mechanism;

[0009] S104: Based on the wind-solar output joint scene, the density peak fast search clustering (CFSFDP) algorithm is combined with the DTW distance to identify extreme scenes and generate a clustering decision diagram.

[0010] In an implementation scheme, the wind and light output data is obtained, a multi-candidate distribution and kernel density estimation are used to model the probability distribution, a goodness-of-fit test and an accuracy evaluation index are used to screen the optimal probability model of each variable, including:

[0011] Obtaining wind and photovoltaic output data of a target area;

[0012] According to the wind and light output data, a plurality of distribution methods are selected, including Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution and kernel density estimation method, and the probability distribution of each variable is modeled respectively;

[0013] Performing goodness-of-fit test and accuracy evaluation, and selecting the optimal probability model.

[0014] In an implementation scheme, based on the fitting result of the optimal marginal distribution, a plurality of Copula functions are used to construct a wind and light output joint probability model, a maximum likelihood estimate and AIC, BIC are used for comprehensive evaluation, and the optimal dependent structure model is screened to generate a joint output scenario considering the correlation of wind and light, including:

[0015] The plurality of copula models include Gaussian copula, t copula, Clayton copula, Gumbel copula, Frank copula, and Independence copula function;

[0016] Taking the maximum likelihood estimate and AIC, BIC as the preferred criteria, the correlation indicators of the above six Copula functions are counted;

[0017] Based on the indicators, the optimal Copula function is selected to establish a wind and light joint probability distribution, and a joint output scenario considering the correlation of wind and light is generated.

[0018] In an implementation scheme, based on the wind and light output joint scenario, an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed, a density peak and variance fusion strategy is used to optimize the selection of initial center, dynamic time warping (DTW) is introduced as a similarity measure, and a typical day joint output scenario set is generated through a Medoids center update mechanism, including:

[0019] The density-variance dynamic time warping clustering (DV-DTW k-Medoids) method includes: initial center scene selection, initial cluster construction, cluster center point update, sample reassignment, and cluster end judgment.

[0020] In an implementation scheme, the improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is characterized in that the density-variance method optimization of initial center scene selection comprises: calculating the density of each scene in the wind power and photovoltaic power scene set, screening a high-density scene set, calculating the variance and standard deviation of each scene in the scene set, screening a scene set with smaller variance, and constructing a clustering center candidate set and selecting an initial clustering center.

[0021] In an implementation scheme, the improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is characterized in that the dynamic time warping (DTW) as a similarity measure updates the clustering center point, which comprises: assigning the remaining scenes to the nearest center scene, obtaining k clusters, calculating the inter-cluster distance and clustering error sum of squares, obtaining the criterion function value, for each scene in the k clusters, calculating the sum of distances of the scene to the remaining scenes in the cluster to which the scene belongs, taking the scene with the minimum sum of distances of the remaining scenes in the cluster as the new clustering center of the cluster, and repeatedly reassigning samples until the clustering ends.

[0022] In an implementation scheme, the DV-DTW k-Medoids method is combined with the DTW distance to identify extreme scenes based on the wind and light output joint scene, and a clustering decision diagram is generated.

[0023] In the second aspect of the present application, a joint scene generation system considering wind-light correlation is provided, and the system comprises:

[0024] A distribution fitting module is configured to obtain wind and light output data, perform probability distribution modeling by using a multi-candidate distribution and kernel density estimation, and screen an optimal probability model of each variable through goodness-of-fit test and precision evaluation index.

[0025] A joint modeling module is configured to construct a wind and light output joint probability model by using a plurality of Copula functions based on the optimal edge distribution fitting result, screen an optimal dependent structure model through maximum likelihood estimate value and AIC and BIC comprehensive evaluation, and generate a joint output scene considering wind-light correlation.

[0026] A time series clustering module is configured to propose an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method based on the wind and light output joint scene, optimize initial center selection by using a density peak and variance fusion strategy, introduce dynamic time warping (DTW) as a similarity measure, and generate a typical day joint output scene set through a Medoids center update mechanism.

[0027] An extreme identification module is used to identify extreme scenes by combining a density peak fast search clustering (CFSFDP) algorithm and a DTW distance based on a wind-solar power output joint scene, and to generate a clustering decision graph.

[0028] In a third aspect, the present application provides an electronic device, comprising a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein the computer program, when executed by the processor, implements the steps of the joint scene generation method considering wind-solar correlation provided in the first aspect of the present application.

[0029] In a fourth aspect, the present application provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program, when executed by a processor, implements the steps of the joint scene generation method considering wind-solar correlation provided in the first aspect of the present application.

[0030] Compared with the prior art, the present application has the following beneficial effects:

[0031] The present application considers the correlation of wind-solar power output, and also considers the influence of spatial density characteristics and time series shape diversity on clustering, and on this basis, proposes an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method to generate typical daily wind-solar joint scenes, and the present application can more accurately reflect the complex spatio-temporal correlation characteristics of wind-solar joint output in actual operation, especially in the case of wind power and photovoltaic power showing violent and uncontrollable fluctuations. By comprehensively considering the influence of different factors on scene generation, more accurate and typical scene generation strategies can be provided for new energy power generation dispatching, the allocation of power resources is optimized, and the reliability of the power system is improved. BRIEF DESCRIPTION OF DRAWINGS

[0032] The accompanying drawings, which are included to provide a further understanding of the embodiments of the present application and constitute a part of this application, do not constitute limitations to the embodiments of the present application. In the drawings:

[0033] Figure 1 A flowchart of a joint scene generation method considering wind-solar correlation provided by the embodiments of the present application is shown in the figure;

[0034] Figure 2 A flowchart of the improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method provided by the embodiments of the present application is shown in the figure;

[0035] Figure 3 A flowchart of the extreme scene identification method combining the density peak fast search clustering (CFSFDP) algorithm and the DTW distance provided by the embodiments of the present application is shown in the figure;

[0036] Figure 4 The wind and light output scene graph considering correlation generation provided by the embodiment of the present application (wherein a is the wind power output scene considering correlation generation, and b is the photovoltaic output scene considering correlation generation) ;

[0037] Figure 5 The wind and light output clustering scene graph provided by the embodiment of the present application (wherein a is the wind power output 4-class typical scene, and b is the photovoltaic output 4-class typical scene) ;

[0038] Figure 6 The wind and light output scene graph and its probability after reduction of the DV-DTW K-medoids algorithm provided by the embodiment of the present application (wherein a is the wind power output 4-class typical scene, and b is the photovoltaic output 4-class typical scene) ;

[0039] Figure 7 The wind and light output joint scene graph after reduction of the DV-DTW K-medoids algorithm provided by the embodiment of the present application;

[0040] Figure 8 The CFSFDP clustering decision schematic diagram provided by the embodiment of the present application;

[0041] Figure 9 The principle block diagram of the joint scene generation system considering wind and light correlation provided by the embodiment of the present application. DETAILED DESCRIPTION

[0042] In order to make the objectives, technical solutions and advantages of the present application clearer, further detailed description will be made to the present application in combination with embodiments and drawings, and the schematic embodiments of the present application and the description thereof are only used to explain the present application, and not as a limitation to the present application.

[0043] It should be noted that the term "include" or "may include" used in various embodiments of the present application indicates the existence of the claimed function, operation or element, and does not limit the addition of one or more functions, operations or elements. In addition, as used in various embodiments of the present application, the terms "include", "have" and their homonyms are only intended to represent a specific feature, number, step, operation, element, component or combination of the foregoing, and should not be understood as first excluding the existence or addition of one or more other features, numbers, steps, operations, elements, components or combinations of the foregoing.

[0044] Please refer to Figure 1 , Figure 1 The flow schematic diagram of the joint scene generation method considering wind and light correlation provided by the embodiment of the present application is shown as Figure 1 shown, the method comprises:

[0045] S101: Obtain wind and light output data, use multi-candidate distribution and kernel density estimation to model probability distribution, and select the optimal probability model of each variable through goodness-of-fit test and precision evaluation index;

[0046] S102: Based on the optimal edge distribution fitting result, use multiple Copula functions to construct wind and light output joint probability model, select the optimal dependent structure model through maximum likelihood estimate value and AIC, BIC comprehensive evaluation, and generate joint output scene considering wind and light correlation;

[0047] S103: Based on the wind and light output joint scene, an improved density-variance dynamic time warping clustering (DV-DTWk-Medoids) method is proposed, which uses density peak value and variance fusion strategy to optimize initial center selection, introduces dynamic time warping (DTW) as similarity measure, and generates typical day joint output scene set through Medoids center update mechanism;

[0048] S104: Based on the wind and light output joint scene, the density peak fast search clustering (CFSFDP) algorithm is combined with the DTW distance to identify extreme scenes and generate clustering decision diagram.

[0049] It should be further pointed out that step S101 specifically includes:

[0050] In this embodiment, the proposed Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution, and kernel density estimation are common edge distribution fitting methods in the field of power systems, so this embodiment will not be described in detail.

[0051] Only the goodness-of-fit and fitting precision test of these candidate edge distributions are described:

[0052] First, obtain wind and light output data, and select Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution, and kernel density estimation method for single-variable probability density fitting.

[0053] Second, use Pearsonχ 2 and Kolmogorov-Smirnov to perform goodness-of-fit test to determine whether the distribution matches the data characteristics and measure the goodness-of-fit of each fitting method to the original data, thereby selecting the best fitting function.

[0054] Specifically, the calculation formula of Pearsonχ 2 statistic is: In the formula, v i is the number of samples in the ith interval; p iThe theoretical value falls in the ith interval.

[0055] When n tends to positive infinity, χ 2 The distribution converges to If a confidence interval α is given, the α distribution point of χ is: In the formula, P(·) represents the probability of an event occurring.

[0056] If the test statistic χ 2 meets , it means that the probability distribution F0(x) at the confidence level α meets the requirements.

[0057] The empirical cumulative distribution function F n (x) is calculated as follows:

[0058] The maximum vertical gap between the theoretical cumulative distribution F0(x) and the empirical cumulative distribution F n (x) is defined as the test statistic D n , and In the formula, i is the ith sampling interval.

[0059] The parameters of the theoretical distribution model can be obtained from actual historical data, and in this case, when a theoretical distribution is rejected in the test, the error generated by the K-S test is relatively small.

[0060] Secondly, the average percentage error and the root mean square error are used to test the fitting accuracy, to quantify the error between the fitted curve and the data, and to measure the fitting accuracy of each fitting method.

[0061] Specifically, the average percentage error is calculated as follows:

[0062] The root mean square error is calculated as follows:

[0063] Where k is the number of intervals; P oi , P gi are the probability of the wind power and photovoltaic standardized output power orthogonal series density distribution and histogram in the ith interval, respectively.

[0064] Further, step S102 specifically includes:

[0065] In this embodiment, first, a plurality of copula models are constructed, including Gaussian copula, t copula, Clayton copula, Gumbel copula, Frank copula, and Independence copula function.

[0066] Secondly, the correlation indicators of the above six Copula functions are counted by using the maximum likelihood estimate value and AIC and BIC as preferred criteria;

[0067] Then, the optimal Copula function is selected based on the indicators, the wind-solar joint probability distribution is established, and the joint output scene considering the correlation between wind and solar is generated.

[0068] It should be further explained that step S103 specifically includes:

[0069] In this embodiment, the selection of the initial clustering center is optimized by using the density-variance method, the clustering center point is updated according to the dynamic time distance (DTW), and the typical day joint scene is generated. Based on step S103, a possible embodiment will be given below to non-restrictively describe the specific implementation scheme, as shown in Figure 2 .

[0070] Step S1031: Calculate the density of each scene in the wind power and photovoltaic scene set.

[0071] The density calculation formula of the scene s is: p(x) = {p e S | dist(x, p) < r} (7), dist(s, p) = ||s-p|| (8), where S is the time series wind power scene set, n is the number of scenes, p is the scene in the scene set S, dist(s, p) is the DTW distance between the scenes s and p, and r is the radius, which is calculated by the average distance between each other. The above formula represents the number of scenes contained in the spherical region with s as the center and r as the radius, so as to display the local information of the spatial distribution of the scene s.

[0072] Step S1032: Select the high-density scene set.

[0073] The average density calculation formula of the time series wind power scene is: The scene s with a density higher than the average density of the scene set is stored in the high-density scene set A: i

[0074] Step S1033: Calculate the variance and standard deviation of each scene in the scene set.

[0075] The standard deviation and variance show the overall information of the spatial distribution of the scene, and the variance calculation formula of each scene in the scene set is: The standard deviation calculation formula of each scene is:

[0076] Step S1034: Select the scene set with small variance.

[0077] The scene variance average calculation formula is:​ Store the scenes s with variance less than the average variance of the scene set in the small variance scene set B:

[0078] Step S1035: Construct the cluster center candidate set and select the initial cluster center.

[0079] Select the intersection of the high density scene set and the small variance scene set as the center candidate set C = A∩B (16), and select the highest density scene in C as the first cluster center C1; delete the scenes in the neighborhood of C1 from C, and select the field with the standard deviation of sample s i as the radius; if the sample distribution is dense, the k cluster centers cannot be selected after excluding the scenes in the neighborhood, a coefficient u can be multiplied to the neighborhood radius, which is larger when the sample distribution is dense and smaller when the sample distribution is sparse, so that the clustering speed and accuracy can be balanced, but the value needs to be obtained through experiments; select the scene farthest from C1 in the candidate set as the second center C2, and delete the scenes in the neighborhood of C2 from C; and so on until C k .

[0080] Step S1036: Construct the initial class cluster.

[0081] Calculate the distance of the remaining scenes in the scene set S to the k cluster center scenes, and assign them to the nearest center scene to obtain the k class clusters. The distance calculation formula between classes is: The cluster error sum of squares calculation formula is: The criterion function value is obtained: fit = E / F (19).

[0082] Step S1037: Update the class cluster center.

[0083] For each scene in the k class clusters, calculate the sum of the DTW distances of the scene to the other scenes in the class cluster to which the scene belongs, and select the scene with the smallest sum of distances to the remaining scenes in the cluster as the new cluster center of the cluster.

[0084] Step S1038: Reassign the samples.

[0085] Reassign the remaining scenes in the scene set except the new center to the nearest center to obtain new class clusters, and calculate the fit value of the new clustering result to reassign the samples.

[0086] Step S1039: Determine whether the clustering is completed.

[0087] If the fit value is greater than or equal to the fit value of the last clustering result, stop and output the clustering result, otherwise continue to update the cluster center.

[0088] It should be further explained that step S104 specifically includes:

[0089] In this embodiment, CFSFDP is combined with DTW distance to perform extreme scene recognition, as shown in FIG. Figure 3 As shown:

[0090] Step S1041: local density calculation.

[0091] Step S1042: Calculate the minimum DTW distance with higher density.

[0092] Calculate each basic scene S i The minimum distance δ to all scenes with a higher density than it i For each basic scenario S i , find the scene set {S h |ρ h >ρ i}(20), h represents the index of the higher density scene. Calculate the scene S i DTW distance d to higher density scenes ih , and select the smallest one from these distances as δ i The value of , that is: For the scene with the highest density S i , calculate its DTW distance d to all other scenes ih . The maximum value among these distances is selected as δ i The value of δ i It reflects the degree of separation between the scene with the highest density and the entire dataset, namely:

[0093] Step S1043: Generate a clustering decision diagram.

[0094] In summary, the embodiment first acquires the output data of wind power and photovoltaic of the target area, selects multiple probability distribution models according to the wind and light output data, including Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution, kernel density estimation, fits these candidate probability distributions for each variable respectively, performs goodness-of-fit and fitting accuracy test, and selects the best probability model; secondly, a variety of copula functions are used, including Gaussian copula, t copula, Clayton copula, Gumbel copula, Frank copula and Independence copula, to construct a wind and light output joint probability model, to take the maximum likelihood estimate and AIC, BIC as the optimization criterion, to statistically analyze the correlation indicators of the above six Copula functions, to screen the optimal dependence structure model, and to generate a joint output scenario considering the correlation of wind and light; further, an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed, a density peak and variance fusion strategy is used to optimize the selection of initial centers, dynamic time warping (DTW) is introduced as a similarity measure, and a typical day joint output scenario set is generated through a Medoids center updating mechanism; finally, the density peak fast search clustering (CFSFDP) algorithm is combined with the DTW distance to identify extreme scenarios and generate a clustering decision diagram.

[0095] The wind and light output scenario generated considering the correlation is shown in Figure 4 The wind and light output clustering scenario is shown in Figure 5 The wind and light output scenario after reduction of the DV-DTW K-medoids algorithm is shown in Figure 6 The wind and light output joint scenario after reduction of the DV-DTW K-medoids algorithm is shown in Figure 7 The CFSFDP clustering decision diagram is shown in Figure 8

[0096] The comparison index results of the existing method are shown in Table 1:

[0097] Table 1

[0098]

[0099]

[0100] ​The example considers the correlation of wind and light output, also considers the spatial density characteristics and time series form diversity to generate wind and light combined typical day scene, the analysis chart can know that, 1) the application realizes the whole process optimization from edge distribution modeling to combined modeling to typical / extreme scene extraction of wind and light output, solves the problem of insufficient description of correlation and time sequence characteristics of traditional method;2) the improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method proposed in the application is improved compared with the traditional clustering method under various indicators, and the clustering effect can meet the accuracy and typicality requirements of the scene;3) the extreme scene recognition method combining the density peak fast search clustering (CFSFDP) algorithm and the DTW distance can clearly identify outliers, and the extreme scene recall rate is obviously improved.

[0101] Please refer to Figure 9 , Figure 9 The principle block diagram of a combined scene generation system considering wind and light correlation provided by the embodiment of the application is shown as Figure 9 The system comprises:

[0102] 901: fitting distribution module, used for obtaining wind and light output data, adopting multiple candidate distributions and kernel density estimation to model probability distribution, and selecting the optimal probability model of each variable through goodness-of-fit test and precision evaluation index;

[0103] 902: combined modeling module, used for constructing wind and light output joint probability model by using multiple Copula functions based on the optimal edge distribution fitting result, screening the optimal dependent structure model through maximum likelihood estimate value and AIC, BIC comprehensive evaluation, and generating wind and light correlation considered joint output scene;

[0104] 903: time series clustering module, used for proposing an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method based on wind and light output joint scene, adopting density peak and variance fusion strategy to optimize initial center selection, introducing dynamic time warping (DTW) as similarity measure, and generating typical day joint output scene set through Medoids center update mechanism;

[0105] 904: extreme recognition module, used for combining density peak fast search clustering (CFSFDP) algorithm and DTW distance to identify extreme scene based on wind and light output joint scene, and generating clustering decision graph.

[0106] In a joint scenario generation system that takes into account the correlation between wind and solar power, provided in an embodiment of the present invention, the correlation between wind and solar power output is taken into account, and the influence of spatial density characteristics and time series morphological diversity on clustering is also considered. On this basis, an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed to generate a typical day wind-solar joint scenario. The present invention can more accurately reflect the complex spatiotemporal correlation characteristics of wind and solar power output in actual operation, especially when wind power and photovoltaic power show drastic and uncontrollable volatility. By comprehensively considering the influence of different factors on scenario generation, a more accurate and typical scenario generation strategy can be provided for new energy power generation scheduling, the allocation of power resources can be optimized, and the reliability of the power system can be improved.

[0107] In some embodiments, the fitting distribution module 901 is specifically used to: obtain wind power and photovoltaic output data in the target area; select multiple distribution methods based on the wind and photovoltaic output data, including Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution and kernel density estimation method, and perform probability distribution modeling for each variable; perform goodness of fit test and accuracy evaluation, and select the optimal probability model.

[0108] In some embodiments, the joint modeling module 902 is specifically used to: based on the optimal marginal distribution fitting results, use multiple Copula functions, including Gaussian copula, t copula, Clayton copula, Gumbel copula, Frank copula, and Independence copula, to construct a joint probability model of wind and solar output; use the maximum likelihood estimate and AIC and BIC as the optimization criteria, count the relevant indicators of the above 6 Copula functions, screen the optimal dependency structure model, and generate a joint output scenario that takes into account the correlation between wind and solar.

[0109] In some embodiments, the time series clustering module 903 is specifically configured to: based on the wind-solar power output joint scene, propose an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method, including: initial center scene selection, initial cluster construction, cluster center point updating, sample reassignment, and cluster end judgment; wherein the initial center scene selection is optimized by the density-variance method, including: calculating the density of each scene in the wind power and photovoltaic power scene set, screening a high-density scene set, calculating the variance and standard deviation of each scene in the scene set, screening a small-variance scene set, and constructing a cluster center candidate set and selecting an initial cluster center; the cluster center point is updated by dynamic time warping (DTW) as a similarity measure, including: assigning the remaining scenes to the nearest center scene, obtaining k clusters, calculating the inter-cluster distance and cluster error sum of squares, obtaining a criterion function value, for each scene in the k clusters, calculating the sum of distances to the remaining scenes in the cluster to which it belongs, and taking the scene with the minimum sum of distances to the remaining scenes in the cluster as the new cluster center of the cluster, repeatedly reassigning samples until the clustering ends.

[0110] In some embodiments, the extreme scenario identification module 904 is specifically configured to: based on the wind-solar power output joint scene, identify extreme scenarios by combining the density peak fast search clustering (CFSFDP) algorithm with the DTW distance, and generate a clustering decision diagram.

[0111] Embodiments of the present application also provide an electronic device. The electronic device includes a processor, a memory, a communication interface, and at least one communication bus for connecting the processor, the memory, and the communication interface. The memory includes, but is not limited to, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (PROM), or a portable read-only memory (CD-ROM), and is used to store related instructions and data.

[0112] The communication interface is used to receive and send data. The processor can be one or more CPUs, and in the case of a single CPU, the CPU can be a single-core CPU or a multi-core CPU. The processor in the electronic device is used to read one or more programs stored in the memory and perform the following operations: obtaining the water penalty cost, the light penalty cost, the wind penalty cost, and the load loss penalty cost of the power system at any time; determining the total operation cost at any time according to the water penalty cost, the light penalty cost, the wind penalty cost, and the load loss penalty cost; obtaining the first operation constraint of the wind power, the light power, and the power system and the second operation constraint of different types of water power; wherein the types of water power include runoff water power, adjustable water power, and cascade water power; based on the first operation constraint and the second operation constraint, solving the optimization scheme of the water-wind-solar coordinated dispatching of the power system when the total operation cost reaches the minimum.

[0113] It should be noted that the specific implementation of each operation can be described above in the method embodiment, and the electronic device can be used to execute the power system water, wind and light coordinated scheduling optimization method of the above method embodiment of the present application, which will not be described in detail here.

[0114] In the embodiments of the present application, a computer readable storage medium is also provided, which is a memory device in a computer device, used to store programs and data. It can be understood that the computer readable storage medium herein can include an internal storage medium in the computer device, and of course can also include an extended storage medium supported by the computer device. The computer readable storage medium provides a storage space which stores an operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer readable storage medium herein can be a high-speed RAM memory, or a non-volatile memory such as at least one disk memory. The one or more instructions stored in the computer readable storage medium can be loaded and executed by the processor to implement the corresponding steps of the power system water, wind and light coordinated scheduling optimization method in the above embodiments. Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program codes.

[0115] The above specific embodiments further illustrate the purpose, technical solutions and beneficial effects of the present application. It should be understood that the above description is only a specific embodiment of the present application and is not used to limit the protection scope of the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A joint scene generation method taking into account the correlation between wind and light, characterized in that: Methods include: S101: Obtain wind and solar power output data, use multiple candidate distributions and kernel density estimation to perform probability distribution modeling, and select the optimal probability model for each variable through goodness of fit tests and accuracy evaluation indicators; S102: Based on the optimal marginal distribution fitting results, a variety of copula functions are used to construct a wind-solar output joint probability model. Through the maximum likelihood estimate and comprehensive evaluation of AIC and BIC, the optimal dependency structure model is selected to generate a joint output scenario that takes into account the correlation between wind and solar power. S103: Based on the wind and solar power combined scenario, an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed. This method uses a density peak and variance fusion strategy to optimize the initial center selection, introduces dynamic time warping (DTW) as a similarity metric, and generates a set of typical daily combined power scenarios through a Medoids center update mechanism. S104: Based on the wind and solar power combined scenario, the fast search clustering for density peaks (CFSFDP) algorithm is combined with the DTW distance to identify extreme scenarios and generate a cluster decision diagram.

2. The method for generating a joint scene taking into account the correlation between wind and light according to claim 1, characterized in that: The wind and solar power output data is obtained, and probability distribution modeling is performed using multiple candidate distributions and kernel density estimation. The optimal probability model of each variable is screened through goodness of fit test and accuracy evaluation index, including: Obtain wind power and photovoltaic output data in the target area; Based on wind and solar power output data, multiple distribution methods are selected, including Beta distribution, normal distribution, lognormal distribution, gamma distribution, Weibull distribution, and kernel density estimation, to perform probability distribution modeling for each variable. Perform goodness of fit test and accuracy assessment and select the optimal probability model.

3. The method for generating a joint scene taking into account the correlation between wind and light according to claim 1, characterized in that: Based on the optimal marginal distribution fitting results, a variety of Copula functions are used to construct a wind-solar output joint probability model. Through the maximum likelihood estimate and AIC and BIC comprehensive evaluation, the optimal dependency structure model is screened to generate a joint output scenario that considers the correlation between wind and solar power, including: The multiple copula models are constructed, including Gaussian copula, t copula, Clayton copula, Gumbel copula, Frank copula, and Independence copula functions; Using maximum likelihood estimation, AIC, and BIC as optimization criteria, the relevant indicators of the above six Copula functions were statistically analyzed; The optimal Copula function is selected based on the indicators, the wind-solar joint probability distribution is established, and a joint output scenario that considers the correlation between wind and solar is generated.

4. The method for generating a joint scene taking into account the correlation between wind and light according to claim 1, characterized in that: Based on the wind and solar power combined scenario, an improved density-variance dynamic time warping clustering (DV-DTW k-Medoids) method is proposed. The density peak and variance fusion strategy is used to optimize the initial center selection, dynamic time warping (DTW) is introduced as a similarity metric, and a typical daily combined power scenario set is generated through the Medoids center update mechanism, including: The density-variance dynamic time warping (DV-DTW k-Medoids) method includes: selecting an initial central scene, constructing an initial cluster, updating the cluster center point, redistributing samples, and determining whether clustering is completed.

5. The improved density-variance dynamic time warping (DV-DTW k-Medoids) method according to claim 4, characterized in that: The density-variance method for optimizing the selection of the initial center scene includes: calculating the density of each scene in the wind power and photovoltaic power scene sets, screening the high-density scene sets, calculating the variance and standard deviation of each scene in the scene set, screening the scene set with smaller variance, constructing a cluster center candidate set and selecting the initial cluster center.

6. The improved density-variance dynamic time warping (DV-DTW k-Medoids) method according to claim 4, characterized in that: The dynamic time warping (DTW) is used as a similarity measure to update the cluster center point, including: assigning the remaining scenes to the nearest central scene to obtain k clusters, calculating the inter-cluster distance and the sum of squares of clustering errors, and obtaining the criterion function value. For each scene in the k clusters, the sum of its distances to the remaining scenes in the cluster to which it belongs is calculated, and the scene with the smallest sum of distances to the remaining scenes in the cluster is used as the new cluster center of the cluster. This process is repeated continuously and samples are redistributed until the clustering is completed.

7. The method for generating a joint scene taking into account the correlation between wind and light according to claim 1, characterized in that: Based on the wind and solar power combined scenario, the fast search clustering for density peaks (CFSFDP) algorithm is combined with the DTW distance to identify extreme scenarios and generate a cluster decision diagram.

8. A joint scene generation system taking into account the correlation between wind and light, characterized in that: The system includes: The distribution fitting module is used to obtain wind and solar power output data, use multiple candidate distributions and kernel density estimation to perform probability distribution modeling, and screen the optimal probability model for each variable through goodness of fit tests and accuracy evaluation indicators; The joint modeling module is used to construct a wind-solar joint output probability model based on the optimal marginal distribution fitting results using multiple Copula functions. Through the maximum likelihood estimate and comprehensive evaluation of AIC and BIC, the optimal dependency structure model is selected to generate a joint output scenario that takes into account the correlation between wind and solar power. The time series clustering module is used to cluster wind and solar power combined scenarios. It proposes an improved density-variance dynamic time warping (DV-DTW k-Medoids) clustering method. It uses a density peak and variance fusion strategy to optimize the initial center selection, introduces dynamic time warping (DTW) as a similarity metric, and generates a set of typical daily combined power scenarios through a Medoids center update mechanism. The extreme identification module is used to identify extreme scenarios based on the combined wind and solar output scenario by combining the fast search for density peaks (CFSFDP) algorithm with the DTW distance to generate a clustering decision diagram.

9. An electronic device, characterized in that: The electronic device includes a processor, a memory, and a computer program stored on the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of a joint scene generation method taking into account wind-light correlation as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the joint scene generation method taking into account the wind-solar correlation as described in any one of claims 1 to 7 are implemented.

Images