A method and system for determining application modes in various application scenarios of renewable energy

Abstract

The application provides a kind of renewable energy application mode determination method and system in each application scene, comprising: based on the characteristic information of seawater desalination, obtain the relevant historical data in island area, generate multiple application scenes;Using clustering method to reduce multiple application scenes;Through analytic hierarchy process, obtain the application mode under each application scene after reduction;The characteristic information of seawater desalination includes: load characteristics, installed capacity, regional water use, wind speed, illumination, tidal energy, population, regional GDP and geographical location.The application maximizes the use of wind, light and other natural resources on the island, in combination with the seawater desalination system, can interact with the island power grid control center, as an effective adjustment means involved in the operation control of power grid to realize the optimization, economic and reliable operation of the whole island power grid, and provides a method basis for exploring the operation mode of the region.

Description

Technical Field

[0001] This invention belongs to the field of comprehensive utilization technology of renewable energy, and specifically relates to a method and system for determining the application mode in various application scenarios of renewable energy. Background Technology

[0002] Currently, the cost of seawater desalination remains relatively high. Although the cost per ton of desalinated water has dropped to around 5 yuan, it is still high compared to the relatively low tap water prices in most cities. This is the most direct and primary factor restricting the development of seawater desalination. The coastal seawater desalination and aquaculture industries are developing rapidly, with increasingly strong demand, especially in areas lacking freshwater resources, where seawater desalination has become a crucial source of demand. Simultaneously, coastal areas possess significant potential for distributed renewable energy sources such as solar, wind, and tidal power. Combining seawater desalination systems with distributed renewable energy can be crucial for solving water resource problems, absorbing distributed renewable energy, balancing and increasing grid load, and promoting comprehensive coastal development. However, the high cost of seawater desalination makes large-scale application difficult without a suitable application model to optimize economic benefits. Furthermore, different application scenarios have varying requirements regarding resource endowment, electricity demand, and water demand when combining seawater desalination systems with distributed renewable energy. Currently, there is a lack of analytical tools to determine which application model is suitable for a given region under current conditions. Summary of the Invention

[0003] To overcome the shortcomings of the prior art, this invention proposes a method for determining application modes in various application scenarios of renewable energy, the improvement of which includes:

[0004] Based on the characteristic information of seawater desalination and disposal, relevant historical data in the island area is obtained to generate multiple application scenarios;

[0005] Clustering methods are used to reduce the number of application scenarios.

[0006] The application patterns under each application scenario after reduction were obtained through the analytic hierarchy process.

[0007] The characteristic information of the seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP, and geographical location.

[0008] Preferably, based on the characteristic information of seawater desalination and utilization, relevant historical data within the island area is obtained to generate multiple application scenarios, including:

[0009] The characteristic information of seawater desalination and consumption is used as a random variable;

[0010] Based on a pre-defined number of scenarios, and using the number of random variables as a dimension, an original scenario set is constructed by combining relevant historical data within the island region.

[0011] Calculate the first n central moments of the correlation matrix of the original scene probability distribution containing multiple random variables;

[0012] Based on the first n moments of each random variable, calculate the transformation parameters required for the cubic transformation of each random variable.

[0013] The correlation coefficient matrix is ​​subjected to Cholesky decomposition to obtain the lower triangular transformation matrix, and a first matrix related to the random variable is obtained based on the lower triangular transformation matrix.

[0014] The first matrix is ​​cubically transformed using the transformation parameters to obtain a second matrix related to the random variable. The elements of each row vector of the second matrix constitute the application scenario in the island mode.

[0015] The value of n is 4, and the first four moments include: mean, variance, skewness, and peak value.

[0016] Preferably, the step of calculating the transformation parameters required for the cubic transformation of each random variable based on the first n moments of each random variable includes:

[0017] A first random number matrix is ​​generated using random numbers, where each column vector follows a standard normal distribution. The number of columns in the first random number matrix is ​​equal to the number of random variables, and the number of rows is greater than or equal to the number of columns.

[0018] Calculate the first M moments of each column vector of the first random number matrix;

[0019] Based on the first M moments and the first n moments of each random variable, calculate the transformation parameters required for the cubic transformation of each random variable.

[0020] Wherein, M = n*(n-1).

[0021] Preferably, obtaining the first matrix related to the random variable based on the lower triangular transformation matrix includes:

[0022] A second random number matrix is ​​generated, where each column vector follows a standard normal distribution. The number of columns in the second random number matrix is ​​equal to the number of random variables, and the number of rows is equal to the number of application scenarios.

[0023] Multiplying the second random number matrix by the transpose of the lower triangular transformation matrix yields the first matrix related to the random variable.

[0024] Preferably, the formula for calculating the cubic transformation is as follows:

[0025]

[0026] In the formula, z i,j Let a represent the element in the i-th row and j-th column of the second matrix. i,j Let a represent the element in the i-th row and j-th column of the first matrix. j b j c j and d j This represents the transformation parameters required when performing a cubic transformation on the j-th random variable.

[0027] Preferably, the method of reducing the number of application scenarios using clustering includes:

[0028] Select multiple cluster centers in the generated application scenario;

[0029] Each application scenario is assigned to a cluster corresponding to its respective cluster center based on the minimum Euclidean distance.

[0030] For each cluster corresponding to a cluster center, update the cluster center of each cluster based on the minimum weighted distance between each application scenario in the cluster and other application scenarios in the same cluster.

[0031] Calculate the weighted sum of distances from all application scenarios to the nearest cluster center. If the weighted sum of distances from all application scenarios to the nearest cluster center is less than a preset threshold or the number of iterations reaches the upper limit, then each cluster center is treated as a reduced application scenario and the process ends. Otherwise, proceed to assign each application scenario to the cluster corresponding to each cluster center based on the minimum Euclidean distance.

[0032] Preferably, the process of obtaining the application patterns for each application scenario after reduction through the analytic hierarchy process includes:

[0033] For each reduced application scenario, a hierarchical structure is established with the optimal application mode corresponding to the application scenario as the target layer, the random variables of the characteristic information of seawater desalination and consumption in the application scenario as the criterion layer, and the multiple application modes to be selected as the scheme layer.

[0034] The judgment matrix of the criterion layer to the target layer is obtained by comparing each random variable pairwise, and the judgment matrix of the scheme layer is obtained by comparing each random variable pairwise to the application mode under the influence of each random variable.

[0035] Calculate the first normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the criterion layer to the target layer in each application scenario, and the second normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the scheme layer.

[0036] Based on the first normalized feature vector and the second normalized feature vector, calculate the weights of each application mode in the target layer relative to the target layer.

[0037] The application mode with the highest weight is selected as the application mode corresponding to the application scenario.

[0038] Preferably, the application modes include: commercial mode, industrial mode, agricultural mode and residential mode.

[0039] Based on the same inventive concept, this application also provides a system for determining application modes in various application scenarios of renewable energy, the improvement of which is that it includes: an application scenario generation module, an application scenario reduction module, and an application mode module.

[0040] The application scenario generation module is used to obtain relevant historical data in the island area based on the characteristic information of seawater desalination and disposal, and generate multiple application scenarios.

[0041] The application scenario reduction module is used to reduce the multiple application scenarios using a clustering method;

[0042] The application mode module is used to obtain the application modes of each application scenario after reduction through the analytic hierarchy process.

[0043] The characteristic information of the seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP, and geographical location.

[0044] Preferably, the application scenario generation module includes: a random variable unit, an original scenario unit, a central moment unit, a transformation parameter unit, a first matrix unit, and an application scenario unit;

[0045] The random variable unit is used to take the characteristic information of seawater desalination and disposal as random variables;

[0046] The original scene unit is used to construct an original scene set based on a pre-set number of scenes, with the number of random variables as the dimension, and in combination with relevant historical data within the island area;

[0047] The central moment unit is used to calculate the first n-th order central moments of the correlation matrix of the original scene probability distribution containing multiple random variables;

[0048] The transformation parameter unit is used to calculate the transformation parameters required for cubic transformation of each random variable based on the first n moments of each random variable.

[0049] The first matrix unit is used to perform Cholesky decomposition on the correlation coefficient matrix to obtain a lower triangular transformation matrix, and to obtain a first matrix related to the random variable based on the lower triangular transformation matrix;

[0050] The application scenario unit is used to perform a cubic transformation on the first matrix using the transformation parameters to obtain a second matrix related to random variables. The elements of each row vector of the second matrix constitute the application scenario in the island mode.

[0051] The value of n is 4, and the first four moments include: mean, variance, skewness, and peak value.

[0052] Compared with the closest existing technology, the present invention has the following beneficial effects:

[0053] This invention provides a method and system for determining application modes in various renewable energy application scenarios, including: acquiring relevant historical data within an island region based on the characteristic information of seawater desalination consumption, generating multiple application scenarios; reducing the number of application scenarios using a clustering method; and obtaining the application mode under each reduced application scenario through an analytic hierarchy process (AHP). The characteristic information of seawater desalination consumption includes: load characteristics, installed capacity, regional water usage, wind speed, solar irradiance, tidal energy, population, regional GDP, and geographical location. This invention addresses the unique climate and natural environment conditions of islands compared to the mainland. By employing a wind-solar hybrid power generation system, it maximizes the utilization of natural resources such as wind and solar power on islands. Combined with a seawater desalination system, it can participate in the operation and control of the power grid as an effective regulatory means through interaction with the island's power grid control center, thereby achieving optimized, economical, and reliable operation of the entire island power grid. This provides a methodological basis for exploring regional operation modes. Attached Figure Description

[0054] Figure 1 A flowchart illustrating a method for determining application modes in various application scenarios of renewable energy provided by the present invention;

[0055] Figure 2 A flowchart illustrating an embodiment of a method for determining application modes in various application scenarios of renewable energy provided by the present invention;

[0056] Figure 3 This is a schematic diagram of the hierarchical structure involved in the present invention;

[0057] Figure 4 A schematic diagram of the basic structure of a system for determining application modes in various application scenarios of renewable energy provided by the present invention;

[0058] Figure 5 This invention provides a detailed structural diagram of a system for determining application modes in various application scenarios of renewable energy. Detailed Implementation

[0059] Given the unique climate and natural environment of islands compared to the mainland, special requirements are placed on the selection of materials, production, and installation of residential power systems on islands. Among these, wind-solar hybrid power generation systems can maximize the utilization of natural resources such as wind and solar power on islands. Combined with seawater desalination systems, these systems can participate in the operation and control of the power grid as an effective regulatory means through interaction with the island's power grid control center, thereby achieving optimized, economical, and reliable operation of the entire island power grid. By screening island pattern characteristic information and using scene analysis technology to generate a large number of scenarios, and then using a scene reduction algorithm to generate multiple application scenarios that approximate the original scenarios, the application modes under each application scenario are obtained through the analytic hierarchy process (AHP), namely four typical application modes: commercial, industrial, agricultural, and residential. This provides a methodological foundation for exploring the operational modes of the region.

[0060] The specific embodiments of the present invention will be further described in detail below with reference to the accompanying drawings.

[0061] Example 1:

[0062] A flowchart illustrating a method for selecting application modes for renewable energy provided by this invention is shown below. Figure 1 As shown, it includes:

[0063] Step 1: Based on the characteristic information of seawater desalination and disposal, obtain relevant historical data in the island area and generate multiple application scenarios;

[0064] Step 2: Use clustering methods to reduce the number of application scenarios;

[0065] Step 3: Obtain the application patterns for each application scenario after reduction using the Analytic Hierarchy Process (AHP).

[0066] The key information for seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP, and geographical location.

[0067] Specifically, the selection method for renewable energy application models is as follows: Figure 2 As shown, it includes:

[0068] Step 101: Obtain historical data for a certain region, including load characteristics, installed capacity, regional water use, wind speed, solar radiation, temperature, tidal energy, population, regional GDP, and geographical location, and define them as random variables.

[0069] Load characteristic B1: Installed capacity of various load types B 11 Typical daily load curves for various load types (B) 12 ;

[0070] Installed capacity B2: Rated installed capacity of wind power generation B 21 Rated installed capacity of photovoltaic power generation B22 Seawater desalination installed capacity B 23 Energy storage installed capacity B 24 ;

[0071] Regional Water Consumption Situation B3: Total Water Consumption by Type of User B 31 Typical daily water consumption curves for various user types (B) 32 ;

[0072] Wind speed B4: Average wind speed B 41 Typical daily wind speed curve B 42 ;

[0073] Light intensity B5: Annual average irradiance B 51 Typical daily irradiance curve B 52 ;

[0074] Temperature B6: Annual average temperature, B 61 Typical daily irradiance curve B 62 ;

[0075] Tidal Energy B7: High and Low Tide Timetable B 71 , High-Tide Curve B 72 ;

[0076] Population size B8: Fixed population size B 81 Number of migrant population B 82 ;

[0077] Regional GDPB9: GDP per capita B 91 GDP annual growth rate B 92 ;

[0078] Geographical Location B 10 Longitude B 10,1 Latitude B 10,2 Altitude B 10,3 .

[0079] Step 102: Since the stochastic programming model contains random variables describing uncertainties, it cannot be directly constructed to solve them. Instead, these uncertainties must be modeled as scenarios, and then appropriate scenario generation methods are used to generate scenarios. A set of scenarios, comprising numerous sub-scenarios and their probabilities of occurrence, is used to simulate the uncertainties. The set of application scenarios under the island model is generated by fitting the first four moments of historical data and calculating the correlation coefficient matrix of each random variable. Assume the number of scenarios to be generated is S, and the dimension of each scenario is T, which represents the number of random variables. Calculate the first four central moments (mean, variance, skewness, and peak value) of the correlation matrix containing the probability distributions of the original scenarios with multiple random variables.

[0080] mean

[0081] variance

[0082] Skewness

[0083] peak

[0084] Where i nt M represents the value of the t-th random variable in the n-th original application scenario, and N represents the total number of historical data samples. t1 M t2 M t3 and M t4 These are the mean, variance, skewness, and peak value of the historical data for each dimension, i.e., each random variable.

[0085] The original application scenario of multiple random variables is denoted as:

[0086]

[0087] I n Represents the nth original application scenario, i NT This represents the value of the T-th random variable in the Nth original application scenario.

[0088] Obtain the correlation coefficient matrix of the original scene probability distribution:

[0089]

[0090] in This is the correlation coefficient between the i-th and j-th random variables, where C(i,j) represents the covariance between the i-th and j-th random variables. Then, the scene set [I1,…I] generated using the moment matching method is used. n ,…I S ] S×T .

[0091] Step 102 specifically includes:

[0092] 102-1: Using MATLAB to perform statistical analysis on historical data, calculate the first four central moments of each dimension, i.e., the historical data of each random variable, which are the mean M. t1 Variance M t2 Skewness M t3 and peak M t4 Find the correlation coefficient between any two random variables and construct the correlation coefficient matrix R. Then, perform Cholesky decomposition on the correlation coefficient matrix to obtain the lower triangular transformation matrix L.

[0093] 102-2: Using MATLAB to randomly generate a The first random number matrix in To make it as large as possible is to better determine the transformation parameters 'a' required for the matrix transformation. j ,b j ,c j ,d j , And ensure that each dimension, i.e., each column vector, follows a standard normal distribution, and then obtain the matrix. The first twelve central moments of each column vector are calculated, and the first twelve central moments of each dimension of sample data, along with the first four central moments of historical data in the same dimension, are substituted into the nonlinear equation system to solve for the parameters required for the cubic transformation of each dimension of data. The nonlinear equation system is as follows:

[0094]

[0095] Wherein: Since the transformation parameters are solved using the above nonlinear equation system for each random variable in the historical data, the subscript j representing the j-th dimension random variable is omitted in the nonlinear equation system. Abbreviated as E[Y i k ],Will Abbreviated as a j Abbreviated as a, b j Abbreviated as b, c j Abbreviated as c, d j Abbreviated as d; E[Y i k ] represents the k-th order central moment of the i-th dimension of historical data. Represents the first random number matrix The central moments of the i-th dimension k-th order random variable in the matrix.

[0096] 102-3: Generate a second random number matrix X of size S×T, and make each column vector follow a standard normal distribution.

[0097] 102-4: Perform matrix transformations to transform the independent data in each column of the second random number matrix X into an initial scene matrix Z, i.e., the first matrix, with a certain correlation coefficient. This includes decomposing the correlation coefficient matrix R into R = LL using Cholesky decomposition. T Multiply the second random number matrix X on the right by the transpose of a lower triangular transformation matrix L, i.e., Z = XL. T .

[0098] 102-5: Perform a cubic transformation on each column of matrix Z so that each dimension of the data has the same first four central moments as the historical data, resulting in a new matrix. That is, the transformation formula for the second matrix is:

[0099]

[0100] Among them, a j b j c j d j Let a represent the transformation parameters required when performing a cubic transformation on the j-th dimension random variable. i,j Let z represent the element in the i-th row and j-th column of matrix Z. i,j Representation matrix The element in the i-th row and j-th column.

[0101] 102-6: Matrix Each element of a row vector can form an initial application scenario sample. Since the sampling process is random, the probability of each scenario is equal, i.e. All scenarios constitute a scenario set, and the scenarios and the probability of their occurrence can serve as a sample to describe the application scenario.

[0102] Step 103: Use the K-medoids clustering method to generate S initial scenes [I1,…I2] from Step 102. n ,…I S ] S×T The original scene set S is replaced by a set B consisting of K scenes. The optimal set B is the set of probability information closest to S, satisfying the following optimization problem.

[0103]

[0104] In the formula It is scene u i and u j The distance, p i It is the probability of scenario i occurring, i.e., in 102-6.

[0105] 103-1: Select K original scenes as the initial class center set B r1 Each center point corresponds to a cluster.

[0106] 103-2: and Among them B r,n This represents the set of class centers after n updates. It is calculated using Euclidean distance d(u i ,uj)=||u i -u j ||2willu i Divide into the cluster that is closest to the center point.

[0107] 103-3: Updating the centroid of class K, i.e., updating the centroid of each cluster separately, requires calculating the centroid u of the cluster in the update algorithm. c This includes calculating the weighted sum of distances between each scene in the cluster and other scenes, and the scene with the smallest sum of distances is the new center point u. c ,Right now

[0108]

[0109] with u c As the new center.

[0110] 103-4: If The iteration ends when the value is less than a given threshold or the maximum number of iterations is reached. r,n This refers to the set of scenes that have been reduced. The threshold is set to 10. -6 .if If the number of iterations is not less than the given threshold and the number of iterations has not reached the upper limit, then proceed to step 103-2.

[0111] Step 104: Select the K centers as K typical scenarios, and choose a suitable application mode for each typical scenario. The selected typical scenarios are categorized using the analytic hierarchy process (AHP). For each scenario, a judgment matrix is ​​constructed based on the specific circumstances of the candidate application modes, and a hierarchical ranking and consistency check are performed. Specifically, this includes:

[0112] Establish evaluation objects such as Figure 3 The hierarchical structure shown is as follows: Target layer A represents the optimal application mode in the Kth scenario; criterion layer B represents the T influencing factors considered; and solution layer C represents the multiple candidate application modes. This embodiment includes four application modes: commercialization, agriculturalization, industrialization, and residential application. In this embodiment, the influencing factors are the random variables in the historical data.

[0113] Construct the judgment matrix for each layer: Normalize the factors in the Kth scenario, assign a value of 1 to the target layer, and then use this value as a weight to assign to different influencing factors. The weight of the corresponding influencing factor represents the importance of the factor in the whole selection process.

[0114] Using a scaling method of 1 to 9 and their reciprocals, different factors are compared pairwise to obtain the judgment matrix of the criterion layer for the target layer. Perform a consistency check on the judgment matrix P, based on the consistency index. This indicates the degree to which the judgment matrix deviates from consistency, expressed as a consistency ratio. The CR indicates whether the consistency of the judgment matrix is ​​acceptable, where the randomness index RI can be obtained from a randomness index table. When CR ≤ 0.10, the judgment matrix is ​​considered to have satisfactory or acceptable consistency. Otherwise, for each element P in the judgment matrix...ij and P ji Make appropriate corrections, where μ max It is the largest eigenvalue of matrix P.

[0115] The randomness index is shown in Table 1:

[0116] Table 1:

[0117] n 1 2 3 4 5 6 7 8 9 10 11 RI 0 0 0.58 0.90 1.12 1.24 1.32 1.41 1.45 1.49 1.51

[0118] The meanings of the 1-9 scales are shown in Table 2:

[0119] Table 2:

[0120]

[0121] Calculate the largest eigenvalue μ of P max The corresponding normalized feature vector W = (w1, w2, ... w T ), where w i It is the weight coefficient of each influencing factor in the criterion layer.

[0122] The decision matrix of the scheme layer is in This refers to the degree of importance of the i-th influencing factor to the application model. A judgment matrix is ​​constructed using a scaling method of 1 to 9 and their reciprocals. A consistency check is performed on the overall hierarchical ranking of the schemes. This indicates the degree to which the judgment matrix deviates from consistency, where It is Q i The largest eigenvalue, the consistency ratio of the overall hierarchical ranking is CI i The RI index is the hierarchical single-rank consistency index of layer C to factor i in layer B, and the randomness index. i The randomness index can be found in the randomness index table. When CR ≤ 0.10, the judgment matrix is ​​considered to have satisfactory or acceptable consistency. Otherwise, for the elements in the judgment matrix... and Make appropriate adjustments. Therefore, the weight of C1 with respect to the target layer A is w. 11 ×w1+w 21 ×w2+...w T1 ×w T , where w 1,i ,w 2,i …w T,i For Q i The eigenvector corresponding to the largest eigenvalue, where W = [w1, w2, ..., w T ] is the largest eigenvalue μ of P maxSimilarly, the weights of C2, C3, and C4 relative to the target layer A can be obtained by sorting the weights of C1, C2, C3, and C4 and selecting the maximum value as the optimal mode.

[0123] In this step, by calculating and verifying the ranking weights of the target-criteria-level judgment matrix P and the ranking weights of the criteria-scheme-level judgment matrix Q, the overall ranking of layer C can be obtained, thereby determining the application mode applicable to the Kth scenario.

[0124] Example 2:

[0125] Based on the same inventive concept, this invention also provides a system for determining the application mode in various application scenarios of renewable energy. Since the principle of these devices in solving technical problems is similar to the method of generating simulation sections of continuous operation of the power grid, the repetitive parts will not be described again.

[0126] The system structure is as follows Figure 4 As shown, it includes: an application scenario generation module, an application scenario reduction module, and an application mode module;

[0127] The application scenario generation module is used to obtain relevant historical data in the island area based on the characteristic information of seawater desalination and disposal, and generate multiple application scenarios.

[0128] The application scenario reduction module is used to reduce multiple application scenarios using clustering methods;

[0129] The application pattern module is used to obtain the application patterns of each application scenario after reduction through the analytic hierarchy process.

[0130] The key information for seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP, and geographical location.

[0131] The detailed structure of the system for determining application modes in various renewable energy application scenarios is as follows: Figure 5 As shown.

[0132] The application scenario generation module includes: a random variable unit, an original scenario unit, a central moment unit, a transformation parameter unit, a first matrix unit, and an application scenario unit.

[0133] Random variable unit, used to take the characteristic information of seawater desalination and consumption as random variables;

[0134] The original scene unit is used to construct an original scene set based on a pre-defined number of scenes, with the number of random variables as the dimension, and in combination with relevant historical data within the island area;

[0135] The central moment unit is used to calculate the first n-th order central moments of the correlation matrix of the original scene probability distribution containing multiple random variables;

[0136] The transformation parameter unit is used to calculate the transformation parameters required for cubic transformation of each random variable based on the first n moments of each random variable.

[0137] The first matrix unit is used to perform Cholesky decomposition on the correlation coefficient matrix to obtain the lower triangular transformation matrix, and to obtain the first matrix related to the random variable based on the lower triangular transformation matrix;

[0138] The application scenario unit is used to perform a cubic transformation on the first matrix using transformation parameters to obtain a second matrix related to random variables. The elements of each row vector of the second matrix constitute the application scenario in the island mode.

[0139] When n is 4, the first 4 moments include: mean, variance, skewness, and peak value.

[0140] The transformation parameter unit includes: a first random number matrix subunit, a central moment subunit, and a transformation parameter subunit;

[0141] The first random number matrix subunit is used to generate a first random number matrix in which each column vector follows a standard normal distribution. The number of columns in the first random number matrix is ​​the number of random variables and the number of rows is greater than or equal to the number of columns.

[0142] The central moment sub-unit is used to calculate the first M moments of each column vector of the first random number matrix;

[0143] The transformation parameter subunit is used to calculate the transformation parameters required for cubic transformation of each random variable based on the first M moments and the first n moments of each random variable.

[0144] Where M = n*(n-1).

[0145] The first matrix unit includes: a second random number matrix subunit and a first matrix calculation subunit;

[0146] The second random number matrix subunit is used to generate a second random number matrix in which each column vector follows a standard normal distribution. The number of columns in the second random number matrix is ​​the number of random variables and the number of rows is the number of application scenarios.

[0147] The first matrix calculation subunit is used to right-multiply the second random number matrix by the transpose of the lower triangular transformation matrix to obtain the first matrix related to the random variable.

[0148] The application scenario reduction module includes: clustering center unit, scenario classification unit, center update unit, and iteration unit;

[0149] Cluster center unit, used to select multiple cluster centers in the generated application scenario;

[0150] The scene classification unit is used to divide each application scene into the cluster corresponding to each cluster center based on the minimum Euclidean distance.

[0151] The central update unit is used to update the cluster center of each cluster according to the minimum weighted distance between each application scenario in the cluster and other application scenarios in the same cluster.

[0152] The iteration unit is used to calculate the weighted sum of distances from all application scenarios to the nearest cluster center. When the weighted sum of distances from all application scenarios to the nearest cluster center is less than a preset threshold or the number of iterations reaches the upper limit, each cluster center is taken as a reduced application scenario and the process ends; otherwise, the scenario classification unit is called.

[0153] The application mode module includes: hierarchical structure unit, judgment matrix unit, feature vector unit, weight unit and application mode unit;

[0154] The hierarchical structure unit is used to establish a hierarchical structure for each reduced application scenario, with the optimal application mode corresponding to the application scenario as the target layer, the random variables of the characteristic information of seawater desalination and consumption in the application scenario as the criterion layer, and the multiple application modes to be selected as the scheme layer.

[0155] The judgment matrix unit is used to obtain the judgment matrix of the criterion layer to the target layer by comparing each random variable pairwise, and to obtain the judgment matrix of the scheme layer by comparing each random variable pairwise to the application mode under the influence of each random variable.

[0156] The feature vector unit is used to calculate the first normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the criterion layer to the target layer in each application scenario, and the second normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the scheme layer.

[0157] The weighting unit is used to calculate the weights of each application mode in the target layer relative to the target layer based on the first normalized feature eigenvector and the second normalized feature eigenvector.

[0158] The application mode unit is used to select the application mode with the highest weight as the application mode corresponding to the application scenario.

[0159] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0160] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0161] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0162] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0163] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and not to limit its protection scope. Although this application has been described in detail with reference to the above embodiments, those skilled in the art should understand that after reading this application, they can still make various changes, modifications or equivalent substitutions to the specific implementation of the application, but these changes, modifications or equivalent substitutions are all within the protection scope of the claims pending approval.

Claims

1. A method for determining application modes in various application scenarios of renewable energy, characterized in that, include: Based on the characteristic information of seawater desalination and disposal, relevant historical data in the island area is obtained to generate multiple application scenarios; Clustering methods are used to reduce the number of application scenarios. The application patterns under each application scenario after reduction were obtained through the analytic hierarchy process. The characteristic information of the seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP and geographical location. Based on the characteristic information of seawater desalination and utilization, relevant historical data within the island area is obtained to generate multiple application scenarios, including: The characteristic information of seawater desalination and consumption is used as a random variable; Based on a pre-defined number of scenarios, and using the number of random variables as a dimension, an original scenario set is constructed by combining relevant historical data within the island region. Calculate the first n central moments of the correlation coefficient matrix of the original scene probability distribution containing multiple random variables; Based on the first n moments of each random variable, calculate the transformation parameters required for the cubic transformation of each random variable. The correlation coefficient matrix is ​​subjected to Cholesky decomposition to obtain the lower triangular transformation matrix, and a first matrix related to the random variable is obtained based on the lower triangular transformation matrix. The first matrix is ​​cubically transformed using the transformation parameters to obtain a second matrix related to the random variable. The elements of each row vector of the second matrix constitute the application scenario in the island mode. The n is 4, and the first 4 moments include: mean, variance, skewness, and peak value; The application patterns obtained through the analytic hierarchy process (AHP) for each application scenario after reduction include: For each reduced application scenario, a hierarchical structure is established with the optimal application mode corresponding to the application scenario as the target layer, the random variables of the characteristic information of seawater desalination and consumption in the application scenario as the criterion layer, and the multiple application modes to be selected as the scheme layer. The judgment matrix of the criterion layer to the target layer is obtained by comparing each random variable pairwise, and the judgment matrix of the scheme layer is obtained by comparing each random variable pairwise to the application mode under the influence of each random variable. Calculate the first normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the criterion layer to the target layer in each application scenario, and the second normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the scheme layer. Based on the first normalized feature vector and the second normalized feature vector, calculate the weights of each application mode in the target layer relative to the target layer. The application mode with the highest weight is selected as the application mode corresponding to the application scenario.

2. The method as described in claim 1, characterized in that, The transformation parameters required for the cubic transformation of each random variable are calculated based on the first n moments of each random variable, including: A first random number matrix is ​​generated using random numbers, where each column vector follows a standard normal distribution. The number of columns in the first random number matrix is ​​equal to the number of random variables, and the number of rows is greater than or equal to the number of columns. Calculate the first M moments of each column vector of the first random number matrix; Based on the first M moments and the first n moments of each random variable, calculate the transformation parameters required for the cubic transformation of each random variable. Wherein, M = n * (n-1).

3. The method as described in claim 1, characterized in that, The process of obtaining the first matrix related to the random variable based on the lower triangular transformation matrix includes: A second random number matrix is ​​generated, where each column vector follows a standard normal distribution. The number of columns in the second random number matrix is ​​equal to the number of random variables, and the number of rows is equal to the number of application scenarios. Multiplying the second random number matrix by the transpose of the lower triangular transformation matrix yields the first matrix related to the random variable.

4. The method as described in claim 1, characterized in that, The formula for calculating the cubic transformation is as follows: In the formula, This represents the element in the i-th row and j-th column of the second matrix. This represents the element in the i-th row and j-th column of the first matrix. , , and This represents the transformation parameters required when performing a cubic transformation on the j-th random variable.

5. The method as described in claim 1, characterized in that, The reduction of the multiple application scenarios using clustering methods includes: Select multiple cluster centers in the generated application scenario; Each application scenario is assigned to a cluster corresponding to its respective cluster center based on the minimum Euclidean distance. For each cluster corresponding to a cluster center, update the cluster center of each cluster based on the minimum weighted distance between each application scenario in the cluster and other application scenarios in the same cluster. Calculate the weighted sum of distances from all application scenarios to the nearest cluster center. If the weighted sum of distances from all application scenarios to the nearest cluster center is less than a preset threshold or the number of iterations reaches the upper limit, then each cluster center is treated as a reduced application scenario and the process ends. Otherwise, proceed to assign each application scenario to the cluster corresponding to each cluster center based on the minimum Euclidean distance.

6. The method as described in claim 1, characterized in that, The application models include: commercialization model, industrialization model, agriculturalization model, and residential model.

7. A system for determining application modes in various application scenarios of renewable energy, characterized in that, include: Application scenario generation module, application scenario reduction module, and application mode module; The application scenario generation module is used to obtain relevant historical data in the island area based on the characteristic information of seawater desalination and disposal, and generate multiple application scenarios. The application scenario reduction module is used to reduce the multiple application scenarios using a clustering method; The application mode module is used to obtain the application modes of each application scenario after reduction through the analytic hierarchy process. The characteristic information of the seawater desalination and utilization includes: load characteristics, installed capacity, regional water use, wind speed, solar radiation, tidal energy, population, regional GDP and geographical location. The application scenario generation module includes: a random variable unit, an original scenario unit, a central moment unit, a transformation parameter unit, a first matrix unit, and an application scenario unit; The random variable unit is used to take the characteristic information of seawater desalination and disposal as random variables; The original scene unit is used to construct an original scene set based on a pre-set number of scenes, with the number of random variables as the dimension, and in combination with relevant historical data within the island area; The central moment unit is used to calculate the first n-th order central moments of the correlation coefficient matrix of the original scene probability distribution containing multiple random variables; The transformation parameter unit is used to calculate the transformation parameters required for cubic transformation of each random variable based on the first n moments of each random variable. The first matrix unit is used to perform Cholesky decomposition on the correlation coefficient matrix to obtain a lower triangular transformation matrix, and to obtain a first matrix related to the random variable based on the lower triangular transformation matrix; The application scenario unit is used to perform a cubic transformation on the first matrix using the transformation parameters to obtain a second matrix related to random variables. The elements of each row vector of the second matrix constitute the application scenario in the island mode. The n is 4, and the first 4 moments include: mean, variance, skewness, and peak value; The application mode module includes: a hierarchical structure unit, a judgment matrix unit, an eigenvector unit, a weight unit, and an application mode unit; The hierarchical structure unit is used to establish a hierarchical structure for each reduced application scenario, with the optimal application mode corresponding to the application scenario as the target layer, the random variables of the characteristic information of seawater desalination and consumption in the application scenario as the criterion layer, and the multiple application modes to be selected as the scheme layer. The judgment matrix unit is used to obtain the judgment matrix of the criterion layer to the target layer by comparing each random variable pairwise, and to obtain the judgment matrix of the scheme layer by comparing each random variable pairwise to the application mode under the influence of each random variable. The feature vector unit is used to calculate the first normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the criterion layer to the target layer in each application scenario, and the second normalized feature vector corresponding to the largest eigenvalue of the judgment matrix of the scheme layer. The weighting unit is used to calculate the weights of each application mode in the target layer relative to the target layer based on the first normalized feature eigenvector and the second normalized feature eigenvector. The application mode unit is used to select the application mode with the highest weight as the application mode corresponding to the application scenario.

Images