Park microgrid load prediction method and system, electronic equipment and storage medium

Abstract

The invention provides a load prediction method and system of a park microgrid, electronic equipment and a storage medium, and belongs to the field of load prediction of the park microgrid. The method comprises the following steps: acquiring historical load data and related influence factor data of the park microgrid; the historical load data and the related influence factor data of the park micro-grid are input to a pre-constructed Fredformer model, a load prediction result of the park micro-grid is obtained, and the Fredformer model comprises a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer. Based on a Transform framework, through discrete Fourier transform, frequency domain optimization, a Transform encoder, a frequency summary layer and inverse discrete Fourier transform framework reconstruction, frequency deviation can be reduced, and neglect of low-amplitude characteristics is prevented.

Description

Technical Field

[0001] The present invention belongs to the technical field of load prediction of a park microgrid, and specifically relates to a load prediction method, system, electronic equipment and storage medium for a park microgrid. Background Art

[0002] Large-scale parks generally have complex energy operation mechanisms, with various types of loads, and are affected by regions, time, weather, and individual changes. Load volatility and uncertainty are increasing, and the uncertainty of loads restricts the flexibility of new energy scheduling, so load forecasting requires higher accuracy and faster speed. Therefore, how to quickly and accurately predict loads is crucial.

[0003] Existing technologies have proposed a variety of prediction methods, such as back propagation neural networks based on modified particle swarm optimization, Markov chain Monte Carlo sampling, long short-term memory networks, and support vector regression. For energy consumption data, which has time series properties, the Transformer-based method is widely used. However, the self-attention mechanism in this method usually gives priority to the low-frequency features of the data and ignores the high-frequency features, which may cause deviations in the model results and lead to information loss. Summary of the invention

[0004] The purpose of the embodiments of the present invention is to provide a load forecasting method, system, electronic device and storage medium for a campus microgrid, which are used to fully or at least partially solve the technical problem that the low-frequency characteristics of the data are given priority while the high-frequency characteristics are ignored in the above-mentioned prior art, causing deviations in the prediction results and leading to information loss.

[0005] In a first aspect, an embodiment of the present application provides a load prediction method for a campus microgrid, comprising: Obtain historical load data of the park microgrid and related influencing factor data; The historical load data of the park microgrid and related influencing factor data are input into the pre-built Fredformer model to obtain the load forecast result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

[0006] Optionally, the historical load data of the campus microgrid and related influencing factor data include but are not limited to time information, weather data, regional information and personal behavior related data.

[0007] Optionally, after obtaining historical load data of the park microgrid and relevant influencing factor data, the load forecasting method of the park microgrid further includes: The outliers and missing values ​​in the historical load data and the data of related influencing factors are removed and normalized; the historical load data and the data of related influencing factors that exceed the preset threshold are regarded as abnormal; the missing values ​​are processed by the mean filling and linear interpolation methods.

[0008] Optionally, a discrete Fourier transform architecture layer in the Fredformer model to decompose the input time series data into frequency components; A frequency domain optimization layer for refining and normalizing frequency components to generate multiple sub-frequency bands; Transformer encoder layer, used to learn the frequency features of each sub-frequency band; Frequency summary layer, used to perform linear transformation on feature information; The inverse discrete Fourier transform layer is used to reconstruct the processed frequency output back into the time domain signal.

[0009] Optionally, the historical load data of the park microgrid and the data of related influencing factors are input into a pre-built Fredformer model to obtain the load forecast result of the park microgrid, including: The historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data are input into the discrete Fourier transform architecture layer to decompose them into frequency components; The frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations through the frequency domain optimization layer to avoid mutual influence between different frequency components; Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing; After learning the features of each sub-frequency band, linear transformation is performed in the frequency summary layer and inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

[0010] Optionally, the discrete Fourier transform architecture layer decomposes the historical load data of the campus microgrid and related influencing factor data into frequency components according to the following formula: k=1,…,L; In the formula, represents the kth frequency component, L represents the lookback window length of each channel, Represents a multivariate time series.

[0011] In a second aspect, an embodiment of the present application further provides a load forecasting system for a campus microgrid, including: An acquisition unit, used to acquire historical load data of the park microgrid and data of related influencing factors; The prediction unit is used to input the historical load data of the park microgrid and related influencing factor data into a pre-built Fredformer model to obtain the load prediction result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

[0012] Optionally, the prediction unit is specifically used for: The historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data are input into the discrete Fourier transform architecture layer to decompose them into frequency components; The frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations through the frequency domain optimization layer to avoid mutual influence between different frequency components; Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing; After learning the features of each sub-frequency band, linear transformation is performed in the frequency summary layer and inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

[0013] In a third aspect, an embodiment of the present application further provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of the load forecasting method for the campus microgrid are implemented.

[0014] In a fourth aspect, an embodiment of the present application further provides a storage medium on which a computer program is stored, and when the computer program is executed by a processor, the steps of the load forecasting method for the campus microgrid described above are implemented.

[0015] It can be seen from the above technical solutions that the present invention has the following advantages: The load forecasting method, system, electronic device and storage medium for the campus microgrid provided in the present application are based on the Transformer framework and can reduce frequency deviation and prevent low-amplitude features from being ignored through discrete Fourier transform, frequency domain optimization, Transformer encoder, frequency summary layer and inverse discrete Fourier transform architecture reconstruction. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solution of the present invention, the accompanying drawings required for use in the description will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0017] Figure 1 A flow chart of an implementation method of a load forecasting method for a campus microgrid provided by an embodiment of the present invention; Figure 2 A schematic diagram of the Fredformer model structure provided by an embodiment of the present invention; Figure 3 A flow chart of load forecasting using the Fredformer model provided in an embodiment of the present invention; Figure 4 A detailed flow chart of a method for load prediction of a campus microgrid provided by an embodiment of the present invention; Figure 5 A schematic diagram of the structure of a load forecasting system for a campus microgrid provided by an embodiment of the present invention; Figure 6 A schematic diagram of the hardware structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In the detailed description below, various embodiments of the present disclosure will be described more fully. The present disclosure may have various embodiments, and adjustments and changes may be made therein. However, it should be understood that there is no intention to limit the various embodiments of the present disclosure to the specific embodiments disclosed herein, but rather the present disclosure should be understood to cover all adjustments, equivalents and / or alternatives that fall within the spirit and scope of the various embodiments of the present disclosure.

[0019] Hereinafter, the terms "include" or "may include" that may be used in various embodiments of the present disclosure indicate the presence of the disclosed functions or operations, and do not limit the addition of one or more functions or operations. In addition, as used in various embodiments of the present disclosure, the terms "include", "have" and their cognates are intended only to indicate specific features, numbers, steps, operations, or combinations of the foregoing items, and should not be understood as first excluding the presence of one or more other features, numbers, steps, operations, or combinations of the foregoing items or the possibility of adding one or more features, numbers, steps, operations, or combinations of the foregoing items.

[0020] In various embodiments of the present disclosure, the expression "or" or "at least one of A or / and B" includes any combination or all combinations of the words listed at the same time. For example, the expression "A or B" or "at least one of A or / and B" may include A, may include B, or may include both A and B.

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] See also Figure 1 The figure is a flow chart of a load prediction method for a park microgrid in a specific embodiment, including the following execution steps: Step 100: Obtain historical load data and related influencing factor data of the campus microgrid.

[0023] Specifically, the historical load data of the park microgrid and related influencing factor data include but are not limited to time information, weather data, regional information and personal behavior related data.

[0024] For example, time information such as hours, days, seasons, etc., weather data such as temperature, humidity, wind speed, sunshine, etc., regional information such as division data of different functional areas and personal behavior related data such as patterns of personnel activities, etc.

[0025] Step 101: input the historical load data of the park microgrid and the relevant influencing factor data into the pre-built Fredformer model to obtain the load forecast result of the park microgrid.

[0026] Among them, see Figure 2 As shown, the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer, and an inverse discrete Fourier transform layer.

[0027] Specifically, the discrete Fourier transform architecture layer (DFT) in the Fredformer model is used to transform the input time series data (X∈R C×L ) is decomposed into frequency components ( ); frequency domain optimization layer (frequency domain modeling), used to refine and normalize frequency components to generate multiple sub-frequency bands ( , ); Transformer encoder layer, used to learn the frequency characteristics of each sub-frequency band ( , ); frequency summary layer, used to summarize feature information ( 、 ) for linear transformation ( , ); Inverse discrete Fourier transform layer, used to reconstruct the processed frequency output back to the time domain signal ( ).

[0028] In this embodiment, based on the Transformer framework, through discrete Fourier transform, frequency domain optimization, Transformer encoder, frequency summarization layer and inverse discrete Fourier transform architecture reconstruction, frequency deviation can be reduced and low-amplitude features can be prevented from being ignored.

[0029] Exemplarily, the Fredformer model consists of four main components: ① dft-idft backbone, ② frequency domain optimization, ③ local frequency independent learning, and ④ global semantic frequency aggregation. The model optimization mainly uses the Nyström approximation method. First, the attention map calculation is simplified. The Nyström approximation is used in Fredformer to reduce the computational complexity of the attention map. By selecting a suitable low-rank approximation method, key information is extracted from the original high-dimensional attention matrix calculation to reduce the amount of calculation. Specifically, a small number of representative points (landmarks) are selected from the input data, and the approximate attention matrix is ​​calculated based on these representative points. According to the simplified attention map calculation method, the parameter update rule of the model is adjusted. During the training process, the historical load data of the park microgrid is used as the training set, the predicted output is compared with the actual load data, the mean square error loss function is used to measure the prediction error, and the model parameters are updated through optimization algorithms such as Adam and SGD to minimize the loss function and improve the prediction accuracy of the model.

[0030] Finally, the model is evaluated and verified. 80% of the collected campus microgrid load data is used as a training set and 20% as a verification set, which are divided into two parts. During the training process, the model is trained using the training set, and the performance of the model is evaluated on the verification set.

[0031] Evaluation indicator selection: Select a variety of evaluation indicators to comprehensively evaluate the performance of the model, including mean square error (MSE), mean absolute error, root mean square error, mean absolute percentage error, etc. The values ​​of these indicators on the validation set are used to determine whether the model is overfitting or underfitting, and to adjust the model's hyperparameters (such as hidden layer size, learning rate, number of training rounds, etc.).

[0032] Model comparison and verification: The Fredformer-based load forecasting model is compared with existing prediction methods such as back propagation neural network based on modified particle swarm optimization, Markov chain Monte Carlo sampling, long short-term memory network and support vector regression on the same data set. By comparing the performance of various methods on evaluation indicators, the superiority of the method of the present invention is verified, that is, it can achieve comparable performance while having less parameter scale and lower computational cost.

[0033] In one embodiment of the present invention, after executing step 100, the following steps are further executed: The outliers and missing values ​​in the historical load data and the data of related influencing factors are removed and normalized; the historical load data and the data of related influencing factors that exceed the preset threshold are regarded as abnormal; the missing values ​​are processed by the mean filling and linear interpolation methods.

[0034] For example, for abnormal values, a reasonable threshold range can be set for judgment and correction. For example, values ​​exceeding ±3 times the standard deviation of the historical load mean are considered abnormal, and can be corrected based on surrounding normal data points or by interpolation. For missing values, mean filling, linear interpolation, and other methods can be used for processing. The cleaned data is processed using the Min-Max normalization method to map all data to the [0,1] or [-1,1] interval.

[0035] In one embodiment of the present invention, refer to Figure 3 As shown, based on step 101, a possible embodiment is given below to illustrate its specific implementation scheme in a non-limiting manner, and the following steps are specifically performed: S1010: Input the historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data into the discrete Fourier transform architecture layer to decompose them into frequency components.

[0036] In some embodiments, represents a multivariate time series consisting of C channels, where each channel records an independent L-length lookback window. The prediction task is to predict future data The H time step in : ; in The prediction function is also based on the Transformer model. The goal is to reduce the learning bias in the Transformer and enhance the prediction result X′, that is, to minimize X′ and Discrete Fourier Transform (DFT), use DFT to analyze X, and the frequency content of X′.

[0037] Specifically, the discrete Fourier transform architecture layer decomposes the historical load data of the park microgrid and related influencing factor data into frequency components according to the following formula: k=1,…,L; In the formula, represents the kth frequency component, The amplitude information of these frequencies, L represents the lookback window length of each channel, Represents a multivariate time series.

[0038] S1011: Divide the frequency components into a plurality of sub-frequency bands by using a non-overlapping patch operation through a frequency domain optimization layer to avoid mutual influence between different frequency components.

[0039] Specifically, the frequency spectrum is subdivided and the frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations to avoid the mutual influence between different frequency components. The frequency patches are normalized to eliminate the proportion differences between different frequency components, ensuring that the model pays equal attention to all key frequency components.

[0040] S1012: Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing.

[0041] Specifically, on the sub-frequency blocks after patch normalization of the frequency domain features, the model deploys a Transformer encoder to independently learn the features of each block. In this way, the model can focus on the cross-channel correlation within the same frequency band, rather than the amplitude difference between different frequency components, thereby achieving debiasing.

[0042] S1013: After learning the features of each sub-frequency band, a linear transformation is performed in the frequency summary layer and an inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

[0043] Specifically, the inverse discrete Fourier transform (IDFT) is , reconstruct the time series data based on the DFT coefficients.

[0044] In one embodiment, Figure 4 This is a detailed flow chart of a load forecasting method for a campus microgrid provided according to an embodiment of the present invention. This embodiment is further optimized and expanded on the basis of the above embodiments.

[0045] S400: Obtain historical load data of the park microgrid and related influencing factor data.

[0046] S401: Remove outliers and missing values ​​from historical load data and related influencing factor data, and perform normalization processing; historical load data and related influencing factor data that exceed preset thresholds are regarded as abnormal; for missing values, mean filling and linear interpolation methods are used for processing.

[0047] S402: Input the historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data into the discrete Fourier transform architecture layer to decompose them into frequency components.

[0048] S403: Divide the frequency components into a plurality of sub-frequency bands by using a non-overlapping patch operation through a frequency domain optimization layer to avoid mutual influence between different frequency components.

[0049] S404: Deploy a Transformer encoder layer on the sub-frequency blocks after the frequency domain features are patch normalized to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing.

[0050] S405: After learning the features of each sub-frequency band, a linear transformation is performed in the frequency summary layer and an inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form a final prediction result.

[0051] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0052] like Figure 5 As shown, the following is an embodiment of the load prediction system for the campus microgrid provided by the embodiment of the present disclosure, which belongs to the same inventive concept as the load prediction method for the campus microgrid in the above-mentioned embodiments. For details not described in detail in the embodiment of the load prediction system for the campus microgrid, please refer to the embodiment of the load prediction method for the campus microgrid mentioned above.

[0053] An acquisition unit 50 is used to acquire historical load data and related influencing factor data of the park microgrid; The prediction unit 51 is used to input the historical load data and related influencing factor data of the park microgrid into a pre-built Fredformer model to obtain the load prediction result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

[0054] Specifically, the prediction unit is specifically used for: The historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data are input into the discrete Fourier transform architecture layer to decompose them into frequency components; The frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations through the frequency domain optimization layer to avoid mutual influence between different frequency components; Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing; After learning the features of each sub-frequency band, linear transformation is performed in the frequency summary layer and inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

[0055] Figure 6 It is a schematic diagram of the hardware structure of an electronic device for implementing various embodiments of the present invention.

[0056] The load forecasting method of the park microgrid provided in the embodiment of the present application can be applied to electronic devices. Those skilled in the art will appreciate that the electronic device structure involved in the embodiment of the present invention does not constitute a limitation on the electronic device, and the electronic device may include more or fewer components than shown, or combine certain components, or arrange different components. In an embodiment of the present invention, the electronic device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of the present application described and / or required herein.

[0057] The electronic device may include a processor, an external memory interface, an internal memory, a universal serial bus (USB) interface, a charging management module, a power management module, a battery, a wireless communication module, an audio module, a speaker, a microphone, a sensor module, buttons, a camera, a display, and a SIM card interface, etc.

[0058] It is to be understood that the structure illustrated in the embodiments of the present application does not constitute a specific limitation on the electronic device. In other embodiments of the present application, the electronic device may include more or fewer components than shown in the figure, or combine certain components, or split certain components, or arrange the components differently. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0059] The processor may include one or more processing units, for example, the processor may include a central processing unit (CPU), an application processor (AP), a modem processor, a graphics processing unit (GPU), an image signal processor (ISP), a controller, a memory, a video codec, a digital signal processor (DSP), a baseband processor, and / or a neural-network processing unit (NPU), etc. Among them, different processing units may be independent devices or integrated into one or more processors.

[0060] The processor can be the nerve center and command center of the electronic device. The controller can generate an operation control signal according to the instruction operation code and timing signal to complete the control of fetching and executing instructions.

[0061] A memory may also be provided in the processor for storing instructions and data. In some embodiments, the memory in the processor is a cache memory. The memory may store instructions or data that the processor has just used or is cyclically used. If the processor needs to use the instruction or data again, it may be directly called from the memory. This avoids repeated access, reduces the waiting time of the processor, and thus improves system efficiency.

[0062] The external memory interface can be used to connect an external memory card, such as a MicroSD card, to expand the storage capacity of the electronic device. The external memory card communicates with the processor through the external memory interface to implement data storage functions. For example, files such as music and videos can be saved in the external memory card.

[0063] The internal memory can be used to store computer executable program codes, which include instructions. The processor executes various functional applications and data processing of the electronic device by running the instructions stored in the internal memory. The internal memory may include a program storage area and a data storage area. The internal memory may include a high-speed random access memory and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, a universal flash storage (UFS), etc.

[0064] The wireless communication function of an electronic device can be realized through an antenna, a wireless communication module, a modem processor, and a baseband processor.

[0065] The wireless communication module can provide wireless communication solutions for electronic devices, including wireless local area networks (WLAN) (such as wireless fidelity (Wi-Fi) networks), Bluetooth (BT), global navigation satellite system (GNSS), frequency modulation (FM), near field communication (NFC), infrared technology (IR), etc.

[0066] Electronic devices can implement audio functions, etc. through audio modules, speakers, receivers, microphones, headphone jacks, and application processors.

[0067] Electronic devices can achieve shooting functions through ISP, camera, video codec, GPU, display and application processor.

[0068] Electronic devices can achieve display functions through GPU, display screen and application processor.

[0069] The GPU is a microprocessor for image processing that connects the display screen and the application processor. The GPU is used to perform mathematical and geometric calculations for graphics rendering. The processor may include one or more GPUs that execute program instructions to generate or change display information.

[0070] The display screen is used to display images, videos, etc. The display screen includes a display panel.

[0071] The above-mentioned electronic device realizes the subject name of the load prediction method for the campus microgrid of the present application, which is to obtain the historical load data of the campus microgrid and the relevant influencing factor data; the historical load data of the campus microgrid and the relevant influencing factor data are input into the pre-built Fredformer model to obtain the load prediction result of the campus microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer technical solution, which achieves the beneficial effect of preventing the model from ignoring the lower amplitude features that are important for accurate prediction and lowering the computational cost.

[0072] The storage medium provided in the present application stores a program product that can implement a load forecasting method for a campus microgrid.

[0073] The load forecasting method of the park microgrid includes: obtaining the historical load data of the park microgrid and the data of related influencing factors; inputting the historical load data of the park microgrid and the data of related influencing factors into the pre-built Fredformer model to obtain the load forecasting result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

[0074] In some possible embodiments, the subject matter of the present disclosure, namely, the method, system, electronic device, and storage medium for load forecasting of a campus microgrid, can be implemented in the form of a program product, which includes a program code. When the program product is run on a terminal device, the program code is used to enable the terminal device to execute the steps described in the above "Exemplary Method" section of this specification according to various exemplary embodiments of the present disclosure.

[0075] The storage medium of the present disclosure can adopt any combination of one or more readable media. The readable medium can be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of readable storage media include: an electrical connection with one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0076] The above description of the disclosed embodiments enables one skilled in the art to implement or use the present invention. Various modifications to these embodiments will be apparent to one skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A load forecasting method for a campus microgrid, characterized in that: include: Obtain historical load data of the park microgrid and related influencing factor data; The historical load data of the park microgrid and related influencing factor data are input into the pre-built Fredformer model to obtain the load forecast result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

2. The load forecasting method for a campus microgrid according to claim 1, characterized in that: The historical load data of the campus microgrid and related influencing factor data include but are not limited to time information, weather data, regional information and personal behavior related data.

3. The load forecasting method for a campus microgrid according to claim 1, characterized in that: After obtaining the historical load data of the park microgrid and the data of related influencing factors, the load forecasting method of the park microgrid further includes: The outliers and missing values ​​in the historical load data and the data of related influencing factors are removed and normalized; the historical load data and the data of related influencing factors that exceed the preset threshold are regarded as abnormal; the missing values ​​are processed by the mean filling and linear interpolation methods.

4. The load forecasting method for a campus microgrid according to claim 1, characterized in that: The discrete Fourier transform architecture layer in the Fredformer model is used to decompose the input time series data into frequency components; A frequency domain optimization layer for refining and normalizing frequency components to generate multiple sub-frequency bands; Transformer encoder layer, used to learn the frequency features of each sub-frequency band; Frequency summary layer, used to perform linear transformation on feature information; The inverse discrete Fourier transform layer is used to reconstruct the processed frequency output back into the time domain signal.

5. The load forecasting method for a campus microgrid according to claim 1, characterized in that: The historical load data of the park microgrid and the data of related influencing factors are input into the pre-built Fredformer model to obtain the load forecast results of the park microgrid, including: The historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data are input into the discrete Fourier transform architecture layer to decompose them into frequency components; The frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations through the frequency domain optimization layer to avoid mutual influence between different frequency components; Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing; After learning the features of each sub-frequency band, linear transformation is performed in the frequency summary layer and inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

6. The load forecasting method for a campus microgrid according to claim 5, characterized in that: The discrete Fourier transform architecture layer decomposes the historical load data of the park microgrid and related influencing factor data into frequency components according to the following formula: k=1,…,L; In the formula, represents the kth frequency component, L represents the lookback window length of each channel, Represents a multivariate time series.

7. A load forecasting system for a campus microgrid, characterized in that: include: An acquisition unit, used to acquire historical load data of the park microgrid and data of related influencing factors; The prediction unit is used to input the historical load data of the park microgrid and related influencing factor data into a pre-built Fredformer model to obtain the load prediction result of the park microgrid, wherein the Fredformer model includes a discrete Fourier transform architecture layer, a frequency domain optimization layer, a Transformer encoder layer, a frequency summary layer and an inverse discrete Fourier transform layer.

8. The load forecasting system for a campus microgrid according to claim 7, characterized in that: The prediction unit is specifically used for: The historical load data of the park microgrid and the time series corresponding to the relevant influencing factor data are input into the discrete Fourier transform architecture layer to decompose them into frequency components; The frequency components are divided into multiple sub-frequency bands using non-overlapping patch operations through the frequency domain optimization layer to avoid mutual influence between different frequency components; Deploy the Transformer encoder layer on the sub-frequency blocks after patch normalization of frequency domain features to independently learn the features of each block to focus on the correlation across channels in the same frequency band to achieve debiasing; After learning the features of each sub-frequency band, linear transformation is performed in the frequency summary layer and inverse discrete Fourier transform is performed in the inverse discrete Fourier transform layer to summarize the feature information and form the final prediction result.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the load forecasting method for the campus microgrid are implemented as described in any one of claims 1 to 6.

10. A storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the load forecasting method for a campus microgrid are implemented as described in any one of claims 1 to 6.

Images