A training method, prediction method and device for energy data prediction model

Abstract

The present application discloses a training method for an energy data prediction model, including: obtaining basic data to be trained from a smart grid; inputting the basic data to be trained into an optimization model to obtain optimized data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model, and a total energy cost model; the optimized data to be trained includes consumer surplus energy data, consumer energy cost data, and total energy cost data; and training the energy data prediction model based on a Gaussian process regression model according to the basic data to be trained and the optimized data to be trained. The present application trains the energy data prediction model based on a Gaussian process regression model by using consumer surplus energy data, consumer energy cost data, and total energy cost data as training data, and taking into account the randomness of renewable energy, thereby solving the problem of supply and demand imbalance in consumer energy production and consumer energy consumption.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to a training method, a prediction method and a device for an energy data prediction model. Background Art

[0002] Over the past few decades, researchers have been studying how to design an integrated energy management model that is beneficial to both the grid and consumers. Over time, energy management techniques and optimization algorithms have been incorporated into electronic management systems to provide reliable, clean, and affordable energy. The use of optimization algorithms for global energy demand and supply in power systems is being explored to meet economic load dispatch, system stability, and service quality.

[0003] However, if the system specifications change, such as variable energy supply due to renewable energy and variable consumer demand, the optimization problem needs to be reformulated to adapt to the new changes. Therefore, the optimized algorithm is not suitable for all parameter changes and is difficult to deploy and test in real time. In addition, the existing energy management system does not take into account the randomness of renewable energy, which will lead to large errors in the predicted energy demand data and energy supply data, making the predicted energy demand data and energy supply data have a large discrepancy with the actual consumer energy demand and consumer energy supply. Summary of the invention

[0004] The embodiments of the present application provide a training method for an energy data prediction model, an energy data prediction method and a device, which can improve the accuracy of predicted energy demand data and energy supply data.

[0005] In order to achieve the above objectives, the present application provides the following technical solutions:

[0006] A training method for an energy data prediction model, comprising:

[0007] Obtain basic data to be trained from the smart grid;

[0008] Inputting the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data;

[0009] According to the basic data to be trained and the optimized data to be trained, the energy data prediction model is trained based on a Gaussian process regression model.

[0010] Preferably, the basic data to be trained includes: E RES,max , ED,max , the consumer surplus energy optimization model includes:

[0011]

[0012] Subject to:

[0013]

[0014]

[0015] in, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, and Represents the consumer's surplus energy data at hour h.

[0016] Preferably, the basic data to be trained includes: p h 、E RES,max , E D,max , the consumer energy optimization cost model includes:

[0017]

[0018] Subject to:

[0019]

[0020]

[0021] Among them, p h represents the consumer energy import rate provided by the smart grid at hour h, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, Represents the consumer energy cost data at hour h.

[0022] Preferably, the basic data to be trained includes: p h 、E RES,max 、E D,max , the total energy cost model includes:

[0023]

[0024] Subject to:

[0025]

[0026] in, is the nominal price provided by the smart grid, is consumer energy imported from utilities, p h represents the consumer energy import rate provided by the smart grid at hour h, E RES,max Represents the maximum consumable energy generated by renewable energy, E D,max represents the maximum consumable energy demand, represents the consumer surplus energy data at hour h, and GR represents the total energy cost.

[0027] Preferably, the Gaussian process regression model includes:

[0028] y=h(x) T β+f(x),f(x)~GP(0,k(x,x)')

[0029] Where x represents the basic data to be trained, y represents the optimized data to be trained, the latent function f(x) represents the Gaussian process of the basic data to be trained based on the covariance k(x,x)', and h(x) T represents the explicit basis function of the basic data to be trained, β represents the coefficient of the basic data to be trained, GP represents the Gaussian process, and k(x,x)' represents the kernel function in the Gaussian process.

[0030] A method for energy data prediction, comprising:

[0031] Obtain basic data to be predicted in smart grid;

[0032] Based on the basic data to be predicted, the prediction results corresponding to the basic data to be predicted are obtained through the energy data prediction management model. The prediction results include consumer surplus energy data, consumer energy cost data, and total energy cost data. The energy data prediction management model is trained using the training method of the above-mentioned energy data prediction model.

[0033] A training device for an energy data prediction model, comprising:

[0034] An acquisition module, used for acquiring basic data to be trained from the smart grid;

[0035] The acquisition module is further used to input the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data;

[0036] The training module is used to train the energy data prediction model based on the Gaussian process regression model according to the basic data to be trained and the optimized data to be trained.

[0037] An energy data prediction device, comprising:

[0038] An acquisition module, used for acquiring basic data to be predicted in the smart grid;

[0039] A prediction module is used to obtain optimized data corresponding to the basic data to be predicted through an energy data prediction management model based on the basic data to be predicted, wherein the optimized data includes consumer surplus energy data, consumer energy cost data, and total energy cost data. The energy data prediction management model is trained using the training method of the above-mentioned energy data prediction model.

[0040] A computer device, comprising: a memory, a transceiver, a processor and a bus system;

[0041] Wherein, the memory is used to store programs;

[0042] The processor is used to execute the program in the memory to implement the various methods described above;

[0043] The bus system is used to connect the memory and the processor so that the memory and the processor can communicate with each other.

[0044] A computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the various methods described above.

[0045] The embodiment of the present application provides a training method for an energy data prediction model, including: obtaining basic data to be trained from a smart grid; inputting the basic data to be trained into an optimization model to obtain optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model, and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data, and total energy cost data; and training the energy data prediction model based on a Gaussian process regression model according to the basic data to be trained and the optimization data to be trained. In the above manner, the consumer surplus energy data, consumer energy cost data, and total energy cost data are used as training data, and the energy data prediction model is trained in combination with a Gaussian process regression model, taking into account the randomness of renewable energy, thereby improving the accuracy of the predicted energy demand data and energy supply data. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0047] Figure 1 A schematic diagram of an embodiment of a method for training an energy data prediction model in an embodiment of the present application;

[0048] Figure 2 A schematic diagram of an embodiment of the energy data prediction method in the embodiment of the present application;

[0049] Figure 3 A schematic diagram of an embodiment of a training device for an energy data prediction model in an embodiment of the present application;

[0050] Figure 4 It is a schematic diagram of an embodiment of an energy data prediction device in an embodiment of the present application. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0052] See also Figure 1 As shown in the figure, the training method of the energy data prediction model in the embodiment of the present application includes:

[0053] 101. Obtain basic data to be trained from the smart grid;

[0054] In this embodiment, basic data to be trained may be obtained from a smart grid, wherein the basic data to be trained may include consumer available energy, energy load, energy pricing, and the like.

[0055] It is understandable that smart grid is the intelligence of power grid, which is based on integrated, high-speed two-way communication network, and realizes the goal of reliable, safe, economical, efficient, environmentally friendly and safe use of power grid through the application of advanced sensing and measurement technology, advanced equipment technology, advanced control method and advanced decision support system technology. The main features of smart grid include self-healing, incentive and protection of users, defense against attacks, providing power quality that meets user needs, allowing access to various forms of power generation, starting power market and optimizing and efficient operation of assets, so the smart grid contains the basic data to be trained for training energy data prediction model.

[0056] 102. Input the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data;

[0057] In this embodiment, the data related to the consumer surplus energy optimization model in the basic data to be trained is input into the consumer surplus energy optimization model to obtain the consumer surplus energy data; the data related to the consumer energy cost optimization model in the basic data to be trained is input into the consumer energy cost optimization model to obtain the consumer energy cost data; the data related to the total energy cost model in the basic data to be trained and the consumer surplus energy data are input into the total energy cost model to obtain the total energy cost data.

[0058] For ease of understanding, the above content is explained as follows:

[0059] (1) Prosumer energy surplus (PES) optimization model

[0060] Most consumers generate electricity using renewable energy. After meeting the energy needs of consumers, the extra available energy is called PES. Consumer-based smart grids enable two-way energy trading between consumers and power companies through service level agreements. Therefore, two-way energy trading helps consumers sell ecosystem service prices at nominal prices. The optimization model of PES is as follows:

[0061]

[0062] Subject to:

[0063]

[0064]

[0065] in, represents the consumable energy generated by renewable energy, such as wind power or photovoltaic arrays at hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, and represents the PES at hour h. The randomness of the basic parameters leads to multiple power demand ranges. The definition is as follows:

[0066]

[0067] in, and represent the maximum, average and minimum PES at hour h, respectively. and represent the maximum, average and minimum consumable energy demand in hour h, and They represent the maximum, average and minimum consumable energy generated by renewable energy at hour h respectively.

[0068] (2) Prosumer energy cost (PEC) optimization model

[0069] Random renewable energy sources and consumers' energy demands may result in consumers' energy production being less than consumers' energy demand. At regular intervals, consumers import energy from utilities to overcome energy shortages. Consumers pay the cost of all energy imported from utilities, called PEC. The optimization model for PEC is expressed as:

[0070]

[0071] Subject to:

[0072]

[0073]

[0074] Among them, p hrepresents the consumer energy import rate provided by the smart grid at hour h, represents the PEC at hour h. The randomness of the basic parameters leads to various ranges of PEC, such as the minimum, average and maximum PEC are defined as:

[0075]

[0076] in, and represent the maximum, average and minimum PEC at hour h, respectively.

[0077] (3) Grid revenue (GR) optimization model

[0078] On the basis of a consumer-based smart grid (rather than a traditional power system), consumers benefit from PES and PEC, while the generation of GR increases. The optimization model of GR is expressed as:

[0079]

[0080] Subject to:

[0081]

[0082] in, is the nominal price provided by the smart grid, is the consumer energy imported from the utility. The randomness of the basic parameters leads to multiple ranges of GR, such as the minimum, mean and maximum GR defined as:

[0083]

[0084] Among them, GR max , GR mean and GR min Represent the maximum, average and minimum GR, respectively.

[0085] It should be noted that the above formulas (5), (10) and (14) will lead to different ranges of results (PES, PEC and GR), ranging from minimum to maximum, depending on the energy demand of consumers. and consumer energy generation

[0086] 103. According to the basic data to be trained and the optimized data to be trained, the energy data prediction model is trained based on a Gaussian process regression model.

[0087] In this embodiment, the energy data prediction model can be trained based on the Gaussian process regression model according to the basic data to be trained and the optimized data to be trained. Specifically, the obtained basic data to be trained and the optimized data to be trained are used as training data and input into the Gaussian process regression model for training, thereby obtaining the energy data prediction model. It can be understood that the energy data prediction model trained by the basic data to be trained and the optimized data to be trained can be used to predict accurate consumer energy data, consumer energy cost data, total energy cost data, etc. according to the basic data obtained in the smart grid.

[0088] It should be noted that machine learning is a technique that enables software applications to predict future responses of dependent variables by learning from a training dataset. A given machine learning method models the interrelationships between the covariates and the response variable of the training dataset. Gaussian process regression is a method in statistical theory where a finite set of random variables distributed over multiple time slots follows a multivariate Gaussian distribution. The Gaussian process regression model is defined as follows:

[0089] y=h(x) T β+f(x),f(x)~GP(0,k(x,x)') (15)

[0090] Where x represents the covariate (renewable energy, load, price), y represents the output (PES, PEC, GR), and the latent function f(x) represents the covariate ( p h , ) is based on a Gaussian process with covariance k(x,x)', h(x) T Explicit basis functions representing covariates ( p h , ) and β represent the coefficients of the covariates ( p h , ), GP represents Gaussian process, and k(x,x)' represents the kernel function in the Gaussian process. Gaussian process regression with zero mean and identical distribution is assumed here. Parameters without covariance are considered hyperparameters. The known hyperparameters are used as the basic assumptions for making inferences. At any time, the output y of the predicted Gaussian process regression model for the training dataset is defined as:

[0091] P(y i |f(x i ),x i )~N(y i |h(x i ) T β+f(x i ),σ 2 ) (16)

[0092] where σ is the standard deviation and σ 2 is the variance of the input covariate. Based on the training data set of the Gaussian process regression model, the model is trained using statistics and the machine learning toolbox. The results of the model are strictly compared with the optimization results of the traditional energy data prediction model based on genetic algorithm and particle swarm algorithm, which proves the effectiveness of the model.

[0093] In an embodiment of the present application, a training method for an energy data prediction model is provided, comprising: obtaining basic data to be trained from a smart grid; inputting the basic data to be trained into an optimization model to obtain optimization data to be trained, wherein the optimization model comprises a consumer surplus energy optimization model, a consumer energy cost optimization model, and a total energy cost model; the optimization data to be trained comprises consumer surplus energy data, consumer energy cost data, and total energy cost data; and training the energy data prediction model based on a Gaussian process regression model according to the basic data to be trained and the optimization data to be trained. In the above manner, the consumer surplus energy data, the consumer energy cost data, and the total energy cost data are used as training data, and the energy data prediction model is trained in combination with a Gaussian process regression model, taking into account the randomness of renewable energy, thereby improving the accuracy of the predicted energy demand data and energy supply data.

[0094] In combination with the above introduction, the method for energy data prediction in the embodiment of the present application will be introduced below. Figure 2 , Figure 2 This is a schematic diagram of an embodiment of the energy data prediction method in the embodiment of the present application. As shown in the figure, the energy data prediction method in the embodiment of the present application includes:

[0095] 201. Obtain basic data to be predicted in the smart grid;

[0096] In this embodiment, basic data to be predicted in the smart grid may be obtained, wherein the basic data to be predicted may include consumer available energy, energy load, energy pricing, and the like.

[0097] It can be understood that smart grid is the intelligence of power grid. It is built on the basis of integrated, high-speed two-way communication network. Through the application of advanced sensing and measurement technology, advanced equipment technology, advanced control methods and advanced decision support system technology, it realizes the goals of reliable, safe, economical, efficient, environmentally friendly and safe use of power grid. Its main features include self-healing, motivating and protecting users, resisting attacks, providing power quality that meets user needs, allowing access to various forms of power generation, starting the power market and optimizing and efficient operation of assets. Therefore, the smart grid contains the basic data to be predicted required for predicting energy data.

[0098] 202. Based on the basic data to be predicted, the prediction results corresponding to the basic data to be predicted are obtained through the energy data prediction management model, and the prediction results include consumer surplus energy data, consumer energy cost data, and total energy cost data. The energy data prediction management model is trained using the solution provided in the above embodiment.

[0099] In this embodiment, the basic data to be predicted can be used as the input of the energy data prediction management model, and then the prediction results corresponding to the basic data to be predicted are output through the energy data prediction management model, where the prediction results include consumer surplus energy data, consumer energy cost data, and total energy cost data.

[0100] Specifically, after obtaining the corresponding consumer surplus energy data, consumer energy cost data, and total energy cost data through the energy forecasting management model, it can be used to analyze consumer energy production and consumer energy consumption.

[0101] In an embodiment of the present application, a method for energy data prediction is provided. By using consumer surplus energy data, consumer energy cost data, and total energy cost data as training data and combining it with a Gaussian process regression model to train an energy data prediction model, the randomness of renewable energy is taken into account, thereby improving the accuracy of the predicted energy demand data and energy supply data.

[0102] In combination with the above introduction, the model training device in this application is described in detail below. Figure 3 , Figure 3 This is a schematic diagram of a training device for an energy data prediction model in an embodiment of the present application. As shown in the figure, the training device 30 for the energy data prediction model includes:

[0103] An acquisition module 301 is used to acquire basic data to be trained from a smart grid;

[0104] The acquisition module 301 is further used to input the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data;

[0105] The training module 302 is used to train the energy data prediction model based on the Gaussian process regression model according to the basic data to be trained and the optimized data to be trained.

[0106] Optionally, the basic data to be trained includes: E RES,max , E D,max, the consumer surplus energy optimization model includes:

[0107]

[0108] Subject to:

[0109]

[0110]

[0111] in, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, and Represents the consumer's surplus energy data at hour h.

[0112] Optionally, the basic data to be trained includes: p h 、E RES,max , E D,max , the consumer energy optimization cost model includes:

[0113]

[0114] Subject to:

[0115]

[0116]

[0117] Among them, p h represents the consumer energy import rate provided by the smart grid at hour h, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, Represents the consumer energy cost data at hour h.

[0118] Optionally, the total energy cost model includes:

[0119]

[0120] Subject to:

[0121]

[0122] in, is the nominal price provided by the smart grid, is consumer energy imported from utilities, p h represents the consumer energy import rate provided by the smart grid at hour h, E RES,max Represents the maximum consumable energy generated by renewable energy, E D,max represents the maximum consumable energy demand, represents the consumer surplus energy data at hour h, and GR represents the total energy cost.

[0123] Optionally, the Gaussian process regression model includes:

[0124] y=h(x) T β+f(x),f(x)~GP(0,k(x,x)')

[0125] Where x represents the basic data to be trained, y represents the optimized data to be trained, the latent function f(x) represents the Gaussian process of the basic data to be trained based on the covariance k(x,x)', and h(x) T represents the explicit basis function of the basic data to be trained, β represents the coefficient of the basic data to be trained, GP represents the Gaussian process, and k(x,x)' represents the kernel function in the Gaussian process.

[0126] In combination with the above introduction, the energy data prediction device in this application is described in detail below. Figure 4 , Figure 4 This is a schematic diagram of an energy data prediction device in an embodiment of the present application. As shown in the figure, the energy data prediction device 40 includes:

[0127] An acquisition module 401 is used to acquire basic data to be predicted in the smart grid;

[0128] The prediction module 402 is used to obtain the prediction results corresponding to the basic data to be predicted through the energy data prediction management model based on the basic data to be predicted, and the prediction results include consumer surplus energy data, consumer energy cost data, and total energy cost data. The energy data prediction management model is trained using the solution provided in the above embodiment.

[0129] The embodiment of the present application also provides a computer device, including: a memory, a transceiver, a processor, and a bus system;

[0130] Wherein, the memory is used to store programs;

[0131] The processor is used to execute the program in the memory to implement the above-mentioned energy data prediction model training method or energy data prediction method;

[0132] The bus system is used to connect the memory and the processor so that the memory and the processor can communicate with each other.

[0133] An embodiment of the present application also provides a computer-readable storage medium storing instructions, which, when executed on a computer, enables the computer to execute a training method for an energy data prediction model or a method for energy data prediction as described above.

[0134] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A training method for an energy data prediction model, characterized in that: include: Obtaining basic data to be trained from the smart grid, wherein the basic data to be trained includes: E RES,max , E D,max 、p h ; Inputting the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data; According to the basic data to be trained and the optimized data to be trained, the energy data prediction model is trained based on a Gaussian process regression model; Wherein, the consumer surplus energy optimization model includes: in, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, and represents the consumer's remaining energy data at hour h; The consumer energy optimization cost model includes: Among them, p h represents the consumer energy import rate provided by the smart grid at hour h, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, represents the consumer energy cost data at hour h; The total energy cost model includes: in, is the nominal price provided by the smart grid, is consumer energy imported from utilities, p h represents the consumer energy import rate provided by the smart grid at hour h, E RES,max Represents the maximum consumable energy generated by renewable energy, E D,max represents the maximum consumable energy demand, represents the consumer surplus energy data at hour h, GR represents the total energy cost; The Gaussian process regression model includes: y=h(x) T β+f(x),f(x)~GP(0,k(x,x)') Where x represents the basic data to be trained, y represents the optimized data to be trained, the latent function f(x) represents the Gaussian process of the basic data to be trained based on the covariance k(x,x)', and h(x) T represents the explicit basis function of the basic data to be trained, β represents the coefficient of the basic data to be trained, GP represents the Gaussian process, and k(x,x)' represents the kernel function in the Gaussian process.

2. A method for energy data prediction, characterized in that: include: Obtain basic data to be predicted in smart grid; Based on the basic data to be predicted, the prediction results corresponding to the basic data to be predicted are obtained through the energy data prediction model, and the prediction results include consumer surplus energy data, consumer energy cost data, and total energy cost data. The energy data prediction management model is trained using the method described in claim 1 above.

3. A training device for an energy data prediction model, characterized in that: include: The acquisition module is used to acquire basic data to be trained from the smart grid, wherein the basic data to be trained includes: E RES,max , E D,max 、p h ; The acquisition module is further used to input the basic data to be trained into the optimization model to obtain the optimization data to be trained, wherein the optimization model includes a consumer surplus energy optimization model, a consumer energy cost optimization model and a total energy cost model; the optimization data to be trained includes consumer surplus energy data, consumer energy cost data and total energy cost data; A training module, used for training the energy data prediction model based on the Gaussian process regression model according to the basic data to be trained and the optimized data to be trained; Wherein, the consumer surplus energy optimization model includes: in, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, and represents the consumer's remaining energy data at hour h; The consumer energy optimization cost model includes: Among them, p h represents the consumer energy import rate provided by the smart grid at hour h, represents the consumable energy generated by renewable energy in hour h, E RES,max Represents the maximum consumable energy produced by renewable energy, represents the consumable energy demand in hour h, E D,max represents the maximum consumable energy demand, represents the consumer energy cost data at hour h; The total energy cost model includes: in, is the nominal price provided by the smart grid, is consumer energy imported from utilities, p h represents the consumer energy import rate provided by the smart grid at hour h, E RES,max Represents the maximum consumable energy generated by renewable energy, E D,max represents the maximum consumable energy demand, represents the consumer surplus energy data at hour h, GR represents the total energy cost; The Gaussian process regression model includes: y=h(x) T β+f(x),f(x)~GP(0,k(x,x)') Where x represents the basic data to be trained, y represents the optimized data to be trained, the latent function f(x) represents the Gaussian process of the basic data to be trained based on the covariance k(x,x)', and h(x) T represents the explicit basis function of the basic data to be trained, β represents the coefficient of the basic data to be trained, GP represents the Gaussian process, and k(x,x)' represents the kernel function in the Gaussian process.

4. An energy data prediction device, characterized in that: include: An acquisition module, used for acquiring basic data to be predicted in the smart grid; A prediction module is used to obtain optimized data corresponding to the basic data to be predicted through an energy data prediction management model based on the basic data to be predicted, wherein the optimized data includes consumer surplus energy data, consumer energy cost data, and total energy cost data, and the energy data prediction management model is trained using the method described in claim 1 above.

5. A computer device, characterized in that: include: memories, transceivers, processors, and bus systems; Wherein, the memory is used to store programs; The processor is used to execute the program in the memory to implement the method of claim 1, or to implement the method of claim 2; The bus system is used to connect the memory and the processor so that the memory and the processor can communicate with each other.

6. A computer-readable storage medium, characterized in that: The device stores instructions which, when executed on a computer, enable the computer to execute the method according to claim 1 or the method according to claim 2.

Images