Ultra-short-term power load prediction method and device, equipment and storage medium

Abstract

The invention provides an ultra-short-term power load prediction method and device, equipment and a storage medium, and relates to the technical field of power load prediction, and the method comprises the steps: obtaining a historical load curve in a preset time period; decomposing the historical load curve to obtain a plurality of components; performing power load prediction on each component by using a preset prediction model to obtain a prediction result corresponding to each component; a predicted load curve is generated using each prediction result. According to the invention, the accuracy of ultra-short-term power load prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric load forecasting, and particularly to an ultra-short-term electric load forecasting method, device, equipment and storage medium. Background Art

[0002] All along, accurate electric load forecasting has been an important part of the operation of power systems. Due to the characteristic that electricity cannot be stored, accurate load is required to determine a more precise power generation plan. With the large-scale grid connection of renewable energy in recent years, it has brought great volatility to the system. Considering the randomness of such power sources, the stable operation of the system also puts forward more precise forecasting requirements for the load side. High-precision electric load forecasting is of great importance.

[0003] Electric load forecasting is divided into long-term load forecasting, medium and long-term load forecasting, short-term load forecasting and ultra-short-term load forecasting according to the time length. Due to more uncertainties, the shorter the time scale, the higher the forecasting accuracy requirement and the greater the implementation difficulty. Ultra-short-term load forecasting is to predict the load curve within the next few dozen minutes, which is of great significance for short-term grid dispatching and control.

[0004] Based on this, how to improve the accuracy of ultra-short-term electric load forecasting has become a technical problem to be solved urgently. Summary of the Invention

[0005] In view of this, in order to solve the above technical problems, the present invention provides an ultra-short-term electric load forecasting method, device, equipment and storage medium.

[0006] The present invention adopts the following technical solutions:

[0007] In the first aspect, the present invention provides an ultra-short-term electric load forecasting method, including:

[0008] Obtaining a historical load curve within a preset time period;

[0009] Decomposing the historical load curve to obtain a plurality of components;

[0010] Using a preset prediction model to respectively perform electric load forecasting on each of the components to obtain prediction results corresponding to each of the components;

[0011] Generating a predicted load curve using each of the prediction results.

[0012] Optionally, the decomposing the historical load curve to obtain a plurality of components specifically includes:

[0013] Using the variational mode decomposition algorithm to decompose the historical load curve to obtain a plurality of components.

[0014] Optionally, the preset prediction model is a prediction model obtained by optimizing the least squares support vector machine using a preset algorithm.

[0015] Optionally, the preset algorithm is the bat algorithm.

[0016] In a second aspect, the present invention provides a very short-term electric load forecasting device, including:

[0017] An acquisition module, configured to acquire a historical load curve within a preset time period;

[0018] A decomposition module, configured to decompose the historical load curve to obtain a plurality of components;

[0019] A prediction module, configured to use a preset prediction model to respectively perform electric load forecasting on each of the components to obtain prediction results corresponding to each of the components;

[0020] A generation module, configured to generate a predicted load curve using each of the prediction results.

[0021] Optionally, the decomposition module is specifically configured to:

[0022] Use the variational mode decomposition algorithm to decompose the historical load curve to obtain a plurality of components.

[0023] Optionally, the preset prediction model is a prediction model obtained by optimizing the least squares support vector machine using a preset algorithm.

[0024] Optionally, the preset algorithm is the bat algorithm.

[0025] In a third aspect, the present invention provides a computer device, including:

[0026] At least one processor; and,

[0027] A memory communicatively connected to the at least one processor; wherein,

[0028] The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to implement each step in the very short-term electric load forecasting method as described above.

[0029] In a fourth aspect, the present invention provides a computer-readable storage medium, on which a computer program is stored, characterized in that when the computer program is executed by a processor, each step in the very short-term electric load forecasting method as described above is implemented.

[0030] The present invention adopts the above technical solution, a method for ultra-short-term electric load forecasting. First, decompose the historical load curve to obtain multiple components. Then, use a preset forecasting model to respectively perform electric load forecasting on each component to obtain the forecasting results corresponding to each component. Finally, use each forecasting result to generate a forecasting load curve. In this way, the accuracy of ultra-short-term electric load forecasting is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0032] Figure 1 is a schematic flowchart of a method for ultra-short-term electric load forecasting provided by an embodiment of the present invention;

[0033] Figure 2 is a schematic structural diagram of an ultra-short-term electric load forecasting device provided by an embodiment of the present invention;

[0034] Figure 3 is a schematic structural diagram of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0035] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will describe the technical solutions of the present invention in detail. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention.

[0036] Embodiment

[0037] Figure 1 is a schematic flowchart of a method for ultra-short-term electric load forecasting provided by an embodiment of the present invention. As Figure 1 shown, this process includes:

[0038] Step 101: Obtain the historical load curve within a preset time period.

[0039] Step 102: Decompose the historical load curve to obtain multiple components.

[0040] Specifically, the Variational Mode Decomposition (VMD) algorithm can be used to decompose the historical load curve into multiple components. As an improved method of the Empirical Mode Decomposition (EMD) algorithm, the VMD algorithm is a signal decomposition method that has better effects in dealing with the mode mixing problem. By introducing constraint conditions and minimizing the cost function, it can effectively decompose complex signals into the linear combination of K mode functions and basis functions, realizing the frequency-domain analysis of non-periodic signals through decomposition, breaking down complex signals into multiple harmonic signals, and enabling the VMD algorithm to reduce the impact of volatility.

[0041] Using the VMD algorithm to decompose the historical load curve specifically includes the following steps:

[0042] (1) For the historical load curve, the constrained variational problem generated by the VMD algorithm is expressed as:

[0043]

[0044] where K is the total number of modal components; {u k} is the set of K modal components, {u k} = {u 1 .u 2 ,...,u k}; {w k} is the center frequency, {w k} = {w 1 .w 2 ,...,w k} is the set of k center frequencies; δ(t) is the Dirac function; t is the sampling point; is the partial derivative with respect to t; j is the imaginary unit; f(t) is the time series.

[0045] (2) With the help of the quadratic penalty term α and the Lagrange operator λ(t), the constraint conditions of the constrained variational problem are eliminated:

[0046]

[0047] (3) Continuously update λ n+1 (w) through the alternating direction multiplier algorithm, and the iterative update formula is:

[0048]

[0049] where n is the number of iterations and τ is the update parameter.

[0050] Through the above conversion, the solution of the variational problem can be completed.

[0051] Step 103: Use the preset prediction model to perform power load prediction on each component respectively, and obtain the prediction results corresponding to each component.

[0052] Among them, the preset prediction model can be a prediction model obtained by optimizing LSSVM (Least Squares Support Vector Machine) using a preset algorithm.

[0053] Traditional SVM (Support Vector Machine) is a classic machine learning algorithm based on a linear model. It transforms the input space into a high-dimensional feature space through a non-linear transformation and finds the optimal linear interface in the new space. To improve the SVM solution speed without reducing the solution accuracy, LS (Least Squares Method) is used for optimization, and then LSSVM is formed.

[0054] The calculation principle of LSSVM is as follows:

[0055] (1) The input data set is expressed as {(x i , y i ) | i = 1, 2,..., I}, and the regression function of SVM can be expressed as:

[0056]

[0057] Among them, x i is the input data (i.e., each component), y i is the approximate value of the output sample data; ω i is the weight; b is the bias, and in actual calculation, it usually takes the average value of the sum of the product of the weight ω i and the support quantity x; K() is the kernel function.

[0058] (2) Select RBF (Radial Basis Function) to construct the SVM kernel function, and the RBF function is expressed as:

[0059]

[0060] Among them, ξ is the width of the kernel function.

[0061] (3) Use LS to optimize SVM to form LSSVM. The optimal decision function of LSSVM in the high-dimensional feature space is:

[0062]

[0063] Among them, is the decision function.

[0064] When solving the regression problem, the formula can be transformed into:

[0065]

[0066] Among them, e m is the error amount; γ is the penalty coefficient, and the value of the penalty coefficient is related to the value of the kernel parameter ξ. If γ is too large, it will cause overlearning of LSSVM, and vice versa, it will cause underlearning of LSSVM.

[0067] In the embodiment of the present invention, the preset algorithm can be BAO (Bat Algorithm Optimization).

[0068] Specifically, BAO is an emerging swarm intelligence optimization algorithm, which searches for the global optimal solution by imitating the ultrasonic echo location predation process of bats. The optimization of BAO for LSSVM is mainly through iteration to optimize the penalty coefficient γ and the kernel parameter ξ. The optimization process of BAO for LSSVM is as follows:

[0069] (1) Set BAO parameters.

[0070] Specifically, set the number of bat populations as i, the population dimension as d, and the maximum number of iterations as T max , the pulse loudness as A(0), the emission rate as r(0), the loudness attenuation coefficient as A f , the pulse emission rate enhancement coefficient as r f , the pulse emission frequency range as [f min , f max .

[0071] (2) Initialize the population, that is, assign initial values to each parameter of BAO.

[0072] Here, it includes calculating the initial position of the bat individual and calculating the fitness value of the current position. Among them, MAPE (Mean Absolute Percentage Error) is used as the function to detect fitness.

[0073] (3) Update the flight speed and position of the bat individual.

[0074] Specifically, it can be expressed as:

[0075] f i = f min + β(f max - f min )......(13)

[0076]

[0077] Among them, f i 、f min and f max are respectively the pulse emission frequency, the minimum pulse emission rate, and the maximum pulse emission rate of the current bat i. β is a random number between (0, 1). is the velocity weight factor. is the velocity of bat i at time t + 1. is the position of bat i at time t. x * is the global optimal position. t is the current iteration number.

[0078] (4) To ensure the local search ability of the algorithm, a solution is randomly selected from the current optimal solution and a random perturbation is applied to it, and then the search is carried out near this solution. The corresponding position update formula in the local search stage is as follows:

[0079]

[0080] Among them, x' * is the new optimal solution after adding the random perturbation; x * represents a randomly selected optimal position solution; θ represents a random number, and θ ∈ [-1, 1]; is the average pulse loudness of all bat individuals at time t.

[0081] During the bat search process, by judging the distance between the prey and itself, the pulse frequency r t i and the loudness A i are updated. When the prey is closer, the loudness will gradually decrease, and the emission frequency of the sound wave will gradually increase.

[0082]

[0083] Among them, A f is the attenuation coefficient of the loudness, generally 0.9; r represents the increase coefficient of the pulse frequency; is the pulse frequency of bat i at time t + 1; represents the maximum pulse frequency of bat i.

[0084] (5) Judge whether the current iteration number reaches the maximum iteration number T max . If the current iteration number reaches the maximum iteration number T max , then end this optimization process, and the obtained optimal position or the loudness and emission frequency corresponding to bat i after reaching the maximum iteration number are the optimization results of the penalty coefficient γ and the kernel parameter ξ of LSSVM; if the current iteration number does not reach the maximum iteration number T max, the above-mentioned step (3) of updating the flight speed and position of the bat individual is executed.

[0085] In addition, after obtaining the preset prediction model by optimizing the LSSVM with BAO, indicators such as MAPE (Mean Absolute Percentage Error), RMSE (Root Mean Square Error), and MAE (Mean Absolute Error) can be combined to detect the effectiveness of the prediction model.

[0086] The calculation formulas for MAPE, RMSE, and MAE are as follows:

[0087]

[0088] where y i is the true value, is the predicted value.

[0089] Step 104: Generate a predicted load curve using each prediction result.

[0090] In the embodiment of the present invention, the above technical solution is adopted. First, the historical load curve is decomposed by VMD to obtain multiple components. Then, the preset prediction model is used to perform power load prediction on each component respectively to obtain the prediction results corresponding to each component. The preset prediction model is a prediction model obtained by optimizing the LSSVM with BAO. Finally, a predicted load curve is generated using each prediction result. In this way, by optimizing the LSSVM with BAO, the prediction accuracy and objectivity of the preset prediction model are improved. And, by decomposing the historical load curve using VMD, the influence of volatility is reduced. In this way, the accuracy of the ultra-short-term power load prediction in the embodiment of the present invention is improved.

[0091] Based on a general inventive concept, the present invention also provides an ultra-short-term power load prediction device. Figure 2 is a schematic structural diagram of an ultra-short-term power load prediction device provided by an embodiment of the present invention. As Figure 2 shown, the device includes:

[0092] An acquisition module 21, configured to acquire a historical load curve within a preset time period.

[0093] A decomposition module 22, configured to decompose the historical load curve to obtain multiple components.

[0094] A prediction module 23, configured to use a preset prediction model to perform power load prediction on each component respectively to obtain the prediction results corresponding to each component.

[0095] A generation module 24 for generating a predicted load curve using each prediction result.

[0096] Optionally, the decomposition module 22 can specifically be used for:

[0097] Decompose the historical load curve using the variational mode decomposition algorithm to obtain multiple components.

[0098] Optionally, the preset prediction model is a prediction model obtained by optimizing the least squares support vector machine using a preset algorithm. The preset algorithm can be the bat algorithm.

[0099] Based on a general inventive concept, the present invention also provides a computer device. Figure 3 It is a schematic structural diagram of a computer device provided by an embodiment of the present invention. As Figure 3 shown, the computer device 300 includes:

[0100] At least one processor 310; and,

[0101] A memory 330 communicatively connected to at least one processor 310; wherein,

[0102] The memory 330 stores instructions 320 executable by at least one processor 310. The instructions 320 are executed by at least one processor 310 so that at least one processor 310 can implement each step in the ultra-short-term power load prediction method described in the above embodiments.

[0103] Based on a general inventive concept, the present invention also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements each step in the ultra-short-term power load prediction method described in the above embodiments.

[0104] It can be understood that the same or similar parts in the above embodiments can be referred to each other. For the content not detailed in some embodiments, reference can be made to the same or similar content in other embodiments.

[0105] It should be noted that in the description of the present invention, terms such as "first" and "second" are only used for descriptive purposes and cannot be construed as indicating or implying relative importance. In addition, in the description of the present invention, unless otherwise specified, the meaning of "a plurality" refers to at least two.

[0106] Any process or method description shown in the process schematic diagram or described in other ways herein can be understood to represent a module, segment, or part of code including one or more executable instructions for implementing a specific logical function or process. The scope of the preferred embodiments of the present invention includes additional implementations, where the functions may be executed in a manner not shown or discussed, including in a substantially simultaneous manner or in a reverse order according to the involved functions, which should be understood by those skilled in the technical field of the embodiments of the present invention.

[0107] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc.

[0108] Those of ordinary skill in the art of this technology can understand that all or part of the steps carried by the method of implementing the above embodiments can be completed by instructing relevant hardware through a program. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0109] In addition, in each embodiment of the present invention, the functional units can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. When the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0110] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disk, etc.

[0111] In the description of this specification, the description referring to terms such as "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0112] Although the embodiments of the present invention have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for ultra-short-term power load forecasting, characterized in that: include: Obtain the historical load curve within a preset period; Decomposing the historical load curve to obtain multiple components; Using a preset prediction model, respectively predicting the power load of each of the components, and obtaining prediction results corresponding to each of the components; A predicted load curve is generated using each of the prediction results.

2. The ultra-short-term power load forecasting method according to claim 1, characterized in that: The historical load curve is decomposed to obtain multiple components, specifically including: The historical load curve is decomposed by using a variational mode decomposition algorithm to obtain multiple components.

3. The ultra-short-term power load forecasting method according to claim 1, characterized in that: The preset prediction model is a prediction model obtained by optimizing the least squares support vector machine using a preset algorithm.

4. The ultra-short-term power load forecasting method according to claim 3, characterized in that: The preset algorithm is the bat algorithm.

5. An ultra-short-term power load forecasting device, characterized in that: include: An acquisition module is used to acquire a historical load curve within a preset period; A decomposition module, used for decomposing the historical load curve to obtain multiple components; A prediction module, used to use a preset prediction model to perform power load prediction on each of the components, and obtain prediction results corresponding to each of the components; A generating module is used to generate a predicted load curve using each of the prediction results.

6. The ultra-short-term power load forecasting device according to claim 5, characterized in that: The decomposition module is specifically used for: The historical load curve is decomposed by using a variational mode decomposition algorithm to obtain multiple components.

7. The ultra-short-term power load forecasting device according to claim 5, characterized in that: The preset prediction model is a prediction model obtained by optimizing the least squares support vector machine using a preset algorithm.

8. The ultra-short-term power load forecasting device according to claim 7, characterized in that: The preset algorithm is the bat algorithm.

9. A computer device, characterized in that: include: at least one processor; as well as, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to implement each step of the ultra-short-term power load forecasting method according to any one of claims 1 to 4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, each step of the ultra-short-term power load forecasting method according to any one of claims 1 to 4 is implemented.

Images