Method and terminal for identifying dominant influence factors of power load change

Abstract

The invention discloses a method and terminal for identifying dominant influence factors of power load change, and the method comprises the steps: obtaining power load data and related to-be-identified influence factor data, calculating an error evaluation index of an incidence relation between the influence factors and the power load data, and obtaining an evidence matrix; according to the evidence matrix, constructing an influence weight calculation model based on consistency sorting so as to calculate the weight of each error evaluation index; based on the weight of each error evaluation index, fusion of each error evaluation index is carried out on each influence factor by using a basic probability distribution function and a DS evidence theory, and a comprehensive evaluation result of each influence factor is obtained; according to the method, a comprehensive evaluation result of multi-feature fusion is formed based on the consistency sorting and DS evidence fusion theory, the sorting of the most critical influence factors for the change of the power load can be obtained, and the identification of the dominant influence factors is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of power load influencing factor analysis, and in particular to a method and a terminal for identifying dominant influencing factors of power load changes. Background Art

[0002] In order to improve the utilization rate of power resources and provide a basis for formulating reasonable power consumption plans, it is necessary to analyze the changes in power load. The changes in power load are affected by many factors, including weather factors, economic factors, etc. The interaction between different factors causes the nonlinearity and dynamics of power load, which makes it more difficult to model the load law.

[0003] Since the factors that affect the change of load data are complicated, if all the influencing factors are taken into account when modeling, the amount of model calculation will increase and the operation efficiency of the model will be affected; at the same time, if the model contains some factors that are less related to load changes, it will affect the model training effect, reduce the accuracy of the model, and cause load prediction deviation; and the factors affecting load changes in different regions are inconsistent. If a single influencing factor is used, the effect of the load model will be poor. However, most of the input factors of the load change analysis model are not screened. In view of the above situation, it is very necessary to identify the key leading factors of load changes in order to improve the load change prediction effect while reducing the model operation pressure. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a method and a terminal for identifying the dominant influencing factors of power load changes, to comprehensively evaluate the influencing factors of power load, and to identify the dominant influencing factors.

[0005] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0006] A method for identifying the dominant influencing factors of power load changes, comprising the steps of:

[0007] S1. Obtain power load data and related data of various influencing factors to be identified, calculate the error evaluation index of the correlation relationship between the influencing factors and the power load data, and obtain an evidence matrix;

[0008] S2. According to the evidence matrix, construct an impact weight calculation model based on consistency ranking to calculate the weight of each error evaluation index;

[0009] S3. Based on the weights of the error evaluation indicators, the basic probability distribution function and DS evidence theory are used to fuse the error evaluation indicators of each influencing factor to obtain a comprehensive evaluation result of each influencing factor.

[0010] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0011] A terminal for identifying dominant influencing factors of power load changes, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method for identifying dominant influencing factors of power load changes described above when executing the computer program.

[0012] The beneficial effects of the present invention are as follows: a method and terminal for identifying the dominant influencing factors of power load changes of the present invention, based on consistency sorting and DS evidence fusion theory, synthesizes different error evaluation indicators of the correlation between various influencing factors and the influence of the power load, and forms a comprehensive evaluation result of multi-feature fusion, which can more accurately reflect the impact of the influencing factors on the power load, thereby obtaining the ranking of the most critical influencing factors for the change of the power load and realizing the identification of the dominant influencing factors. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 A flowchart of a method for identifying dominant influencing factors of power load changes according to an embodiment of the present invention;

[0014] Figure 2 A structural diagram of a terminal for identifying dominant influencing factors of power load changes according to an embodiment of the present invention;

[0015] Figure 3 This is an example diagram of the process steps of a method for identifying dominant influencing factors of power load changes according to an embodiment of the present invention;

[0016] Description of labels:

[0017] 1. A terminal for identifying dominant influencing factors of power load changes; 2. A processor; 3. A memory. DETAILED DESCRIPTION

[0018] In order to explain the technical content, achieved objectives and effects of the present invention in detail, the following is an explanation in combination with the implementation modes and the accompanying drawings.

[0019] Please refer to Figure 1 as well as Figure 3 , a method for identifying the dominant influencing factors of power load changes, comprising the steps of:

[0020] S1. Obtain power load data and related data of various influencing factors to be identified, calculate the error evaluation index of the correlation relationship between the influencing factors and the power load data, and obtain an evidence matrix;

[0021] S2. According to the evidence matrix, construct an impact weight calculation model based on consistency ranking to calculate the weight of each error evaluation index;

[0022] S3. Based on the weights of the error evaluation indicators, the basic probability distribution function and DS evidence theory are used to fuse the error evaluation indicators of each influencing factor to obtain a comprehensive evaluation result of each influencing factor.

[0023] From the above description, it can be seen that the beneficial effects of the present invention are: a method for identifying the dominant influencing factors of power load changes in the present invention, based on consistency sorting and DS evidence fusion theory, synthesizes different error evaluation indicators of the correlation between various influencing factors and the influence of the power load, and forms a comprehensive evaluation result of multi-feature fusion, thereby obtaining the ranking of the most critical influencing factors for the change of power load and realizing the identification of the dominant influencing factors.

[0024] Furthermore, the evidence matrix Ω is constructed as follows:

[0025]

[0026] Among them, l represents the number of influencing factors, p represents the number of error evaluation indicators, α lp It represents the p-th error evaluation index of the l-th influencing factor.

[0027] From the above description, it can be seen that for each influencing factor, the error evaluation index is calculated through a variety of correlation measurement algorithms to construct an evidence matrix of the association relationship.

[0028] Furthermore, the influence weight calculation model based on consistency sorting is expressed as:

[0029]

[0030] Among them, β j (A i ) represents the number of parameters of the jth error evaluation index of the i-th influencing factor that is located after the index in the consistency sorting, β j (A) represents the total number of parameters in the ranking of the jth error evaluation index, δ j represents the weight of the j-th error evaluation index, and k represents the number of error evaluation indicators.

[0031] From the above description, it can be seen that an influence weight calculation model based on consistency sorting is constructed to calculate the weights of various error evaluation indicators of influencing factors.

[0032] Further, step S3 comprises the steps of:

[0033] S31. According to the weight of the error evaluation index, the corresponding error evaluation index in the evidence matrix Ω is updated to obtain an updated evidence matrix:

[0034]

[0035] Among them, δ j represents the weight of the jth error evaluation index, A represents the influencing factor, B j represents the error evaluation index in the jth column of the evidence matrix;

[0036] S32, establish a recognition framework O, expressed as:

[0037] Ο={Ο1,Ο2,…,Ο i ,…,Ο l};

[0038] Wherein, l represents the number of influencing factors, O i Represents the i-th basic element in the recognition frame;

[0039] S33, basic probability distribution of events under the recognition framework, basic probability distribution function m:2 Ο →[0,1] satisfies:

[0040]

[0041] Among them, K represents the degree of conflict between evidence values, m(A) is the comprehensive evaluation result of the influencing factors, and m i (A i ) represents the basic probability distribution of the proposition calculation of the i-th influencing factor on the identification framework, and O represents the identification framework.

[0042] From the above description, it can be seen that based on the DC evidence theory, the evidence matrix is ​​updated according to the calculated weights of various error evaluation indicators, and an identification framework is established. By performing basic probability allocation on events under the identification framework, the comprehensive evaluation results of various influencing factors are determined.

[0043] Furthermore, the error evaluation indexes for each item of the influencing factor data include at least two, and the selection of the error evaluation indexes includes the maximum mutual information coefficient MIC, the Pearson correlation coefficient, the Spearman correlation coefficient and the Kendall correlation coefficient.

[0044] From the above description, it can be seen that the selection of error evaluation indicators includes but is not limited to the maximum mutual information coefficient MIC, the Pearson correlation coefficient, the Spearman correlation coefficient and the Kendall correlation coefficient.

[0045] Furthermore, the maximum mutual information coefficient MIC is calculated as follows:

[0046]

[0047] Where X represents the power load variable, Y represents the influencing factor variable, a represents the number of grids divided in the X direction, b represents the number of grids divided in the Y direction, I*(D| G ,a,b) represents the maximum mutual information under multiple grids, B(n) represents a predefined threshold, which is usually selected as the square root or fractional power of the number of data points to control the complexity of the grid.

[0048] From the above description, it can be seen that the calculation of the maximum mutual information coefficient is as shown above.

[0049] Furthermore, the Pearson correlation coefficient r xy The calculation is as follows:

[0050]

[0051] Where N represents x i represents the ith observation value of the power load variable, y i represents the i-th observed value of the influencing factor variable, represents the mean value of the power load variable, Represents the mean of the influencing factor variable.

[0052] From the above description, it can be seen that the calculation of the Pearson correlation coefficient is as shown above.

[0053] Furthermore, the calculation of the Spearman correlation coefficient ρ is as follows:

[0054]

[0055] Where n represents the number of samples, d i Represents the difference vector obtained after sorting the power load and influencing factors.

[0056] From the above description, it can be seen that the calculation of the Spearman correlation coefficient is as shown above.

[0057] Furthermore, the calculation of Kendall correlation coefficient τ is as follows:

[0058]

[0059] Among them, C represents the number of consistent pairs, D represents the number of inconsistent pairs, n represents the number of samples, T1 represents the number of pairs with the same ranking in the first variable, and T2 represents the number of pairs with the same ranking in the second variable.

[0060] From the above description, it can be seen that the calculation of Kendall's correlation coefficient is as shown above.

[0061] Please refer to Figure 2 A terminal for identifying dominant influencing factors of power load changes includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps in the method for identifying dominant influencing factors of power load changes described above are implemented.

[0062] The present invention provides a method and terminal for identifying dominant influencing factors of power load changes, which are suitable for power load influencing factor analysis.

[0063] Please refer to Figure 1 and Figure 3 , Embodiment 1 of the present invention is:

[0064] A method for identifying the dominant influencing factors of power load changes, comprising the steps of:

[0065] S1. Obtain power load data and related data of various influencing factors to be identified, calculate the error evaluation index of the correlation relationship between the influencing factors and the power load data, and obtain an evidence matrix;

[0066] The error evaluation indexes for each item of the influencing factor data include at least two, and the selection of the error evaluation indexes includes the maximum mutual information coefficient MIC, the Pearson correlation coefficient, the Spearman correlation coefficient and the Kendall correlation coefficient;

[0067] The evidence matrix Ω is constructed as follows:

[0068]

[0069] Among them, l represents the number of influencing factors, p represents the number of error evaluation indicators, α lp It represents the p-th error evaluation index of the l-th influencing factor.

[0070] In this embodiment, the maximum mutual information coefficient MIC, Pearson correlation coefficient, Spearman correlation coefficient and Kendall correlation coefficient are selected as four error evaluation indicators to construct the evidence matrix Ω:

[0071]

[0072] Among them, the i-th row vector A i (1≤i≤l) represents the evaluation index results contained in different influencing factors, and l represents the number of items of the influencing factors. j (1≤j≤4) represents different evaluation indicators.

[0073] Among them, the calculation of the maximum mutual information coefficient MIC is specifically as follows:

[0074]

[0075] Where X represents the power load variable, Y represents the influencing factor variable, a represents the number of grids divided in the X direction, b represents the number of grids divided in the Y direction, I*(D| G ,a,b) represents the maximum mutual information under multiple grids, B(n) represents a predefined threshold, which is usually selected as the square root or fractional power of the number of data points to control the complexity of the grid.

[0076] Pearson correlation coefficient r xy The calculation is as follows:

[0077]

[0078] Where N represents x i represents the ith observation value of the power load variable, y i represents the i-th observation value of the influencing factor variable, x represents the mean of the power load variable, and y represents the mean of the influencing factor variable.

[0079] The calculation of Spearman correlation coefficient ρ is as follows:

[0080]

[0081] Where n represents the number of samples, d i Represents the difference vector obtained after sorting the power load and influencing factors.

[0082] The calculation of Kendall correlation coefficient τ is as follows:

[0083]

[0084] Among them, C represents the number of consistent pairs, D represents the number of inconsistent pairs, n represents the number of samples, T1 represents the number of pairs with the same ranking in the first variable, and T2 represents the number of pairs with the same ranking in the second variable.

[0085] In this embodiment, the daily maximum temperature, minimum temperature, average temperature, maximum relative humidity, minimum relative humidity, average relative humidity, maximum wind speed, minimum wind speed, average wind speed, rainfall, and load average data of a certain area from February 1, 2023 to October 31, 2023, a total of 273 days, are selected as analysis samples to analyze the correlation characteristics of power load influencing factors. A quantitative relationship evidence matrix of the maximum mutual information coefficient MIC, Pearson correlation coefficient, Spearman correlation coefficient, and Kendall correlation coefficient between the above multidimensional influencing factors and power load is established, and an example of the evidence matrix Ω is shown in the following Table 1:

[0086] Table 1

[0087]

[0088] That is, the quantitative results of the correlation of factors affecting electricity load.

[0089] S2. According to the evidence matrix, construct an impact weight calculation model based on consistency ranking to calculate the weight of each error evaluation index;

[0090] The influence weight calculation model based on consistency sorting is expressed as:

[0091]

[0092] Among them, β j (A i ) represents the number of parameters of the jth error evaluation index of the i-th influencing factor that is located after the index in the consistency sorting, β j (A) represents the total number of parameters in the ranking of the jth error evaluation index, δ j represents the weight of the j-th error evaluation index, and k represents the number of error evaluation indicators.

[0093] S3. Based on the weights of the error evaluation indicators, the basic probability distribution function and DS evidence theory are used to integrate the error evaluation indicators of the influencing factors to obtain a comprehensive evaluation result of the influencing factors.

[0094] Step S3 includes the steps of:

[0095] S31. According to the weight of the error evaluation index, the corresponding error evaluation index in the evidence matrix Ω is updated to obtain an updated evidence matrix:

[0096]

[0097] Among them, δ jrepresents the weight of the jth error evaluation index, A represents the influencing factor, B j represents the error evaluation index in the jth column of the evidence matrix;

[0098] S32, establish a recognition framework O, expressed as:

[0099] Ο={Ο1,Ο2,…,Ο i ,…,Ο l};

[0100] Wherein, l represents the number of influencing factors, O i Represents the i-th basic element in the recognition frame;

[0101] S33, basic probability distribution of events under the recognition framework, basic probability distribution function m:2 Ο →[0,1] satisfies:

[0102]

[0103]

[0104] Among them, K represents the degree of conflict between evidence values, m(A) is the comprehensive evaluation result of the influencing factors, and m i (A i ) represents the basic probability distribution of the proposition calculation of the i-th influencing factor on the identification framework, and O represents the identification framework.

[0105] From the relationship matrix shown in Table 1 above, when a single evaluation index cannot fully display the correlation between power load and variables, the different evaluation directions of different indicators lead to misunderstandings in the determination of the dominant influencing factors. At this time, DS evidence fusion is used to synthesize all evaluation results to form a combined evaluation result of multi-feature fusion, and the most critical influencing factors for load changes are ranked, thereby improving the effectiveness of the correlation results.

[0106] In this embodiment, the results of each evaluation index after consistency sorting and the final comprehensive evaluation result are shown in the following Table 2.

[0107] Table 2

[0108]

[0109]

[0110] It can be seen that for changes in electricity load, the top three influencing factors are average temperature, minimum temperature and maximum temperature. Therefore, changes in temperature are very important in affecting human life and activities.

[0111] Please refer to Figure 2 , Embodiment 2 of the present invention is:

[0112] A terminal 1 for identifying dominant influencing factors of power load changes, comprising a processor 2, a memory 3, and a computer program stored in the memory 3 and executable on the processor 2, wherein the processor 2 implements the steps of a method for identifying dominant influencing factors of power load changes described in the first embodiment above when executing the computer program.

[0113] In summary, the present invention provides a method and terminal for identifying the dominant influencing factors of power load changes. Based on consistency sorting and DS evidence fusion theory, different error evaluation indicators of the correlation between various influencing factors and the influence of the power load are synthesized to form a comprehensive evaluation result of multi-feature fusion, which can more accurately reflect the impact of the influencing factors on the power load, thereby obtaining the ranking of the most critical influencing factors for the change of the power load and realizing the identification of the dominant influencing factors.

[0114] The dominant factor identification method for power load change established by the present invention fully explores the correlation between load and influencing factors, screens dominant factors from multiple influencing factors, and can reasonably select input influencing factors of load change analysis model, and verify the feasibility of the method through actual load and influencing factor data. While reducing the complexity of model calculation, it also reduces the error caused by low-correlation factors to model training results, which has certain practical significance for realizing load change analysis under multiple influencing factors.

[0115] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent transformations made using the contents of the present invention's specification and drawings, or directly or indirectly applied in related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for identifying the dominant influencing factors of power load changes, characterized in that: Includes steps: S1. Obtain power load data and related data of various influencing factors to be identified, calculate the error evaluation index of the correlation relationship between the influencing factors and the power load data, and obtain an evidence matrix; S2. According to the evidence matrix, construct an impact weight calculation model based on consistency ranking to calculate the weight of each error evaluation index; S3. Based on the weights of the error evaluation indicators, the basic probability distribution function and DS evidence theory are used to fuse the error evaluation indicators of each influencing factor to obtain a comprehensive evaluation result of each influencing factor.

2. A method for identifying the dominant influencing factors of power load changes according to claim 1, characterized in that: The evidence matrix Ω is constructed as follows: Among them, l represents the number of influencing factors, p represents the number of error evaluation indicators, α lp It represents the p-th error evaluation index of the l-th influencing factor.

3. The method for identifying the dominant influencing factors of power load changes according to claim 1 is characterized in that: The influence weight calculation model based on consistency sorting is expressed as: Among them, β j (A i ) represents the number of parameters of the jth error evaluation index of the i-th influencing factor that is located after the index in the consistency sorting, β j (A) represents the total number of parameters in the ranking of the jth error evaluation index, δ j represents the weight of the j-th error evaluation index, and k represents the number of error evaluation indicators.

4. The method for identifying the dominant influencing factors of power load changes according to claim 1 is characterized in that: Step S3 includes the steps of: S31. According to the weight of the error evaluation index, the corresponding error evaluation index in the evidence matrix Ω is updated to obtain an updated evidence matrix: Among them, δ j represents the weight of the jth error evaluation index, A represents the influencing factor, B j represents the error evaluation index in the jth column of the evidence matrix; S32, establish a recognition framework O, expressed as: O={O1,O2,…,O i ,…,THE l }; Wherein, l represents the number of influencing factors, O i Represents the i-th basic element in the recognition frame; S33, basic probability distribution of events under the recognition framework, basic probability distribution function m:2 Ο →[0,1] satisfies: Among them, K represents the degree of conflict between evidence values, m(A) is the comprehensive evaluation result of the influencing factors, and m i (A i ) represents the basic probability distribution of the proposition calculation of the i-th influencing factor on the identification framework, and O represents the identification framework.

5. The method for identifying the dominant influencing factors of power load changes according to claim 1 is characterized in that: The error evaluation indexes for each item of the influencing factor data include at least two, and the selection of the error evaluation indexes includes the maximum mutual information coefficient MIC, the Pearson correlation coefficient, the Spearman correlation coefficient and the Kendall correlation coefficient.

6. A method for identifying the dominant influencing factors of power load changes according to claim 5, characterized in that: The calculation of the maximum mutual information coefficient MIC is as follows: Where X represents the power load variable, Y represents the influencing factor variable, a represents the number of grids divided in the X direction, b represents the number of grids divided in the Y direction, I*(D| G ,a,b) represents the maximum mutual information under multiple grids, B(n) represents a predefined threshold, which is usually selected as the square root or fractional power of the number of data points to control the complexity of the grid.

7. A method for identifying the dominant influencing factors of power load changes according to claim 5, characterized in that: Pearson correlation coefficient r xy The calculation is as follows: Where N represents x i represents the ith observation value of the power load variable, y i represents the i-th observed value of the influencing factor variable, represents the mean value of the power load variable, Represents the mean of the influencing factor variable.

8. A method for identifying the dominant influencing factors of power load changes according to claim 5, characterized in that: The calculation of Spearman correlation coefficient ρ is as follows: Where n represents the number of samples, d i Represents the difference vector obtained after sorting the power load and influencing factors.

9. A method for identifying the dominant influencing factors of power load changes according to claim 5, characterized in that: The calculation of Kendall correlation coefficient τ is as follows: Among them, C represents the number of consistent pairs, D represents the number of inconsistent pairs, n represents the number of samples, T1 represents the number of pairs with the same ranking in the first variable, and T2 represents the number of pairs with the same ranking in the second variable.

10. A terminal for identifying dominant influencing factors of power load changes, comprising a processor, a memory, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method for identifying the dominant influencing factors of power load changes described in any one of claims 1 to 9 are implemented.

Images