KR102521807B1 - Method for predicting energy production and consumption data using double sequence deep learning model, and apparatus thereof - Google Patents

Abstract

The present invention relates to a method performed by an energy prediction device, and more specifically, to a method and device for predicting energy production and consumption data using a double-sequence deep learning model, wherein the method includes the steps of: loading energy production data and consumption data; preprocessing energy production data and consumption data; inputting preprocessed data into a learned double-sequence deep learning model and outputting the predicted values regarding the energy production and consumption data; and updating the deep learning model by inputting the predicted values into the deep learning model.

Description

Method and device for predicting energy production and consumption data using double sequence deep learning model

The present invention relates to a method and apparatus for predicting renewable energy production and consumption by utilizing a dual sequence deep learning model, and more specifically to predicting renewable energy production and consumption by utilizing a deep learning model learned from preprocessed data It relates to a method and its device.

As response strategies for climate change and carbon neutrality issues are emphasized, the importance of renewable energy is increasing day by day. In particular, with the introduction of a specific execution framework for ESG management, the demand for renewable energy is rapidly increasing, and major companies around the world are also busily moving to secure renewable energy sources through self-installation or planned power purchase contracts.

Currently, the small electricity brokerage market operated by the Korea Power Exchange operates a forecast inventory settlement system that compensates brokerages based on forecast accuracy. The forecast inventory settlement is a method of paying expenses according to the error rate as an incentive for forecast accuracy. In addition, if the average error rate is calculated for a certain period of time and exceeds a certain rate, a penalty is imposed that restricts participation in the brokerage market.

Therefore, when installing renewable energy for business purposes, the accuracy of forecasting becomes one of the most important issues because it is directly related to profitability.

To this end, it is necessary to predict renewable energy more accurately with a very low error rate and discuss practical technologies that can utilize it.

The above-described background art is technical information that the inventor possessed for derivation of the present invention or acquired during the derivation process of the present invention, and cannot necessarily be said to be known to the general public prior to filing the present invention.

An object to be solved by some embodiments of the present invention is to provide a method and apparatus for more efficiently predicting production and consumption of renewable energy.

In addition, the problem to be solved through the present invention is to provide a method and apparatus utilizing a deep learning algorithm to predict the production and consumption of renewable energy.

An object to be solved by the present invention is to provide a method and apparatus for predicting energy production and consumption with a lower error rate by using a deep learning algorithm using a dual sequence deep learning model.

In the method performed by the energy prediction device according to some embodiments of the present invention for solving the above technical problem, loading energy production data and consumption data, preprocessing energy production data and consumption data, learning outputting predicted values for energy production and consumption data through inputting preprocessed data to the double-sequence deep learning model, and updating the deep learning model by inputting the predicted values into the deep learning model.

In one embodiment, the dual sequence deep learning model is composed of at least one core block, and the core block may be composed of four convolutional layers and a ReLu function as an activation function.

In one embodiment, at least one core block may further include a pooling layer in which hyper parameters are shared.

In one embodiment, the dual sequence deep learning model may include at least one Denso layer.

In one embodiment, the step of preprocessing the energy production data and consumption data may be characterized by removing outlier data and redundant data from the energy production data and consumption data, and performing a data normalization process.

According to the above-described problem solving means of the present invention, it is possible to control randomness and strong variability of energy production and consumption data and provide a dual sequence deep learning algorithm with a higher prediction rate than existing models as a model.

In addition, according to the problem solving means of the present invention, by providing a deep learning algorithm model with a significantly reduced error rate compared to existing models, it can be actively used in the smart grid industry by greatly improving energy production and consumption prediction efficiency, including renewable energy. .

In addition, according to the problem solving means of the present invention, by predicting energy production and consumption with a low error rate, it is possible to secure more profitability in the market utilizing renewable energy and to improve corporate profits.

1 illustrates an exemplary environment to which an energy prediction apparatus according to some embodiments of the present disclosure may be applied.
2 is a flowchart illustrating a method of predicting energy production and consumption data that may be performed in an energy prediction device according to some embodiments of the present disclosure.
3 is a diagram for explaining in detail a step of pre-processing energy production and consumption data.
4 is a diagram for detailed description of a method of learning a deep learning model of the present invention and outputting a predicted value from the learned deep learning model.
5 is a diagram for explaining the architecture of the deep learning model of the present invention.
6 is a diagram of an exemplary computing device in which devices and/or systems in accordance with various embodiments of the present disclosure may be implemented.

Hereinafter, preferred embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Advantages and features of the present disclosure, and methods of achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the technical spirit of the present disclosure is not limited to the following embodiments and may be implemented in various different forms, and only the following embodiments complete the technical spirit of the present disclosure, and the technical field to which the present disclosure belongs It is provided to fully inform those skilled in the art of the scope of the present disclosure, and the technical spirit of the present disclosure is only defined by the scope of the claims.

In adding reference numerals to the components of each drawing, it should be noted that the same components have the same numerals as much as possible, even if they are displayed on different drawings. In addition, in describing the present disclosure, if it is determined that a detailed description of a related known configuration or function may obscure the gist of the present disclosure, the detailed description will be omitted.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with a meaning that can be commonly understood by those of ordinary skill in the art to which this disclosure belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined. Terminology used herein is for describing the embodiments and is not intended to limit the present disclosure. In this specification, the singular also includes the plural unless specifically stated in the phrase.

Also, terms such as first, second, A, B, (a), and (b) may be used in describing the components of the present disclosure. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. When an element is described as being “connected,” “coupled to,” or “connected” to another element, that element is or may be directly connected to the other element, but there is another element between the elements. It should be understood that elements may be “connected”, “coupled” or “connected”.

As used herein, "comprises" and/or "comprising" means that a stated component, step, operation, and/or element is the presence of one or more other components, steps, operations, and/or elements. or do not rule out additions.

Hereinafter, various embodiments of the present disclosure will be described in detail with reference to the accompanying drawings.

Also, terms such as first, second, A, B, (a), and (b) may be used in describing the components of the present invention. These terms are only used to distinguish the component from other components, and the nature, order, or order of the corresponding component is not limited by the term. Throughout the specification, when a part 'includes' or 'includes' a certain component, it means that it may further include other components without excluding other components unless otherwise stated. . Also, terms such as 'unit' and 'module' described in the specification refer to a unit that processes at least one function or operation, and may be implemented by hardware, software, or a combination of hardware and software.

1 illustrates an exemplary environment to which an energy prediction apparatus according to some embodiments of the present disclosure may be applied. Through a system including the electronic device 100 and the energy prediction device 200 shown in FIG. 1 , it is possible to predict renewable energy production and consumption. Through this, the imbalance in renewable energy utilization can be resolved, and a schedule for renewable energy production and consumption can be established. It can be determined that the efficiency of the energy predicting device 200 is high as the predicted values of renewable energy production and consumption are accurate.

Hereinafter, operations of the components shown in FIG. 1 related to the energy prediction operation provided through the above-described system will be described in more detail.

1 shows an example in which the electronic device 100 and the energy prediction device 200 are connected through a network, but this is only for convenience of understanding, and the number of devices that can be connected to the network may vary. can

Meanwhile, FIG. 1 only illustrates a preferred embodiment for achieving the object of the present disclosure, and some components may be added or deleted as necessary. Hereinafter, the components shown in FIG. 1 will be described in more detail.

The energy prediction device 200 may collect and analyze various information about energy production and consumption generated by the electronic device 100 . Here, the energy prediction device 200 may process various collected and analyzed information.

The electronic device 100 may be an electronic product that uses electrical energy in daily life. More specifically, it may be a device that consumes energy through power and a device that produces power energy. For example, a device that consumes power energy is an air conditioner, a television (TV), a computer (computer), and a light bulb. (light), workstation (Workstation), laptop (Laptop), and the like. As another example, the device for generating power energy may include a solar collector plate for generating solar heat energy.

Power energy may refer to renewable energy, and renewable energy may include solar energy, hydraulic energy, wind energy, and the like, and renewable energy within a range that can be understood by a person skilled in the art may be included, , It is not construed as being limited to the above-described example.

The energy prediction device 200 may predict energy production and consumption by using a deep learning algorithm based on various collected information.

In order to exclude redundant descriptions, various operations performed by the energy prediction device 200 will be described in more detail later with reference to the drawings below in FIG. 2 .

Meanwhile, the energy prediction device 200 may be implemented with one or more computing devices. For example, all functions of the energy prediction device 200 may be implemented in a single computing device. As another example, the first function of the energy prediction device 200 may be implemented in a first computing device, and the second function may be implemented in a second computing device. Here, the computing device may be a notebook, a desktop, or a laptop, but is not limited thereto and may include any type of device equipped with a computing function. However, it may be preferable that the energy prediction device 200 be implemented as a high-performance server-class computing device. An example of a computing device will be described with reference to FIG. 6 .

In some embodiments, components included in the environment to which the energy prediction device 200 is applied may communicate through a network. The network may be implemented as all types of wired/wireless networks such as a Local Area Network (LAN), a Wide Area Network (WAN), a mobile radio communication network, and Wibro (Wireless Broadband Internet). can

On the other hand, the environment shown in FIG. 1 shows that the electronic device 100 and the energy prediction device 200 are connected through a network, but the scope of the present disclosure is not limited thereto, and the electronic device 100 ) should be noted that they may be connected through P2P (Peer to Peer).

So far, with reference to FIG. 1 , an exemplary environment to which the energy prediction device 200 according to some embodiments of the present disclosure can be applied has been described. Hereinafter, methods according to various embodiments of the present disclosure will be described in detail with reference to the drawings below in FIG. 2 .

Each step of the methods described below may be performed by a computing device. In other words, each step of the methods may be implemented as one or more instructions executed by a processor of a computing device. All of the steps involved in these methods could be executed by a single physical computing device, but first steps of the method are performed by a first computing device and second steps of the method are performed by a second computing device. may be performed.

In FIG. 2 , description will be continued on the assumption that each step of the methods is performed by the energy estimation device 200 illustrated in FIG. 1 . However, for convenience of description, the description of the subject of operation of each step included in the methods may be omitted.

2 is a flowchart illustrating a method of predicting energy production and consumption data that may be performed in an energy prediction device according to some embodiments of the present disclosure.

In step S100, the energy prediction device 200 may load energy consumption and production data.

As described above in FIG. 1 , the energy consumption and production data may be data communicated from an electronic device through a network. It goes without saying that the format of such data is not limited to any format. For example, the energy consumption and production data may be the amount of energy consumed by an electronic device during a unit period or the amount of energy produced by an energy generating device during a unit period.

In step S200, the energy prediction device 200 may pre-process energy consumption and production data. The process will be described in more detail with reference to FIG. 3 .

3 is a diagram for explaining in detail a step of pre-processing energy production and consumption data.

In step S210, the energy prediction device 200 may process outlier data and missing data in the energy production and consumption data.

It may include removing abnormal data such as outliers, missing values, and redundant values from energy consumption and production data. Such abnormal data may be data due to a failure or short circuit in the electronic device shown in FIG. 1 . In addition, in terms of energy production data, abnormal data may be data due to unpredictable meteorological changes in the case of an energy production device among electronic devices.

In step S220, the energy prediction device 200 may normalize energy production and consumption data. In step S220, the energy prediction device 200 may normalize data. The energy prediction device 200 may perform a normalization process because data values of energy consumption and production data are different from each other due to power consumption units of electronic devices.

In the energy estimation apparatus 200, the preprocessing processes of steps S210 and S220 may be performed sequentially or dynamically in parallel.

Returning to FIG. 2 , in step S300 , the energy prediction device 200 may output predicted values of energy consumption and production data by utilizing the learned deep learning model.

The trained deep learning model can predict energy consumption and production, which can predict energy consumption and production by outputting the predicted value through time-series analysis when data preprocessed by the deep learning model is input. A detailed description of this will be made in detail with reference to FIG. 4 .

4 is a diagram for detailed description of a method of learning a deep learning model of the present invention and outputting a predicted value from the learned deep learning model.

In step S310, the energy prediction device 200 may extract time features by inputting preprocessing data. In step S320, the energy prediction device 200 may extract spatial features by inputting preprocessing data.

In the present invention, the deep learning model is a dual sequence model, and may be composed of an architecture of two core blocks and a dense layer. The two core blocks consist of four convolutional layers, an activation function ReLU, and a pooling layer in which hyperparameters are shared across these blocks. The convolutional layer of the Core block serves as a feature extractor with activation functions, which are iteratively convolved with the cross-channels of the next layer contributing to extraction of advanced information in the temporal dimension. Also, a pooling layer can be integrated with the core block to facilitate overfitting.

In step S330, the energy prediction apparatus 200 may learn the dual sequence model based on temporal and spatial characteristics.

After the dual sequence model is learned, prediction values related to energy production and consumption data may be output.

Returning to FIG. 2 , in step S400 , the energy prediction device 200 may update the deep learning model by inputting a prediction value to the deep learning model.

Here, the energy prediction device 200 may determine the performance of the deep learning model. More specifically, the above-described test data may be used to determine the performance of the learned deep learning model.

The energy prediction device may calculate an error value between an actual value of the test data and a predicted value predicted through a deep learning model, and may determine a difference between the error value and a preset threshold value.

In step S400, the energy prediction device 200 may generate the learned deep learning model as a final model when the error value is smaller than a preset threshold value. Generating the final model may mean stopping the learning process and generating the deep learning model generated at the stopped point as the final model.

On the other hand, if the error value is smaller than a preset threshold value, deep learning model learning may continue. Through this, the process of improving the performance of the deep learning model can be continued. The specific numerical value of the preset threshold value is not limited to a specific numerical value, and it is natural that it can be any numerical value.

In addition, the preset threshold value is a value that may vary depending on the external environment. More specifically, the external environment may be an external environment that changes according to the electronic device shown in FIG. 1 or may be an external environment that changes according to the meteorological environment. For example, the preset threshold may adjust the threshold when the deviation of the renewable energy excavation value per unit period is equal to or greater than a predetermined reference value according to external environmental conditions. This adjustment may adjust the performance of the deep learning model of the present invention due to large variations in changes in external environmental conditions, thereby creating a deep learning model that can better adapt to changes in the external environment.

In the above-described adjustment of the threshold value, the threshold value may be adjusted lower as the external environment change deviation increases, or may be adjusted larger as the external environment change deviation increases.

Actual values of the above-described test data may be energy consumption and production, and these test data may be previously generated test data.

In order to determine the performance of the deep learning model, the error value may be, for example, a mean absolute error (MSE) value or a root mean square error (RMSE) value, but the interpretation is not limited thereto. Hereinafter, ESN, which is a type of regression neural network, will be described.

5 is a diagram for explaining the architecture of the deep learning model of the present invention.

In the present invention, the deep learning model is a dual sequence model, and may be composed of an architecture of two core blocks and a dense layer. The two core blocks consist of four convolutional layers, an activation function ReLU, and a pooling layer in which hyperparameters are shared across these blocks. The convolutional layer of the Core block serves as a feature extractor with activation functions, which are iteratively convolved with the cross-channels of the next layer contributing to extraction of advanced information in the temporal dimension. Also, a pooling layer can be integrated with the core block to facilitate overfitting.

The dual sequence model of the present invention can play a role of learning spatial and temporal information. Core blocks can be a combination of different core blocks and fully connected layers. The first core block of our dual sequence model is initialized with an input layer, followed by a convolutional layer and a pooling layer. The output of each Core block is fed into other Core blocks. In each core block, the kernel size of the convolution layer is 3, and the default stride and filter size can be 8. Here, is initialized to 1 and doubled in each core block. In each core block, the pooling layer is set to maximum pooling with a kernel size of 3, and there may be fully connected layers of sizes 32 and 12 following the core block.

In the present invention, only two core blocks are shown, but this corresponds to an example, and it is natural that the number can be added.

Hereinafter, with reference to FIG. 6, a system and apparatus in which the present invention may be disclosed will be described in detail.

6 is a diagram of an exemplary computing device in which devices and/or systems in accordance with various embodiments of the present disclosure may be implemented.

The computing device 1500 includes one or more processors 1510, a bus 1550, a communication interface 1570, a memory 1530 for loading a computer program 1591 executed by the processor 1510, A storage 1590 for storing the computer program 1591 may be included. However, only components related to the embodiment of the present disclosure are shown in FIG. 6 . Accordingly, those skilled in the art to which the present disclosure belongs may know that other general-purpose components may be further included in addition to the components shown in FIG. 6 .

The processor 1510 controls the overall operation of each component of the computing device 1500 . The processor 1510 includes a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), a graphics processing unit (GPU), or any type of processor well known in the art of the present disclosure. It can be. Also, the processor 1510 may perform an operation for at least one application or program for executing a method according to embodiments of the present disclosure. Computing device 1500 may include one or more processors.

Memory 1530 stores various data, commands and/or information. Memory 1530 may load one or more programs 1591 from storage 1590 to execute a method according to embodiments of the present disclosure. The memory 1530 may be implemented as a volatile memory such as RAM, but the technical scope of the present disclosure is not limited thereto.

The bus 1550 provides a communication function between components of the computing device 1500 . The bus 1550 may be implemented as various types of buses such as an address bus, a data bus, and a control bus.

The communication interface 1570 supports wired and wireless Internet communication of the computing device 1500 . Also, the communication interface 1570 may support various communication methods other than Internet communication. To this end, the communication interface 1570 may include a communication module well known in the art of the present disclosure.

According to some embodiments, communication interface 1570 may be omitted.

The storage 1590 may non-temporarily store the one or more programs 1591 and various data.

The storage 1590 may be a non-volatile memory such as read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, etc., a hard disk, a removable disk, or It may be configured to include any well-known computer-readable recording medium in any form.

Computer program 1591 may include one or more instructions that when loaded into memory 1530 cause processor 1510 to perform methods/operations in accordance with various embodiments of the present disclosure. That is, the processor 1510 may perform methods/operations according to various embodiments of the present disclosure by executing the one or more instructions.

So far, various embodiments of the present disclosure and effects according to the embodiments have been described with reference to FIGS. 1 to 6 . Effects according to the technical idea of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description of the specification.

The technical idea of the present disclosure described with reference to FIGS. 1 to 6 so far may be implemented as computer readable code on a computer readable medium. The computer-readable recording medium may be, for example, a removable recording medium (CD, DVD, Blu-ray disc, USB storage device, removable hard disk) or a fixed recording medium (ROM, RAM, computer-equipped hard disk). can The computer program recorded on the computer-readable recording medium may be transmitted to another computing device through a network such as the Internet, installed in the other computing device, and thus used in the other computing device.

In the above, even though all the components constituting the embodiments of the present disclosure have been described as being combined or operated as one, the technical idea of the present disclosure is not necessarily limited to these embodiments. That is, within the scope of the purpose of the present disclosure, all of the components may be selectively combined with one or more to operate.

Although actions are shown in a particular order in the drawings, it should not be understood that the actions must be performed in the specific order shown or in a sequential order, or that all shown actions must be performed to obtain a desired result. In certain circumstances, multitasking and parallel processing may be advantageous. Moreover, the separation of the various components in the embodiments described above should not be understood as requiring such separation, and the described program components and systems may generally be integrated together into a single software product or packaged into multiple software products. It should be understood that there is

Although the embodiments of the present disclosure have been described with reference to the accompanying drawings, those skilled in the art to which the present disclosure pertains may be implemented in other specific forms without changing the technical spirit or essential features. understand that you can. Therefore, the embodiments described above should be understood as illustrative in all respects and not limiting. The scope of protection of the present disclosure should be construed by the following claims, and all technical ideas within the equivalent scope should be construed as being included in the scope of rights of the technical ideas defined by the present disclosure.

Claims (10)

In the method performed by the energy prediction device,
loading energy production data and consumption data;
pre-processing the energy production data and consumption data;
outputting predicted values related to energy production and consumption data by inputting preprocessed data to the learned dual sequence deep learning model; and
Including; updating the deep learning model by inputting a prediction value to the deep learning model;
The dual sequence deep learning model is composed of at least one core block, and the core block is composed of four convolutional layers and a ReLu function as an activation function.
A method for predicting energy production and consumption data using a dual-sequence deep learning model. delete According to claim 1,
The at least one core block further comprises a pooling layer in which hyperparameters are shared.
A method for predicting energy production and consumption data using a dual-sequence deep learning model. According to claim 3,
The dual sequence deep learning model includes at least one Denso layer
A method for predicting energy production and consumption data using a dual-sequence deep learning model. According to claim 4,
The step of preprocessing the energy production data and consumption data removes outlier data and redundant data from the energy production data and consumption data, and performs a data normalization process.
A method for predicting energy production and consumption data using a dual-sequence deep learning model. processor;
network interface;
Memory; and
A computer program loaded into the memory and executed by the processor,
the processor,
instructions for loading energy production data and consumption data;
instructions for preprocessing energy production data and consumption data;
an instruction for inputting preprocessed data to the learned dual sequence deep learning model and outputting predicted values for energy production and consumption data; and
an instruction for updating the deep learning model by inputting a predicted value to the deep learning model; including,
The dual sequence deep learning model is composed of at least one core block, and the core block is composed of four convolutional layers and a ReLu function as an activation function.
A device for predicting energy production and consumption data using a dual sequence deep learning model. delete According to claim 6,
The at least one core block further comprises a pooling layer in which hyperparameters are shared.
A device for predicting energy production and consumption data using a dual-sequence deep learning model. According to claim 8,
The dual sequence deep learning model includes at least one Denso layer
A device for predicting energy production and consumption data using a dual-sequence deep learning model. According to claim 9,
The instruction for preprocessing the energy production data and consumption data removes outlier data and redundant data from the energy production data and consumption data, and performs a data normalization process.
A device for predicting energy production and consumption data using a dual-sequence deep learning model.

Images