CN119848801B - A time-sharing electricity decomposition method based on sensitivity analysis - Google Patents
A time-sharing electricity decomposition method based on sensitivity analysis Download PDFInfo
- Publication number
- CN119848801B CN119848801B CN202510315935.7A CN202510315935A CN119848801B CN 119848801 B CN119848801 B CN 119848801B CN 202510315935 A CN202510315935 A CN 202510315935A CN 119848801 B CN119848801 B CN 119848801B
- Authority
- CN
- China
- Prior art keywords
- time
- electric quantity
- sharing electric
- decomposition
- sharing
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/27—Regression, e.g. linear or logistic regression
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Supply And Distribution Of Alternating Current (AREA)
Abstract
本发明提出了一种基于敏感性分析的分时电量分解方法，涉及电量分解技术领域，获取目标区域的历史和实时分时电量数据、分时气象数据，网格化处理后得到气象因子和联合数据集，筛选出关键气象因子集合；用历史分时电量数据构建分位数回归的基础分时电量分解模型，输入实时数据得基础分量并获分时电量残差，对残差进行二次模态分解，得到气象敏感模态分量集合；构建 MOEA/D 多目标优化模型，以气象敏感模态分量集合为输入，用切比雪夫分解法求解，基于关键气象因子更新权重获修正气象电量分量；最后将基础分量与修正分量叠加得到最终分解结果。本发明通过多模型结合提高电量分解精度，为电力调度、需求预测提供有力支持。
The present invention proposes a time-sharing electricity decomposition method based on sensitivity analysis, which relates to the technical field of electricity decomposition. The historical and real-time time-sharing electricity data and time-sharing meteorological data of the target area are obtained, and the meteorological factors and joint data sets are obtained after gridding processing, and the key meteorological factor set is screened out; the basic time-sharing electricity decomposition model of quantile regression is constructed with the historical time-sharing electricity data, and the real-time data is input to obtain the basic components and the time-sharing electricity residuals, and the residuals are subjected to secondary modal decomposition to obtain the meteorological sensitive modal component set; the MOEA/D multi-objective optimization model is constructed, and the meteorological sensitive modal component set is used as input, and the Chebyshev decomposition method is used to solve it, and the weights are updated based on the key meteorological factors to obtain the corrected meteorological electricity components; finally, the basic components and the corrected components are superimposed to obtain the final decomposition result. The present invention improves the accuracy of electricity decomposition by combining multiple models, and provides strong support for power dispatching and demand forecasting.
Description
Technical Field
The invention relates to a time-sharing electric quantity decomposition method based on sensitivity analysis, and belongs to the technical field of electric quantity decomposition.
Background
In the field of power load prediction and management, the influence of meteorological factors on power demand is increasingly emphasized. In recent years, with advances in data analysis technology and meteorology, researchers have begun to explore the inclusion of meteorological factors into power decomposition and prediction models, thereby improving the accuracy and reliability of predictions.
The Chinese patent publication No. CN109299814A discloses a method for predicting the weather-induced electric quantity of a system. The method systematically analyzes the influence of meteorological factors on the power load through the steps of setting basic months, calculating the weather correlation, calculating the growth rate of the meteorological influence electric quantity and the like. Specifically, the method divides one year into winter and summer, calculates correlations respectively, and derives the importance of meteorological factors from the obtained data. This process helps to significantly improve the accuracy of load prediction and reduce errors.
However, the above patent adopts seasonal division (winter/summer) and linear growth rate calculation methods, only can process the linear relation between a single meteorological factor and electric quantity, and cannot effectively capture nonlinear coupling effects among multiple meteorological factors (temperature, humidity, wind speed and the like), and the method for screening based on pearson correlation coefficients is difficult to identify meteorological factor combinations with obvious interaction, so that the prediction accuracy of the model under complex climatic conditions is limited.
Disclosure of Invention
In order to solve the problems in the prior art, the invention provides a time-sharing electric quantity decomposition method based on sensitivity analysis.
The technical scheme of the invention is as follows:
the invention provides a time-sharing electric quantity decomposition method based on sensitivity analysis, which comprises the following steps:
acquiring time-sharing electric quantity related data of a target area, wherein the time-sharing electric quantity related data comprises historical time-sharing electric quantity data, real-time-sharing electric quantity data and time-sharing meteorological data, and performing gridding processing on the time-sharing electric quantity related data to acquire meteorological factors and a time-sharing electric quantity-meteorological joint data set;
Calculating sensitivity indexes of all weather factors to the time-sharing electric quantity based on the time-sharing electric quantity-weather combined data set, and screening weather factors with sensitivity indexes larger than a sensitivity threshold value to obtain a key weather factor set;
A basic time-sharing electric quantity decomposition model is built based on the historical time-sharing electric quantity data, and real-time-sharing electric quantity data is input into the basic time-sharing electric quantity decomposition model to obtain basic time-sharing electric quantity components;
Performing secondary modal decomposition on the time-sharing electric quantity residual by using an improved fully-adaptive noise set decomposition algorithm ICEEMDAN to obtain a weather-sensitive modal component set;
Establishing an MOEA/D multi-target optimization model, taking a weather sensitive modal component set as input, and solving the MOEA/D multi-target optimization model by using a Chebyshev decomposition method to obtain a modified weather electric quantity component, wherein the MOEA/D multi-target optimization model carries out weight vector update based on a key weather factor set;
and superposing the basic time-sharing electric quantity component and the corrected meteorological electric quantity component to obtain a final time-sharing electric quantity decomposition result.
As a preferred embodiment of the present invention, the time-sharing weather data includes temperature, humidity, wind speed, solar intensity and precipitation.
In a preferred embodiment of the present invention, the meshing process is performed on the time-division electric quantity related data to obtain a meteorological factor and a time-division electric quantity-meteorological joint data set specifically includes:
Carrying out gridding treatment on the time-sharing meteorological data according to the city dimension to obtain gridding meteorological indexes, namely meteorological factors, carrying out gridding treatment on the historical time-sharing electric quantity data and the real-time-sharing electric quantity data according to the city dimension and the industry dimension, and constructing a time-sharing electric quantity-meteorological combined data set by combining the meteorological factors, wherein the gridding resolution is 1km multiplied by 1km;
and carrying out data preprocessing on the time-sharing electric quantity-weather combined data set, wherein the data preprocessing comprises data alignment, outlier rejection and feature standardization.
As a preferred embodiment of the invention, the sensitivity index of each meteorological factor to the time-sharing electric quantity is calculated based on the time-sharing electric quantity-meteorological combined data set, and specifically comprises the following steps:
The global sensitivity corresponding to the time-sharing electric quantity-meteorological joint data set is calculated through a Sobol index, and expressed as follows by a formula:
;
In the formula, Is the firstGlobal sensitivity to individual meteorological factors; Is the first Weather factors; to remove Other meteorological factors; Is the variance; is a conditional expectation; is a time-sharing electric quantity output variable; To give at the first Individual meteorological factorsTime-sharing electric quantity output variable when taking valueRegarding the removal ofOther meteorological factorsIs not limited to the desired one; To expect to Regarding the firstIndividual meteorological factorsSolving variance; Is a time-sharing electric quantity output variable Is a contrast of (3);
Calculating mutual information of meteorological factors and time-sharing electric quantity by using nuclear density estimation Expressed as:
;
In the formula, Is thatAndIs a joint probability density of (2); And Respectively isAndIs a boundary probability density of (1); And Respectively isAndSpecific values of (2);
based on global sensitivity And mutual informationCalculating to obtain a sensitivity index, wherein the sensitivity index is expressed as the following formula:
;
In the formula, Is the firstSensitivity indexes of individual meteorological factors; Is a weight coefficient; The information entropy of the time-sharing electric quantity.
As a preferred embodiment of the present invention, the construction of the basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data is specifically as follows:
the basic time-sharing electric quantity decomposition model is a quantile regression model, and is expressed as follows:
;
In the formula, Is thatThe number of time-division digits isIn the training process,Is thatThe number of time-division digits isIs used for monitoring the historical time-sharing electric quantity in real time,Is thatThe number of time-division digits isIs a real-time-sharing electric quantity; And Is thatThe number of time-division digits isRegression coefficients of (a); And Respectively represent the firstMarket and firstIndustry by industry; And Respectively the total number of the ground cities and the total number of the industries; Is that The number of time-division digits isIs the first of (2)A ground variable coefficient; Is that The number of time-division digits isIs the first of (2)Industry variable coefficients for individual industries; Is that The number of time-division digits isError terms of (2); is quantile;
and training the basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data to obtain a trained basic time-sharing electric quantity decomposition model, and inputting the real-time-sharing electric quantity data into the basic time-sharing electric quantity decomposition model to obtain real-time basic time-sharing electric quantity components.
As a preferred embodiment of the present invention, the time-sharing electric quantity residual obtained based on the basic time-sharing electric quantity component is expressed as:
;
In the formula, Is thatTime-sharing electric quantity residual error at moment; Is that Real-time-sharing electric quantity at moment.
As a preferred embodiment of the present invention, the improved fully adaptive noise set decomposition algorithm ICEEMDAN is used to perform a secondary modal decomposition on the time-sharing electric quantity residual to obtain a weather-sensitive modal component set, specifically:
Performing first ICEEMDAN decomposition on the time-divided electric quantity residual error to obtain An intrinsic mode function and a residual component, expressed as:
;
In the formula, For the first decomposition to obtainA natural mode function; Residual components obtained by the first decomposition;
and (3) selecting all the natural mode functions obtained by the first decomposition to carry out the second decomposition, wherein the second decomposition is expressed as follows:
;
In the formula, Representation pairThe second decomposition is carried out to obtain the firstThe function of the individual natural modes,The total number of the intrinsic mode functions obtained by the second decomposition is the total number of the intrinsic mode functions; Residual components obtained by the second decomposition;
combining all the natural mode functions obtained by the second decomposition to obtain a weather sensitive mode component set Expressed as:
。
as a preferred embodiment of the invention, the built MOEA/D multi-objective optimization model is expressed as follows:
;
In the formula, Minimizing a target for residual fitting errors; Is the first Individual meteorological factors are atA nonlinear mapping function of time; the length of the real-time-sharing electric quantity signal is; Is the first The weight coefficient of each meteorological factor; maximizing a target for consistency of the weight and the sensitivity coefficient; Is the total number of meteorological factors.
As a preferred embodiment of the invention, the MOEA/D multi-objective optimization model is solved by using a Chebyshev decomposition method, which is specifically as follows:
Definition of Chebyshev distance Expressed as:
;
In the formula, Is a decision variable, used to refer to the firstThe weight coefficient of each meteorological factor; For a multi-objective set of trade-off weights, ;Indexing for a target; As a set of ideal points, ;As a decision variableA corresponding objective function value;
generating a group of initial weight vector sets which are uniformly distributed according to the key meteorological factor sets, and expressing the initial weight vector sets as follows by a formula:
;
In the formula, Is an initial set of weight vectors; bit index for initial weight vector; for the total number of initial weight vectors, each initial weight vector corresponds to one sub-problem;
calculating weight vectors in an initial set of weight vectors Selecting initial weight vectors with the distances meeting the distance threshold to form corresponding neighborhoods;
Randomly generating an initial population setExpressed as:
;
In the formula, Is corresponding to the initial weight vectorSolution of (2);
From the neighborhood Randomly selecting two solutions, carrying out genetic operation on the two selected solutions to obtain a new solution, and calculating an objective function value of the new solution;
Performing iterative solution, including respectively calculating chebyshev distances of the current solution and the new solution relative to the multi-objective weighing weight and the ideal point for each sub-problem in the neighborhood in each iteration, and if the chebyshev distance corresponding to the new solution is smaller than or equal to the chebyshev distance corresponding to the current solution, replacing the current solution with the new solution;
If the objective function value is smaller than the ideal point, the ideal point is equal to the objective function value;
and repeating the iterative solving process until a preset termination condition is met, stopping iteration, wherein the solution in the population is the approximate optimal solution of the MOEA/D multi-objective optimization model, and obtaining the corrected meteorological electric quantity component based on the approximate optimal solution.
As a preferred embodiment of the present invention, the final time-sharing electric quantity decomposition result is expressed as:
;
In the formula, The final time-sharing electric quantity decomposition result; and the corrected meteorological electric quantity component.
The invention has the following beneficial effects:
1. The invention relates to a time-sharing electric quantity decomposition method based on sensitivity analysis, which is characterized in that the sensitivity index of each meteorological factor to the time-sharing electric quantity is calculated, key meteorological factors with the sensitivity index larger than a threshold value are screened out, the meteorological factors with larger influence on the time-sharing electric quantity are focused, the interference of irrelevant information is reduced, the calculation efficiency and accuracy of a subsequent model are improved, and the model can decompose the electric quantity more pertinently;
2. The invention relates to a time-sharing electric quantity decomposition method based on sensitivity analysis, which utilizes historical time-sharing electric quantity data to construct a basic time-sharing electric quantity decomposition model (quantile regression model), inputs real-time-sharing electric quantity data into the model to obtain basic time-sharing electric quantity components, and the quantile regression model can consider electric quantity distribution conditions under different quantiles to comprehensively reflect the change characteristics of electric quantity and calculate the basic electric quantity components more accurately;
3. The invention relates to a time-sharing electric quantity decomposition method based on sensitivity analysis, which is characterized in that a time-sharing electric quantity residual error is subjected to secondary modal decomposition by adopting an improved complete self-adaptive noise set decomposition algorithm (ICEEMDAN) to obtain a weather sensitive modal component set, weather sensitive information in the residual error is further separated and refined, weather influencing factors hidden in the residual error are excavated, more detailed information is provided for subsequently correcting the weather electric quantity component, and the accuracy of electric quantity decomposition is improved;
4. The invention discloses a time-sharing electric quantity decomposition method based on sensitivity analysis, which is characterized in that an MOEA/D multi-objective optimization model is constructed, a Chebyshev decomposition method is utilized to solve the MOEA/D multi-objective optimization model, and a weight vector is updated based on a key meteorological factor set, so that a corrected meteorological electric quantity component is obtained.
Drawings
FIG. 1 is a flow chart of the method of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
It should be understood that the step numbers used herein are for convenience of description only and are not limiting as to the order in which the steps are performed.
It is to be understood that the terminology used in the description of the invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used in this specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise.
The terms "comprises" and "comprising" indicate the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof.
The term "and/or" refers to any and all possible combinations of one or more of the associated listed items, and includes such combinations.
Example 1:
Referring to fig. 1, the embodiment provides a time-sharing electric quantity decomposition method based on sensitivity analysis, which includes the following steps:
S1, acquiring time-sharing electric quantity related data of a target area through a marketing business system and a data center, wherein the time-sharing electric quantity related data comprises historical time-sharing electric quantity data, real-time-sharing electric quantity data and time-sharing meteorological data, and performing gridding processing on the time-sharing electric quantity related data to acquire a meteorological factor and a time-sharing electric quantity-meteorological combined data set;
s11, the time-sharing meteorological data comprise temperature, humidity, wind speed, sunlight intensity and precipitation;
S12, carrying out gridding treatment on time-sharing meteorological data according to the city dimension to obtain gridding meteorological indexes, namely meteorological factors, carrying out gridding treatment on historical time-sharing electric quantity data and real-time-sharing electric quantity data according to the city dimension and the industry dimension, and constructing a time-sharing electric quantity-meteorological combined data set by combining the meteorological factors, wherein the gridding resolution is 1km multiplied by 1km;
further, in this embodiment, the city dimension is divided by administrative area, the industry dimension includes industry, commerce, agriculture, pasture, fishery, and urban and rural resident classifications;
S13, carrying out data preprocessing on the time-sharing electric quantity-weather combined data set, wherein the data preprocessing comprises data alignment, outlier rejection and characteristic standardization, and using the preprocessed time-sharing electric quantity-weather combined data set for subsequent decomposition calculation;
S2, calculating sensitivity indexes of all weather factors to time-sharing electric quantity based on the time-sharing electric quantity-weather combined data set, and screening weather factors with sensitivity indexes larger than a sensitivity threshold value to obtain a key weather factor set;
S21, calculating global sensitivity corresponding to the time-sharing electric quantity-meteorological joint data set through a Sobol index, wherein the global sensitivity is expressed as follows:
;
In the formula, Is the firstGlobal sensitivity to individual meteorological factors; Is the first Weather factors; to remove Other meteorological factors; Is the variance; is a conditional expectation; is a time-sharing electric quantity output variable; To give at the first Individual meteorological factorsTime-sharing electric quantity output variable when taking valueRegarding the removal ofOther meteorological factorsIs not limited to the desired one; To expect to Regarding the firstIndividual meteorological factorsSolving variance; Is a time-sharing electric quantity output variable Is a contrast of (3);
Calculating mutual information of meteorological factors and time-sharing electric quantity by using nuclear density estimation Expressed as:
;
In the formula, Is thatAndIs a joint probability density of (2); And Respectively isAndIs a boundary probability density of (1); And Respectively isAndSpecific values of (2);
based on global sensitivity And mutual informationCalculating to obtain a sensitivity index, wherein the sensitivity index is expressed as the following formula:
;
In the formula, Is the firstSensitivity indexes of individual meteorological factors; Is a weight coefficient; information entropy of time-sharing electric quantity;
s22, dynamically determining the sensitivity threshold through Monte Carlo cross-validation, which is a conventional technical means in the field and is not described in detail herein;
S3, constructing a basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data, inputting the real-time-sharing electric quantity data into the basic time-sharing electric quantity decomposition model to obtain a basic time-sharing electric quantity component;
S31, in the embodiment, historical time-sharing electric quantity data in winter (12 months-2 months) and summer (6 months-8 months) can be further selected, a fractional regression model is respectively built for 24 time-sharing points of a day, namely a basic time-sharing electric quantity decomposition model, and the model is expressed as follows:
;
In the formula, Is thatThe number of time-division digits isIn the training process,Is thatThe number of time-division digits isIs used for monitoring the historical time-sharing electric quantity in real time,Is thatThe number of time-division digits isIs a real-time-sharing electric quantity; And Is thatThe number of time-division digits isRegression coefficients of (a); And Respectively represent the firstMarket and firstIndustry by industry; And Respectively the total number of the ground cities and the total number of the industries; Is that The number of time-division digits isIs the first of (2)A ground variable coefficient; Is that The number of time-division digits isIs the first of (2)Industry variable coefficients for individual industries; Is that The number of time-division digits isError terms of (2); is quantile;
The reasons for selecting winter and summer include 1) that the two seasons are the seasons with the most obvious temperature fluctuation, and the influence of temperature and humidity on the electricity consumption is particularly prominent, 2) that the temperature and the electricity consumption respectively show negative correlation and positive correlation in the two seasons, and are representative;
the selection of the quantile is also important, and regression results obtained by different quantiles are quite different, and the method is specific:
1) When (when) =0.1 AndWhen=0.9, the method is used for capturing extreme electric field scenes:
=0.1 reflecting the lowest power demand under extremely low temperature or low humidity conditions (such as winter chills and summer storms), corresponding to the "bottom-guard" load of the power system;
=0.9, reflecting peak electricity demand under extremely high temperature or high humidity conditions (e.g. summer heat, winter heating peak), corresponding to "overload" risk points of the grid;
When (when) =0.25At=0.75, the normal fluctuation range is analyzed:
=0.25, a lower interval characterizing the power distribution, which can be used to evaluate the base load in mild weather;
=0.75, the higher interval of the characteristic electric quantity distribution corresponds to the electricity utilization increase when the common high temperature or humidity rises;
When (when) When=0.5 (median), for robust base power estimation:
=0.5 as an intermediate value of the power distribution, insensitive to abnormal values (e.g. holidays, equipment failure), used to calculate the base power, separating the reference line of meteorological effects;
In the present embodiment, since the median of the power distribution is selected to provide a robust baseline load, interference of extreme events to the model results is avoided, the score is selected as ;
S32, training the basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data to obtain a trained basic time-sharing electric quantity decomposition model, and inputting the real-time-sharing electric quantity data into the basic time-sharing electric quantity decomposition model to obtain a real-time basic time-sharing electric quantity component;
S33, obtaining a time-sharing electric quantity residual error based on the basic time-sharing electric quantity component, wherein the time-sharing electric quantity residual error is expressed as the following formula:
;
In the formula, Is thatTime-sharing electric quantity residual error at moment; Is that Real-time-sharing electric quantity at moment;
S4, performing secondary modal decomposition on the time-sharing electric quantity residual by using an improved complete self-adaptive noise set decomposition algorithm ICEEMDAN to obtain a weather-sensitive modal component set, and specifically:
S41, performing first ICEEMDAN decomposition on the bisection electric quantity residual error to obtain An intrinsic mode function and a residual component, expressed as:
;
In the formula, For the first decomposition to obtainA natural mode function; Residual components obtained by the first decomposition;
S42, selecting all natural mode functions obtained by the first decomposition to perform the second decomposition, wherein the second decomposition is expressed as follows:
;
In the formula, Representation pairThe second decomposition is carried out to obtain the firstThe function of the individual natural modes,The total number of the intrinsic mode functions obtained by the second decomposition is the total number of the intrinsic mode functions; Residual components obtained by the second decomposition;
s43, combining all the natural mode functions obtained by the second decomposition to obtain a weather sensitive mode component set Expressed as:
;
S5, constructing a MOEA/D multi-target optimization model based on a decomposition multi-target evolutionary algorithm, taking a weather sensitive modal component set as input, and solving the MOEA/D multi-target optimization model by using a Chebyshev decomposition method to obtain a corrected weather electric quantity component, wherein the MOEA/D multi-target optimization model carries out weight vector update based on a key weather factor set;
S51, in the embodiment, two targets are set, including minimization of residual fitting errors and maximization of consistency of weights and sensitivity coefficients, and the built MOEA/D multi-target optimization model is expressed as follows:
;
In the formula, Minimizing a target for residual fitting errors; Is the first Individual meteorological factors are atA nonlinear mapping function of time; the length of the real-time-sharing electric quantity signal is; Is the first The weight coefficient of each meteorological factor; maximizing a target for consistency of the weight and the sensitivity coefficient; is the total number of meteorological factors;
The method comprises the steps of obtaining a data driving model, wherein the data driving model comprises a residual error fitting error, a weight and sensitivity coefficient consistency, a residual error fitting error and a data driving model, wherein the residual error fitting error is minimized in order to enable a combination of the meteorological factors to be capable of maximally fitting a meteorological sensitive signal in the residual error, the weight and sensitivity coefficient consistency is maximized in order to prevent the optimized weight from deviating from physical significance completely (for example, a meteorological factor with small actual influence is given to an excessively high weight), and the two targets utilize the residual error signal in the data and respect the physical influence mechanism of the meteorological factor, so that the overfitting risk of the pure data driving model is avoided;
S52, solving the MOEA/D multi-objective optimization model by using a Chebyshev decomposition method, wherein the method specifically comprises the following steps:
Definition of Chebyshev distance Expressed as:
;
In the formula, Is a decision variable, used to refer to the firstThe weight coefficient of each meteorological factor; for the multi-objective weighting set, in particular to the present embodiment, since the objective of the present embodiment is two, corresponding ;Indexing for a target; As a set of ideal points, ;As a decision variableA corresponding objective function value;
generating a group of initial weight vector sets which are uniformly distributed according to the key meteorological factor sets, and expressing the initial weight vector sets as follows by a formula:
;
In the formula, Is an initial set of weight vectors; bit index for initial weight vector; for the total number of initial weight vectors, each initial weight vector corresponds to one sub-problem;
calculating weight vectors in an initial set of weight vectors Selecting initial weight vectors with the distances meeting the distance threshold to form corresponding neighborhoods;
Randomly generating an initial population setExpressed as:
;
In the formula, Is corresponding to the initial weight vectorSolution of (2);
From the neighborhood Randomly selecting two solutions, carrying out genetic operation on the two selected solutions to obtain a new solution, and calculating an objective function value of the new solution;
Performing iterative solution, including respectively calculating chebyshev distances of the current solution and the new solution relative to the multi-objective weighing weight and the ideal point for each sub-problem in the neighborhood in each iteration, and if the chebyshev distance corresponding to the new solution is smaller than or equal to the chebyshev distance corresponding to the current solution, replacing the current solution with the new solution;
If the objective function value is smaller than the ideal point, the ideal point is equal to the objective function value;
repeating the iterative solving process until a preset termination condition is met, stopping iteration, wherein the solution in the population is the approximate optimal solution of the MOEA/D multi-objective optimization model, and obtaining the corrected meteorological electric quantity component based on the approximate optimal solution;
S6, superposing the basic time-sharing electric quantity component and the corrected meteorological electric quantity component to obtain a final time-sharing electric quantity decomposition result, wherein the final time-sharing electric quantity decomposition result is expressed as follows by a formula:
;
In the formula, The final time-sharing electric quantity decomposition result; and the corrected meteorological electric quantity component.
In summary, the present embodiment provides a time-sharing electric quantity decomposition method based on sensitivity analysis.
In the embodiments of the present invention, "at least one" means one or more, and "a plurality" means two or more. "and/or", describes an association relation of association objects, and indicates that there may be three kinds of relations, for example, a and/or B, and may indicate that a alone exists, a and B together, and B alone exists. Wherein A, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of the following" and the like means any combination of these items, including any combination of single or plural items. For example, at least one of a, b and c may represent a, b, c, a and b, a and c, b and c, or a and b and c, wherein a, b, c may be single or plural.
Those of ordinary skill in the art will appreciate that the various elements and algorithm steps described in the embodiments disclosed herein can be implemented as a combination of electronic hardware, computer software, and electronic hardware. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the solution. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.
In several embodiments provided by the present invention, any of the functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, comprising several instructions for causing a computer device (which may be a personal computer, a server, a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. The storage medium includes various media capable of storing program codes, such as a U disk, a removable hard disk, a Read-Only Memory (ROM), a random access Memory (Random Access Memory, RAM), a magnetic disk, or an optical disk.
The foregoing description is only illustrative of the present invention and is not intended to limit the scope of the invention, and all equivalent structures or equivalent processes or direct or indirect application in other related technical fields are included in the scope of the present invention.
Claims (8)
1. A time-sharing electric quantity decomposition method based on sensitivity analysis, which is characterized by comprising the following steps:
acquiring time-sharing electric quantity related data of a target area, wherein the time-sharing electric quantity related data comprises historical time-sharing electric quantity data, real-time-sharing electric quantity data and time-sharing meteorological data, and performing gridding processing on the time-sharing electric quantity related data to acquire meteorological factors and a time-sharing electric quantity-meteorological joint data set;
the sensitivity index of each meteorological factor to the time-sharing electric quantity is calculated based on the time-sharing electric quantity-meteorological combined data set, and specifically comprises the following steps:
The global sensitivity corresponding to the time-sharing electric quantity-meteorological joint data set is calculated through a Sobol index, and expressed as follows by a formula:
;
In the formula, Is the firstGlobal sensitivity to individual meteorological factors; Is the first Weather factors; to remove Other meteorological factors; Is the variance; is a conditional expectation; is a time-sharing electric quantity output variable; To give at the first Individual meteorological factorsTime-sharing electric quantity output variable when taking valueRegarding the removal ofOther meteorological factorsIs not limited to the desired one; To expect to Regarding the firstIndividual meteorological factorsSolving variance; Is a time-sharing electric quantity output variable Is a contrast of (3);
Calculating mutual information of meteorological factors and time-sharing electric quantity by using nuclear density estimation Expressed as:
;
In the formula, Is thatAndIs a joint probability density of (2); And Respectively isAndIs a boundary probability density of (1); And Respectively isAndSpecific values of (2);
based on global sensitivity And mutual informationCalculating to obtain a sensitivity index, wherein the sensitivity index is expressed as the following formula:
;
In the formula, Is the firstSensitivity indexes of individual meteorological factors; Is a weight coefficient; information entropy of time-sharing electric quantity;
Screening meteorological factors with sensitivity indexes larger than a sensitivity threshold value to obtain a key meteorological factor set;
A basic time-sharing electric quantity decomposition model is built based on the historical time-sharing electric quantity data, and real-time-sharing electric quantity data is input into the basic time-sharing electric quantity decomposition model to obtain basic time-sharing electric quantity components;
Performing secondary modal decomposition on the time-sharing electric quantity residual by using an improved fully-adaptive noise set decomposition algorithm ICEEMDAN to obtain a weather-sensitive modal component set, wherein the method specifically comprises the following steps of:
Performing first ICEEMDAN decomposition on the time-divided electric quantity residual error to obtain An intrinsic mode function and a residual component, expressed as:
;
In the formula, For the first decomposition to obtainA natural mode function; Residual components obtained by the first decomposition;
and (3) selecting all the natural mode functions obtained by the first decomposition to carry out the second decomposition, wherein the second decomposition is expressed as follows:
;
In the formula, Representation pairThe second decomposition is carried out to obtain the firstThe function of the individual natural modes,The total number of the intrinsic mode functions obtained by the second decomposition is the total number of the intrinsic mode functions; Residual components obtained by the second decomposition;
combining all the natural mode functions obtained by the second decomposition to obtain a weather sensitive mode component set Expressed as:
;
Establishing an MOEA/D multi-target optimization model, taking a weather sensitive modal component set as input, and solving the MOEA/D multi-target optimization model by using a Chebyshev decomposition method to obtain a modified weather electric quantity component, wherein the MOEA/D multi-target optimization model carries out weight vector update based on a key weather factor set;
and superposing the basic time-sharing electric quantity component and the corrected meteorological electric quantity component to obtain a final time-sharing electric quantity decomposition result.
2. The time-sharing power decomposition method based on sensitivity analysis of claim 1, wherein said time-sharing weather data includes temperature, humidity, wind speed, solar intensity and precipitation.
3. The method for time-sharing electric quantity decomposition based on sensitivity analysis according to claim 1, wherein the step of performing gridding processing on the time-sharing electric quantity related data to obtain a meteorological factor and a time-sharing electric quantity-meteorological joint data set specifically comprises:
Carrying out gridding treatment on the time-sharing meteorological data according to the city dimension to obtain gridding meteorological indexes, namely meteorological factors, carrying out gridding treatment on the historical time-sharing electric quantity data and the real-time-sharing electric quantity data according to the city dimension and the industry dimension, and constructing a time-sharing electric quantity-meteorological combined data set by combining the meteorological factors, wherein the gridding resolution is 1km multiplied by 1km;
and carrying out data preprocessing on the time-sharing electric quantity-weather combined data set, wherein the data preprocessing comprises data alignment, outlier rejection and feature standardization.
4. The time-sharing electric quantity decomposition method based on sensitivity analysis of claim 1, wherein the construction of the basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data is specifically as follows:
the basic time-sharing electric quantity decomposition model is a quantile regression model, and is expressed as follows:
;
In the formula, Is thatThe number of time-division digits isIn the training process,Is thatThe number of time-division digits isIs used for monitoring the historical time-sharing electric quantity in real time,Is thatThe number of time-division digits isIs a real-time-sharing electric quantity; And Is thatThe number of time-division digits isRegression coefficients of (a); And Respectively represent the firstMarket and firstIndustry by industry; And Respectively the total number of the ground cities and the total number of the industries; Is that The number of time-division digits isIs the first of (2)A ground variable coefficient; Is that The number of time-division digits isIs the first of (2)Industry variable coefficients for individual industries; Is that The number of time-division digits isError terms of (2); is quantile;
and training the basic time-sharing electric quantity decomposition model based on the historical time-sharing electric quantity data to obtain a trained basic time-sharing electric quantity decomposition model, and inputting the real-time-sharing electric quantity data into the basic time-sharing electric quantity decomposition model to obtain real-time basic time-sharing electric quantity components.
5. The time-sharing power decomposition method according to claim 4, wherein the time-sharing power residual obtained based on the basic time-sharing power component is expressed as:
;
In the formula, Is thatTime-sharing electric quantity residual error at moment; Is that Real-time-sharing electric quantity at moment.
6. The time-sharing electric quantity decomposition method based on sensitivity analysis according to claim 5, wherein the built MOEA/D multi-objective optimization model is expressed as:
;
In the formula, Minimizing a target for residual fitting errors; Is the first Individual meteorological factors are atA nonlinear mapping function of time; the length of the real-time-sharing electric quantity signal is; Is the first The weight coefficient of each meteorological factor; maximizing a target for consistency of the weight and the sensitivity coefficient; Is the total number of meteorological factors.
7. The time-sharing electric quantity decomposition method based on sensitivity analysis according to claim 6, wherein the method for solving the MOEA/D multi-objective optimization model by using chebyshev decomposition method is specifically as follows:
Definition of Chebyshev distance Expressed as:
;
In the formula, Is a decision variable, used to refer to the firstThe weight coefficient of each meteorological factor; For a multi-objective set of trade-off weights, ;Indexing for a target; As a set of ideal points, ;As a decision variableA corresponding objective function value;
generating a group of initial weight vector sets which are uniformly distributed according to the key meteorological factor sets, and expressing the initial weight vector sets as follows by a formula:
;
In the formula, Is an initial set of weight vectors; bit index for initial weight vector; for the total number of initial weight vectors, each initial weight vector corresponds to one sub-problem;
calculating weight vectors in an initial set of weight vectors Selecting initial weight vectors with the distances meeting the distance threshold to form corresponding neighborhoods;
Randomly generating an initial population setExpressed as:
;
In the formula, Is corresponding to the initial weight vectorSolution of (2);
From the neighborhood Randomly selecting two solutions, carrying out genetic operation on the two selected solutions to obtain a new solution, and calculating an objective function value of the new solution;
Performing iterative solution, including respectively calculating chebyshev distances of the current solution and the new solution relative to the multi-objective weighing weight and the ideal point for each sub-problem in the neighborhood in each iteration, and if the chebyshev distance corresponding to the new solution is smaller than or equal to the chebyshev distance corresponding to the current solution, replacing the current solution with the new solution;
If the objective function value is smaller than the ideal point, the ideal point is equal to the objective function value;
and repeating the iterative solving process until a preset termination condition is met, stopping iteration, wherein the solution in the population is the approximate optimal solution of the MOEA/D multi-objective optimization model, and obtaining the corrected meteorological electric quantity component based on the approximate optimal solution.
8. The time-sharing electric quantity decomposition method based on sensitivity analysis of claim 7, wherein the final time-sharing electric quantity decomposition result is expressed as:
;
In the formula, The final time-sharing electric quantity decomposition result; and the corrected meteorological electric quantity component.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510315935.7A CN119848801B (en) | 2025-03-18 | 2025-03-18 | A time-sharing electricity decomposition method based on sensitivity analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202510315935.7A CN119848801B (en) | 2025-03-18 | 2025-03-18 | A time-sharing electricity decomposition method based on sensitivity analysis |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN119848801A CN119848801A (en) | 2025-04-18 |
| CN119848801B true CN119848801B (en) | 2025-05-27 |
Family
ID=95356389
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202510315935.7A Active CN119848801B (en) | 2025-03-18 | 2025-03-18 | A time-sharing electricity decomposition method based on sensitivity analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN119848801B (en) |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109299814A (en) * | 2018-08-30 | 2019-02-01 | 国网江苏电力设计咨询有限公司 | A Weather-Influenced Electricity Decomposition Prediction Method |
| CN110543929A (en) * | 2019-08-29 | 2019-12-06 | 华北电力大学（保定） | A Wind Speed Interval Prediction Method and System Based on Lorenz System |
Family Cites Families (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP7190846B2 (en) * | 2018-08-27 | 2022-12-16 | 中部電力株式会社 | Power Consumption Usage Decomposition System and Program |
| CN115994605A (en) * | 2022-10-08 | 2023-04-21 | 三峡大学 | Multi-data fusion photovoltaic power prediction algorithm for comprehensive meteorological factor data |
| CN117239842A (en) * | 2023-09-18 | 2023-12-15 | 河北工业大学 | Flexible substation distribution network optimization dispatching method based on MOEA/D adaptively modified weight vector |
| CN117290712A (en) * | 2023-10-10 | 2023-12-26 | 北京林业大学 | Statistical downscaling method for frequency decomposition of meteorological input data based on EEMD method |
| CN117556692A (en) * | 2023-11-08 | 2024-02-13 | 南昌大学 | A surrogate model-assisted MOEA/D method for multi-objective expensive constrained optimization problems |
| CN117728386A (en) * | 2023-11-15 | 2024-03-19 | 北京理工大学 | A prediction method for renewable energy power generation in green buildings |
| CN118194602A (en) * | 2024-04-28 | 2024-06-14 | 上海工程技术大学 | A multi-objective evolutionary energy system design method based on adaptive weight adjustment |
| CN118676909A (en) * | 2024-06-18 | 2024-09-20 | 山东大学 | Deep learning power load prediction method, medium and equipment for adaptive modal decomposition |
-
2025
- 2025-03-18 CN CN202510315935.7A patent/CN119848801B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109299814A (en) * | 2018-08-30 | 2019-02-01 | 国网江苏电力设计咨询有限公司 | A Weather-Influenced Electricity Decomposition Prediction Method |
| CN110543929A (en) * | 2019-08-29 | 2019-12-06 | 华北电力大学（保定） | A Wind Speed Interval Prediction Method and System Based on Lorenz System |
Also Published As
| Publication number | Publication date |
|---|---|
| CN119848801A (en) | 2025-04-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Dong et al. | Hourly energy consumption prediction of an office building based on ensemble learning and energy consumption pattern classification | |
| Moody | Economic forecasting: Challenges and neural network solutions | |
| Guo et al. | Prediction of global wheat cultivation distribution under climate change and socioeconomic development | |
| WO2015150747A1 (en) | A computer implemented method of deriving performance from a local model | |
| Hong et al. | Locational marginal price forecasting using deep learning network optimized by mapping-based genetic algorithm | |
| CN119918023B (en) | Tobacco number prediction method based on deep learning | |
| CN116611588B (en) | Precipitation multi-driving factor segmented calibrated optimization forecast method and system | |
| CN113537645B (en) | Soybean yield prediction method based on machine learning fusion satellite and weather data | |
| CN115641175A (en) | Medium-and-long-term power transaction assistant decision-making determination method and device for new energy power station | |
| CN117909888B (en) | Intelligent artificial intelligence climate prediction method | |
| WO2023245399A1 (en) | Rice production potential simulation method based on land system and climate change coupling | |
| CN120124819A (en) | Dynamic prediction method of carbon sink in ecological restoration area based on time series remote sensing | |
| Beyer et al. | Deep spatial-temporal graph modeling for efficient NDVI forecasting | |
| CN120655450A (en) | Intelligent farmland soil humidity monitoring system based on Internet of things | |
| Ma et al. | Short-Term Household Load Forecasting Based on Attention Mechanism and CNN-ICPSO-LSTM. | |
| CN117973916A (en) | A method for calculating and managing optimal allocation of carbon quotas | |
| CN120409837B (en) | Agricultural production area monitoring big data analysis method and system | |
| CN119848801B (en) | A time-sharing electricity decomposition method based on sensitivity analysis | |
| Minsan et al. | Enhancing decomposition and Holt-Winters weekly forecasting of pm2. 5 concentrations in Thailand’s eight northern provinces using the cuckoo search algorithm | |
| CN120410595A (en) | A method and system for predicting electricity spot prices based on similar day probability correction | |
| CN113792480A (en) | An early warning method for a comprehensive early warning system of extreme climate and compound extreme climate | |
| CN119863126A (en) | Agricultural financing guarantee-oriented wind control model based on intelligent insurance multidimensional dynamic scoring | |
| Feng et al. | Comparative Analysis of Risk Factor Weighting GBDT Methods for Enhancing the Accuracy of Flood Risk Assessment | |
| CN119787294A (en) | Photovoltaic output combination prediction method based on error reciprocal method | |
| Wijayaningrum et al. | Advanced Predictive Analytics for Agricultural Weather Forecasting Using Machine Learning |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |