Price prediction method, system and device based on lso-vmd-lstm and medium
By decomposing and optimizing electricity price data using the LSO-VMD-LSTM model, the problem of high cost and low efficiency in existing electricity price forecasting methods is solved, achieving higher accuracy and greater adaptability in electricity price forecasting, thus meeting the trading strategy needs of power companies.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUBEI ELECTRIC POWER TRADING CENT CO LTD
- Filing Date
- 2024-12-18
- Publication Date
- 2026-06-19
Smart Images
Figure CN122243588A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of power economics technology, and in particular relates to a method, system, electronic device and storage medium for electricity price prediction based on LSO-VMD-LSTM. Background Technology
[0002] In the day-ahead electricity market, to address the significant start-up and shutdown inertia of traditional thermal power units and the need for advance production planning decisions, coordinated trading between the market and generation plans has emerged. Day-ahead electricity prices have a significant impact on power companies' electricity trading strategies; therefore, accurate and effective forecasting of day-ahead electricity prices in the spot market is crucial. Although existing forecasting methods largely rely on manual operation, leading to high costs and low efficiency, artificial intelligence models can provide more accurate forecasts. However, these models may rely too heavily on simple linear combinations and fail to achieve optimal forecasting. Selecting a suitable electricity price forecasting algorithm requires comprehensive consideration of factors such as data characteristics, forecasting objectives, and accuracy to ensure that power companies can rationally formulate their trading strategies in the electricity market based on the forecast results.
[0003] Choosing a suitable electricity price forecasting algorithm is complex, and existing artificial intelligence models may not provide optimal solutions, exhibiting problems such as overly simplistic linear combinations and insufficient generalization ability. To improve the accuracy and efficiency of forecasting, a new forecasting method is needed that can effectively process historical data, automatically optimize model parameters, and adapt to market changes. Summary of the Invention
[0004] To address the aforementioned issues, this invention provides an electricity price forecasting method, system, electronic device, and storage medium based on LSO-VMD-LSTM. The optimal VMD parameters are determined through the self-optimization method of the LSO algorithm, thus improving the complexity of selecting a suitable electricity price forecasting algorithm.
[0005] This invention provides an electricity price forecasting method based on LSO-VMD-LSTM, the method comprising:
[0006] Obtain the raw electricity price data and preprocess the raw electricity price data;
[0007] The raw electricity price data was decomposed into different IMF components using the LSO-VMD algorithm;
[0008] Each IMF component is decomposed into a training set, a validation set, and a test set according to a preset ratio;
[0009] An LSTM model is constructed for each of the IMF components. The LSTM model is trained using the training set data. The parameters of the LSTM model are adjusted using the validation set and test set data to optimize the hyperparameters.
[0010] By using the hyperparameter-optimized LSTM model and real-time electricity price data, the prediction results of the LSTM model for each IMF component are precisely added together to obtain the predicted electricity price.
[0011] Furthermore, the original electricity price data includes:
[0012] Historical day-ahead hourly electricity load, planned external transmission volume, renewable energy output, downstream reserve, non-market-based output, and day-ahead electricity price data.
[0013] Further, the preprocessing of the raw electricity price data includes:
[0014] Median interpolation is performed on outliers and missing values in the original electricity price data;
[0015] All the original electricity price data are standardized to transform them into a unified unit of measurement.
[0016] Furthermore, the step of using the LSO-VMD algorithm to decompose the original electricity price data into different IMF components includes:
[0017] The LSO-VMD algorithm can automatically adjust the decomposition parameters according to the characteristics of the data to obtain representative IMF components.
[0018] Furthermore, the precise summation of the LSTM model prediction results of each of the IMF components includes:
[0019] During the addition process, the weight and importance of each IMF component need to be considered to ensure that the prediction result can fully reflect the overall characteristics of the original electricity price data.
[0020] Furthermore, the LSTM model is calibrated using mean absolute percentage error and mean square percentage error. If the error values are all within the preset range, the model training is successful.
[0021] The present invention also provides an electricity price forecasting system based on LSO-VMD-LSTM, the system comprising: a data preprocessing module, an LSO-VMD decomposition module, a data partitioning module, an LSTM forecasting module, and a result integration module;
[0022] The data preprocessing module is used to acquire raw electricity price data and preprocess the raw electricity price data;
[0023] The LSO-VMD decomposition module is used to decompose the original electricity price data into different IMF components using the LSO-VMD algorithm.
[0024] The data partitioning module is used to decompose each IMF component into a training set, a validation set, and a test set according to a preset ratio.
[0025] The LSTM prediction module is used to construct an LSTM model for each IMF component, train the LSTM model using training set data, and adjust the LSTM model parameters using validation set and test set data to perform hyperparameter optimization.
[0026] The result integration module uses the hyperparameter-optimized LSTM model and real-time electricity price data to accurately add up the LSTM model prediction results of each IMF component to obtain the predicted electricity price value.
[0027] Furthermore, the data preprocessing module is also used to perform median interpolation on outliers and missing values in the original electricity price data;
[0028] The data preprocessing module is also used to standardize all the original electricity price data and convert the original electricity price data into a unified unit of measurement.
[0029] The present invention also provides an electronic device, characterized in that it comprises:
[0030] processor;
[0031] And a memory storing computer-executable instructions, which, when executed, cause the processor to perform the method described in any of the above embodiments.
[0032] The present invention also provides a computer storage medium, characterized in that the computer storage medium stores one or more programs, which, when executed by a processor, implement the method described in any of the above embodiments.
[0033] Compared with the prior art, the present invention has the following advantages:
[0034] 1. Reduce costs and improve efficiency: By automatically acquiring data through a regional power trading market platform, manual operations are reduced, overcoming the high cost and low efficiency problems caused by manual forecasting methods in existing technologies.
[0035] 2. Improved Prediction Accuracy: Outlier and missing value handling, as well as data standardization, improved data quality and thus enhanced prediction accuracy. The LSO-VMD-LSTM model was employed, combining the powerful adaptive decomposition capabilities of LSO-VMD (Local Search Optimization Variational Mode Decomposition) with the excellent time series processing capabilities of LSTM (Long Short-Term Memory Network). LSO-VMD can decompose complex historical electricity price sequences into multiple relatively simple IMFs (Modal Components), effectively reducing data complexity and nonlinearity. LSTM, on the other hand, can fully capture the time series characteristics of each IMF component, modeling long-term and short-term time dependencies through memory units and gating mechanisms.
[0036] 3. Adapting to market changes: The model can be continuously trained and optimized based on acquired historical data, enabling it to adapt to market changes and meet the needs of power companies to rationally formulate their trading strategies in the power market based on forecast results.
[0037] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description
[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0039] Figure 1 A schematic diagram of the electricity price forecasting method based on LSO-VMD-LSTM of the present invention is shown;
[0040] Figure 2 The diagram shows the structure of the electricity price prediction system based on LSO-VMD-LSTM of this invention.
[0041] Figure 3 A schematic diagram of an electronic device structure according to the present invention is shown. Detailed Implementation
[0042] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0043] The same reference numerals in the accompanying drawings denote the same or similar elements, components, or parts, and therefore, repeated descriptions of the same or similar elements, components, or parts may be omitted below. It should also be understood that although terms such as first, second, third, etc., indicating numbers may be used herein to describe various devices, elements, components, or parts, these devices, elements, components, or parts should not be limited by these terms. That is, these terms are only used to distinguish one from another. For example, a first device may also be referred to as a second device, without departing from the essential technical solution of the invention. Furthermore, the terms "and / or" and "and / or" refer to all combinations including any one or more of the listed items.
[0044] Figure 1 This is a schematic diagram of the electricity price forecasting method based on LSO-VMD-LSTM provided in an embodiment of the present invention, specifically including:
[0045] S11: Obtain the raw electricity price data and preprocess the raw electricity price data.
[0046] The specific data to be collected is as follows:
[0047] By acquiring historical electricity load, transmission plans, renewable energy output, downstream reserves, non-market-based output, day-ahead electricity prices, and thermal power spatial data from electricity spot market transactions, a historical dataset of the electricity spot market is obtained.
[0048] The historical datasets of day-ahead hourly electricity load, planned external transmission volume, renewable energy output, downstream reserve, non-market output, and day-ahead electricity price are obtained from the historical datasets of the electricity spot market. These historical datasets are used to prepare for subsequent model training.
[0049] The specific steps are as follows:
[0050] (1) Obtain historical daily electricity load and day-ahead electricity price datasets for more than three months through the electricity market data platform. Perform median interpolation on outliers and missing values to ensure data integrity and accuracy. By employing advanced data cleaning techniques, outlier data points can be effectively identified and corrected, and missing values can be appropriately filled, thereby improving data quality. The specific processing methods are as follows:
[0051] Sn =0.5S n-1 +0.5S n+1
[0052] Among them, S n For missing or outlier values to be filled, S n-1 and S n+1 S n The actual values at the preceding and following time points.
[0053] (2) Standardize all data to improve model stability. Many machine learning algorithms, especially those based on gradient descent, are sensitive to data scale. Transforming data into a uniform scale facilitates subsequent model training and analysis. Standardization eliminates dimensional differences between different data sets, enabling the model to better learn the intrinsic characteristics of the data. Standardization can also accelerate the convergence speed of these algorithms. The formula is as follows:
[0054]
[0055] Among them, v t (n i Let x be the initial value of the i-th feature on day t in the historical timeline. t (n i ) represents the standardized value of the i-th feature on day t in the historical data.
[0056] S2: The original electricity price data is decomposed into different IMF components using the LSO-VMD algorithm.
[0057] The specific breakdown details are as follows:
[0058] The Local Search Optimized Variational Mode Decomposition (LSO-VMD) algorithm is used to adaptively decompose historical electricity price sequences into different IMFs (modal components). The LSO-VMD algorithm has strong adaptive capabilities and can automatically adjust the decomposition parameters according to the characteristics of the data, thereby obtaining more representative IMF components.
[0059] Each IMF component reflects the characteristics of historical electricity price series at different frequencies and time scales, providing rich information for subsequent forecasting.
[0060] The specific steps are as follows:
[0061] (1) Input the original signal: First, the historical electricity price sequence is input into the LSO-VMD algorithm as the original signal.
[0062] (2) Add white noise: Add a Gaussian white noise sequence to the original signal to improve the reliability and accuracy of the decomposition.
[0063] (3) Signal decomposition: The LSO-VMD algorithm is used to decompose the signal with added white noise into several IMF components and a residual sequence. Each IMF component represents a specific frequency component in the original signal.
[0064] (4) Repeat the decomposition process: Repeat the above decomposition steps using different white noise sequences until the maximum number of iterations of the algorithm is reached. This can improve the stability and accuracy of the decomposition results.
[0065] (5) Calculate the mean: Calculate the mean of all IMF components and residual sequences obtained by repeated decomposition to obtain the final decomposition sequence of the original signal.
[0066] (6) Output IMF components: Finally, the original historical electricity price series is decomposed into several IMF components and a residual series, which can be further used for analysis or as input to other models.
[0067] S3: Decompose each of the IMF components into a training set, a validation set, and a test set according to a preset ratio.
[0068] In this embodiment, each IMF component can be decomposed into a training set, a validation set, and a test set in a 6:2:2 ratio. This scientifically sound division ensures that the model has sufficient data to learn during training, while also allowing for effective evaluation and adjustment of the model's performance through the validation and test sets.
[0069] S4: Construct an LSTM model for each IMF component, train the LSTM model using the training set data, and adjust the LSTM model parameters using the validation set and test set data to optimize the hyperparameters.
[0070] The specific details of the model construction are as follows:
[0071] For each IMF component, a Long Short-Term Memory (LSTM) network model is used for prediction. As an advanced deep learning model, the LSTM model has excellent time series processing capabilities and can fully capture the time series features in the IMF components.
[0072] Through memory units and gating mechanisms, LSTM models can effectively handle long-term and short-term time dependencies, providing accurate prediction results for electricity price forecasting.
[0073] The specific steps are as follows:
[0074] (1) Construct an LSTM model for each IMF component. The LSTM model is a special type of recurrent neural network (RNN) that is good at processing and predicting time series data.
[0075] (2) Determine the architecture of the LSTM model, including the number of layers, the number of units in each layer, activation functions, etc.
[0076] (3) Train the LSTM model using the training set data.
[0077] (4) Adjust the model parameters using validation set data to optimize hyperparameters.
[0078] During training, the model's performance can be monitored in real time using the validation set, allowing for timely adjustments to the training strategy to improve the model's training effectiveness.
[0079] S5: Using the hyperparameter-optimized LSTM model and real-time electricity price data, the LSTM model prediction results of each IMF component are precisely added together to obtain the electricity price prediction value.
[0080] The specific integration details are as follows:
[0081] The LSTM prediction results of each IMF component are precisely summed. During the summation process, the weight and importance of each IMF component need to be considered to ensure that the final prediction result can fully reflect the overall characteristics of the historical electricity price series.
[0082] The results integration module can fully integrate the prediction information of each IMF component, thereby improving the overall prediction accuracy and stability.
[0083] The specific steps for calculating model error are as follows:
[0084] (1) The calculation method for model calibration using Mean Absolute Percentage Error (MAPE) and Mean Square Percentage Error (RMSPE) is as follows: Mean Absolute Percentage Error (MAPE) is:
[0085]
[0086] Among them, y i For predicted values, This is the actual value.
[0087] (2) The mean square percentage error (RMSPE) is:
[0088]
[0089] In this embodiment, the error standard range can be within 10%.
[0090] Figure 2 The structure diagram of the electricity price prediction system based on LSO-VMD-LSTM provided in this embodiment of the invention specifically includes:
[0091] The data preprocessing module 11 is used to acquire the original electricity price data and preprocess the original electricity price data.
[0092] LSO-VMD decomposition module 12 is used to decompose the original electricity price data into different IMF components using the LSO-VMD algorithm.
[0093] The data partitioning module 13 is used to decompose each of the IMF components into a training set, a validation set, and a test set according to a preset ratio.
[0094] The LSTM prediction module 14 is used to construct an LSTM model for each of the IMF components, train the LSTM model using training set data, and adjust the LSTM model parameters using validation set and test set data to perform hyperparameter optimization.
[0095] The result integration module 15 is used to accurately add up the LSTM model prediction results of each IMF component to obtain the electricity price prediction value.
[0096] like Figure 3 As shown, this embodiment of the invention provides an electronic device, including a processor 1110, a communication interface 1120, a memory 1130, and a communication bus 1140, wherein the processor 1110, the communication interface 1120, and the memory 1130 communicate with each other through the communication bus 1140.
[0097] Memory 1130 is used to store computer programs;
[0098] When the processor 1110 executes the program stored in the memory 1130, it implements any of the above-described determination methods.
[0099] The electronic device provided in this embodiment of the invention includes a processor 1110 that executes a program stored in a memory 1130 to obtain the fluid flow rate of each branch under different switching states and determines the initial volumetric flow rate of each branch; it corrects the initial volumetric flow rate based on the pipe parameters and fluid parameters of each branch when it is in operating condition and standard condition to obtain the standard condition volumetric flow rate of each branch; it obtains multiple total standard condition volumetric flow rates based on the standard condition volumetric flow rates of each branch under different switching states, and determines the optimal switching state of each branch by using the switching state of each branch when the total standard condition volumetric flow rate reaches a preset target.
[0100] The communication bus 1140 mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. This communication bus 1140 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, it is shown in the figure with only one thick line, but this does not indicate that there is only one bus or one type of bus.
[0101] The communication interface 1120 is used for communication between the above-mentioned electronic device and other devices.
[0102] The memory 1130 may include random access memory (RAM) or non-volatile memory, such as at least one disk storage device. Optionally, the memory 1130 may also be at least one storage device located remotely from the aforementioned processor 1110.
[0103] The processor 1110 mentioned above can be a general-purpose processor 1110, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components.
[0104] This invention provides a computer-readable storage medium storing one or more programs that can be executed by one or more processors 1110 to implement the determination method of any of the above embodiments.
[0105] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).
[0106] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for electricity price forecasting based on LSO-VMD-LSTM, characterized in that, The method includes: Obtain the raw electricity price data and preprocess the raw electricity price data; The raw electricity price data was decomposed into different IMF components using the LSO-VMD algorithm; Each IMF component is decomposed into a training set, a validation set, and a test set according to a preset ratio; An LSTM model is constructed for each of the IMF components. The LSTM model is trained using the training set data. The parameters of the LSTM model are adjusted using the validation set and test set data to optimize the hyperparameters. By using the hyperparameter-optimized LSTM model and real-time electricity price data, the prediction results of the LSTM model for each IMF component are precisely added together to obtain the predicted electricity price.
2. The prediction method according to claim 1, characterized in that, The original electricity price data includes: Historical day-ahead hourly electricity load, planned external transmission volume, renewable energy output, downstream reserve, non-market-based output, and day-ahead electricity price data.
3. The prediction method according to claim 1, characterized in that, Preprocessing the raw electricity price data includes: Median interpolation is performed on outliers and missing values in the original electricity price data; All the original electricity price data are standardized to transform them into a unified unit of measurement.
4. The prediction method according to claim 1, characterized in that, The step of using the LSO-VMD algorithm to decompose the raw electricity price data into different IMF components includes: The LSO-VMD algorithm can automatically adjust the decomposition parameters according to the characteristics of the data to obtain representative IMF components.
5. The prediction method according to claim 1, characterized in that, The precise addition of the LSTM model prediction results of each of the IMF components includes: During the addition process, the weight and importance of each IMF component need to be considered to ensure that the prediction result can fully reflect the overall characteristics of the original electricity price data.
6. The prediction method according to claim 1, characterized in that, The LSTM model is calibrated using mean absolute percentage error and mean square percentage error. If the error values are all within the preset range, the model training is successful.
7. A power price forecasting system based on LSO-VMD-LSTM, characterized in that, The system includes: a data preprocessing module, an LSO-VMD decomposition module, a data partitioning module, an LSTM prediction module, and a result integration module; The data preprocessing module is used to acquire raw electricity price data and preprocess the raw electricity price data; The LSO-VMD decomposition module is used to decompose the original electricity price data into different IMF components using the LSO-VMD algorithm. The data partitioning module is used to decompose each IMF component into a training set, a validation set, and a test set according to a preset ratio. The LSTM prediction module is used to construct an LSTM model for each IMF component, train the LSTM model using training set data, and adjust the LSTM model parameters using validation set and test set data to perform hyperparameter optimization. The result integration module uses the hyperparameter-optimized LSTM model and real-time electricity price data to accurately add up the LSTM model prediction results of each IMF component to obtain the predicted electricity price value.
8. The prediction system according to claim 7, characterized in that, include: The data preprocessing module is also used to perform median interpolation on outliers and missing values in the original electricity price data; The data preprocessing module is also used to standardize all the original electricity price data and convert the original electricity price data into a unified unit of measurement.
9. An electronic device, characterized in that, include: processor; And a memory storing computer-executable instructions, which, when executed, cause the processor to perform the method according to any one of claims 1-6.
10. A computer storage medium, characterized in that, in, The computer storage medium stores one or more programs that, when executed by a processor, implement the method of any one of claims 1-6.