Power load prediction method, device, equipment and medium

Through the variational modal decomposition method and the dung beetle algorithm, the power load prediction model is optimized, and the traditional short-term power load prediction accuracy is solved, achieving higher prediction accuracy and accuracy.

CN119070272APending Publication Date: 2024-12-03NANJING UNIV OF POSTS & TELECOMM
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202410987117.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

Traditional short-term power load prediction is difficult to achieve high accuracy, especially under the structure of diversified power system, it is difficult to accurately predict future power loads.

Method used

The variational modal decomposition method is used to perform quadratic decomposition on high-frequency eigenmodal function components, and the decomposition number is determined through the dung beetle algorithm, and combined with Piecewise chaotic mapping, triangle walking strategy and cross-crossing strategy to optimize the power load prediction model to reduce the High frequency data complexity and improve prediction accuracy.

Benefits of technology

The prediction accuracy of complex power load sequences and the accuracy of power load prediction models are significantly improved, and the complexity of high-frequency data is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119070272A_ABST
    Figure CN119070272A_ABST
Patent Text Reader

Abstract

The invention discloses a power load prediction method, device and equipment and a medium, and belongs to the technical field of power load prediction. The method comprises the following steps: decomposing power load data to obtain a first intrinsic mode function component; dividing the first intrinsic mode component into a high-frequency component, a low-frequency component and a trend component; performing secondary decomposition on the first high-frequency intrinsic mode function component based on a variational mode decomposition method, determining a decomposition number through a dung beetle algorithm, and obtaining a second intrinsic mode function component; and inputting the first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component into a pre-established power load prediction model to obtain a power load prediction result. According to the method, through secondary decomposition of the high-frequency power load data, the complexity and nonlinearity of the original data are reduced, and the prediction precision of the power load sequence with relatively high complexity is improved; and the accuracy of the prediction model is further improved by introducing the dung beetle algorithm.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention belongs to the technical field of power load forecasting, and relates to a power load forecasting method, device, equipment and medium. Background Art

[0002] Power load forecasting is one of the important tasks of the power sector. Accurate load forecasting can economically and reasonably arrange the start and stop of power generation units within the power grid, maintain the safety and stability of power grid operation, reduce unnecessary investment, and improve economic and social benefits. In addition to considering the historical data of power load, power load forecasting must also formulate appropriate analysis plans based on various uncertain factors such as relative temperature, relative humidity, rainfall, temperature difference and average temperature at that time. The forecasting work is to comprehensively analyze various influencing factors to predict the power load in a specific time period in the future, so as to formulate appropriate deployment work and ensure that the power dispatching department works in an orderly manner.

[0003] Power load forecasting is divided into short-term, medium-term and long-term according to the length of the forecast time. Among them, short-term power load forecasting is an important basis for the start and stop, scheduling and future power generation plan of generator sets. Faced with the current diversified power system structure, traditional short-term power load forecasting is difficult to achieve prediction accuracy, so a more accurate prediction model is needed. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies in the prior art and to provide a method, device, equipment and medium for predicting power load, so as to improve the accuracy of short-term power load prediction.

[0005] To achieve the above object, the present invention is implemented by adopting the following technical solutions:

[0006] In a first aspect, the present invention provides a method for predicting power load, comprising:

[0007] Obtain power load data;

[0008] Decomposing the power load data to obtain a plurality of first intrinsic mode function components;

[0009] By calculating the sample entropy of the first eigenmode component, the plurality of first eigenmode components are divided into a first high-frequency eigenmode function component, a first low-frequency eigenmode function component and a first trend eigenmode function component;

[0010] Based on the variational mode decomposition method, the first high-frequency intrinsic mode function component is secondary decomposed, and the decomposition quantity is determined by the dung beetle algorithm to obtain the second intrinsic mode function component;

[0011] The first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component are input into a pre-established power load prediction model to obtain a power load prediction result.

[0012] Furthermore, it also includes: preprocessing the power load data, including: filling in missing data and correcting abnormal data; and arranging the power load data in chronological order.

[0013] Furthermore, the power load data is decomposed to obtain a plurality of first intrinsic mode function components, including: decomposing the power load data based on an adaptive noise complete set empirical mode decomposition method to obtain a plurality of first intrinsic mode function components.

[0014] Furthermore, the number of decompositions is determined by a dung beetle algorithm, including: globally optimizing the penalty factor and the number of decompositions of the variational mode decomposition method by using the dung beetle algorithm.

[0015] Furthermore, a Piecewise chaotic map is introduced to initialize the position of the dung beetle population in the dung beetle algorithm, including:

[0016] ,

[0017] in, Indicates the first The initial positions of dung beetles, wherein the dung beetle population includes ball-rolling dung beetles, brooding balls, small dung beetles and thieving dung beetles; Indicated by The chaotic sequence value generated is the value of the first The initial position of each dung beetle; Represents the preset chaotic mapping parameter, which is a random number uniformly distributed in the range of (0, 1).

[0018] Furthermore, the method further includes: after the position of the dung beetle population is updated, the position of the dung beetle population is updated again by using a triangle wandering strategy, including:

[0019] Get the distance between the dung beetle population location and the optimal location , and the range of walking steps of dung beetle populations , and The calculation formula is as follows:

[0020] ,

[0021] ,

[0022] in express The updated position of the dung beetle population at each moment, express The local best position at the moment, Indicates generating a random number;

[0023] Defines the direction the dung beetle walks :

[0024] ,

[0025] The second updated position of the dung beetle population obtained through the triangle walking strategy The calculation formula is:

[0026]

[0027] in, Represents a random number uniformly distributed in the range (0, 1);

[0028] And disturbing the entire dung beetle population through cross-strategies, including:

[0029] The individuals of the dung beetle population are randomly paired, and the two paired individuals of the dung beetle population are horizontally crossed, including:

[0030] and represents the randomly paired parent individuals of the dung beetle population, and Indicates the and The offspring individuals of the dung beetle population, and The calculation formula for the offspring individuals generated by horizontal crossover in the jth dimension is:

[0031] ,

[0032] ,

[0033] in, and Indicates the and The offspring individuals produced by horizontal crossover in the jth dimension; and Indicates the and The jth dimension of and Represents a random number uniformly distributed in the range (0, 1); and It means a random number uniformly distributed in the range of (-1, 1);

[0034] The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population;

[0035] A vertical crossover of dung beetle population individuals was conducted, including:

[0036] Randomly select the parent individuals of the dung beetle population The two dimensions d1 and d2 are vertically crossed, and their offspring individuals The calculation formula for the d1th dimension is:

[0037] ,

[0038] in, express The d1th dimension of express The d2th dimension of Represents a random number uniformly distributed in the range (0, 1);

[0039] Other dimensions of the offspring individuals Stay consistent;

[0040] The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population.

[0041] Furthermore, during the training phase of the power load forecasting model, the learning rate, the number of neurons in the hidden layer and the number of training times of the power load forecasting model are optimized by using the dung beetle algorithm.

[0042] In a second aspect, the present invention further provides a power load forecasting device, the device comprising:

[0043] A power load data acquisition module, used to acquire power load data;

[0044] A first intrinsic mode function component acquisition module, used for decomposing the power load data to obtain a plurality of first intrinsic mode function components;

[0045] A first intrinsic mode function component division module, configured to divide the plurality of first intrinsic mode function components into a first high-frequency intrinsic mode function component, a first low-frequency intrinsic mode function component and a first trend intrinsic mode function component by calculating the sample entropy of the first intrinsic mode function component;

[0046] A second intrinsic mode function component acquisition module is used to perform secondary decomposition on the first high-frequency intrinsic mode function component based on the variational mode decomposition method, and determine the decomposition quantity through the dung beetle algorithm to obtain the second intrinsic mode function component;

[0047] The power load forecast result acquisition module is used to input the first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component into a pre-established power load forecast model to obtain the power load forecast result for obtaining power load data.

[0048] In a third aspect, the present invention further provides a computer device, comprising:

[0049] Memory for storing computer programs;

[0050] A processor is used to execute the computer program to implement the steps of the above-mentioned power load forecasting method.

[0051] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, the program being executed by a processor to implement the steps of the above-mentioned power load forecasting method.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The power load forecasting method provided by the present invention performs secondary decomposition of the first high-frequency intrinsic mode function component by the variational mode decomposition method to reduce the complexity of high-frequency data; the penalty factor and the number of decompositions of the variational mode decomposition method are globally optimized by the dung beetle algorithm, and the Piecewise chaotic mapping, triangle walk strategy and vertical and horizontal cross strategy are introduced to optimize the dung beetle algorithm, thereby improving the prediction accuracy of the original power load sequence with high complexity. In addition, the dung beetle algorithm is introduced to optimize the hyperparameters of the power load forecasting model network, thereby further improving the accuracy of the power load forecasting model. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] Figure 1 A schematic diagram of a flow chart of a power load forecasting method provided by an embodiment of the present invention;

[0055] Figure 2 A schematic diagram of the structure of a power load forecasting method provided by an embodiment of the present invention;

[0056] Figure 3 It is the first intrinsic mode function component result obtained by the adaptive noise complete set empirical mode decomposition method in an embodiment of the present invention;

[0057] Figure 4 It is a schematic diagram of the iterative process of the dung beetle algorithm;

[0058] Figure 5 is the result of the second intrinsic mode function component obtained by the variational mode decomposition method in an embodiment of the present invention;

[0059] Figure 6It is a schematic diagram of the structure of the long short-term memory neural network model;

[0060] Figure 7 The power load prediction result obtained by the power load prediction method in the embodiment of the present invention;

[0061] Figure 8 A schematic diagram of the structure of a power load prediction device provided by an embodiment of the present invention;

[0062] Fig. 9 An internal structure diagram of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The technical solution of the present invention is described in detail below through the accompanying drawings and specific embodiments. The same reference numerals in the accompanying drawings indicate the same or similar components or parts. It should be understood by those skilled in the art that these drawings are not necessarily drawn to scale. The embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations on the technical solution of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0064] The term "and / or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0065] Embodiment 1:

[0066] like Figures 1 to 7 As shown, an embodiment of the present invention provides a method for predicting power load. Figure 1 The flowchart of the power load forecasting method is shown in FIG. 1 . This flowchart only shows the logical sequence of the method described in this embodiment. In other possible embodiments of the present invention, different methods may be used without conflict. Figure 1 The steps shown or described are accomplished in the order shown.

[0067] The circuit board defect detection method provided in this embodiment can be applied to a terminal, and can be executed by a circuit board defect detection device, which can be implemented by software and / or hardware, and the device can be integrated in the terminal.

[0068] See also Figure 1 The method of the embodiment of the present invention specifically includes the following steps 1 to 5. Among them:

[0069] Step 1: Obtain power load data.

[0070] Among them, the acquired power load data is a time series.

[0071] In addition, the acquired power load data needs to be preprocessed, including: for missing data, the embodiment of the present invention fills the missing position data by adopting the average value of the first two data and the last two data of the data; and correcting the abnormal data.

[0072] Step 2: Decompose the power load data to obtain a plurality of first intrinsic mode function components.

[0073] like Figure 2 As shown, in an embodiment of the present invention, the pre-processed power load data is decomposed by Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) to obtain multiple first intrinsic mode function components (IMF).

[0074] The adaptive noise complete ensemble empirical mode decomposition method is an improvement on the empirical mode decomposition (EMD) method, and borrows the idea of ​​adding Gaussian noise and canceling the noise by multiple superposition and averaging in the ensemble empirical mode decomposition (EEMD).

[0075] The specific steps of decomposing the power load data by using the adaptive noise complete set empirical mode decomposition method to obtain multiple first intrinsic mode function components are:

[0076] Step 201: Add Gaussian white noise to the pre-processed power load data to generate a new data sequence;

[0077] Step 202: Decompose the new data sequence by using the empirical mode decomposition method to obtain a first intrinsic mode function component (IMF) and a residual;

[0078] Step 23: Add Gaussian white noise to the obtained residual and decompose it by empirical mode decomposition to obtain a first intrinsic mode function component and a residual;

[0079] Step 24: Repeat the idea of ​​step 23 until the decomposed residual is a monotonic function that cannot be further decomposed. The iteration ends and several first intrinsic mode function components are finally obtained.

[0080] Figure 3It shows the results of multiple first intrinsic mode function components obtained by the adaptive noise complete set empirical mode decomposition method in an embodiment of the present invention.

[0081] Step 3: dividing the multiple first eigenmode components into a first high-frequency eigenmode function component, a first low-frequency eigenmode function component and a first trend eigenmode function component by calculating the sample entropy of the first eigenmode component.

[0082] Among them, sample entropy can evaluate the system complexity of time series data. Generally speaking, the larger the sample entropy of a sequence, the more complex the sequence is. Integrating intrinsic mode function components with similar sample entropy values ​​can appropriately reduce the amount of calculation and reduce overfitting.

[0083] The plurality of first eigenmode components are divided into a first high-frequency eigenmode function component, a first low-frequency eigenmode function component and a first trend eigenmode function component in descending order of sample entropy of the first eigenmode components.

[0084] Step 4: Based on the variational mode decomposition method, the first high-frequency intrinsic mode function component is decomposed twice, and the decomposition quantity is determined by the dung beetle algorithm to obtain the second intrinsic mode function component.

[0085] Back to Figure 2 In order to reduce the complexity of the first high-frequency intrinsic mode function component, the present invention uses the variational mode decomposition (VMD) method to perform a secondary decomposition on it.

[0086] Specifically, the first high-frequency intrinsic mode function component is secondary decomposed based on the variational mode decomposition method, and its decomposition model includes:

[0087] { min { z k },{ v k } { ∑ k = 1 K ‖ ∂ t { [ [ δ ( t ) + j π t ] z k ( t ) } e − jv k t ‖ 2 2 } s . t . ∑ k = 1 K z k = f ( t ) ,

[0088] in, represents a set of the second high frequency eigenmode function components; represents the number of the second high frequency eigenmode function component; express Moment The second eigenmode function component; A set of center frequencies representing the second high frequency eigenmode function components; express The original input signal at time t, i.e., the first high-frequency intrinsic mode function component; represents the impulse function; Indicates the moment Find the differential; represents a second normal form function;

[0089] A quadratic penalty factor and a Lagrange multiplication operator are introduced to convert the constrained variational problem of the quadratic decomposition model into an unconstrained variational problem;

[0090] The secondary decomposition model can be equivalent to:

[0091] ,

[0092] in, represents the quadratic penalty factor, represents the inner product, express Lagrange multiplication operator of the moment;

[0093] Through alternating updating calculation, the second intrinsic mode function component after secondary decomposition and its corresponding center frequency are obtained;

[0094] ,

[0095] ,

[0096] in, Indicates The first iteration The second eigenmode function component; Indicates that the data is from the time domain Transform to frequency domain; express Frequency domain function of express Frequency domain function of Indicates The second eigenmode function component; Indicates The iteration The center frequency of the second eigenmode function component.

[0097] The present invention uses the dung beetle algorithm to and the number of the second high frequency eigenmode function components Perform global optimization.

[0098] Dung Beetle Optimizer (DBO) is a swarm intelligence optimization algorithm. Its iterative process is as follows: Figure 5 The specific steps are as follows:

[0099] Step 41: Initialize the parameters of the dung beetle algorithm and the dung beetle population;

[0100] The present invention initializes the position of the dung beetle population of the dung beetle algorithm by introducing Piecewise chaotic mapping.

[0101] ,

[0102] in, Indicates the first The initial position of dung beetles, wherein the dung beetle population includes four sub-populations, namely, rolling ball dung beetles, brooding balls, small dung beetles and stealing dung beetles; Indicated by The chaotic sequence value generated is the value of the first The initial position of each dung beetle; represents the preset chaotic mapping parameter, which is a random number uniformly distributed in the range of (0, 1); the initial position of the first dung beetle is a random number in the range of (0, 1).

[0103] Step 42: Establish a fitness calculation function for variational mode decomposition, using the average value of the sample entropy of the second eigenmode function component as the fitness function, with the goal of finding a set of quadratic penalty factors and the number of modal component decompositions to minimize the average value of the sample entropy of each modal component after decomposition;

[0104] Step 43: by calculating the fitness value, the positions of all dung beetles are updated, wherein the dung beetle population is divided into four sub-populations, namely, rolling ball dung beetles, brooding balls, small dung beetles, and stealing dung beetles;

[0105] Specifically, when there are no obstacles on the route, the rolling dung beetle uses the sun as a navigation guide and rolls. The calculation formula is as follows:

[0106] ,

[0107] ,

[0108] in: Indicates Dung beetle The position information in the iteration, represents the disturbance coefficient, Represents a random number between 0 and 1. Indicates the worst position, Indicates the intensity of light, Takes -1 or 1.

[0109] When the ball-rolling dung beetle encounters an obstacle, it adjusts its position by dancing. The calculation formula is as follows:

[0110]

[0111] in: Indicates Dung beetle The position information in the iteration, represents the deflection angle, When it is 0, π / 2 or π, the position of the dung beetle is not updated;

[0112] The brooding ball simulates the egg-laying position of the female dung beetle through the boundary selection strategy, and dynamically updates the position during the iteration process. The calculation formula is as follows:

[0113] ,

[0114] in: Indicates The brooding ball is in The position information in the iteration, represents the current local best position, and is a one-dimensional random vector; and Indicates the upper and lower limits of the spawning location;

[0115] The dung beetle needs to find an optimal foraging area and forage for food, and update its position during the foraging process. The calculation formula is as follows:

[0116] ,

[0117] in, Indicates A little dung beetle is in the The position information in the iteration, is a random number that follows a normal distribution, is a one-dimensional random vector belonging to (0,1), and Indicates the upper and lower limits of the optimal foraging area.

[0118] A thieving dung beetle will steal dung balls from other dung beetles. The position of a thieving dung beetle is calculated as follows:

[0119] ,

[0120] in, Indicates A thieving dung beetle The position information in the iteration, is a one-dimensional random vector that follows a normal distribution, is a constant.

[0121] After the position of the dung beetle population is updated, the present invention updates the position of the dung beetle population for a second time through a triangle walking strategy, and the specific steps are as follows:

[0122] Get the distance between the dung beetle population location and the optimal location , and the range of walking steps of dung beetle populations , and The calculation formula is as follows:

[0123] ,

[0124] ,

[0125] in express The location of the dung beetle population at the time, express The local best position at the moment, Indicates generating a random number;

[0126] Defines the direction the dung beetle walks :

[0127] ,

[0128] The second updated position of the dung beetle population after walking through the triangle The calculation formula is:

[0129]

[0130] in, Represents a random number uniformly distributed in the range (0, 1);

[0131] In addition, the present invention disturbs the entire dung beetle population through a vertical and horizontal cross strategy, including:

[0132] The individuals of the dung beetle population are randomly paired, and the two paired individuals of the dung beetle population are horizontally crossed, including:

[0133] and represents the randomly paired parent individuals of the dung beetle population, and Indicates the and The offspring individuals of the dung beetle population, and The calculation formula for the offspring individuals generated by horizontal crossover in the jth dimension is:

[0134] ,

[0135] ,

[0136] in, and Indicates the and The offspring individuals produced by horizontal crossover in the jth dimension; and Indicates the and The jth dimension of and Represents a random number uniformly distributed in the range (0, 1); and It means a random number uniformly distributed in the range of (-1, 1);

[0137] The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population;

[0138] A vertical crossover of dung beetle population individuals was conducted, including:

[0139] Randomly select the parent individuals of the dung beetle population The two dimensions d1 and d2 are vertically crossed, and their offspring individuals The calculation formula for the d1th dimension is:

[0140] ,

[0141] in, express The d1th dimension of express The d2th dimension of Represents a random number uniformly distributed in the range (0, 1);

[0142] Other dimensions of the offspring individuals Stay consistent;

[0143] The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population.

[0144] In order not to increase the complexity of the algorithm, the embodiment of the present invention introduces a vertical and horizontal cross strategy factor ,when Greater than When , we adopt the cross-cutting strategy. Less than When , this iteration skips the vertical and horizontal cross strategy. It is expressed as:

[0145] ,

[0146] in, Indicates the current iteration number, Indicates the total number of iterations.

[0147] Step 44: Establish a boundary checking function to determine whether each dung beetle is beyond the boundary.

[0148] Step 45: Update the current optimal solution and fitness value.

[0149] Step 46: Repeat the above steps until the maximum number of iterations is reached.

[0150] Figure 5 It shows the results of multiple second eigenmode function components obtained by the variational mode decomposition method in an embodiment of the present invention.

[0151] Step 5: Input the first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component into a pre-established power load prediction model to obtain a power load prediction result.

[0152] In the embodiment of the present invention, the power load prediction model uses a long short-term memory neural network model (Long Short-Term Memory, LSTM), which is a time-recurrent neural network. Figure 6 The structure of the long short-term memory neural network model is shown, which consists of three control gates: forget gate, input gate and output gate. express The model input at time express The hidden layer state at any moment, For update Remember cell status at all times.

[0153] In the training stage of the power load forecasting model, the present invention also optimizes the learning rate, the number of hidden layer neurons and the number of training times of the long short-term memory neural network through the above-mentioned dung beetle algorithm. By establishing a fitness calculation function for power load forecasting, the mean square error between the power load forecast value and the true value of the verification set is used as the fitness function, with the aim of finding a set of hyperparameters that minimizes the error of the power load forecasting network.

[0154] The absolute percentage error MAPE, root mean square error RMSE, mean absolute error MAE and goodness of fit are used. The power load forecasting model is evaluated and the calculation formula is as follows:

[0155] ,

[0156] ,

[0157] ,

[0158] ,

[0159] in, and are the true value and predicted value of the test set, respectively. is the mean value of the test set.

[0160] The first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component are sequentially input into the trained power load prediction model, and the prediction results of each intrinsic mode function component are superimposed and reconstructed to obtain the final power load prediction result. Figure 7 The comparison results between the predicted values ​​and the true values ​​of the power load forecasting model at different time steps are shown.

[0161] Embodiment 2:

[0162] Based on the same inventive concept as Example 1, the embodiment of the present invention also provides a power load prediction device for implementing the above-mentioned image restoration method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above-mentioned method, so the specific limitations in the power load prediction device embodiment provided below can refer to the limitations of the image restoration method above, and will not be repeated here.

[0163] like Figure 8 As shown, an embodiment of the present invention provides a power load prediction device, comprising:

[0164] A power load data acquisition module, used to acquire power load data;

[0165] A first intrinsic mode function component acquisition module, used for decomposing the power load data to obtain a plurality of first intrinsic mode function components;

[0166] A first intrinsic mode function component division module, configured to divide the plurality of first intrinsic mode function components into a first high-frequency intrinsic mode function component, a first low-frequency intrinsic mode function component and a first trend intrinsic mode function component by calculating the sample entropy of the first intrinsic mode function component;

[0167] A second intrinsic mode function component acquisition module is used to perform secondary decomposition on the first high-frequency intrinsic mode function component based on the variational mode decomposition method, and determine the decomposition quantity through the dung beetle algorithm to obtain the second intrinsic mode function component;

[0168] The power load forecast result acquisition module is used to input the first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component into a pre-established power load forecast model to obtain the power load forecast result for obtaining power load data.

[0169] Embodiment 3:

[0170] The embodiment of the present invention further provides a computer device, which may be a server, and its internal structure diagram may be as shown in FIG. Fig. 9 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, the power load forecasting method in the aforementioned embodiment is implemented.

[0171] Those skilled in the art will understand that Fig. 9 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0172] Embodiment 4:

[0173] The embodiment of the present invention further provides a computer-readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the steps of the following method are implemented:

[0174] Obtain power load data;

[0175] Decomposing the power load data to obtain a plurality of first intrinsic mode function components;

[0176] By calculating the sample entropy of the first eigenmode component, the plurality of first eigenmode components are divided into a first high-frequency eigenmode function component, a first low-frequency eigenmode function component and a first trend eigenmode function component;

[0177] Based on the variational mode decomposition method, the first high-frequency intrinsic mode function component is secondary decomposed, and the decomposition quantity is determined by the dung beetle algorithm to obtain the second intrinsic mode function component;

[0178] The first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component are input into a pre-established power load prediction model to obtain a power load prediction result.

[0179] It should be understood by those skilled in the art that the embodiments of the present invention may be provided as methods, systems, or computer program products, and therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

[0180] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0181] These computer program instructions may also be stored in a computer readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture including an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0182] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0183] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the enlightenment of the present invention, ordinary technicians in this field can also make many forms without departing from the scope of protection of the purpose of the present invention and the claims, which all fall within the protection of the present invention.

Claims

1. A method for predicting power load, characterized in that: include: Obtain power load data; Decomposing the power load data to obtain a plurality of first intrinsic mode function components; By calculating the sample entropy of the first eigenmode component, the plurality of first eigenmode components are divided into a first high-frequency eigenmode function component, a first low-frequency eigenmode function component and a first trend eigenmode function component; Based on the variational mode decomposition method, the first high-frequency intrinsic mode function component is secondary decomposed, and the decomposition quantity is determined by the dung beetle algorithm to obtain the second intrinsic mode function component; The first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component are input into a pre-established power load prediction model to obtain a power load prediction result.

2. The power load forecasting method according to claim 1, characterized in that: Also includes: Preprocessing the power load data includes: filling in missing data and correcting abnormal data; and arranging the power load data in chronological order.

3. The power load forecasting method according to claim 1, characterized in that: Decomposing the power load data to obtain a plurality of first intrinsic mode function components includes: decomposing the power load data based on an adaptive noise complete set empirical mode decomposition method to obtain a plurality of first intrinsic mode function components.

4. The power load forecasting method according to claim 1, characterized in that: Determining the number of decompositions by using the dung beetle algorithm includes: performing global optimization on the penalty factor and the number of decompositions of the variational mode decomposition method by using the dung beetle algorithm.

5. The power load forecasting method according to claim 4, characterized in that: The Piecewise chaotic map is introduced to initialize the position of the dung beetle population in the dung beetle algorithm, including: , in, Indicates the first The initial positions of dung beetles, wherein the dung beetle population includes ball-rolling dung beetles, brooding balls, small dung beetles and thieving dung beetles; Indicated by The chaotic sequence value generated is the value of the first The initial position of each dung beetle; Represents the preset chaotic mapping parameter, which is a random number uniformly distributed in the range of (0, 1).

6. The power load forecasting method according to claim 5, characterized in that: Also includes: After the position of the dung beetle population is updated, the position of the dung beetle population is updated again through the triangle wandering strategy, including: Get the distance between the dung beetle population location and the optimal location , and the range of walking steps of dung beetle populations , and The calculation formula is as follows: , , in express The updated position of the dung beetle population at each moment, express The local best position at the moment, Represents a random number; Defines the direction the dung beetle walks : , The second updated position of the dung beetle population obtained through the triangle wandering strategy The calculation formula is: , in, Represents a random number uniformly distributed in the range (0, 1); And disturbing the entire dung beetle population through cross-strategies, including: The individuals of the dung beetle population are randomly paired, and the two paired individuals of the dung beetle population are horizontally crossed, including: and represents the randomly paired parent individuals of the dung beetle population, and Indicates the and The offspring individuals of the dung beetle population, and The calculation formula for the offspring individuals generated by horizontal crossover in the jth dimension is: , , in, and Indicates the and The offspring individuals produced by horizontal crossover in the jth dimension; and Indicates the and The jth dimension of and Represents a random number uniformly distributed in the range (0, 1); and It means a random number uniformly distributed in the range of (-1, 1); The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population; A vertical crossover of dung beetle population individuals was conducted, including: Randomly select the parent individuals of the dung beetle population The two dimensions d1 and d2 are vertically crossed, and their offspring individuals The calculation formula for the d1th dimension is: , in, express The d1th dimension of express The d2th dimension of Represents a random number uniformly distributed in the range (0, 1); Other dimensions of the offspring individuals Stay consistent; The generated offspring individuals of the dung beetle population are compared with their parent individuals, and the individuals with smaller fitness values ​​are retained and added to the dung beetle population.

7. The power load forecasting method according to claim 6, characterized in that: During the training phase of the power load forecasting model, the learning rate, the number of neurons in the hidden layer and the number of training times of the power load forecasting model are optimized by using the dung beetle algorithm.

8. A power load forecasting device, characterized in that: include: A power load data acquisition module, used to acquire power load data; A first intrinsic mode function component acquisition module, used for decomposing the power load data to obtain a plurality of first intrinsic mode function components; A first intrinsic mode function component division module, configured to divide the plurality of first intrinsic mode function components into a first high-frequency intrinsic mode function component, a first low-frequency intrinsic mode function component and a first trend intrinsic mode function component by calculating the sample entropy of the first intrinsic mode function component; A second intrinsic mode function component acquisition module is used to perform secondary decomposition on the first high-frequency intrinsic mode function component based on the variational mode decomposition method, and determine the decomposition quantity through the dung beetle algorithm to obtain the second intrinsic mode function component; The power load forecast result acquisition module is used to input the first low-frequency intrinsic mode function component, the first trend intrinsic mode function component and the second intrinsic mode function component into a pre-established power load forecast model to obtain the power load forecast result for obtaining power load data.

9. A computer device, characterized in that: include: Memory for storing computer programs; A processor, configured to execute the computer program to implement the steps of the power load forecasting method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the power load forecasting method according to any one of claims 1 to 7 are implemented.

Citation Information

Cited By

  • Settlement prediction method and system fused with multi-scale optimization, terminal and storage medium

    CN120338211A

  • Power dispatching method based on prediction model

    CN121119277A

  • Electric vehicle load prediction method, device, equipment and medium

    CN121291215A