Load prediction method and device, nonvolatile storage medium and computer equipment

Abstract

The invention discloses a load prediction method and device, a nonvolatile storage medium and computer equipment. The method comprises the following steps: acquiring a ciphertext corresponding to load data in a target area and a ciphertext corresponding to operation data of a charging station in the target area; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; extracting features in the decrypted load data and features in the decrypted operation data to obtain load data features and operation data features; and inputting the load data features and the operation data features into a federated learning model to obtain a load prediction result of the target transformer area. The technical problem that the prediction result is not accurate enough due to the fact that only load data of the transformer area is considered in load prediction is solved.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a load forecasting method, device, non-volatile storage medium and computer equipment. Background Art

[0002] At present, when conducting power grid planning and dispatching, it is necessary to predict load, but when predicting load, more attention is paid to the load at the substation dimension. In addition, when conducting research related to power grid load, data from multiple parties is often involved, but it is usually necessary to centrally store and process the original data from all parties, which may cause data privacy and security issues.

[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0004] The embodiments of the present invention provide a load forecasting method, apparatus, non-volatile storage medium and computer equipment to at least solve the technical problem that load forecasting only considers load data of a substation, resulting in inaccurate forecasting results.

[0005] According to one aspect of an embodiment of the present invention, a load forecasting method is provided, including: obtaining ciphertext corresponding to load data in a target substation and ciphertext corresponding to operation data of charging stations in the target substation; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features; inputting the load data features and operation data features into a federated learning model to obtain a load forecasting result for the target substation.

[0006] Optionally, the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data are decrypted to obtain the decrypted load data and the decrypted operation data, including: obtaining a first key for the ciphertext corresponding to the load data and a second key for the ciphertext corresponding to the operation data; decrypting the first key and the second key based on an asymmetric encryption algorithm to obtain a plaintext of the first key and a plaintext of the second key; using the plaintext of the first key to decrypt the ciphertext corresponding to the load data based on a key encryption algorithm to obtain the decrypted load data; using the plaintext of the second key to decrypt the ciphertext corresponding to the operation data based on a key encryption algorithm to obtain the decrypted operation data.

[0007] Optionally, extracting features from the decrypted load data and features from the decrypted operating data to obtain load data features and operating data features includes: determining missing values ​​in the decrypted load data and missing values ​​in the decrypted operating data; inputting time series data corresponding to the missing values ​​in the decrypted load data and time series data corresponding to the missing values ​​in the decrypted operating data into the generator network, respectively, to obtain padded load data and padded operating data; determining load data based on the padded load data and the decrypted load data; determining operating data based on the padded operating data and the decrypted operating data; extracting features from the load data and features from the operating data to obtain load data features and operating data features.

[0008] Optionally, features in the decrypted load data and features in the decrypted operation data are extracted to obtain load data features and operation data features, including: load data features include at least one of the following: load peak moment, load valley moment, and load capacity of the target substation in the historical process; operation data features include at least one of the following: peak moment, valley moment, and load capacity of the charging and discharging load of the electric vehicle corresponding to the charging station in the historical process.

[0009] Optionally, the federated learning model is obtained by training the original federated learning model using a sample training set, wherein the sample training set includes sample load data in the target area, sample operation data of the charging station, and sample load results corresponding to the sample load data and the sample operation data.

[0010] Optionally, the load forecast result includes at least one of the following: a maximum load value in the target substation, a daily load peak time and a daily load valley time in the target substation.

[0011] Optionally, the load forecast result is sent to a power supply service end corresponding to the target area and a service end corresponding to the charging station, wherein the power supply service end and the service end corresponding to the charging station regulate the power in the target area based on the load forecast result.

[0012] According to another aspect of an embodiment of the present invention, a load prediction device is also provided, including: an acquisition module, used to obtain the ciphertext corresponding to the load data in the target substation and the ciphertext corresponding to the operation data of the charging station in the target substation; a decryption module, used to decrypt the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data; an extraction module, used to extract the features in the decrypted load data and the features in the decrypted operation data to obtain the load data features and the operation data features; a prediction module, used to input the load data features and the operation data features into a federated learning model to obtain the load prediction result of the target substation.

[0013] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided. The non-volatile storage medium includes a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute any one of the above-mentioned load prediction methods.

[0014] According to another aspect of the embodiments of the present invention, a computer device is provided. The computer device includes a processor, and the processor is used to run a program. When the program is run, any one of the above-mentioned load forecasting methods is executed.

[0015] According to another aspect of an embodiment of the present invention, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, any one of the above load forecasting methods is implemented.

[0016] In an embodiment of the present invention, a load prediction method is adopted, by obtaining the ciphertext corresponding to the load data in the target substation and the ciphertext corresponding to the operation data of the charging station in the target substation; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data; extracting the features in the decrypted load data and the features in the decrypted operation data to obtain the load data features and the operation data features; inputting the load data features and the operation data features into a federated learning model to obtain the load prediction result of the target substation, thereby achieving the purpose of predicting the load of the target substation in combination with the data of the charging station, thereby realizing the technical effect of improving the accuracy of the predicted load results, and further solving the technical problem that the load prediction only considers the load data of the substation, resulting in inaccurate prediction results. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a load forecasting method is shown;

[0019] Figure 2 is a flow chart of a load forecasting method provided according to an embodiment of the present invention;

[0020] Figure 3 It is a structural block diagram of a load prediction device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0023] According to an embodiment of the present invention, a method embodiment of a load forecasting method is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0024] The method embodiment provided in the first embodiment of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computer terminal for implementing a load forecasting method. Figure 1 As shown, the computer terminal 10 may include one or more (102a, 102b, ..., 102n are used to illustrate) processors (the processor may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1Different configurations are shown.

[0025] It should be noted that the one or more processors and / or other data processing circuits described above may generally be referred to herein as "data processing circuits". The data processing circuits may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuit may be a single independent processing module, or may be incorporated in whole or in part into any of the other components in the computer terminal 10. As described in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0026] The memory 104 can be used to store software programs and modules of application software, such as program instructions / data storage devices corresponding to the load forecasting method in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, realizing the load forecasting method of the above-mentioned application program. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0027] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 .

[0028] Figure 2 is a flow chart of a load forecasting method according to an embodiment of the present invention. Figure 2 As shown, the method comprises the following steps:

[0029] Step S202, obtaining the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area.

[0030] In this step, the load data in the target area can be the power data provided by the power supplier in the target area, including area archive data, load conditions, etc. The operation data of the charging station can come from the operator of the charging station, wherein the operation data can include the capacity of the charging station, the charging load and other data. In order to avoid data leakage and the like during data transmission, the ciphertext corresponding to the load data and the operation data can be obtained to avoid the security risks caused by data leakage. The charging mode of the electric vehicle corresponding to the charging station can be divided into disordered charging, unidirectional orderly charging and bidirectional orderly charging and discharging, that is, different working modes of the charging system. Among them, disordered charging means that when the electric vehicle is charging, the load of the power grid or the charging needs of other electric vehicles are not considered, that is, the electric vehicle can be connected to the charging pile for charging at any time without prior arrangement or scheduling, and the user can go to the charging station for charging at any time when needed. Unidirectional orderly charging means that the charging system is scheduled according to the load condition or other priorities of the power grid to ensure efficiency and balance during charging. In this case, the electric vehicle needs to make an appointment for charging time in advance, and the charging pile will arrange the charging order according to the appointment condition and the load condition of the power grid to avoid overload or other problems of the power grid. Bidirectional orderly charging and discharging can not only charge the electric vehicle, but also allow the electric vehicle to discharge to the power grid as an energy storage device. In this way, electric vehicles can release stored electricity during peak hours according to grid demand to balance grid load or provide backup energy. This charging method requires charging piles with bidirectional energy conversion function and electric vehicles that support bidirectional energy transmission, commonly known as V2G technology. Charging station operators can count the disordered charging load data, unidirectional orderly charging load data, bidirectional orderly charging load data, and bidirectional orderly discharge load data of their operating charging stations based on charging piles.

[0031] Step S204, decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data.

[0032] In this step, the ciphertext corresponding to the acquired load data and the ciphertext corresponding to the operation data can be decrypted. Specifically, the load data and the operation data can be encrypted using the RSA encryption algorithm (asymmetric encryption algorithm) and the DES encryption algorithm (key encryption algorithm). Specifically, the obtained ciphertext can generate a random number as a key for each participant (power data party, charging station operator), and the key will be used to encrypt and match the area field. Each participant uses the Shamir secret sharing technology to divide its key into multiple shares and distribute these shares to other participants. Then each participant uses the key share it holds to encrypt its own area field. An encryption method combining the RSA encryption algorithm with the DES algorithm is used. The plaintext data can be encrypted with the DES algorithm, and then the DES key is encrypted with RSA, and the ciphertext and the encrypted key are sent to the recipient together. That is, to decrypt the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data, the decryption key can be used first. Since the encryption and decryption keys of DES are the same, the DES key is obtained, and then the ciphertext is decrypted to obtain the plaintext.

[0033] Step S206, extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features.

[0034] In this step, features can be extracted from the decrypted load data and operation data. Among them, the power data side, that is, the features used for the daily maximum load prediction of the target area in the load data, can include the maximum value, minimum value, mean, variance, kurtosis, skewness, 75% percentile of the daily load of the area, and the capacity of the area, and the daily average temperature of the area. The power data side, that is, the features used for the daily load peak time and daily load valley time of the area in the load data, can include the peak time, valley time, peak period, valley period of the daily load of the area, and the capacity of the area, and the daily average temperature of the area. The features used for the daily maximum load prediction of the area in the operation data of the charging station operator, that is, the charging station, are: the maximum value, minimum value, mean, variance, kurtosis, skewness, 75% percentile of the daily dimension of the four types of electric vehicle charging and discharging loads (disordered charging load, unidirectional ordered charging load, bidirectional ordered charging load, and bidirectional ordered discharging load) of the area, and the capacity of the charging station in the area. The characteristics of the daily load peak time and daily load valley time in the charging station operator, that is, the charging station operation data, include: the peak time, valley time, peak period, and valley period of the four types of electric vehicle charging and discharging loads in the area (disordered charging load, unidirectional ordered charging load, bidirectional ordered charging load, and bidirectional ordered discharging load) in the daily dimension, and the capacity of the charging stations in the area.

[0035] Step S208: input the load data characteristics and the operation data characteristics into the federated learning model to obtain the load forecast result of the target substation.

[0036] In this step, the federated learning model is a distributed machine learning model that aims to train machine learning models without centralizing all data in a single location. For example, in user devices or different data centers, the model is only updated locally, and then the encrypted model parameters are aggregated to the central server to obtain a global model update, which can effectively protect data privacy. By inputting the load data characteristics and operation data characteristics into the federated learning model, the load forecast results of the target substation can be obtained. The goal of the federated learning model here is to jointly optimize the load forecasting model through the data characteristics of both parties, and realize the accurate prediction of the daily maximum load of the substation that integrates the data of both the power data party and the charging station operator. The use of a federated learning model can improve the security of the prediction process and avoid data leakage.

[0037] Through the above steps, the purpose of predicting the load of the target area by combining the data of the charging station is achieved, thereby achieving the technical effect of improving the accuracy of the predicted load results, and further solving the technical problem that the load prediction only considers the load data of the area, resulting in inaccurate prediction results.

[0038] As an optional embodiment, decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data includes: obtaining a first key for the ciphertext corresponding to the load data and a second key for the ciphertext corresponding to the operation data; decrypting the first key and the second key based on an asymmetric encryption algorithm to obtain a plaintext of the first key and a plaintext of the second key; using the plaintext of the first key to decrypt the ciphertext corresponding to the load data based on a key encryption algorithm to obtain the decrypted load data; using the plaintext of the second key to decrypt the ciphertext corresponding to the operation data based on a key encryption algorithm to obtain the decrypted operation data.

[0039] Optionally, the load data and operation data can be encrypted based on the RSA encryption algorithm (asymmetric encryption algorithm) and the DES encryption algorithm (key encryption algorithm). For example, the obtained ciphertext can generate a random number as a key for each participant (power data party, charging station operator), and the key will be used to encrypt and match the area field. Each participant uses the Shamir secret sharing technique to divide its key into multiple shares and distribute these shares to other participants. Then each participant uses the key share it holds to encrypt its own area field. An encryption method combining the RSA encryption algorithm with the DES algorithm is used. The plaintext data can be encrypted with the DES algorithm, and then the DES key is encrypted with RSA, and the ciphertext and the encrypted key are sent to the recipient together. Correspondingly, the decryption process can be for the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data. The decryption key can be used first. Since the encryption and decryption keys of DES are the same, the DES key is obtained, and then the ciphertext is decrypted to obtain the plaintext.

[0040] The load prediction here can be based on the federated learning model, that is, there is no need to concentrate the data in the same place, and the model can be calculated and updated in different data centers, and finally summarized to the central server. In the encryption process, each participant (power data party, charging station operator) generates a random number as a key, which will be used to encrypt and match the area field. Each participant uses the Shamir secret sharing technology to divide its key into multiple shares and distribute these shares to other participants. Each participant uses the key share it holds to encrypt its own area field. An encryption method combining the RSA encryption algorithm with the DES algorithm is used. Each participant uses the MPC protocol (multi-party computing protocol) to calculate on the key share it holds to check whether there is a matching area field. Then determine the matching sample, and only when both parties agree can the matching of the area field be determined. If the calculation results of both parties indicate that there is a matching area, the matching area can be determined. Otherwise, no information about the match will be leaked.

[0041] As an optional embodiment, extracting features from the decrypted load data and features from the decrypted operating data to obtain load data features and operating data features includes: determining missing values ​​in the decrypted load data and missing values ​​in the decrypted operating data; inputting time series data corresponding to the missing values ​​in the decrypted load data and time series data corresponding to the missing values ​​in the decrypted operating data into the generator network respectively to obtain padded load data and padded operating data; determining load data based on the padded load data and the decrypted load data; determining operating data based on the padded operating data and the decrypted operating data; extracting features from the load data and features from the operating data to obtain load data features and operating data features.

[0042] Optionally, in the process of collecting data, due to various reasons such as communication failure, equipment failure, data transmission error, environmental factors, etc., there will generally be more missing values ​​in the acquired data, which will affect the prediction effect. The generative adversarial network model architecture can be adopted, and the generator can be constructed using the long short-term memory network and the recurrent gating unit. In this way, the powerful generation ability of the generative adversarial network and the advantages of the recurrent neural network in processing time and sequence data can be combined to achieve accurate filling of missing values ​​in the load data and improve data quality. The time series data corresponding to the missing values ​​can be input into the generator network to obtain filled operation data and filled load data. Specifically, the load data containing missing values ​​can be preprocessed first, and the missing values ​​can be marked as a specific placeholder NaN. Then the generator network is designed, and the time series data of the missing values ​​are respectively used as input, and the recurrent neural network structure such as LSTM (long short-term memory network) or GRU (recurrent gating unit) is used to generate complete data with filled missing values. The goal of the generator is to generate realistic data so that the filled data is as close to the real data as possible. Then the discriminator network can be designed to distinguish the generated filled data from the real complete data. The goal of the discriminator is to identify as accurately as possible whether the generated data is real data, and then use it to judge the quality of the filled data. The generator network and the discriminator network are combined for adversarial training. The generator tries to generate realistic data to deceive the discriminator, while the discriminator tries to distinguish between generated data and real data. Through alternating training, the generator and the discriminator gradually improve their capabilities. After the generator training is completed, the generator network is used to fill in the load data with missing values. The generator receives the load data with missing values ​​as input and generates complete load data with the missing values ​​filled in.

[0043] As an optional embodiment, features in the decrypted load data and features in the decrypted operation data are extracted to obtain load data features and operation data features, including: the load data features include at least one of the following: load peak time, load valley time, and load capacity of the target area in the historical process; the operation data features include at least one of the following: peak time, valley time, and load capacity of the charging and discharging load of the electric vehicle corresponding to the charging station in the historical process.

[0044] Optionally, the power data side, that is, the decrypted load data, has the following features for predicting the daily maximum load of the substation: the daily maximum value, minimum value, mean, variance, kurtosis, skewness, 75% percentile of the substation load, and the substation capacity, and the daily average temperature of the substation. The data dimension is the daily dimension. The features for the daily load peak time and daily load valley time of the substation are: the daily peak time, valley time, peak period, valley period of the substation load, and the substation capacity, and the daily average temperature of the substation. The data dimension is the daily dimension, and the load data of the past 30 days can be selected. The charging station operator, that is, the decrypted operation data, has the following features for predicting the daily maximum load of the substation: the maximum value, minimum value, mean, variance, kurtosis, skewness, 75% percentile of the daily dimension of the four types of electric vehicle charging and discharging loads in the substation (disordered charging load, unidirectional ordered charging load, bidirectional ordered charging load, and bidirectional ordered discharging load), and the capacity of the charging stations in the substation. The characteristics used for the daily load peak time and daily load valley time of the substation are: the peak time, valley time, peak period, and valley period of the four types of electric vehicle charging and discharging loads in the substation (disordered charging load, unidirectional ordered charging load, bidirectional ordered charging load, and bidirectional ordered discharging load) in the substation in the daily dimension, and the charging station capacity in the substation. Among them, the load kurtosis is used to describe the peak degree of the load data distribution. The load skewness is used to describe the skewness of the load data distribution. The load percentile is used to describe the value at a certain percentage in the load data, where the 75% load percentile means that 75% of the data in the load data is less than or equal to this value, and 25% of the data is greater than this value. The peak period refers to the high load interval in the load data, and the peak period is determined according to the 95% percentile of the load data. The valley period refers to the low load interval in the load data, and the valley period is determined according to the 5% percentile of the load data.

[0045] As an optional embodiment, the federated learning model is obtained by training the original federated learning model using a sample training set, wherein the sample training set includes sample load data in the target area, sample operation data of the charging station, and sample load results corresponding to the sample load data and the sample operation data.

[0046] Optionally, the original federated learning model can be trained using sample load data, sample operation data of charging stations, and labels. Among them, the label can be the time corresponding to the sample load data and sample operation data, the load change situation, and the corresponding future sample data. Specifically, using the federated learning model to predict the load results, the power data party can be defined as the active party (Active Party), which holds both data and labels; the charging station operator is defined as the passive party (Passive Party), which has data but no labels. Since label information is essential for supervised learning, there must be an active party that can access the label Lable. In federated learning, the active party, the power data party, assumes the responsibility of the main server. In the process of building a new tree, the power data party directly calculates the information gain of the split point for the features on its own side because it does not need to interact with the charging station operator or the vehicle platform, without synchronizing to the other two parties. For the features of the charging station operator, the power data party needs to calculate the first-order and second-order derivatives of the loss function based on the current predicted value and Label, and semi-homomorphically encrypt the first-order and second-order derivatives (using the Paillier library), and then send them to the party to which the features belong. Then, each party builds a gradient (first-order and second-order derivative) histogram based on its own feature data, and the charging station operator sends the encrypted gradient histogram to the power data party. At this time, the power data party has the histogram information of all the features and can use the semi-homomorphic encryption method to decrypt the histogram sent by the charging station operator. After the power data party completes the decryption of the other party's gradient histogram, it has the gradient histogram of all the features of the two parties. According to the split gain calculation formula, each feature is enumerated based on the histogram to calculate the optimal solution and find the global optimal split point. If the optimal split point belongs to the power data party, there is no need to pass the split information to the other party. If the optimal split point belongs to the charging station operator, the power data party needs to return the split information (excluding information gain) to the charging station operator for analysis. The party with the optimal split point splits the sample on the tree node into two left and right sub-data sets, and sends the split result to the other two parties as the split sample of the child node. The power data party can calculate the predicted value of the leaf node according to the calculation formula; but it will not be synchronized, so the charging station operator cannot know the weight of the leaf node. According to the predicted value of the leaf node, the predicted value of the maximum load of the substation on day T is optimized. The use of the federated learning model can ensure that data privacy and security issues are avoided.

[0047] As an optional embodiment, the load forecast result includes at least one of the following: a maximum load value in the target area, a daily load peak time and a daily load valley time in the target area.

[0048] As an optional embodiment, the load forecast result is sent to the power supply service end corresponding to the target area and the service end corresponding to the charging station, wherein the power supply service end and the service end corresponding to the charging station regulate the power in the target area based on the load forecast result.

[0049] Optionally, the load forecast results can be sent to the power supply service end and the service end corresponding to the charging station, and the power supply service end and the service end of the charging station can adjust the power based on the load forecast results. For example, according to the peak time in the load forecast results, it can be considered to appropriately adjust the power price at that time to achieve the effect of reducing the load at the peak time, avoid excessive load and ensure the smooth operation of the power grid.

[0050] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the present invention.

[0051] Through the description of the above implementation methods, those skilled in the art can clearly understand that the load forecasting method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a magnetic disk, or an optical disk), and includes a number of instructions for enabling a terminal device (which can be a mobile phone, a computer, a server, or a network device, etc.) to execute the methods described in each embodiment of the present invention.

[0052] According to an embodiment of the present invention, a load forecasting device for implementing the above load forecasting method is also provided. Figure 3 is a structural block diagram of a load prediction device provided according to an embodiment of the present invention. Figure 3 As shown, the load prediction device includes: an acquisition module 32, a decryption module 34, an extraction module 36 and a prediction module 38. The load prediction device is described below.

[0053] The acquisition module 32 is used to acquire the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area.

[0054] The decryption module 34 is connected to the acquisition module 32 and is used for decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data.

[0055] The extraction module 36 is connected to the decryption module 34 and is used to extract features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features.

[0056] The prediction module 38 is connected to the extraction module 36 and is used to input the load data characteristics and the operation data characteristics into the federated learning model to obtain the load prediction result of the target substation.

[0057] It should be noted that the acquisition module 32, decryption module 34, extraction module 36 and prediction module 38 correspond to steps S202 to S208 in the embodiment, and the examples and application scenarios implemented by the multiple modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment. It should be noted that the above modules, as part of the device, can be run in the computer terminal 10 provided in the embodiment.

[0058] An embodiment of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one network device among multiple network devices of a computer network. The computer device includes a memory and a processor.

[0059] Among them, the memory can be used to store software programs and modules, such as program instructions / modules corresponding to the load forecasting method and device in the embodiment of the present invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned load forecasting method. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories can be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0060] The processor can call the information and application programs stored in the memory through the transmission device to execute the following steps: obtain the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area; decrypt the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data; extract the features in the decrypted load data and the features in the decrypted operation data to obtain the load data features and the operation data features; input the load data features and the operation data features into the federated learning model to obtain the load prediction result of the target area.

[0061] Optionally, the processor may also execute the following program code: decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the running data to obtain the decrypted load data and the decrypted running data, including: obtaining the first key of the ciphertext corresponding to the load data and the second key of the ciphertext corresponding to the running data; decrypting the first key and the second key based on an asymmetric encryption algorithm to obtain the plaintext of the first key and the plaintext of the second key; using the plaintext of the first key to decrypt the ciphertext corresponding to the load data based on the key encryption algorithm to obtain the decrypted load data; using the plaintext of the second key to decrypt the ciphertext corresponding to the running data based on the key encryption algorithm to obtain the decrypted running data.

[0062] Optionally, the processor may also execute the following program code: extracting features from the decrypted load data and features from the decrypted operating data to obtain load data features and operating data features, including: determining missing values ​​in the decrypted load data and missing values ​​in the decrypted operating data; inputting the time series data corresponding to the missing values ​​in the decrypted load data and the time series data corresponding to the missing values ​​in the decrypted operating data into the generator network respectively to obtain filled load data and filled operating data; determining load data based on the filled load data and the decrypted load data; determining operating data based on the filled operating data and the decrypted operating data; extracting features from the load data and features from the operating data to obtain load data features and operating data features.

[0063] Optionally, the processor may also execute the program code of the following steps: extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features, including: the load data features include at least one of the following: load peak moments, load valley moments, and load capacity of the target area in the historical process; the operation data features include at least one of the following: peak moments, valley moments, and load capacity of the charging and discharging loads of the electric vehicles corresponding to the charging station in the historical process.

[0064] Optionally, the processor may also execute the program code of the following steps: the federated learning model is obtained by training the original federated learning model using a sample training set, wherein the sample training set includes sample load data in the target area, sample operation data of the charging station, and sample load results corresponding to the sample load data and the sample operation data.

[0065] Optionally, the processor may also execute program code of the following steps: the load forecast result includes at least one of the following: the maximum load value in the target area, the daily load peak time and the daily load valley time in the target area.

[0066] Optionally, the processor may also execute the program code of the following steps: sending the load forecast result to the power supply service end corresponding to the target area and the service end corresponding to the charging station, wherein the power supply service end and the service end corresponding to the charging station regulate the power in the target area based on the load forecast result.

[0067] The embodiment of the present invention provides a solution, thereby achieving the purpose and solving the technical problems in the related art.

[0068] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a non-volatile storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0069] The embodiment of the present invention further provides a non-volatile storage medium. Optionally, in this embodiment, the non-volatile storage medium can be used to store the program code executed by the load prediction method provided in the above embodiment.

[0070] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group.

[0071] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: obtaining the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features; inputting the load data features and operation data features into the federated learning model to obtain the load forecast result for the target area.

[0072] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data, including: obtaining a first key for the ciphertext corresponding to the load data and a second key for the ciphertext corresponding to the operation data; decrypting the first key and the second key based on an asymmetric encryption algorithm to obtain a plaintext of the first key and a plaintext of the second key; using the plaintext of the first key to decrypt the ciphertext corresponding to the load data based on a key encryption algorithm to obtain the decrypted load data; using the plaintext of the second key to decrypt the ciphertext corresponding to the operation data based on a key encryption algorithm to obtain the decrypted operation data.

[0073] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: extracting features from the decrypted load data and features from the decrypted operating data to obtain load data features and operating data features, including: determining missing values ​​in the decrypted load data and missing values ​​in the decrypted operating data; inputting time series data corresponding to the missing values ​​in the decrypted load data and time series data corresponding to the missing values ​​in the decrypted operating data into the generator network respectively to obtain padded load data and padded operating data; determining load data based on the padded load data and the decrypted load data; determining operating data based on the padded operating data and the decrypted operating data; extracting features from the load data and features from the operating data to obtain load data features and operating data features.

[0074] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features, including: the load data features include at least one of the following: load peak moments, load valley moments, and load capacity of the target area in the historical process; the operation data features include at least one of the following: peak moments, valley moments, and load capacity of the charging and discharging loads of the electric vehicles corresponding to the charging station in the historical process.

[0075] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: the federated learning model is obtained by training the original federated learning model using a sample training set, wherein the sample training set includes sample load data in the target area, sample operation data of the charging station, and sample load results corresponding to the sample load data and the sample operation data.

[0076] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: the load forecast result includes at least one of the following: the maximum load value in the target area, the daily load peak time and the daily load valley time in the target area.

[0077] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for executing the following steps: sending the load forecast result to the power supply service end corresponding to the target area and the service end corresponding to the charging station, wherein the power supply service end and the service end corresponding to the charging station regulate the power in the target area based on the load forecast result.

[0078] An embodiment of the present invention also provides a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can achieve: obtaining the ciphertext corresponding to the load data in the target substation and the ciphertext corresponding to the operation data of the charging station in the target substation; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features; inputting the load data features and operation data features into a federated learning model to obtain a load forecast result for the target substation.

[0079] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0080] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0081] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0082] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0083] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0084] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0085] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A load forecasting method, characterized in that: include: Obtaining the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area; decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; Extracting features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features; The load data characteristics and the operation data characteristics are input into a federated learning model to obtain a load forecast result for the target substation.

2. The method according to claim 1, characterized in that The decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain the decrypted load data and the decrypted operation data includes: Obtaining a first key of the ciphertext corresponding to the load data and a second key of the ciphertext corresponding to the operation data; Decrypting the first key and the second key based on an asymmetric encryption algorithm to obtain a plaintext of the first key and a plaintext of the second key; Decrypting the ciphertext corresponding to the payload data using the plaintext of the first key based on a key encryption algorithm to obtain the decrypted payload data; The plain text of the second key is used to decrypt the ciphertext corresponding to the operation data based on the key encryption algorithm to obtain the decrypted operation data.

3. The method according to claim 1, characterized in that The extracting the features in the decrypted load data and the features in the decrypted operation data to obtain the load data features and the operation data features includes: determining missing values ​​in the decrypted load data and missing values ​​in the decrypted operation data; Inputting the time series data corresponding to the missing values ​​in the decrypted load data and the time series data corresponding to the missing values ​​in the decrypted operation data into the generator network respectively to obtain the filled load data and the filled operation data; determining the payload data according to the padded payload data and the decrypted payload data; determining the operating data according to the padded operating data and the decrypted operating data; The features in the load data and the features in the operation data are extracted to obtain the load data features and the operation data features.

4. The method according to claim 1, characterized in that: The extracting of features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features includes: the load data features include at least one of the following: load peak moments, load valley moments, and load capacity of the target area in the historical process; the operation data features include at least one of the following: peak moments, valley moments, and load capacity of the charging and discharging loads of the electric vehicles corresponding to the charging station in the historical process.

5. The method according to claim 1, characterized in that The federated learning model is obtained by training the original federated learning model using a sample training set, wherein the sample training set includes sample load data in the target area, sample operation data of the charging station, and sample load results corresponding to the sample load data and the sample operation data.

6. The method according to any one of claims 1 to 5, characterized in that: The load forecast result includes at least one of the following: the maximum load value in the target substation, the daily load peak time and the daily load valley time in the target substation.

7. The method according to any one of claims 1 to 5, characterized in that: Also includes: The load forecast result is sent to a power supply service end corresponding to the target area and a service end corresponding to the charging station, wherein the power supply service end and the service end corresponding to the charging station regulate the power in the target area based on the load forecast result.

8. A load forecasting device, characterized in that: include: An acquisition module, used to acquire the ciphertext corresponding to the load data in the target area and the ciphertext corresponding to the operation data of the charging station in the target area; A decryption module, used for decrypting the ciphertext corresponding to the load data and the ciphertext corresponding to the operation data to obtain decrypted load data and decrypted operation data; An extraction module, used to extract features from the decrypted load data and features from the decrypted operation data to obtain load data features and operation data features; The prediction module is used to input the load data characteristics and the operation data characteristics into the federated learning model to obtain the load prediction result of the target substation.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute the load prediction method according to any one of claims 1 to 7.

10. A computer device, characterized in that: include: Memory and processor, The memory stores a computer program; The processor is used to execute the computer program stored in the memory, and when the computer program is running, the processor executes the load forecasting method according to any one of claims 1 to 7.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the load forecasting method according to any one of claims 1 to 7 is implemented.

Images