Electrical load prediction method and device for multiple industries, and storage medium

Abstract

The invention discloses an electrical load prediction method and device for multiple industries and a storage medium, and the method comprises the steps: collecting an electrical load curve corresponding to each industry, and determining a plurality of first electrical load vectors corresponding to each historical moment; determining a first electrical load influence factor corresponding to each industry, and determining a plurality of first influence factor vectors corresponding to each historical moment; inputting the plurality of first electrical load vectors into a first LSTM model, outputting a plurality of second electrical load vectors, inputting the plurality of first influence factor vectors into a second LSTM model, and outputting a plurality of second influence factor vectors; based on the plurality of second electrical load vectors and the plurality of second influence factor vectors, predicting a maximum electrical load moment by using a preset prediction model; and determining a confidence interval of the maximum electrical load according to the maximum electrical load moment and by using the plurality of electrical load curves corresponding to the industries.

Description

Technical Field

[0001] The present application relates to the technical field of electricity forecasting, and in particular to a method, device and storage medium for forecasting electricity load for multiple industries. Background Art

[0002] At present, with the continuous expansion of electricity consumption, accurate prediction of electricity load plays a vital role in the normal operation of the power system. Electricity load forecasting technology determines the electricity load data at a specific time or time period in the future while meeting certain accuracy requirements. That is, it infers future values ​​based on historical electricity load data.

[0003] Although, with the continuous maturity and widespread application of machine learning technology, relevant models for power load forecasting in various industries have been constructed, and the power load of various industries have been predicted accordingly using relevant models. However, since the existing technology predicts the power load of each industry separately, and does not predict the power load of multiple industries, the existing technology can only use the maximum power load of each industry to predict the maximum power load of multiple industries, and cannot directly predict the maximum power load of multiple industries. As a result, it may be impossible to accurately determine the maximum power load of multiple industries.

[0004] With regard to the technical problem that the above-mentioned prior art can only predict the maximum power load of multiple industries by using the maximum power load of each industry, but cannot directly predict the maximum power load of multiple industries, resulting in the inability to accurately determine the maximum power load of multiple industries, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present disclosure provide a method, device and storage medium for predicting electricity load for multiple industries, so as to at least solve the technical problem in the prior art that the maximum electricity load of multiple industries can only be predicted by using the maximum electricity load of each industry, but the maximum electricity load of multiple industries cannot be directly predicted, resulting in the inability to accurately determine the maximum electricity load of multiple industries.

[0006] According to one aspect of an embodiment of the present disclosure, a method for predicting power load for multiple industries is provided, including: collecting power load curves corresponding to each industry, and determining multiple first power load vectors corresponding to each historical moment based on multiple power load curves corresponding to each industry, wherein the first power load vector corresponds to the power load of each industry; determining the first power load influencing factor corresponding to each industry, and determining multiple first influencing factor vectors corresponding to each historical moment based on the multiple first power load influencing factors corresponding to each industry, wherein the first influencing factor vector includes the first power load influencing factor corresponding to each industry; inputting the multiple first power load vectors into a first LSTM model, and outputting multiple second power load vectors, inputting the multiple first influencing factor vectors into a second LSTM model, and outputting multiple second influencing factor vectors; predicting the maximum power load moment based on the multiple second power load vectors and the multiple second influencing factor vectors and using a pre-set prediction model; and determining the confidence interval of the maximum power load according to the maximum power load moment and using the multiple power load curves corresponding to each industry.

[0007] According to another aspect of the embodiment of the present disclosure, there is also provided a power load prediction device for multiple industries, including: a first power load vector determination module, used to collect power load curves corresponding to each industry, and based on the multiple power load curves corresponding to each industry, determine multiple first power load vectors corresponding to each historical moment, wherein the first power load vector corresponds to the power load of each industry; a first influencing factor vector determination module, used to determine the first power load influencing factor corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determine multiple first influencing factor vectors corresponding to each historical moment, wherein the first influencing factor vector The invention discloses a method for predicting the moment of maximum power load based on the plurality of second power load vectors and the plurality of second power load vectors, and a confidence interval determination module, which is used to determine the confidence interval of the maximum power load according to the moment of maximum power load and using the plurality of power load curves corresponding to the respective industries.

[0008] According to another aspect of an embodiment of the present disclosure, a storage medium is further provided, the storage medium including a stored program, wherein when the program is running, a processor executes any one of the methods described above.

[0009] According to another aspect of the embodiment of the present disclosure, there is also provided a power load prediction device for multiple industries, including: a processor; and a memory connected to the processor, for providing the processor with instructions for processing the following processing steps: collecting power load curves corresponding to each industry, and based on the multiple power load curves corresponding to each industry, determining multiple first power load vectors corresponding to each historical moment, wherein the first power load vector corresponds to the power load of each industry; determining the first power load influencing factor corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determining multiple first influencing factor vectors corresponding to each historical moment, wherein the first influencing factor vector includes the first power load influencing factor corresponding to each industry; inputting the multiple first power load vectors into a first LSTM model, and outputting multiple second power load vectors, inputting the multiple first influencing factor vectors into a second LSTM model, and outputting multiple second influencing factor vectors; based on the multiple second power load vectors and the multiple second influencing factor vectors, and using a pre-set prediction model, predicting the maximum power load moment; and determining the confidence interval of the maximum power load according to the maximum power load moment and using the multiple power load curves corresponding to each industry.

[0010] The present application provides a method for predicting power load for multiple industries. First, the processor collects power load curves corresponding to each industry, and determines multiple first power load vectors corresponding to each historical moment based on multiple power load curves corresponding to each industry. Then, the processor determines the first power load influencing factors corresponding to each industry, and determines multiple first influencing factor vectors corresponding to each historical moment based on multiple first power load influencing factors corresponding to each industry. Further, the processor inputs multiple first power load vectors into a first LSTM model and outputs multiple second power load vectors, and the processor inputs multiple first influencing factor vectors into a second LSTM model and outputs multiple second influencing factor vectors. After that, the processor predicts the maximum power load moment based on multiple second power load vectors and multiple second influencing factor vectors, and uses a pre-set prediction model. Finally, the processor determines the confidence interval of the maximum power load according to the maximum power load moment and using multiple power load curves corresponding to each industry.

[0011] Since the present application collects the power load curves corresponding to each industry and determines multiple first power load vectors corresponding to each historical moment (where the first power load vector corresponds to the power load of each industry), the maximum power load moment predicted by the final prediction model corresponds to multiple industries. And finally, based on the maximum power load moment and the power load curves corresponding to each industry, the confidence interval of the maximum power load corresponding to multiple industries can be determined. Therefore, unlike the prior art, the present application can achieve the technical effect of improving the prediction efficiency and accurately determining the maximum power load of multiple industries by directly predicting the maximum power load of multiple industries.

[0012] In addition, since the present application determines the first electricity load influencing factor corresponding to each industry, and simultaneously uses multiple first influencing factor vectors and multiple first electricity load vectors corresponding to each historical moment to predict the confidence interval of the maximum electricity load corresponding to multiple industries, the final prediction result is associated with the first electricity load influencing factor of each industry (that is, the electricity load influencing factor with a higher degree of correlation with each industry), which further improves the accuracy of the prediction.

[0013] This solves the technical problem that the existing technology can only use the maximum power load of each industry to predict the maximum power load of multiple industries, but cannot directly predict the maximum power load of multiple industries, resulting in the inability to accurately determine the maximum power load of multiple industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] The drawings described herein are used to provide a further understanding of the present disclosure and constitute a part of the present application. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure. In the drawings:

[0015] Figure 1 is a hardware structure block diagram of a computing device for implementing the method according to Embodiment 1 of the present application;

[0016] Figure 2 is a schematic diagram of a power load forecasting system for multiple industries according to Example 1 of the present application;

[0017] Figure 3 is a flow chart of the power load forecasting method for multiple industries according to Example 1 of the present application;

[0018] Figure 4 is a schematic diagram of the prediction model according to Example 1 of the present application;

[0019] Figure 5 is a schematic diagram of a power load forecasting system for multiple industries according to Example 2 of the present application;

[0020] Figure 6 It is a schematic diagram of the electricity load forecasting system for multiple industries described in Example 3 of the present application. DETAILED DESCRIPTION

[0021] In order to enable those skilled in the art to better understand the technical solutions of the present disclosure, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only embodiments of a part of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without making creative work should fall within the scope of protection of the present disclosure.

[0022] It should be noted that the terms "first", "second", etc. in the specification and claims of the present disclosure 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 interchangeable where appropriate, so that the embodiments of the present disclosure 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] Example 1

[0024] According to this embodiment, a method embodiment for electricity load forecasting for multiple industries 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.

[0025] The method embodiment provided in this embodiment can be executed in a mobile terminal, a computer terminal, a server or a similar computing device. Figure 1 FIG. 1 shows a hardware structure block diagram of a computing device for implementing a method for predicting power load for multiple industries. Figure 1As shown, the computing device may include one or more 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 for storing data, a transmission device for communication functions, and an input / output interface. The memory, the transmission device, and the input / output interface are connected to the processor via a bus. In addition, it may also include: a display, a keyboard, and a cursor control device connected to the input / output interface. A person skilled in the art can understand 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 1 Different configurations shown.

[0026] 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 circuitry". The data processing circuitry may be embodied in whole or in part as software, hardware, firmware, or any other combination thereof. In addition, the data processing circuitry may be a single independent processing module, or may be incorporated in whole or in part into any of the other components of the computing device. As involved in the embodiments of the present disclosure, the data processing circuitry acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0027] The memory can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the power load forecasting method for multiple industries in the embodiment of the present disclosure. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, the power load forecasting method for multiple industries of the above-mentioned application program is realized. 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 may be connected to the computing device 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.

[0028] The transmission device is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computing device. In one example, the transmission device includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device can be a radio frequency (Radio Frequency, RF) module, which is used to communicate with the Internet wirelessly.

[0029] The display may be, for example, a touch screen liquid crystal display (LCD) that may enable a user to interact with a user interface of the computing device.

[0030] It should be noted that, in some optional embodiments, the above Figure 1 The computing device shown may include hardware elements (including circuits), software elements (including computer code stored on a computer-readable medium), or a combination of both hardware elements and software elements. Figure 1 This is merely one example of a particular embodiment and is intended to illustrate the types of components that may be present in the computing devices described above.

[0031] Figure 2 Schematic diagram of the power load forecasting system for multiple industries according to this embodiment. Figure 2 As shown, the system includes: a data acquisition module, a first feature extraction module, a second feature extraction module, a power load prediction module and a confidence interval determination module.

[0032] The data acquisition module is connected to the first feature extraction module and the second feature extraction module respectively, and is used to collect the power load curve corresponding to each industry from the data center, and determine multiple first influencing factor vectors corresponding to each historical moment. The data acquisition module is also used to determine the first power load influencing factor corresponding to each historical moment from the data center, and determine multiple first influencing factor vectors corresponding to each historical moment.

[0033] The first feature extraction module is connected to the prediction module for extracting features from the first power load vectors and outputting the second power load vectors. The second feature extraction module is connected to the prediction module for extracting features from the first influencing factor vectors and outputting the second influencing factor vectors.

[0034] The power load prediction module is connected to the confidence interval determination module, and is used to predict the maximum power load moment based on a plurality of second power load vectors and a plurality of second influencing factor vectors and using a prediction model.

[0035] The confidence interval determination module is used to determine the confidence interval of the maximum power load according to the maximum power load moment and the power load curve corresponding to each industry.

[0036] Under the above operating environment, according to the first aspect of this embodiment, a method for predicting power load for multiple industries is provided. Figure 2 The electricity load forecasting system shown in is implemented. Figure 3 A schematic diagram showing the process of the method is shown in FIG. Figure 3As shown, the method includes:

[0037] S302: Collecting power load curves corresponding to various industries, and determining a plurality of first power load vectors corresponding to various historical moments based on the plurality of power load curves corresponding to various industries, wherein the first power load vectors correspond to the power loads of various industries;

[0038] S304: determining a first power load influencing factor corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determining multiple first influencing factor vectors corresponding to each historical moment, wherein the first influencing factor vectors include the first power load influencing factors corresponding to each industry;

[0039] S306: inputting a plurality of first power load vectors into a first LSTM model, and outputting a plurality of second power load vectors, inputting a plurality of first influencing factor vectors into a second LSTM model, and outputting a plurality of second influencing factor vectors;

[0040] S308: predicting the maximum power load moment based on the plurality of second power load vectors and the plurality of second influencing factor vectors and using a preset prediction model; and

[0041] S310: Determine a confidence interval of the maximum power load according to the maximum power load moment and using a plurality of power load curves corresponding to each industry.

[0042] Specifically, first, the data collection module collects the power load curves corresponding to each industry from the data center, and based on the multiple power load curves corresponding to each industry, determines multiple first power load vectors corresponding to each historical moment (S302). Among them, the power load curves corresponding to each industry collected by the data collection module include, for example, the power load curve corresponding to agriculture, the power load curve corresponding to the construction industry, the power load curve corresponding to industry, and the power load curve corresponding to the accommodation and catering industry, etc., which will not be repeated here.

[0043] In addition, the power load curve in this embodiment is used to indicate the relationship between each moment in a day and the active power. That is, the active power corresponding to each moment. Therefore, the data acquisition module can determine the third time series data corresponding to the power load of each industry at each historical moment based on the power load curve corresponding to each industry.

[0044] For example, the data acquisition module acquires the power load curve corresponding to the first industry, and determines the third time series data Q corresponding to the power load of the first industry. 1，1 ~Q 1，n Among them, the power load Q corresponding to time T1 is 1，1 , and the power load Q corresponding to time T21，2 , and the power load Q corresponding to time T3 1，3 , ..., and T n The power load Q corresponding to the moment 1，n .

[0045] The data acquisition module acquires the power load curve corresponding to the second industry and determines the third time series data Q corresponding to the power load of the second industry. 2，1 ~Q 2，n Among them, the power load Q corresponding to time T1 is 2，1 , and the power load Q corresponding to time T2 2，2 , and the power load Q corresponding to time T3 2，3 , ..., with T n The power load Q corresponding to the moment 2，n .

[0046] And so on.

[0047] The data acquisition module collects the power load curve corresponding to the mth industry and determines the third time series data Q corresponding to the power load of the mth industry. m，1 ~Q m，n Among them, the power load Q corresponding to time T1 is m，1 , and the power load Q corresponding to time T2 m，2 , and the power load Q corresponding to time T3 m，3 , ..., with T n The power load Q corresponding to the moment m，n .

[0048] Then, the data acquisition module determines a plurality of first power load vectors corresponding to each historical moment based on the third time series data corresponding to the power load of each industry. For example, the first power load vector Q1 corresponding to the moment T1 = [Q 1，1 , Q 2，1 , ..., Q m，1 ], and the first power load vector Q2 corresponding to time T2 = [Q 1，2 , Q 2，2 , ..., Q m，2 ], and the first power load vector Q3 corresponding to time T3 = [Q 1，3 , Q 2，3 , ..., Q m，3 ], ..., and T n The first power load vector Q corresponding to the time n =[Q 1，n , Q 2，n , ..., Q m，n]. Therefore, it can be known from the above-mentioned multiple first power load vectors that each first power load vector includes the power load of each industry corresponding to the corresponding historical moment.

[0049] Similarly, the data acquisition module collects the first power load influencing factors corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determines multiple first influencing factor vectors corresponding to each historical moment (S304). Among them, the data acquisition module, for example, collects multiple second power load influencing factors corresponding to each industry from the data center, and further determines multiple first power load influencing factors among the multiple second power load influencing factors. Among them, the first power load influencing factor is used to indicate the influencing factor with a higher degree of correlation with the corresponding industry among the multiple second power load influencing factors. The above content will be described in detail later, so it will not be repeated here.

[0050] It is worth noting that in this embodiment, the first power load influencing factors corresponding to each industry are different. For example, the first power load influencing factors corresponding to agriculture are weather, holidays, and agricultural product prices. The first power load influencing factors corresponding to industry are industrial added value and raw material demand. The first power load influencing factors corresponding to the accommodation and catering industry are the number of people traveling on holidays and the consumer price index. I will not go into details here.

[0051] Therefore, when the data acquisition module determines the first power load influencing factor corresponding to each industry, the second time series data corresponding to the first power load influencing factor of each industry can be determined.

[0052] For example, the data acquisition module determines a plurality of first power load influencing factors corresponding to the first industry, and determines the second time series data corresponding to the plurality of first power load influencing factors of the first industry. Among them, the first first power load influencing factor value corresponding to time T1 is The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0053] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0054] And so on.

[0055] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 T n The value of the first power load influencing factor corresponding to the kth time

[0056] Similarly, the data acquisition module determines multiple first power load influencing factors corresponding to the second industry, and determines the second time series data corresponding to the first power load influencing factor of the second industry. Among them, the first first power load influencing factor value corresponding to time T1 is The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0057] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0058] And so on.

[0059] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 T n The value of the first power load influencing factor corresponding to the kth time

[0060] Thus, the data acquisition module can determine multiple first power load influencing factors corresponding to the mth industry, and determine the second time series data corresponding to the first power load influencing factor of the mth industry. Among them, the first first power load influencing factor value corresponding to time T1 The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0061] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0062] And so on.

[0063] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 The value of the first power load influencing factor corresponding to the kth time

[0064] The data acquisition module then determines multiple first influencing factor vectors corresponding to each historical moment based on the second time series data corresponding to the first power load influencing factor of each industry. For example, the first influencing factor vector corresponding to time T1 is The first influencing factor vector corresponding to time T2 With T n The first influencing factor vector corresponding to the moment Therefore, it can be known from the above-mentioned multiple first influencing factor vectors that each first influencing factor vector includes the first power load influencing factor value of each industry corresponding to the corresponding historical moment.

[0065] Afterwards, the data acquisition module sends the plurality of first power load vectors to the first feature extraction module, and sends the plurality of first influencing factor vectors to the second feature extraction module. Thus, the first feature extraction module inputs the plurality of first power load vectors into the first LSTM model, and outputs a plurality of second power load vectors. The second feature extraction module inputs the plurality of first influencing factor vectors into the second LSTM model, and outputs a plurality of second influencing factor vectors (S306).

[0066] Then, the power load prediction module inputs the received multiple second power load vectors and multiple second influencing factor vectors into a preset prediction model to predict the maximum power load moment (S308). The prediction model includes a fully connected layer and multiple softmax classifiers. And the number of the multiple softmax classifiers corresponds to the number of moments to be predicted. Figure 4 is a schematic diagram of a prediction model according to an embodiment of the present application. Figure 4 As shown, for example, if the power load probability corresponding to 24 hours needs to be determined, the number of softmax classifiers is 24.

[0067] Thus, firstly, the power load prediction module uses the MLP perceptron to perform feature splicing and fusion on multiple second power load vectors and multiple second influencing factor vectors, and generates a first feature matrix G. Further, the prediction model inputs the first feature matrix G into the prediction model. Thus, the prediction model can output the probabilities corresponding to each moment to be predicted. The above-mentioned probability includes a first sub-probability indicating that it is the moment of maximum power load and a second sub-probability that is not the moment of maximum power load. Then, the prediction model determines the largest first sub-probability, and uses the moment to be predicted corresponding to the largest first sub-probability as the moment of maximum power load.

[0068] refer to Figure 4 As shown, for example, after the prediction model inputs the first feature matrix G into the prediction model, the prediction model outputs that the first sub-probability corresponding to 1:00 is 5%, and the second sub-probability is 95%; the first sub-probability corresponding to 2:00 is 10%, and the second sub-probability is 90%; the first sub-probability corresponding to 3:00 is 1%, and the second sub-probability is 99%; ...; the first sub-probability corresponding to 8:00 is 60%, and the second sub-probability is 40%; ...; the first sub-probability corresponding to 0:00 is 3%, and the second sub-probability is 97%. Thus, the prediction model determines that the maximum first sub-probability is 60%, and 8:00 is used as the maximum power load time. That is, at 8:00, the power load of multiple industries reaches the maximum value.

[0069] Finally, the confidence interval determination module determines the confidence interval of the maximum power load according to the maximum power load moment and using multiple power load curves corresponding to each industry (S310). f In the case of maximum power load time T f Determine the power load Q corresponding to the power load curve of each industry 1，f ~Q m，f For example, if the confidence interval determination module determines that the maximum power load moment is T3, then the confidence interval determination module can further determine the power load Q corresponding to the power load curve of the first industry at time T3. 1，3 The confidence interval determination module can further determine the power load Q corresponding to the power load curve of the second industry at time T3. 2，3 And so on. The confidence interval determination module can further determine the power load Q corresponding to the power load curve of the mth industry at time T3 m，3

[0070] Furthermore, the confidence interval determination module determines the maximum power load time T f Corresponding multiple power loads Q 1，f ~Q m，f In the case of multiple power loads Q1，f ~Q m，f Perform fitting and determine the corresponding probability distribution curve.

[0071] Furthermore, the confidence interval determination module determines the multiple power loads Q 1，f ~Q m，f The corresponding average value and variance. 1，f ~Q m,f The corresponding average value is calculated as follows:

[0072]

[0073] in, Represents the average value corresponding to multiple power loads, Q i，f Represents the power load corresponding to the maximum power load moment of the i-th industry.

[0074] Among them, with multiple power loads Q 1，f ~Q m，f The corresponding variance calculation formula is as follows:

[0075]

[0076] Among them, Q 2 represents the variance corresponding to multiple power loads, Q i，f represents the power load corresponding to the maximum power load moment of the i-th industry, Indicates the average value corresponding to multiple power loads.

[0077] Finally, the confidence interval determination module determines the confidence interval of the maximum power load based on the above determined mean value and variance and using the probability distribution curve diagram. The above content will be described in detail later, so it will not be repeated here.

[0078] As described in the background technology, with the continuous maturity and widespread application of machine learning technology, relevant models for power load forecasting in various industries have been constructed, and the power load of various industries have been predicted accordingly using relevant models. However, since the existing technology predicts the power load of each industry separately, and does not predict the power load of multiple industries, the existing technology can only use the maximum power load of each industry to predict the maximum power load of multiple industries, and cannot directly predict the maximum power load of multiple industries. This may make it impossible to accurately determine the maximum power load of multiple industries.

[0079] In view of this, the present application provides a method for predicting power load for multiple industries. And since the present application collects power load curves corresponding to each industry and determines multiple first power load vectors corresponding to each historical moment (where the first power load vector corresponds to the power load of each industry), the maximum power load moment predicted by the final prediction model corresponds to multiple industries. And finally, based on the maximum power load moment and the power load curves corresponding to each industry, the confidence interval of the maximum power load corresponding to multiple industries can be determined. Therefore, unlike the prior art, the present application can achieve the technical effect of improving the prediction efficiency and accurately determining the maximum power load of multiple industries by directly predicting the maximum power load of multiple industries.

[0080] In addition, since the present application determines the first electricity load influencing factor corresponding to each industry, and simultaneously uses multiple first influencing factor vectors and multiple first electricity load vectors corresponding to each historical moment to predict the confidence interval of the maximum electricity load corresponding to multiple industries, the final prediction result is associated with the first electricity load influencing factor of each industry (that is, the electricity load influencing factor with a higher degree of correlation with each industry), which further improves the accuracy of the prediction.

[0081] This solves the technical problem that the existing technology can only use the maximum power load of each industry to predict the maximum power load of multiple industries, but cannot directly predict the maximum power load of multiple industries, resulting in the inability to accurately determine the maximum power load of multiple industries.

[0082] Optionally, the operation of determining the confidence interval of the maximum power load based on the maximum power load moment and using multiple power load curves corresponding to each industry includes: determining the corresponding power load in each power load curve based on the maximum power load moment; fitting multiple power loads and determining the corresponding probability distribution curve graph; determining the mean value and variance corresponding to the multiple power loads; and determining the confidence interval of the maximum power load based on the mean value and variance and using the probability distribution curve graph.

[0083] Specifically, the confidence interval determination module determines the maximum power load time T f In the case of maximum power load time T f Determine the power load Q corresponding to the power load curve of each industry 1，f ~Q m，f For example, if the confidence interval determination module determines that the maximum power load moment is T3, then the confidence interval determination module can further determine the power load Q corresponding to the power load curve of the first industry at time T3. 1，3 The confidence interval determination module can further determine the power load Q corresponding to the power load curve of the second industry at time T3. 2，3And so on. The confidence interval determination module can further determine the power load Q corresponding to the power load curve of the mth industry at time T3 m，3

[0084] Furthermore, the confidence interval determination module determines the maximum power load time T f Corresponding multiple power loads Q 1，f ~Q m，f In the case of multiple power loads Q 1，f ~Q m，f Perform fitting and determine the corresponding probability distribution curve.

[0085] Furthermore, the confidence interval determination module determines the multiple power loads Q 1，f ~Q m，f The corresponding average value and variance. 1，f ~Q m，f The corresponding average value is calculated as follows:

[0086]

[0087] in, Represents the average value corresponding to multiple power loads, Q i，f Represents the power load corresponding to the maximum power load moment of the i-th industry.

[0088] Among them, with multiple power loads Q 1，f ~Q m，f The corresponding variance calculation formula is as follows:

[0089]

[0090] Among them, Q 2 represents the variance corresponding to multiple power loads, Q i，f represents the power load corresponding to the maximum power load moment of the i-th industry, Indicates the average value corresponding to multiple power loads.

[0091] Finally, the confidence interval determination module is based on the above determined mean value and variance Q 2 , and use the probability distribution curve to determine the confidence interval of the maximum power load.

[0092] For example, the confidence interval determination module selects a confidence level of 95% and determines the Z value corresponding to the confidence level. Then the confidence interval determination module determines the standard deviation based on the variance determined above. Finally, the confidence interval determination module calculates the upper and lower limits of the confidence interval based on the mean, Z value and standard deviation. Among them, the upper limit of the confidence interval is used to indicate the maximum value of the power load of multiple industries corresponding to the moment of maximum power load, and the lower limit of the confidence interval is used to indicate the minimum value of the power load of multiple industries corresponding to the moment of maximum power load. It is worth noting that the above-mentioned minimum power load value refers to the minimum value corresponding to the moment of maximum power load, not the minimum value corresponding to the moment of minimum power load.

[0093] Optionally, based on multiple second power load vectors and multiple second influencing factor vectors, and using a preset prediction model, the operation of outputting the maximum power load moment includes: using MLP to perform feature concatenation and fusion on multiple second power load vectors and multiple second influencing factor vectors, and generate a first feature matrix; and inputting the first feature matrix into the prediction model, and outputting the maximum power load moment. Further optionally, the operation of inputting the first feature matrix into the prediction model and outputting the maximum power load moment includes: inputting the first feature matrix into the prediction model, and determining the probabilities corresponding to each moment to be predicted, wherein the probability includes a first sub-probability for indicating that it is the maximum power load moment and a second sub-probability for not being the maximum power load moment; and determining the maximum first sub-probability, and taking the moment to be predicted corresponding to the maximum first sub-probability as the maximum power load moment.

[0094] Specifically, when the first feature extraction module outputs multiple second power load vectors and the second feature extraction module outputs multiple second influencing factor vectors, the MLP perceptron is used to perform feature splicing and fusion on the multiple second power load vectors and the multiple second influencing factor vectors, and generate a first feature matrix G. Furthermore, the prediction model inputs the first feature matrix G into the prediction model, so that the prediction model can output the probabilities corresponding to each moment to be predicted. The above-mentioned probability includes a first sub-probability for indicating that it is the moment of maximum power load and a second sub-probability for not being the moment of maximum power load. The prediction model then determines the largest first sub-probability and uses the moment to be predicted corresponding to the largest first sub-probability as the moment of maximum power load T. f .

[0095] For example, after the prediction model inputs the first feature matrix G into the prediction model, the prediction model outputs that the first sub-probability corresponding to 1:00 is 5%, and the second sub-probability is 95%; the first sub-probability corresponding to 2:00 is 10%, and the second sub-probability is 90%; the first sub-probability corresponding to 3:00 is 1%, and the second sub-probability is 99%; ...; the first sub-probability corresponding to 8:00 is 60%, and the second sub-probability is 40%; .... Thus, the prediction model determines that the maximum first sub-probability is 60%, and 8:00 is used as the maximum power load time. That is, at 8:00, the power load of multiple industries reaches the maximum value.

[0096] Optionally, the operation of determining the target influencing factors corresponding to each industry includes: determining the second electricity load influencing factors corresponding to each industry, and determining the first time series data corresponding to the second electricity load influencing factors of each industry at each historical moment; based on multiple first time series data corresponding to each industry, determining a second characteristic matrix including multiple characteristic vectors, wherein the characteristic vectors correspond to the second electricity load influencing factors at each historical moment, and the second characteristic matrix corresponds to each industry; normalizing the second characteristic matrix corresponding to each industry, and generating a third characteristic matrix corresponding to each industry; calculating the sample correlation matrix corresponding to each third characteristic matrix, and eigendecomposing the sample correlation matrix, so as to determine multiple eigenvalues ​​corresponding to each industry; and determining the number of variance contribution rates of multiple eigenvalues ​​reaching a preset threshold, and determining the first electricity load influencing factors corresponding to each industry.

[0097] Specifically, first, the data acquisition module determines the second power load influencing factor corresponding to each industry, and determines the first time series data corresponding to the second power load influencing factor of each industry at each historical moment.

[0098] For example, the data acquisition module determines multiple second power load influencing factors corresponding to the first industry, and determines the first time series data corresponding to the second power load influencing factor of the first industry. Among them, the first second power load influencing factor value corresponding to time T1 The first second power load influencing factor value corresponding to time T2 T n The first and second power load influencing factor values ​​corresponding to the time

[0099] The second power load influencing factor value corresponding to time T1 The second power load influencing factor value corresponding to time T2 T n The second power load influencing factor value corresponding to the time

[0100] And so on.

[0101] The hth second power load influencing factor value corresponding to time T1 The hth second power load influencing factor value corresponding to time T2 T n The value of the second power load influencing factor corresponding to the hth time It is worth noting that the number of the second power load influencing factors is greater than the number of the first power load influencing factors, that is, h>k.

[0102] Similarly, the data acquisition module determines multiple second power load influencing factors corresponding to the second industry, and determines the first time series data corresponding to the second power load influencing factors of the second industry. Among them, the first second power load influencing factor corresponding to time T1 is The first second power load influencing factor corresponding to time T2 T n The first and second power load influencing factors corresponding to the time

[0103] The second power load influencing factor corresponding to time T1 The second power load influencing factor corresponding to time T2 T n The second power load influencing factor corresponding to the time

[0104] And so on.

[0105] The hth second power load influencing factor corresponding to time T1 The hth second power load influencing factor corresponding to time T2 T n The influencing factor of the hth second power load corresponding to the time

[0106] Thus, the data acquisition module can determine multiple second power load influencing factors corresponding to the mth industry, and determine the first time series data corresponding to the second power load influencing factors of the mth industry. Among them, the first second power load influencing factor corresponding to time T1 The first second power load influencing factor corresponding to time T2 T n The first and second power load influencing factors corresponding to the time

[0107] The second power load influencing factor corresponding to time T1 The second power load influencing factor corresponding to time T2 T n The second power load influencing factor corresponding to the time

[0108] And so on.

[0109] The hth second power load influencing factor corresponding to time T1 The hth second power load influencing factor corresponding to time T2 T n The influencing factor of the hth second power load corresponding to the time

[0110] The data collection module then determines multiple feature vectors corresponding to each historical moment based on the first time series data corresponding to the second power load influencing factor of each industry. For example, the feature vector corresponding to moment T1 is The eigenvector corresponding to time T2 With T n The eigenvector corresponding to the moment

[0111] Furthermore, a plurality of feature vectors S1 to S n Afterwards, it is determined that multiple feature vectors S1~S n The second characteristic matrix X of .

[0112] Then, the data acquisition module normalizes the second characteristic matrix X corresponding to each industry and generates the third characteristic matrix Y corresponding to each industry. The calculation formula is as follows:

[0113]

[0114] Where i = 1 to e, e represents the second power load influencing factor corresponding to m industries, and e = m×h. j = 1 to n. y i，j Represents the element value in the third feature matrix Y after normalization. i，j Represents the element values ​​in the second feature matrix X that have not been normalized. represents the average value of each eigenvector. represents the variance of each eigenvector.

[0115] And the specific calculation formula of the average value of each eigenvector is as follows:

[0116]

[0117] The specific calculation formula for the variance of each eigenvector is as follows:

[0118]

[0119] Furthermore, when the data acquisition module determines the third characteristic matrix Y corresponding to each industry, the sample correlation matrix corresponding to the third characteristic matrix is ​​calculated, and the sample correlation matrix is ​​eigen-decomposed to determine a plurality of eigenvalues ​​corresponding to each industry.

[0120] Among them, the specific calculation formula of the sample correlation matrix is ​​as follows:

[0121]

[0122] Among them, R represents the sample correlation matrix and Y represents the third characteristic matrix.

[0123] The data acquisition module performs eigendecomposition on the sample correlation matrix R and determines the calculation formulas of multiple eigenvalues ​​as follows:

[0124] |R-λI|=0

[0125] Among them, R represents the sample correlation matrix, I represents the identity matrix, and λ represents the eigenvalue.

[0126] Thus, the data acquisition module can calculate the e eigenvalues ​​λ1, λ2, ..., λ of the sample correlation matrix based on the above formula. e . And among them, λ1≥λ2≥…≥λ e .

[0127] Then, the data acquisition module determines the e eigenvalues ​​λ1, λ2, ..., λ of the sample correlation matrix R. e In the case of , the number of variance contribution rates of multiple eigenvalues ​​reaching the preset threshold is further determined, and the target influencing factors corresponding to each industry are determined.

[0128] The specific calculation formula of variance contribution rate is as follows:

[0129]

[0130] Among them, λ k Represents the variance contribution rate of the kth eigenvalue among multiple eigenvalues, represents the sum of e eigenvalues, η k Represents the variance contribution rate.

[0131]

[0132] in, represents the sum of the variance contributions of k eigenvalues, represents the sum of k eigenvalues, represents the sum of e eigenvalues.

[0133] Finally, the data acquisition module determines multiple second power load influencing factors based on the preset threshold and the sum of the variance contribution rate. For example, based on the above content, it can be known that λ1≥λ2≥...≥λ e , so in this embodiment, the preset threshold is taken as 80%, then when the sum of the variance contribution rates of the k eigenvalues ​​is greater than or equal to 80%, the value of k is determined. That is, determine the value of k related to λ1, λ2, λ3, ..., λ k The corresponding first power load influencing factor is the second power load influencing factor, wherein the second power load influencing factor is used to indicate the first power load influencing factor that is highly associated with each industry.

[0134] For example, the determined k The corresponding first power load influencing factor is the second influencing factor, which means that there are k influencing factors with a high degree of correlation with the corresponding industry. And the eigenvector corresponding to time T1 The eigenvector corresponding to time T2 With T n The eigenvector corresponding to the moment

[0135] Thus, by determining a plurality of second power load influencing factors that are highly correlated with the corresponding industries and using the second power load influencing factors to predict the operation at the time of maximum power load, the technical effect of improving the accuracy of the prediction results is achieved.

[0136] Optionally, the operation of determining the first electricity load influencing factor corresponding to each industry, and determining multiple first influencing factor vectors corresponding to each historical moment based on the multiple first electricity load influencing factors corresponding to each industry, includes: determining the first electricity load influencing factor corresponding to each industry, and determining the second time series data corresponding to the first electricity load influencing factor of each industry at each historical moment; and determining multiple first influencing factor vectors corresponding to each historical moment based on the second time series data.

[0137] Specifically, when the data acquisition module determines a plurality of first power load influencing factors corresponding to each industry, the first time series data corresponding to the first power load influencing factors of each industry at each historical moment is further determined.

[0138] For example, the data acquisition module determines multiple first power load influencing factors corresponding to the first industry, and determines the second time series data corresponding to the first power load influencing factor of the first industry. Among them, the first first power load influencing factor value corresponding to time T1 The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0139] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0140] And so on.

[0141] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 T n The value of the first power load influencing factor corresponding to the kth time

[0142] Similarly, the data acquisition module determines the first power load influencing factor corresponding to the second industry, and determines the second time series data corresponding to the first power load influencing factor of the second industry. Among them, the first first power load influencing factor value corresponding to time T1 is The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0143] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0144] And so on.

[0145] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 T n The value of the first power load influencing factor corresponding to the kth time

[0146] Thus, the data acquisition module can determine the first power load influencing factor corresponding to the mth industry, and determine the second time series data corresponding to the first power load influencing factor of the mth industry. Among them, the first first power load influencing factor value corresponding to time T1 is The first power load influencing factor value corresponding to time T2 T n The first power load influencing factor value corresponding to the time

[0147] The second first power load influencing factor value corresponding to time T1 The second first power load influencing factor value corresponding to time T2 T n The second first power load influencing factor value corresponding to the time

[0148] And so on.

[0149] The kth first power load influencing factor value corresponding to time T1 The kth first power load influencing factor value corresponding to time T2 T n The value of the first power load influencing factor corresponding to the kth time

[0150] The data acquisition module then determines a plurality of first influencing factor vectors corresponding to each historical moment based on the second time series data corresponding to the first power load influencing factor of each industry. For example, the first influencing factor vector corresponding to time T1 is The first influencing factor vector corresponding to time T2 With T n The first influencing factor vector corresponding to the moment Therefore, it can be known from the above-mentioned multiple first influencing factor vectors that each first influencing factor vector includes the first power load influencing factor value of each industry corresponding to the corresponding historical moment.

[0151] Optionally, the operation of collecting electricity load curves corresponding to each industry, and determining multiple first electricity load vectors corresponding to each historical moment based on the multiple electricity load curves corresponding to each industry, includes: collecting electricity load curves corresponding to each industry, and determining third time series data corresponding to the electricity load of each industry at each historical moment; and determining multiple first electricity load vectors corresponding to each historical moment based on the third time series data.

[0152] Specifically, first, the data collection module collects the power load curves corresponding to each industry. Then, the data collection module determines the third time series data corresponding to the power load of each industry at each disadvantageous moment.

[0153] For example, the data acquisition module acquires the power load curve corresponding to the first industry, and determines the third time series data Q corresponding to the power load of the first industry. 1，1 ~Q 1，n Among them, the power load Q corresponding to time T1 is 1，1 , and the power load Q corresponding to time T2 1，2 , and the power load Q corresponding to time T3 1，3 , ..., with T n The power load Q corresponding to the moment 1，n .

[0154] The data acquisition module acquires the power load curve corresponding to the second industry and determines the third time series data Q corresponding to the power load of the second industry. 2，1 ~Q 2，n Among them, the power load Q corresponding to time T1 is 2，1 , and the power load Q corresponding to time T2 2，2 , and the power load Q corresponding to time T3 2，3 , ..., with T n The power load Q corresponding to the moment 2，n .

[0155] And so on.

[0156] The data acquisition module collects the power load curve corresponding to the mth industry and determines the third time series data Q corresponding to the power load of the mth industry. m，1 ~Q m，n Among them, the power load Q corresponding to time T1 is m，1 , and the power load Q corresponding to time T2 m，2 , and the power load Q corresponding to time T3 m，3 , ..., with T n The power load Q corresponding to the moment m，n .

[0157] Then, the data acquisition module determines a plurality of first power load vectors corresponding to each historical moment based on the third time series data. For example, the first power load vector Q1 corresponding to the moment T1 = [Q 1，1 , Q 2，1 , ..., Q m，1 ], and the first power load vector Q2 corresponding to time T2 = [Q 1，2 , Q 2，2 , ..., Q m，2], and the first power load vector Q3 corresponding to time T3 = [Q 1，3 , Q 2，3 , ..., Q m，3] , ..., with T n The first power load vector Q corresponding to the time n =[Q 1，n , Q 2，n , ..., Q m，n ]. Therefore, it can be known from the above-mentioned multiple first power load vectors that each first power load vector includes the power load of each industry corresponding to the corresponding historical moment.

[0158] Thus, according to the first aspect of this embodiment, the technical effect of improving prediction efficiency and accurately determining the maximum power load of multiple industries is achieved.

[0159] In addition, reference Figure 1 As shown, according to the second aspect of this embodiment, a storage medium is provided, wherein the storage medium includes a stored program, wherein when the program is run, a processor executes any one of the above methods.

[0160] Therefore, according to this embodiment, the technical effect of improving prediction efficiency and accurately determining the maximum power load of multiple industries is achieved.

[0161] 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.

[0162] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course 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.

[0163] Example 2

[0164] Figure 5FIG. 5 shows a device 500 for predicting power load for multiple industries according to this embodiment, and the device 500 corresponds to the method according to Embodiment 1. Figure 5 As shown, the device 500 includes: a first power load vector determination module 510, which is used to collect power load curves corresponding to each industry, and based on the multiple power load curves corresponding to each industry, determine multiple first power load vectors corresponding to each historical moment, wherein the first power load vector corresponds to the power load of each industry; a first influencing factor vector determination module 520, which is used to determine the first power load influencing factor corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determine multiple first influencing factor vectors corresponding to each historical moment, wherein the first influencing factor vector includes the first power load influencing factor corresponding to each industry. Load influencing factors; a feature extraction module 530, which is used to input multiple first power load vectors into a first LSTM model and output multiple second power load vectors, and input multiple first influencing factor vectors into a second LSTM model and output multiple second influencing factor vectors; a maximum power load moment prediction module 540, which is used to predict the maximum power load moment based on multiple second power load vectors and multiple second influencing factor vectors and using a pre-set prediction model; and a confidence interval determination module 550, which is used to determine the confidence interval of the maximum power load according to the maximum power load moment and using multiple power load curves corresponding to each industry.

[0165] Optionally, the confidence interval determination module 550 includes: a power load determination module, which is used to determine the corresponding power load in each power load curve based on the maximum power load moment; a fitting module, which is used to fit multiple power loads and determine the corresponding probability distribution curve; a parameter determination module, which is used to determine the average value and variance corresponding to multiple power loads; and a confidence interval determination submodule, which is used to determine the confidence interval of the maximum power load based on the average value and variance and using the probability distribution curve.

[0166] Optionally, the maximum power load moment prediction module 540 includes: a first feature matrix generation module, used to use MLP to perform feature splicing and fusion of multiple second power load vectors and multiple second influencing factor vectors, and generate a first feature matrix; and a maximum power load moment output module, used to input the first feature matrix into the prediction model and output the maximum power load moment.

[0167] Optionally, the maximum power load moment output module includes: a probability determination module, which is used to input the first characteristic matrix into the prediction model and determine the probabilities corresponding to each moment to be predicted, wherein the probability includes a first sub-probability for indicating that it is the maximum power load moment and a second sub-probability for not being the maximum power load moment; and a maximum power load moment determination module, which is used to determine the maximum first sub-probability and use the moment to be predicted corresponding to the maximum first sub-probability as the maximum power load moment.

[0168] Optionally, the first influencing factor vector determination module 520 includes: a first time series data determination module, which is used to determine the second power load influencing factor corresponding to each industry, and determine the first time series data corresponding to the second power load influencing factor of each industry at each historical moment; a second characteristic matrix determination module, which is used to determine a second characteristic matrix containing multiple characteristic vectors based on multiple first time series data corresponding to each industry, wherein the characteristic vectors correspond to the second power load influencing factor at each historical moment, and the second characteristic matrix corresponds to each industry; a normalization processing module, which is used to normalize the second characteristic matrix corresponding to each industry, and generate a third characteristic matrix corresponding to each industry; an eigenvalue determination module, which is used to calculate the sample correlation matrix corresponding to the third characteristic matrix, and perform eigendecomposition on the sample correlation matrix, so as to determine multiple eigenvalues ​​corresponding to each industry; and a quantity determination module, which is used to determine the number of variance contribution rates of multiple eigenvalues ​​that reach a preset threshold, and determine the first power load influencing factor corresponding to each industry.

[0169] Optionally, the first influencing factor vector determination module 520 includes: a second time series data determination module, used to determine the first electricity load influencing factor corresponding to each industry, and determine the second time series data corresponding to the first electricity load influencing factor of each industry at each historical moment; and a first vector determination submodule, used to determine multiple first influencing factor vectors corresponding to each historical moment based on the second time series data.

[0170] Optionally, the first power load vector determination module 510 includes: a third time series data determination module, which is used to collect power load curves corresponding to each industry, and determine the third time series data corresponding to the power load of each industry at each historical moment; and a second vector determination submodule, which is used to determine multiple first power load vectors corresponding to each historical moment based on the third time series data.

[0171] Therefore, according to this embodiment, the technical effect of improving prediction efficiency and accurately determining the maximum power load of multiple industries is achieved.

[0172] Example 3

[0173] Figure 6FIG. 6 shows a device 600 for predicting power load for multiple industries according to this embodiment, which corresponds to the method according to Embodiment 1. Figure 6 As shown, the device 600 includes: a processor 610; and a memory 620, connected to the processor 610, for providing the processor 610 with instructions for processing the following processing steps: collecting power load curves corresponding to various industries, and based on the multiple power load curves corresponding to the various industries, determining multiple first power load vectors corresponding to various historical moments, wherein the first power load vector corresponds to the power load of each industry; determining the first power load influencing factors corresponding to each industry, and based on the multiple first power load influencing factors corresponding to each industry, determining the multiple first influencing factors corresponding to each historical moment. factor vector, wherein the first influencing factor vector includes the first power load influencing factor corresponding to each industry; multiple first power load vectors are input into the first LSTM model, and multiple second power load vectors are output; multiple first influencing factor vectors are input into the second LSTM model, and multiple second influencing factor vectors are output; based on the multiple second power load vectors and the multiple second influencing factor vectors, and using a pre-set prediction model, the maximum power load moment is predicted; and according to the maximum power load moment and using multiple power load curves corresponding to each industry, the confidence interval of the maximum power load is determined.

[0174] Optionally, the operation of determining the confidence interval of the maximum power load based on the maximum power load moment and using multiple power load curves corresponding to each industry includes: determining the corresponding power load in each power load curve based on the maximum power load moment; fitting multiple power loads and determining the corresponding probability distribution curve graph; determining the mean value and variance corresponding to the multiple power loads; and determining the confidence interval of the maximum power load based on the mean value and variance and using the probability distribution curve graph.

[0175] Optionally, based on multiple second power load vectors and multiple second influencing factor vectors, and using a pre-set prediction model, the operation of outputting the maximum power load moment includes: using MLP to perform feature splicing and fusion on multiple second power load vectors and multiple second influencing factor vectors, and generating a first feature matrix; and inputting the first feature matrix into the prediction model, and outputting the maximum power load moment.

[0176] Optionally, the operation of inputting the first characteristic matrix into the prediction model and outputting the moment of maximum power load includes: inputting the first characteristic matrix into the prediction model, and determining the probabilities corresponding to each moment to be predicted, wherein the probability includes a first sub-probability for indicating that it is the moment of maximum power load and a second sub-probability for not being the moment of maximum power load; and determining the maximum first sub-probability, and taking the moment to be predicted corresponding to the maximum first sub-probability as the moment of maximum power load.

[0177] Optionally, the operation of determining the target influencing factors corresponding to each industry includes: determining the second electricity load influencing factors corresponding to each industry, and determining the first time series data corresponding to the second electricity load influencing factors of each industry at each historical moment; based on multiple first time series data corresponding to each industry, determining a second characteristic matrix including multiple characteristic vectors, wherein the characteristic vectors correspond to the second electricity load influencing factors at each historical moment, and the second characteristic matrix corresponds to each industry; normalizing the second characteristic matrix corresponding to each industry, and generating a third characteristic matrix corresponding to each industry; calculating the sample correlation matrix corresponding to the third characteristic matrix, and eigendecomposing the sample correlation matrix to determine multiple eigenvalues ​​corresponding to each industry; and determining the number of variance contribution rates of the multiple eigenvalues ​​that reach a preset threshold, and determining the first electricity load influencing factors corresponding to each industry.

[0178] Optionally, the operation of determining the first electricity load influencing factor corresponding to each industry, and determining multiple first influencing factor vectors corresponding to each historical moment based on the multiple first electricity load influencing factors corresponding to each industry, includes: determining the first electricity load influencing factor corresponding to each industry, and determining the second time series data corresponding to the first electricity load influencing factor of each industry at each historical moment; and determining multiple first influencing factor vectors corresponding to each historical moment based on the second time series data.

[0179] Optionally, the operation of collecting electricity load curves corresponding to each industry, and determining multiple first electricity load vectors corresponding to each historical moment based on the multiple electricity load curves corresponding to each industry, includes: collecting electricity load curves corresponding to each industry, and determining third time series data corresponding to the electricity load of each industry at each historical moment; and determining multiple first electricity load vectors corresponding to each historical moment based on the third time series data.

[0180] Therefore, according to this embodiment, the technical effect of improving prediction efficiency and accurately determining the maximum power load of multiple industries is achieved.

[0181] 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.

[0182] 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.

[0183] 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 is only 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.

[0184] 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 network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0185] 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.

[0186] 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 computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, 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, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a 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.

[0187] 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 method for predicting power load for multiple industries, characterized in that: include: Collecting power load curves corresponding to various industries, and determining a plurality of first power load vectors corresponding to various historical moments based on the plurality of power load curves corresponding to the various industries, wherein the first power load vectors correspond to the power loads of the various industries; Determine the first power load influencing factors corresponding to the industries, and determine a plurality of first influencing factor vectors corresponding to the respective historical moments based on the plurality of first power load influencing factors corresponding to the industries, wherein the first influencing factor vectors include the first power load influencing factors corresponding to the industries; Input the plurality of first power load vectors into a first LSTM model, and output a plurality of second power load vectors, input the plurality of first influencing factor vectors into a second LSTM model, and output a plurality of second influencing factor vectors; Based on the plurality of second power load vectors and the plurality of second influencing factor vectors, and using a preset prediction model, predicting the maximum power load moment; as well as A confidence interval of the maximum power load is determined according to the maximum power load moment and using a plurality of power load curves corresponding to the various industries.

2. The method according to claim 1, characterized in that The operation of determining the confidence interval of the maximum power load according to the maximum power load moment and using a plurality of power load curves corresponding to the various industries includes: Based on the maximum power load moment, determining the corresponding power load in each power load curve; Fitting the multiple power loads and determining corresponding probability distribution curves; determining an average value and a variance corresponding to the plurality of electrical loads; and Based on the mean value and the variance, and using the probability distribution curve, a confidence interval of the maximum power load is determined.

3. The method according to claim 2, characterized in that The operation of outputting the maximum power load moment based on the plurality of second power load vectors and the plurality of second influencing factor vectors and using a preset prediction model includes: Using MLP to perform feature concatenation and fusion on the plurality of second power load vectors and the plurality of second influencing factor vectors, and generate a first feature matrix; and The first characteristic matrix is ​​input into the prediction model, and the maximum power load moment is output.

4. The method according to claim 3, characterized in that The operation of inputting the first characteristic matrix into the prediction model and outputting the maximum power load time includes: Inputting the first feature matrix into the prediction model, and determining the probability corresponding to each time to be predicted, wherein the probability includes a first sub-probability indicating that it is the time of maximum power load and a second sub-probability indicating that it is not the time of maximum power load; and The maximum first sub-probability is determined, and the to-be-predicted time corresponding to the maximum first sub-probability is used as the maximum power load time.

5. The method according to claim 1, characterized in that Determine the operations of the target influencing factors corresponding to each of the industries, including: Determine the second power load influencing factors corresponding to the industries, and determine the first time series data corresponding to the second power load influencing factors of the industries at each historical moment; Based on a plurality of first time series data corresponding to the respective industries, determining a second characteristic matrix including a plurality of characteristic vectors, wherein the characteristic vectors correspond to the second power load influencing factors at the respective historical moments, and the second characteristic matrix corresponds to the respective industries; Normalizing the second characteristic matrices corresponding to the industries, and generating third characteristic matrices corresponding to the industries; Calculating a sample correlation matrix corresponding to the third characteristic matrix, and performing characteristic decomposition on the sample correlation matrix, thereby determining a plurality of eigenvalues ​​corresponding to the various industries; and Determine the number of variance contribution rates of the multiple characteristic values ​​that reach a preset threshold, and determine the first power load influencing factor corresponding to each industry.

6. The method according to claim 1, characterized in that The operation of determining the first power load influencing factors corresponding to the industries, and determining a plurality of first influencing factor vectors corresponding to the respective historical moments based on the plurality of first power load influencing factors corresponding to the industries, includes: Determining the first power load influencing factors corresponding to the industries, and determining the second time series data corresponding to the first power load influencing factors of the industries at the respective historical moments; and Based on the second time series data, a plurality of first influencing factor vectors corresponding to the respective historical moments are determined.

7. The method according to claim 1, characterized in that The operation of collecting power load curves corresponding to each industry, and determining a plurality of first power load vectors corresponding to each historical moment based on the plurality of power load curves corresponding to the respective industries, includes: Collecting power load curves corresponding to each industry, and determining third time series data corresponding to the power load of each industry at each historical moment; and Based on the third time series data, a plurality of first power load vectors corresponding to the respective historical moments are determined.

8. A storage medium, characterized in that: The storage medium includes a stored program, wherein when the program is run, the processor executes the method according to any one of claims 1 to 7.

9. A power load forecasting device for multiple industries, characterized in that: include: A first power load vector determination module is used to collect power load curves corresponding to various industries, and determine a plurality of first power load vectors corresponding to various historical moments based on a plurality of power load curves corresponding to the various industries, wherein the first power load vectors correspond to the power loads of the various industries; a first influencing factor vector determining module, configured to determine the first power load influencing factors corresponding to the industries, and determine a plurality of first influencing factor vectors corresponding to the respective historical moments based on the plurality of first power load influencing factors corresponding to the industries, wherein the first influencing factor vectors include the first power load influencing factors corresponding to the industries; A feature extraction module, used for inputting the plurality of first power load vectors into a first LSTM model and outputting a plurality of second power load vectors, inputting the plurality of first influencing factor vectors into a second LSTM model and outputting a plurality of second influencing factor vectors; A maximum power load moment prediction module, configured to predict the maximum power load moment based on the plurality of second power load vectors and the plurality of second influencing factor vectors and using a preset prediction model; and The confidence interval determination module is used to determine the confidence interval of the maximum power load according to the maximum power load moment and using a plurality of power load curves corresponding to the various industries.

10. A power load forecasting device for multiple industries, characterized in that: include: processor; as well as A memory, connected to the processor, configured to provide the processor with instructions for processing the following processing steps: Collecting power load curves corresponding to various industries, and determining a plurality of first power load vectors corresponding to various historical moments based on the plurality of power load curves corresponding to the various industries, wherein the first power load vectors correspond to the power loads of the various industries; Determine the first power load influencing factors corresponding to the industries, and determine a plurality of first influencing factor vectors corresponding to the respective historical moments based on the plurality of first power load influencing factors corresponding to the industries, wherein the first influencing factor vectors include the first power load influencing factors corresponding to the industries; Input the plurality of first power load vectors into a first LSTM model, and output a plurality of second power load vectors, input the plurality of first influencing factor vectors into a second LSTM model, and output a plurality of second influencing factor vectors; Based on the plurality of second power load vectors and the plurality of second influencing factor vectors, and using a preset prediction model, predicting the maximum power load moment; as well as A confidence interval of the maximum power load is determined according to the maximum power load moment and using a plurality of power load curves corresponding to the various industries.

Images