An optimization method for intelligent power load prediction model

Abstract

The present invention discloses an optimization method for an intelligent prediction model of electric load, comprising: establishing an intelligent prediction model, setting required optimization parameters and the optimization range of each parameter; initializing the parameters and population of a global sparrow optimization algorithm; iteratively updating the sparrow positions to obtain the optimal parameters; and applying the obtained optimal parameters to the intelligent prediction model for load prediction. The present invention initializes the population using a tent chaotic sequence, which can make the population more evenly distributed in the solution space and increase the global search capability of the population in the early stage; the present invention adds a step-size control factor, which gives the population a larger search range in the early stage of iteration, enhancing the global search capability in the early stage. As the number of iterations increases, the step-size control factor decreases, allowing for better local search and improving the convergence speed; the present invention improves the prediction capability of the model by optimizing the parameters of the intelligent prediction model.

Description

Technical Field

[0001] The present invention relates to the technical field of power load forecasting, and in particular to an optimization method for a power load intelligent forecasting model. Background Art

[0002] Load forecasting is an important part of smart grid systems. With the development of technologies such as machine learning and neural networks, more and more load forecasting methods have emerged. The main problems faced by load forecasting are: (1) prediction accuracy; (2) the ability to process large-scale data; and (3) the ability to process irregular data. With the emergence of optimization algorithms, how to combine intelligent prediction models with optimization algorithms to perform load forecasting has become an important research topic.

[0003] Current load forecasting methods based on traditional machine learning, such as regression analysis and time series methods, suffer from limited learning capabilities and exhibit poor prediction accuracy when processing nonlinear data or data with weak regularities. Subsequently, load forecasting methods based on intelligent methods such as artificial neural networks (ANNs) and support vector machines (SVMs) have become widely used. However, these methods require manual parameter setting, which can lead to low prediction accuracy and make them unsuitable for long-term load forecasting. In recent years, deep learning has become a popular research method in the load forecasting field due to its excellent performance in processing large-scale, high-dimensional data. However, like intelligent methods such as ANNs and SVMs, some parameters in deep learning methods require manual specification, and different parameters have varying impacts on forecast results. While optimization algorithms can be used to find appropriate parameters, these algorithms can suffer from limited search ranges, slow convergence, and a tendency to fall into local optima, resulting in poor optimization results. Summary of the Invention

[0004] The purpose of this section is to summarize some aspects of the embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the abstract and title of this application to avoid obscuring the purpose of this section, the abstract and the title of the invention, and such simplifications or omissions should not be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the technical problem solved by the present invention is: how to improve the intelligent prediction effect of power load.

[0007] To solve the above technical problems, the present invention provides the following technical solutions: establishing an intelligent prediction model, setting the required optimization parameters and the optimization range of each parameter; initializing the parameters and population of the global sparrow optimization algorithm; iteratively updating the sparrow positions to obtain the optimal parameters; and applying the obtained optimal parameters to the intelligent prediction model for load forecasting.

[0008] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the process of initializing the parameters and population of the global sparrow optimization algorithm includes initializing the parameters of the global sparrow optimization algorithm, wherein the parameters of the global sparrow optimization algorithm include the population size, the number of iterations, the ratio of discoverers to scouts, and the warning value; initializing the population according to the improved Tent chaos mapping formula; carrier-transferring the sequence X to the original solution space; calculating the fitness value of each sparrow, and taking the sparrows with the best and worst fitness as the global best and worst individuals.

[0009] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the formula for converting the sequence X carrier to the original solution space includes:

[0010] W j =min j +(max j -min j )×X j

[0011] Among them, min j and max j is the minimum and maximum value of the j-th dimension vector, W is the obtained sparrow position, X j is a j-dimensional sequence, W j is the position of the carrier after arriving at the original solution space.

[0012] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the improved Tent chaos mapping formula is:

[0013]

[0014] Among them, N is the total number of sparrows in the sequence, rand(0,1) is a random number in [0,1], x i is a value in the sequence X, x i+1 To pass the xth i The value of is calculated.

[0015] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, wherein: calculating the fitness value of each sparrow, and taking the sparrows with the best and worst fitness as the global best and worst individuals includes using the position of the i-th sparrow to assign values to the parameters of the deep extreme learning machine; inputting a training set to train the intelligent prediction model; inputting a test set after reaching the number of iterations to obtain a predicted value, and calculating the fitness value of each sparrow by the following formula:

[0016]

[0017] in, with y i are the predicted value and true value of the test set respectively;

[0018] Based on the calculation results, the sparrows with the best and worst fitness are regarded as the global best and worst individuals.

[0019] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the iterative updating of the sparrow positions to obtain the optimal parameters includes increasing the step size control factor; updating the position of the discoverer; updating the position of the follower; updating the position of the scout; calculating the new fitness value of the sparrow according to the fitness value calculation step of each sparrow, and updating the global optimal and worst individuals.

[0020] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the formula for increasing the step size control factor includes:

[0021]

[0022] Where t is the number of iterations.

[0023] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, the formula for updating the position of the discoverer includes:

[0024]

[0025] in, is the position of the i-th sparrow in the t-th iteration and the j-th dimension, α is a random number between (0, 1], R2∈[0, 1] and ST∈[0.5, 1] are the warning value and the safety value respectively, Q is a random number that obeys the normal distribution, and L is a 1×d matrix with each element being 1.

[0026] As a preferred solution of the optimization method of the power load intelligent prediction model described in the present invention, wherein: the updating of the follower position includes:

[0027]

[0028] Among them, x p with x worst are the best position and the global worst position of the finder, n is the population size, A is a 1×d matrix, each element is 1 or -1, and A + =A T (AA T ) -1 .

[0029] As a preferred solution of the optimization method of the power load intelligent prediction model of the present invention, wherein: the updating of the position of the scout includes:

[0030]

[0031] Among them, x best is the global optimal position, β is a normally distributed random number with a mean of 0 and a variance of 1, k is a random number between [-1, 1], and f i is the current sparrow’s fitness value, f g With f w are the current global optimal and worst fitness values, respectively, and ε is the minimum constant.

[0032] As a preferred solution of the optimization method of the electric power load intelligent prediction model described in the present invention, wherein: applying the obtained optimal parameters to the intelligent prediction model for load prediction includes reading the obtained optimal parameters and applying the optimal parameters to the intelligent prediction model; reading the processed training data set to train the intelligent prediction model; and reading the test data set for prediction.

[0033] The beneficial effects of the present invention are as follows: the present invention initializes the population through the Tent chaotic sequence, which can make the population more evenly distributed in the solution space. Compared with the original random population generation method, this method can increase the global search ability of the early population; the present invention adds a step size control factor, so that the population has a larger search range in the early stage of iteration, and enhances the early global search ability. As the number of iterations increases, the step size control factor decreases, and better local search can be performed, and the convergence speed is also improved; the present invention improves the prediction ability of the model by optimizing the parameters of the intelligent prediction model. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. Those skilled in the art can also derive other drawings based on these drawings without inventive effort. Among them:

[0035] Figure 1 A schematic diagram of the basic flow of an optimization method for an intelligent power load prediction model provided by one embodiment of the present invention;

[0036] Figure 2 A schematic diagram showing a performance comparison of a global sparrow optimization algorithm on a unimodal function for an optimization method of an intelligent power load prediction model provided by one embodiment of the present invention;

[0037] Figure 3 A schematic diagram showing a performance comparison of a global sparrow optimization algorithm on a multimodal function for an optimization method of an intelligent power load prediction model provided by one embodiment of the present invention;

[0038] Figure 4 A schematic diagram showing a performance comparison of a global sparrow optimization algorithm of an optimization method for an intelligent power load prediction model provided by one embodiment of the present invention on a multimodal function of a fixed dimension;

[0039] Figure 5 A schematic diagram showing a comparison of prediction results on weekdays of an optimization method for an intelligent power load prediction model provided by one embodiment of the present invention and other models;

[0040] Figure 6 A schematic diagram showing a comparison of prediction results on rest days between the present invention and other models for an optimization method of an intelligent power load prediction model provided by one embodiment of the present invention;

[0041] Figure 7 A schematic diagram showing a comparison of prediction results during holidays between an optimization method for an intelligent power load prediction model provided by the present invention and other models according to an embodiment of the present invention. DETAILED DESCRIPTION

[0042] To make the above-mentioned objects, features, and advantages of the present invention more clearly understood, the following detailed description of the specific embodiments of the present invention is given in conjunction with the accompanying drawings. It is obvious that the described embodiments are only part of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in this field without creative work should fall within the scope of protection of the present invention.

[0043] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0044] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0045] The present invention is described in detail with reference to schematic diagrams. For ease of illustration, cross-sectional views of device structures may be partially enlarged and not to scale when describing embodiments of the present invention. Furthermore, the schematic diagrams are merely illustrative and should not limit the scope of the present invention. Furthermore, in actual production, the three-dimensional dimensions of length, width, and depth should be included.

[0046] In the description of the present invention, it should be noted that the terms "upper, lower, inner, and outer" and other references to orientations or positional relationships are based on the orientations or positional relationships shown in the accompanying drawings and are intended solely to facilitate and simplify the description of the present invention. They are not intended to indicate or imply that the devices or components referred to must have, be constructed, or operate in a specific orientation, and therefore should not be construed as limitations on the present invention. Furthermore, the terms "first, second, or third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0047] In this disclosure, unless otherwise specified or limited, the terms "mounted," "connected," and "connected" should be interpreted broadly. For example, they may refer to fixed, removable, or integral connections. They may also refer to mechanical, electrical, or direct connections, indirect connections through an intermediary, or internal communication between two components. Those skilled in the art will understand the specific meanings of these terms in this disclosure.

[0048] Example 1

[0049] Reference Figure 1 , as one embodiment of the present invention, provides an optimization method for an intelligent power load prediction model, comprising:

[0050] S1: Establish an intelligent prediction model, set the required optimization parameters and the optimization range of each parameter.

[0051] It should be noted that the steps are detailed as follows:

[0052] (1) Establish an intelligent prediction model;

[0053] (2) Set the parameters to be optimized and the optimization range of the parameters.

[0054] S2: Initialize the parameters and population of the global sparrow optimization algorithm.

[0055] It should be noted that the process of initializing the parameters and population of the global sparrow optimization algorithm includes:

[0056] (1) Initializing the global sparrow optimization algorithm parameters, which include population size, number of iterations, ratio of discoverers to scouts, and warning value;

[0057] (2) Initialize the population according to the improved Tent chaotic mapping formula, which is:

[0058]

[0059] Among them, N is the total number of sparrows in the sequence, rand(0,1) is a random number in [0,1], x i is a value in the sequence X, x i+1 To pass the xth i The value of is calculated.

[0060] (3) Carry the sequence X to the original solution space:

[0061] W j =min j +(max j -min j )×X j

[0062] Among them, min j with max j is the minimum and maximum value of the j-th dimension vector, W is the obtained sparrow position, X j is a j-dimensional sequence, W j is the position of the carrier after it arrives in the original solution space.

[0063] (4) Calculate the fitness value of each sparrow and take the sparrows with the best and worst fitness as the global best and worst individuals.

[0064] Furthermore, step (4) is further refined as follows:

[0065] Use the position of the i-th sparrow to assign values to the parameters of the deep extreme learning machine;

[0066] Input the training set to train the intelligent prediction model;

[0067] After reaching the number of iterations, the test set is input to obtain the predicted value, and the fitness value of each sparrow is calculated using the following formula:

[0068]

[0069] in, with y i are the predicted value and true value of the test set respectively;

[0070] Based on the calculation results, the sparrows with the best and worst fitness are regarded as the global best and worst individuals.

[0071] S3: Iteratively update the sparrow position to obtain the optimal parameters.

[0072] It should be noted that the detailed steps of this step are as follows:

[0073] (1) Increasing the step size control factor improves the global and local search capabilities of the algorithm and increases the convergence speed. According to the formula:

[0074]

[0075] Here, t is the number of iterations. In the early stage of iteration, the step size control factor is larger, the sparrow search range is wider, and the global search capability in the early stage is enhanced. As the control factor is adaptively reduced, better local search can be performed and the convergence speed is improved.

[0076] (2) Update the discoverer's location according to the formula:

[0077]

[0078] in, is the position of the i-th sparrow in the t-th iteration and the j-th dimension, α is a random number between (0, 1], R2∈[0, 1] and ST∈[0.5, 1] are the warning value and the safety value respectively, Q is a random number that obeys the normal distribution, and L is a 1×d matrix with each element being 1.

[0079] (3) Update the follower's position according to the formula:

[0080]

[0081] Among them, x p with x worst are the best position and the global worst position of the finder, n is the population size, A is a 1×d matrix, each element is 1 or -1, and A + =A T (AA T ) -1 .

[0082] (4) Update the scout's position according to the formula:

[0083]

[0084] Among them, x best is the global optimal position, β is a normally distributed random number with a mean of 0 and a variance of 1, which is used to control the step size, k is a random number between [-1, 1], and f iis the current sparrow’s fitness value, f g With f w are the current global optimal and worst fitness values, respectively, and ε is the minimum constant used to avoid the denominator being 0.

[0085] According to step (4) in S2, the new fitness value of the sparrow is calculated and the global optimal and worst individuals are updated.

[0086] S4: Apply the obtained optimal parameters to the intelligent prediction model for load forecasting.

[0087] It should be noted that the specific details of this step are as follows:

[0088] Read the obtained optimal parameters and apply them to the intelligent prediction model;

[0089] Read the processed training data set to train the intelligent prediction model.

[0090] Read the test dataset for prediction.

[0091] The present invention initializes the population through the Tent chaotic sequence, which can make the population more evenly distributed in the solution space. Compared with the original random population generation method, this method can increase the global search ability of the early population; the present invention adds a step size control factor, which gives the population a larger search range in the early stage of iteration, enhances the early global search ability, and as the number of iterations increases, the step size control factor decreases, which can perform better local search and improve the convergence speed; the present invention improves the prediction ability of the model by optimizing the parameters of the intelligent prediction model.

[0092] Example 2

[0093] Reference Figures 2 to 7 This is another embodiment of the present invention. Different from the first embodiment, this embodiment provides a verification test of an optimization method for an intelligent prediction model of electric load. In order to verify and illustrate the technical effects adopted in this method, this embodiment adopts a traditional technical solution and the method of the present invention for comparative testing, and compares the test results by means of scientific demonstration to verify the real effect of this method.

[0094] The experimental data consists of electricity load data from a region in southern China from 2012 to 2014. This dataset is collected every 15 minutes, with 96 data points per day, totaling 104,609 data points. Date types include year, month, day, and hour (1-96), and weather data includes daily average temperature, relative humidity, rainfall, daily minimum temperature, and daily maximum temperature. Missing values in the historical load are filled using a random forest method, which first uses the actual data as labels and then trains and fits new load values using other attributes of the dataset. The parameters selected for the global sparrow optimization algorithm are: a sparrow population of 20, 100 iterations, a ratio of 20% of the population for each detector and scout, and a warning value of 0.8. The Deep Extreme Learning Machine (DEML) is selected as the power load forecasting model, where the activation function is the sigmoid function. The number of network nodes of the model is 300-150-100-50-9. The global sparrow optimization algorithm is used to optimize the weights of the input layer and the bias of the hidden layer in each ELM autoencoder (ELM-AutoEncoder, ELM-AE) of DELM. The data from 2012 to 2013 are used as the training set to train the model, and the data from 2014 is used as the test set to verify the accuracy of the model.

[0095] This paper verifies the superiority of GSSA over SSA by using different types of benchmark functions. Single-modal test functions can be used to evaluate the convergence and development capabilities of the function, while multi-modal test functions can be used to evaluate the global and local search capabilities of the algorithm. The benchmark functions used are as follows:

[0096] Table 1: Benchmark function table.

[0097]

[0098] from Figure 2-4 It can be seen that whether it is a single-modal or multi-modal function, GSSA is superior to SSA in search accuracy and convergence speed, which verifies that GSSA has better search capabilities.

[0099] This paper analyzes load forecast results for three different date types: weekdays, weekends, and holidays. It then compares the results with three different forecasting models: a back propagation neural network (BPNN), a deep extreme learning machine (DELM), and a deep extreme learning machine optimized with a standard sparrow search algorithm (SSA-DELM). Figure 5 、 Figure 6 、 Figure 7The following are prediction images corresponding to three types of dates. According to image analysis, the present invention can better predict the load among the three types of dates. The errors of the three types of dates are as follows:

[0100] Table 2: Error results for three date types.

[0101]

[0102] It should be appreciated that embodiments of the present invention can be implemented or practiced by computer hardware, a combination of hardware and software, or by computer instructions stored in a non-transitory computer-readable memory. The methods can be implemented in a computer program using standard programming techniques, including a non-transitory computer-readable storage medium configured with a computer program, wherein the storage medium so configured causes the computer to operate in a specific and predefined manner, according to the methods and figures described in the specific embodiments. Each program can be implemented in a high-level procedural or object-oriented programming language to communicate with the computer system. However, if desired, the program can be implemented in assembly or machine language. In any case, the language can be a compiled or interpreted language. In addition, the program can be run on a programmed application-specific integrated circuit for this purpose.

[0103] Furthermore, the operations of the processes described herein may be performed in any suitable order unless otherwise indicated herein or otherwise clearly contradicted by the context. The processes described herein (or variations and / or combinations thereof) may be performed under the control of one or more computer systems configured with executable instructions and may be implemented as code (e.g., executable instructions, one or more computer programs, or one or more applications) that is executed collectively on one or more processors, by hardware, or a combination thereof. The computer program includes a plurality of instructions that can be executed by one or more processors.

[0104] Furthermore, the methods can be implemented in any type of computing platform operably connected to a suitable computer, including but not limited to a personal computer, minicomputer, mainframe, workstation, network or distributed computing environment, standalone or integrated computer platform, or in communication with a charged particle tool or other imaging device, etc. Aspects of the present invention can be implemented as machine-readable code stored on a non-transitory storage medium or device, whether removable or integrated into a computing platform, such as a hard drive, optical read and / or write storage medium, RAM, ROM, etc., such that it can be read by a programmable computer, and when the storage medium or device is read by the computer, can be used to configure and operate the computer to perform the processes described herein. Furthermore, the machine-readable code, or portions thereof, can be transmitted over wired or wireless networks. When such media includes instructions or programs for implementing the steps described above in conjunction with a microprocessor or other data processor, the invention described herein includes these and other different types of non-transitory computer-readable storage media. The invention also includes the computer itself, when programmed according to the methods and techniques described herein. The computer program can be applied to input data to perform the functions described herein, thereby converting the input data to generate output data that is stored in non-volatile memory. The output information can also be applied to one or more output devices, such as a display. In a preferred embodiment of the present invention, the converted data represents a physical and tangible object, including a specific visual depiction of the physical and tangible object produced on a display.

[0105] As used in this application, the terms "component", "module", "system" and the like are intended to refer to a computer-related entity, which can be hardware, firmware, a combination of hardware and software, software, or software in operation. For example, a component can be, but is not limited to: a process running on a processor, a processor, an object, an executable file, a thread in execution, a program and / or a computer. As an example, both an application running on a computing device and the computing device can be a component. One or more components can exist in an executing process and / or thread, and a component can be located in a computer and / or distributed between two or more computers. In addition, these components can be executed from various computer-readable media having various data structures thereon. These components can communicate in the form of local and / or remote processes, such as based on signals having one or more data packets (e.g., data from a component that interacts with another component in a local system, a distributed system, and / or interacts with other systems in the form of signals over a network such as the Internet).

[0106] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A method for optimizing an intelligent power load prediction model, characterized in that: include: Establish an intelligent prediction model, set the required optimization parameters and the optimization range of each parameter; Initialize the parameters and population of the global sparrow optimization algorithm; Iteratively update the sparrow position to obtain the optimal parameters; Applying the obtained optimal parameters to the intelligent prediction model to perform load forecasting; Power load data includes date type and weather type data. Date type includes year, month, day, and hour. Weather type data includes daily average temperature, relative humidity, rainfall, daily minimum temperature, and daily maximum temperature. The process of initializing the parameters and population of the global sparrow optimization algorithm includes: Initializing global sparrow optimization algorithm parameters, wherein the global sparrow optimization algorithm parameters include population number, number of iterations, ratio of discoverers to scouts, and warning value; Initializing the population according to an improved Tent chaotic mapping formula; Carry the sequence x to the original solution space; Calculate the fitness value of each sparrow, and take the sparrows with the best and worst fitness as the global best and worst individuals; The formula for converting the sequence x carrier to the original solution space includes: IN j =min j +(max j -min j )×X j Among them, min j and max j is the minimum and maximum value of the j-th dimension vector, W is the obtained sparrow position, X j is a j-dimensional sequence, W j is the position of the carrier after it arrives at the original solution space; The improved Tent chaos mapping formula is: Among them, N is the total number of sparrows in the sequence, rand(0,1) is a random number in [0,1], x i is a value in the sequence X, x i+1 To pass the xth i The value of is calculated; Calculate the fitness value of each sparrow, and include the sparrows with the best and worst fitness as the global best and worst individuals. Use the position of the i-th sparrow to assign values to the parameters of the deep extreme learning machine; Inputting a training set to train the intelligent prediction model; After reaching the number of iterations, the test set is input to obtain the predicted value, and the fitness value of each sparrow is calculated using the following formula: in, with y i are the predicted value and true value of the test set respectively; Based on the calculation results, the sparrows with the best and worst fitness are regarded as the global best and worst individuals; The iterative update of the sparrow position to obtain the optimal parameters includes: Increase the step size control factor; Update the finder's location; Update the follower's position; Update the scout's location; Calculate the new fitness value of each sparrow according to the fitness value calculation step, and update the global optimal and worst individuals; The formula for increasing the step size control factor includes: Where t is the number of iterations; The formula for updating the finder's position includes, in, is the position of the i-th sparrow in the t-th iteration and the j-th dimension, α is a random number between (0, 1], R2∈[0, 1] and ST∈[0.5, 1] are the warning value and the safety value respectively, Q is a random number that obeys the normal distribution, and L is a 1×d matrix with each element being 1.

2. The method for optimizing the power load intelligent prediction model according to claim 1, wherein: Said updating of the follower's position comprises, Among them, x p with x worst are the best position and the global worst position of the finder, n is the population size, A is a 1×d matrix, each element is 1 or -1, and A + =A T (AA T ) -1 .

3. The optimization method of the power load intelligent prediction model according to claim 1, characterized in that: The updating of the scout's location includes, Among them, x best is the global optimal position, β is a normally distributed random number with a mean of 0 and a variance of 1, k is a random number between [-1, 1], and f i is the current sparrow’s fitness value, f g With f w are the current global optimal and worst fitness values, respectively, and ε is the minimum constant.

Images