System and method for providing optimal digital twin operation platform for improving reliability of smart energy community

Abstract

A system and method for providing an optimal digital twin operation platform, according to an embodiment, secure system stability for an industrial complex target operation platform, thereby improving power quality reliability. In addition, the system and method for providing an optimal digital twin operation platform for improving the reliability of a smart energy community, according to an embodiment, can quickly respond to the operation and management of the smart energy community by using a digital twin. In addition, in an embodiment, it is possible to simulate various situations by using the same data as an actual operating situation through modeling using the digital twin.

Description

A system and method for providing an optimal digital twin operation platform to improve the reliability of smart energy communities.

[0001] The present disclosure relates to a system and method for providing a digital twin optimal operation platform, and more specifically, to a system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community by analyzing data collected from various resources and facilities and providing a digital twin optimal operation service based on the analysis results.

[0002] Unless otherwise indicated herein, the materials described in this section are not prior art to the claims of this application, and their inclusion in this section is not intended to be admitted as prior art.

[0003] A digital twin is a virtual model created by digitally replicating an actual physical object or process. Digital twins are updated in real time using technologies such as sensor data, big data, artificial intelligence, and machine learning, and are used to predict and analyze interactions and behaviors with real-world objects.

[0004] A smart community is a local community that efficiently utilizes renewable energy within various social systems, including housing, buildings, transportation, and lifestyles. The existing Virtual Power Plant (VPP) distributed energy management and operation system used in smart communities operates individually according to the system's purpose and the types of resources involved, resulting in low operational efficiency and resource flexibility.

[0005] Furthermore, existing systems can only operate and evaluate under steady-state conditions, making them unable to respond to rapid changes in the grid. To overcome this, power system operation technologies that utilize real-time data-driven collective resource models are being proposed. Power system operation technologies can be categorized into steady-state-based and dynamic model-based technologies. Among these, steady-state-based technologies only consider ideal conditions, making them difficult to handle highly volatile data and potentially unable to address issues arising from transient conditions.

[0006] A transient state is the transition period between a random temporal change caused by a change in input and the return of output to a steady state. Most domestic power systems operate on AC power, with fixed tolerances for each component. Therefore, problems arising from transients during resource control and fault conditions can threaten system stability. Stability issues can degrade power quality even in demanding environments like industrial complexes. Operating platforms targeting industrial complexes require technologies that enhance power quality reliability by ensuring system stability. This is driving a growing demand for technologies that enable smart communities to utilize renewable energy and operate power systems more efficiently.

[0007] As the existing centralized energy supply system, centered on large-scale power generation, transitions to a smart energy supply system centered on regionally distributed energy sources, the development of distributed energy-centered grid operation technologies is urgently needed. Furthermore, existing smart community analysis and operation methods suffer from numerous issues with data quality and analysis speed, limiting their application at the microgrid level. For example, in the event of a fault or accident within a conventional power system, the omission or delay of specific data can hinder system recovery, potentially posing a risk to the power system. Furthermore, prolonged dynamic load transients can lead to more serious secondary accidents. Furthermore, even with the previously defined real-time aggregate load model, the massive amount of data within the system makes rapid response to faults and accidents challenging.

[0008] The system and method for providing a digital twin optimal operation platform according to an embodiment secures system stability for an operation platform targeting an industrial complex, thereby improving power quality reliability.

[0009] In addition, the system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment enable rapid response to the operation and management of a smart energy community through a digital twin.

[0010] Additionally, in the embodiment, modeling through digital twins enables simulation of various situations using data identical to actual operating situations.

[0011] In addition, through examples, we provide a federated learning framework in a distributed environment that enables artificial intelligence model learning with minimal data transfer without the need for raw data transfer.

[0012] The system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment more accurately analyzes the impact of a grid linked with distributed energy and expands the acceptance of distributed energy through precise modeling of energy resources.

[0013] In addition, it can contribute to deriving a stable and optimal operation and design plan for smart energy communities.

[0014] Furthermore, the embodiment operates the system as a dynamic model of the target customer's resources, reflecting real-time data and controlling individual resources through transient analysis to respond to transient events. Furthermore, this system can contribute to enhancing reliability by improving the quality of power supply within industrial complexes.

[0015] In this embodiment, a digital twin system and operating platform for a smart energy community are built, and various energy resource precision modeling and SILS (Software in the Loop Simulation) verification tools are provided.

[0016] Additionally, through examples, we perform system and resource operations to expand distributed energy acceptance, and modeling to ensure data security and privacy preservation.

[0017] Additionally, the embodiment implements sensor malfunction monitoring technology and unsafe data processing. Furthermore, data security is considered during the process of collecting data from various stakeholders. Data acquired from different measuring devices and environments is time-synchronized to the same time zone, generating and analyzing input data.

[0018] The system, which provides a digital twin optimal operation platform for improving the reliability of smart energy communities, collects monitoring data from communication objects included in the smart energy community, builds a digital twin for the smart energy community using the collected monitoring data, and creates an optimal operation process for the smart energy community through the digital twin.

[0019] In this embodiment, a simulation is performed using data identical to the operating status of a smart energy community through a digital twin, and the simulation results are analyzed.

[0020] The system and method for providing a digital twin optimal operation platform for improving the reliability of smart energy communities, as described above, enable power system operators to utilize small-scale distributed energy resources that could not be considered in the power supply process before the introduction of digital twin-based operation technology through aggregation technology, thereby maintaining a more stable supply-demand balance in the power system.

[0021] In addition, by using distributed energy resources installed in the region to stabilize regional voltage at the distribution system level as well as frequency for the entire power system through examples, it is possible to avoid investment in new facilities for future transmission and distribution stabilization.

[0022] In addition, the system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment guarantees profit diversity by enabling distributed energy owners to participate in the market as an emergency power generation source, unlike when they participated in the market with only simple power generation or load reduction amounts.

[0023] Additionally, by registering as a participant resource on the platform, you reduce the management burden, including monitoring, of the resources you own.

[0024] Furthermore, in this embodiment, by utilizing the digital twin-based operating platform developed from the perspective of new investors in distributed energy resources, even if they do not currently possess distributed energy resources, it becomes possible to identify in advance the capacity that can be linked to the connection point in the relevant area, thereby significantly reducing the cost of prior review.

[0025] Furthermore, applying collective operations can alleviate difficulties in the market participation process for revenue generation. Furthermore, by providing a federated learning framework in a distributed environment that enables AI model training with minimal data transfer required (raw data), the platform itself secures data, protects privacy, and prevents network overload.

[0026] In addition, the performance of the AI ​​model is improved through clustering of sensors based on physical distance / data similarity and personalization of the AI ​​model based on the learned AI model based on the federated learning framework.

[0027] In addition, through examples, we secure graph-based embedding technology and improve the reliability of the platform itself through automatic sensor / system malfunction detection technology.

[0028] In addition, it enables rapid response to unexpected disturbances in the power system, such as load changes and failures, and improves the accuracy performance of the platform itself through statistical and artificial intelligence-based algorithms for processing lost and unstable data.

[0029] The effects of the present invention are not limited to the effects described above, and should be understood to include all effects that can be inferred from the detailed description of the present invention or the composition of the invention described in the claims.

[0030] Figure 1 is a diagram illustrating a system for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment.

[0031] Figure 2 is a block diagram showing a server according to an embodiment.

[0032] Figure 3 is a diagram showing a data processing configuration of a processor according to an embodiment.

[0033] Figure 4 is a diagram showing the process of enhancing the operating scenario through abnormal waveform analysis according to an embodiment.

[0034] Figure 5 is a diagram showing the optimal operation process of a smart energy community using a digital twin according to an embodiment.

[0035] Hereinafter, the embodiments disclosed in this specification will be described in detail with reference to the attached drawings. Regardless of the drawing numbers, identical or similar components will be given the same reference numbers, and redundant descriptions thereof will be omitted. The suffixes "module" and "part" used for components in the following description are assigned or used interchangeably only for the convenience of writing the specification, and do not in themselves have distinct meanings or roles. In addition, when describing the embodiments disclosed in this specification, if it is determined that a specific description of a related known technology may obscure the gist of the embodiments disclosed in this specification, a detailed description thereof will be omitted. In addition, the attached drawings are only intended to facilitate easy understanding of the embodiments disclosed in this specification, and the technical ideas disclosed in this specification are not limited by the attached drawings, and should be understood to include all modifications, equivalents, and substitutes included in the spirit and technical scope of the present invention.

[0036] Terms that include ordinal numbers, such as first, second, etc., may be used to describe various components, but the components are not limited by these terms. These terms are used solely to distinguish one component from another.

[0037] When a component is referred to as being "connected" or "connected" to another component, it should be understood that it may be directly connected or connected to that other component, but that there may be other components intervening. Conversely, when a component is referred to as being "directly connected" or "connected" to another component, it should be understood that there are no other components intervening.

[0038] In this application, terms such as “include” or “have” are intended to specify the presence of a feature, number, step, operation, component, part or combination thereof described in the specification, but should be understood not to exclude in advance the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts or combinations thereof.

[0039] In this specification, the term "unit" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Furthermore, a single unit may be realized using two or more pieces of hardware, and two or more units may be realized by a single piece of hardware.

[0040] Some of the operations or functions described herein as being performed by a terminal, apparatus, or device may instead be performed by a server connected to the terminal, apparatus, or device. Similarly, some of the operations or functions described herein as being performed by a server may also be performed by a terminal, apparatus, or device connected to the server.

[0041] Hereinafter, the present invention will be described in detail with reference to the attached drawings.

[0042] Figure 1 is a diagram illustrating a system for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment.

[0043] Referring to FIG. 1, a system for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment may be configured to include a smart energy community (200) and a server (100). The smart energy community (200) is a local community that efficiently utilizes renewable energy within various social systems such as housing, buildings, transportation, and lifestyles. The smart energy community according to the embodiment is an eco-friendly society that reduces waste by highly managing the flow of natural resources, energy, and waste, and is a system that includes communication objects such as housing, buildings, and vehicles.

[0044] The server (100) collects monitoring data from the smart energy community (200) and uses this to design and create a digital twin that simulates the smart energy community. In an embodiment, the server (100) simulates monitoring data using a digital twin model to enable optimal operation and management of the smart energy community. In addition, the server (100) monitors the smart energy community through the digital twin, thereby predicting the safety and soundness of the smart energy community (200) and responding to and preventing the occurrence of abnormal events.

[0045] Fig. 2 is a block diagram showing a server according to an embodiment.

[0046] Referring to FIG. 2, the interview server (100) according to the embodiment may include a communication unit (110), a memory (120), and a processor (130). Here, the configuration of the server (100) illustrated in FIG. 2 is merely a simplified example. In one embodiment, the server (100) may include other components for implementing the computing environment of the server (100), and only some of the disclosed components may constitute the server (100).

[0047] The communication unit (110) may be configured regardless of the communication mode, such as wired or wireless, and may be configured with various communication networks, such as a personal area network (PAN) and a wide area network (WAN). In addition, the communication unit (110) may operate based on the well-known World Wide Web (WWW), and may also utilize a wireless transmission technology used for short-distance communication, such as infrared (IrDA: Infrared Data Association) or Bluetooth. For example, the communication unit (110) may be responsible for transmitting and receiving data required to perform a technique according to an embodiment of the present disclosure.

[0048] The memory (120) may refer to any type of storage medium. For example, the memory (120) may include at least one type of storage medium among a flash memory type, a hard disk type, a multimedia card micro type, a card type memory (e.g., SD or XD memory, etc.), a RAM (Random Access Memory), a SRAM (Static Random Access Memory), a ROM (Read-Only Memory), an EEPROM (Electrically Erasable Programmable Read-Only Memory), a PROM (Programmable Read-Only Memory), a magnetic memory, a magnetic disk, and an optical disk. This memory (120) may also constitute a database required for digital twin design, monitoring data analysis, simulation, etc.

[0049] The memory (120) can store at least one instruction that can be executed by the processor (130). In addition, the memory (120) can store any form of information generated or determined by the processor (130) and any form of information received by the interview feedback providing device (100).

[0050] The memory (120) may store various types of modules or models. A model in this specification may refer to any form of computer program that operates based on a network function, an artificial neural network, and / or a neural network. Throughout this specification, the terms model, neural network, network function, and neural network may be used interchangeably. A neural network is a network in which one or more nodes are interconnected through one or more links to form an input node and an output node relationship within the neural network. The characteristics of the neural network may be determined based on the number of nodes and links within the neural network, the correlation between the nodes and links, and the weight value assigned to each link. A neural network may be composed of a set of one or more nodes. A subset of the nodes constituting the neural network may constitute a layer.

[0051] A deep neural network (DNN) may refer to a neural network that includes multiple hidden layers in addition to an input layer and an output layer. A deep neural network may include a convolutional neural network (CNN), a recurrent neural network (RNN), an autoencoder, a generative adversarial network (GAN), a restricted boltzmann machine (RBM), a deep belief network (DBN), a Q network, a U network, a Siamese network, a generative adversarial network (GAN), a transformer, and the like. The description of the above-described deep neural network is merely an example, and the present disclosure is not limited thereto.

[0052] Neural networks can learn through at least one of the following methods: supervised learning, unsupervised learning, semi-supervised learning, self-supervised learning, or reinforcement learning. Neural network learning can be the process of applying knowledge to the neural network to perform a specific action.

[0053] Neural networks can be trained to minimize output errors. This process involves repeatedly inputting training data into the neural network, calculating the neural network output and target error for the training data, and backpropagating the neural network error from the output layer to the input layer to update the weights of each node in the neural network to reduce the error. In supervised learning, labeled data is used for each training data, while unsupervised learning uses unlabeled data. The amount of change in the connection weights of each updated node can be determined by the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate can vary depending on the number of iterations in the neural network's training cycle. Additionally, to prevent overfitting, methods such as increasing the learning data, regularization, dropout that disables some nodes, and batch normalization layers can be applied.

[0054] In one embodiment, the model may borrow at least a portion of a transformer. The transformer may be composed of an encoder that encodes embedded data and a decoder that decodes the encoded data. The transformer may have a structure that receives a series of data and outputs a series of data of different types through encoding and decoding steps. In one embodiment, the series of data may be processed into a form operable by the transformer. The process of processing the series of data into a form operable by the transformer may include an embedding process. Expressions such as data tokens, embedding vectors, and embedding tokens may refer to data embedded in a form operable by the transformer.

[0055] To encode and decode a series of data, a transformer can utilize an attention algorithm to process the encoders and decoders within the transformer. An attention algorithm can refer to an algorithm that, for a given query, calculates the similarity for one or more keys, reflects this similarity in the values ​​corresponding to each key, and then weights and sums the values ​​to which the similarity is reflected to calculate an attention value.

[0056] Depending on how the query, key, and value are set, various types of attention algorithms can be categorized. For example, if attention is obtained by setting the query, key, and value all to the same value, this could be a self-attention algorithm. If attention is obtained by reducing the dimensionality of the embedding vector to process a series of input data in parallel and then generating individual attention heads for each segmented embedding vector, this could be a multi-head attention algorithm.

[0057] In one embodiment, the transformer may be composed of modules that perform multiple multi-head self-attention algorithms or multi-head encoder-decoder algorithms. In one embodiment, the transformer may also include additional components other than attention algorithms, such as embedding, normalization, and softmax. Methods for constructing a transformer using attention algorithms may include methods disclosed in Vaswani et al., Attention Is All You Need, 2017 NIPS, which is incorporated herein by reference.

[0058] Neural networks can be trained to minimize output errors. This process involves repeatedly inputting training data into the neural network, calculating the neural network output and target error for the training data, and backpropagating the neural network error from the output layer to the input layer to update the weights of each node in the neural network to reduce the error. In supervised learning, labeled data is used for each training data, while unsupervised learning uses unlabeled data. The amount of change in the connection weights of each updated node can be determined by the learning rate. The neural network's calculation of input data and backpropagation of errors can constitute a learning cycle (epoch). The learning rate can vary depending on the number of iterations in the neural network's training cycle. Additionally, to prevent overfitting, methods such as increasing the learning data, regularization, dropout that disables some nodes, and batch normalization layers can be applied.

[0059] In one embodiment, the model may borrow at least a portion of a transformer. The transformer may be composed of an encoder that encodes embedded data and a decoder that decodes the encoded data. The transformer may have a structure that receives a series of data and outputs a series of data of different types through encoding and decoding steps. In one embodiment, the series of data may be processed into a form operable by the transformer. The process of processing the series of data into a form operable by the transformer may include an embedding process. Expressions such as data tokens, embedding vectors, and embedding tokens may refer to data embedded in a form operable by the transformer.

[0060] To encode and decode a series of data, a transformer can utilize an attention algorithm to process the encoders and decoders within the transformer. An attention algorithm can refer to an algorithm that, for a given query, calculates the similarity for one or more keys, reflects this similarity in the values ​​corresponding to each key, and then weights and sums the values ​​to which the similarity is reflected to calculate an attention value.

[0061] Depending on how the query, key, and value are set, various types of attention algorithms can be categorized. For example, if attention is obtained by setting the query, key, and value all to the same value, this could be a self-attention algorithm. If attention is obtained by reducing the dimensionality of the embedding vector to process a series of input data in parallel and then generating individual attention heads for each segmented embedding vector, this could be a multi-head attention algorithm.

[0062] In one embodiment, the transformer may be composed of modules that perform multiple multi-head self-attention algorithms or multi-head encoder-decoder algorithms. In one embodiment, the transformer may also include additional components other than attention algorithms, such as embedding, normalization, and softmax. Methods for constructing a transformer using attention algorithms may include methods disclosed in Vaswani et al., Attention Is All You Need, 2017 NIPS, which is incorporated herein by reference.

[0063] Transformers can be applied to various data domains, such as embedded segmented image data and audio waveforms, to transform a series of input data into a series of output data. To transform data from various data domains into a series of data that can be input to the transformer, the transformer can embed the data. The transformer can process additional data that expresses the relative positional or phase relationship between the series of input data. Alternatively, vectors expressing the relative positional or phase relationship between the input data can be additionally reflected in the series of input data to embed the series of input data.

[0064] In one embodiment, the estimation model may include a Recurrent Neural Network (RNN), a Long Short Term Memory (LSTM) network, a Bidirectional Encoder Representations from Transformers (BERT), or a Generative Pre-trained Transformer (GPT).

[0065] In one embodiment, the estimation model may be a model trained using transfer learning. Transfer learning, in this context, refers to a learning method that pre-trains a large amount of unlabeled training data using semi-supervised or self-learning methods to obtain a pre-trained model for a first task, then fine-tunes the pre-trained model to suit a second task and trains it using supervised learning on labeled training data to implement a target model.

[0066] The processor (130) according to the embodiment can perform technical features according to the embodiments of the present disclosure to be described later by executing at least one instruction stored in the memory (120). In one embodiment, the processor (130) can be configured with at least one core and can include a processor for data analysis and / or processing, such as a central processing unit (CPU), a general purpose graphics processing unit (GPGPU), a tensor processing unit (TPU), etc. of the computer device (100).

[0067] Fig. 3 is a diagram showing a data processing configuration of a processor according to an embodiment.

[0068] Referring to FIG. 3, the processor (130) of the server according to the embodiment may be configured to include a collection unit (131), a preprocessing unit (132), a deep learning unit (133), a digital twin design unit (134), a simulation unit (135), an analysis unit (136), a process extraction unit (137), an abnormal situation analysis unit (138), and a feedback unit (139). The term 'unit' used in this specification should be interpreted to include software, hardware, or a combination thereof, depending on the context in which the term is used. For example, the software may be machine language, firmware, embedded code, and application software. As another example, the hardware may be a circuit, a processor, a computer, an integrated circuit, an integrated circuit core, a sensor, a MEMS (Micro-Electro-Mechanical System), a passive device, or a combination thereof.

[0069] The collection unit (131) collects monitoring data from communication objects included in the smart energy community. In an embodiment, the communication objects may include vehicles, buildings, houses, communication devices, etc. of the smart energy community, and the monitoring data is a series of data generated from each communication object. For example, the monitoring data includes, but is not limited to, power usage data, energy production data, building internal environment data, power quality data, charging infrastructure data, and energy usage pattern data of the smart energy community. In addition, in an embodiment, the collection unit (131) collects training data of a deep learning model including a digital twin creation model, a simulation model, a result analysis model, an optimal process extraction model, and an abnormality judgment model.

[0070] The preprocessing unit (132) preprocesses the collected AI learning data to remove biased or discriminatory data. In an embodiment, the preprocessing unit (132) preprocesses the collected data and processes it into a form suitable for AI model learning. For example, the preprocessing unit (132) may perform processes such as noise removal, outlier removal, and missing value processing. In addition, the preprocessing unit (132) may normalize the data, remove outliers, or adjust the data scale through data preprocessing to prevent the model from learning unnecessary patterns.

[0071] Additionally, the preprocessing unit (132) performs preprocessing of the collected monitoring data. In an embodiment, the preprocessing unit (132) can classify monitoring data collected from the smart energy community according to the providing object and time-synchronize monitoring data collected from other objects to generate model input data.

[0072] For example, the preprocessing unit (132) receives various monitoring data from the collection unit (131) and classifies the transmitted data into various types such as energy usage, production volume, environmental conditions, and power quality. Specifically, the preprocessing unit (132) appropriately classifies the monitoring data according to the providing object. For example, power usage data can be classified into an energy management system, and environmental data can be classified into an internal building system. Since there may be multiple communication objects in a smart energy community, and each of these can collect data at different times, time synchronization is a process of accurately synchronizing the time of each data in order to use the data collected from other objects as model input data. In the embodiment, time synchronization of the collected data enables integrated utilization of data collected at different times. In addition, the preprocessing unit (132) performs preprocessing tasks such as missing value processing, outlier removal, and normalization of the monitoring data, thereby configuring the input data according to the characteristics of the data and the requirements of the model, thereby generating learning data for optimal model learning.

[0073] The deep learning unit (133) trains a deep learning neural network with learning data to implement a deep learning model. In an embodiment, the deep learning model may include a digital twin design model, a simulation model, a simulation result analysis model, an optimal process extraction model, an abnormal situation prediction model for a smart energy community, an abnormal data detection model, etc. In an embodiment, the digital twin design model is an artificial neural network model that designs a digital replica model of an object or system, accurately models a smart energy community with digital data and information, and digitally simulates and analyzes the operation, performance, status, etc. of the object. The simulation model is a model that simulates a digital twin using various monitoring data collected from a smart energy community, and is an artificial neural network model that creates a simulation scenario using actual monitoring data and applies it to the digital twin. The simulation result analysis model is a model that analyzes the simulation results of the digital twin, extracts characteristic data for each simulation, and predicts actual information related to the simulation. The optimal process extraction model is an artificial neural network model that extracts an optimal response process for abnormal situations and events based on simulation results. In an embodiment, the optimal process extraction model may include a model that extracts various control processes for optimal operation, such as improving power efficiency, energy efficiency, and management efficiency, of a smart energy community. The abnormal situation prediction model for a smart energy community is a model that predicts abnormal situations in a smart energy community based on monitoring data and data analysis results generated from a digital twin. In an embodiment, the abnormal situations include, but are not limited to, energy usage abnormalities, energy production abnormalities, communication abnormalities, environmental sensor abnormalities, and power quality abnormalities. The abnormal data detection model is a deep learning model that detects abnormal data associated with an abnormal situation from sensor data. In an embodiment, the abnormal data detection model can detect data outside the normal range as abnormal data through waveform analysis.

[0074] The digital twin design unit (134) creates a digital twin for the smart energy community using monitoring data from the smart energy community and a digital twin creation model. In one embodiment, the digital twin design unit (134) models and simulates monitoring data collected from the smart energy community to build a digital twin model that simulates the smart energy community.

[0075] In an embodiment, the digital twin design unit (134) designs the structure of a digital twin model based on preprocessed data. The digital twin design unit (134) can configure the model by considering the characteristics of the smart energy community's system components, energy production system, and energy usage system. Then, the designed digital twin model is executed in a simulation environment to simulate operations similar to those of an actual smart energy community. The digital twin design unit (134) simulates the energy flow, performance, status, etc. of the smart energy community through the digital twin model.

[0076] In the embodiment, the digital twin design unit (134) verifies the accuracy and validity of the digital twin model by comparing simulation results with data from an actual smart energy community. The digital twin design unit (134) may perform tasks such as adjusting model parameters to reduce errors in the digital twin model. Furthermore, since the smart energy community operates in a real-time, constantly changing environment, the digital twin model also continuously updates data to reflect the latest information.

[0077] The simulation unit (135) performs a simulation using data identical to the operating status of a smart energy community through a digital twin. To this end, the simulation unit (135) prepares data to be input into the digital twin model. In an embodiment, the data to be input into the digital twin model is data from the same point in time as the operating status of the smart energy community, and includes information such as energy usage, production volume, environmental conditions, and power quality. Thereafter, the simulation unit (135) digitally simulates the operating status of the smart energy community through the digital twin model and the input data. The model calculates the energy flow, performance, and status of the system based on the input data and executes the simulation.

[0078] The analysis unit (136) analyzes the simulation results to evaluate the operational status of the smart energy community. In an embodiment, the analysis unit (136) inputs monitoring data of the smart energy community into a digital twin model, analyzes the energy flow and status of the digital twin, and evaluates the operational status of the energy community based on the analysis results. For example, the analysis unit (136) analyzes information on the smart energy community's energy usage, production, efficiency, etc. to identify the strengths and weaknesses of its operational status.

[0079] Specifically, the analysis unit (136) analyzes the energy usage data of the digital twin to identify energy usage patterns by time zone, day, and month. This allows the analysis unit to identify major causes of energy usage, peak hours, and potential for energy savings. Furthermore, the analysis unit (136) analyzes the production data of the digital twin's energy production system (solar power, wind power, etc.) to identify the difference between expected and actual production. This allows the analysis unit to evaluate the efficiency and performance of the production system. Furthermore, the analysis unit (136) compares the energy usage and production of the digital twin to assess the system's energy efficiency. For example, it analyzes whether the produced energy is being utilized efficiently and whether there is energy loss.

[0080] Additionally, the analysis unit (136) evaluates operating costs and revenue based on the energy usage and production data of the digital twin. In the embodiment, the digital twin can be used to analyze economic feasibility by considering the operating costs and energy purchase costs of the production system.

[0081] Additionally, in the embodiment, the analysis unit (136) visualizes the analysis results and presents them in an easily understandable format. Graphs, charts, heat maps, and the like are utilized to visually understand the data's characteristics and patterns. Furthermore, the analysis unit (136) provides a digital twin utilization solution that visualizes collected monitoring data, model input data, and simulation result data using 3D, AR (Augmented Reality), and VR (Virtual Reality) technologies.

[0082] Afterwards, the analysis unit (136) analyzes the simulation results of the digital twin to identify the strengths and weaknesses of the smart energy community's operational status. For example, the analysis unit (136) determines whether energy usage is being managed efficiently and whether the production system's performance is stable.

[0083] In addition, in the embodiment, the analysis unit (136) compares the simulation results with actual operation data to verify accuracy and validity. To this end, the analysis unit (136) collects actual operation data of the smart energy community. This data includes information such as energy usage, production volume, environmental conditions, and power quality, and is data from the same point in time as the input data used by the simulation unit. The collected actual operation data is then compared with the simulation data. Data normalization, scaling, and missing value processing are performed to ensure data quality is consistent. The analysis unit (136) then compares the simulation results of the digital twin model executed by the simulation unit with the actual operation data. In the embodiment, statistical analysis or visual comparison can be performed to assess the similarity of the two data. In the embodiment, the analysis unit (136) verifies the accuracy of the model by evaluating the differences between the simulation results and the actual data. At this stage, the analysis analyzes how closely the model's predictions match the actual data and the level of error. Furthermore, the analysis unit (136) evaluates the validity of the digital twin model by assessing whether the model accurately reflects the actual situation. To this end, the validity of the model is verified by comparing simulation results under various scenarios and conditions with actual data. The analysis unit (136) then analyzes the results to draw conclusions about the accuracy and validity of the model. If the model's performance is excellent and reliable, energy management strategies can be established or optimal operating plans can be derived based on the simulation results. Furthermore, the analysis unit (136) performs model improvement work if the model's performance is insufficient or errors occur through comparison of simulation results with actual data. Model accuracy and validity are enhanced through tasks such as adjusting model parameters, collecting additional data, and modifying the model structure.

[0084] The process extraction unit (137) explores ways to optimize the operation of the smart energy community based on the simulation results of the digital twin. Furthermore, it analyzes and compares the results of different operating conditions simulated under various scenarios, thereby generating improvement measures to strengthen the strengths and address weaknesses of the smart energy community operation. For example, the process extraction unit (137) can propose energy conservation policies, production system upgrades, and energy efficiency improvement measures.

[0085] Additionally, the process extraction unit (137) analyzes information on the smart energy community's energy usage, production, and efficiency to identify operational strengths and weaknesses. This allows it to suggest more efficient and intelligent operational strategies, contributing to improving the performance of the smart energy community.

[0086] Additionally, the process extraction unit (137) generates an operation schedule for the smart energy community based on the simulation data of the digital twin. For example, the process extraction unit (137) can analyze the simulation results and derive an optimal operation schedule for the smart energy community based on the analysis results. In an embodiment, the optimal operation schedule may include a smart energy community control sequence that maintains a balance between energy production and consumption and maximizes the stability and efficiency of the power grid. In an embodiment, the process extraction unit (137) continuously updates the digital twin model during actual smart energy community operation and monitors the operating status by comparing it with real-time data to derive an optimal operation schedule. Thereafter, the smart energy community operation schedule can be adjusted and optimized based on the operation status monitoring results.

[0087] The abnormal situation analysis unit (138) collects abnormal situation data related to the occurrence of an abnormal situation from among the collected monitoring data, analyzes the abnormal situation data, and extracts an optimal process for the abnormal situation. In an embodiment, the abnormal situation data includes the object in which the abnormal situation occurred, the type of abnormal event, and monitoring data of the object for a certain period of time prior to the occurrence of the abnormal event.

[0088] In addition, in the embodiment, the abnormal situation analysis unit (138) generates an optimal operation scenario through abnormal analysis and diagnosis technology using a deep learning model. In addition, in the embodiment, the abnormal situation analysis unit (138) calculates market participation capacity through abnormal waveform analysis, and provides a system stability operation service within the industrial complex based on the abnormal waveform analysis results. In addition, the abnormal situation analysis unit (138) performs sensor malfunction detection to improve the reliability of the smart energy community. In the embodiment, the analysis unit (138) detects sensor malfunction (Anomaly Detection) based on normal data characteristics through a real-time time series data characteristic analysis module. In addition, the abnormal situation analysis unit (138) optimizes the update cycle to reflect spatiotemporal changes in normal data characteristics, and performs loss and unstable data processing to improve platform accuracy. For example, the abnormal situation analysis unit (138) generates statistics for loss and unstable data correction. Additionally, we provide AI algorithms for data correction and low-dimensional embedding techniques that preserve only essential information from lost and incomplete data. Furthermore, we enable the implementation of robust AI models based on data correction and low-dimensional embedding.

[0089] Figure 4 is a diagram illustrating the process of enhancing an operational scenario through abnormal waveform analysis by an abnormal situation analysis unit (138) according to an embodiment. Referring to Figure 4, in the embodiment, when abnormal data is detected among sensor data, the abnormal situation analysis unit (138) determines whether a sensor malfunction has occurred. If a sensor malfunction has not occurred, it analyzes the trend before the abnormal situation occurred. Thereafter, the trend analysis results are reflected in the enhancement of the operational scenario.

[0090] The feedback unit (139) evaluates the learned artificial neural network model and deep learning model. In an embodiment, the feedback unit (139) can evaluate the artificial neural network model through at least one of accuracy, precision, and recall. Accuracy is an index that measures how much the results predicted by the artificial neural network model match the actual results. Precision is an index that measures the ratio of actual positives among the results predicted as positive. Recall is an index that measures the ratio of actual positives predicted by the model as positives. In an embodiment, the feedback unit (139) can calculate the accuracy, precision, and recall of the artificial neural network model, and evaluate the artificial neural network model based on at least one of the calculated indices.

[0091] In an embodiment, the feedback unit (139) can measure the accuracy of the artificial neural network model using an evaluation dataset. The evaluation dataset consists of data that the model did not use for training and is used to objectively evaluate the model's performance. In an embodiment, the feedback unit (139) executes the artificial neural network model using the evaluation dataset and compares the predicted value of the artificial neural network model for each input data with the actual correct answer value of the corresponding data. Thereafter, the accuracy of the model's predictions can be measured based on the comparison results. For example, the accuracy in the feedback unit (139) can be calculated as the ratio of data correctly predicted by the model among the entire data.

[0092] In addition, the feedback unit (139) can calculate the F1 score, which is an index indicating the balance of precision and recall, which is an index calculated as the harmonic mean of precision and recall, evaluate the artificial neural network model based on the calculated F1 score, generate an AUC-ROC curve, which is an index that visualizes the performance of the classification model in a graph, and evaluate the artificial neural network model based on the generated AUC-ROC curve. In an embodiment, the feedback unit (139) can evaluate that the performance of the model is better as the area under the ROC curve (AUC) is closer to 1.

[0093] In addition, the feedback unit (139) can evaluate the interpretability of the artificial neural network model. In an embodiment, the feedback unit (139) evaluates the interpretability of the artificial neural network model through SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-agnostic Explanations) methods. SHAP (SHapley Additive exPlanations) is a library that provides an interpretation of the results predicted by the model, and the feedback unit (139) extracts SHAP values ​​from the library. In an embodiment, the feedback unit (139) can predict how much the characteristic information input to the model influenced the model prediction through the SHAP value extraction.

[0094] The Local Interpretable Model-agnostic Explanations (LIME) method is a method for explaining model predictions for individual samples. In one embodiment, the feedback unit (139) uses the LIME method to approximate a sample as an interpretable model and calculates the importance of each characteristic. Furthermore, the feedback unit (139) can estimate the influence of each characteristic variable by analyzing the model's internal weights and bias values.

[0095] The feedback unit (139) performs improvement work when the fairness of the artificial neural network model is low or shows discrimination. In an embodiment, the feedback unit (139) collects additional data representing the specific group when the data for the specific group is insufficient by a certain level or more, and performs a data preprocessing process. In an embodiment, the feedback unit (139) performs a data preprocessing process including data normalization, outlier removal, and data scaling to prevent the model from learning unnecessary patterns. In addition, in an embodiment, the feedback unit (139) can prevent discrimination or ensure fairness by adding specific conditions to the model learning algorithm.

[0096] In the embodiment, the feedback unit (139) evaluates the performance of the model by comparing the model's predicted results with the actual results through confusion matrix analysis to ensure fairness. The confusion matrix is ​​a matrix that evaluates the classification performance of the model in supervised learning. The confusion matrix displays the classification results by comparing the model's predicted results with the actual results. In the embodiment, the feedback unit (139) can evaluate the performance of the model by calculating the accuracy and misclassification rate for each class through confusion matrix analysis.

[0097] Additionally, in the embodiment, the feedback unit (139) enables the distribution of data to be confirmed through visual analysis of learning data. For example, in the case of image data, image samples for each class can be visualized to evaluate the diversity and fairness of the data.

[0098] Additionally, the feedback unit (139) performs bias verification of the artificial neural network model. In an embodiment, the feedback unit (139) verifies bias of learning information to determine whether the model is biased toward a specific class or attribute. To this end, the feedback unit (139) compares the number of samples for each class or evaluates the classification performance for each class.

[0099] In addition, the feedback unit (139) verifies the fairness and diversity of the learning data and improves the artificial neural network model through fairness verification and evaluation index calculation. In an embodiment, fairness verification is to check whether the artificial neural network model shows discrimination for specific data attributes with respect to the learning information. In an embodiment, the feedback unit (139) can check whether discrimination for specific attributes is present by comparing the number of samples for each attribute or evaluating the classification performance for each attribute.

[0100] Additionally, the feedback unit (139) calculates various indicators to evaluate the performance of the artificial neural network model. For example, model performance can be evaluated by calculating indicators such as accuracy, precision, recall, and F1 score. At this time, indicators for each class can be calculated to evaluate the fairness and diversity of the model.

[0101] In addition, the feedback unit (139) collects feedback on problems that arise when the artificial neural network model is used in an actual environment, and continuously improves the artificial neural network model by reflecting the collected feedback in the artificial neural network model.

[0102] Below, we sequentially describe the optimal operation of a smart energy community using digital twins. Since the operation (function) of the method according to the embodiment is essentially the same as the system's function, any description that overlaps with that in Figures 1 through 4 will be omitted.

[0103] Figure 5 is a diagram showing the optimal operation process of a smart energy community using a digital twin according to an embodiment.

[0104] Referring to Figure 5, in step S100, a digital twin replicating the smart energy community is created using smart energy community monitoring data. In step S200, the created digital twin is used to input data identical to the smart energy community's operational status and a simulation is performed. In step S300, the simulation results are analyzed. In step S400, the optimal smart energy community operation process is extracted based on the analysis results.

[0105] The system and method for providing a digital twin optimal operation platform according to an embodiment secures grid stability for an operation platform targeting an industrial complex, thereby improving power quality reliability. Furthermore, the system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment enable rapid response to the operation and management of the smart energy community through digital twins. Furthermore, the system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment utilize data identical to actual operating situations through modeling using digital twins, enabling simulations of various situations. The system and method for providing a digital twin optimal operation platform for improving the reliability of a smart energy community according to an embodiment more accurately analyzes the impact of distributed energy on the grid, thereby expanding the acceptance of distributed energy through precise modeling of energy resources. Furthermore, the system can contribute to the development of stable and optimal operation and design plans for a smart energy community. Furthermore, the system operates based on a dynamic model of the target customer's resources, reflecting real-time data and controlling individual resources through transient state analysis to respond to transient conditions. Furthermore, the system can contribute to enhancing reliability by improving the quality of power supply within an industrial complex.

[0106] In this embodiment, a digital twin system and operating platform for a smart energy community are built, and various energy resource precision modeling and SILS (Software in the Loop Simulation) verification tools are provided.

[0107] The disclosed content is merely an example, and various modifications and implementations can be made by a person skilled in the art without departing from the gist of the claims claimed in the patent, so the scope of protection of the disclosed content is not limited to the specific embodiments described above.

Claims

1. Memory that stores at least one instruction; and Includes a processor, By executing at least one instruction by the processor, An optimal operation platform providing server characterized by constructing a digital twin for the smart energy community using smart energy community monitoring data and creating an optimal operation process for the smart energy community through the digital twin.

2. In the first paragraph, the optimal operating platform providing server; A server that provides an optimal operating platform, featuring simulations using data identical to the operational status of a smart energy community through digital twins and analysis of the simulation results.

3. In the first paragraph, the optimal operating platform providing server; An optimal operation platform providing server characterized by implementing a deep learning model including a simulation model, a simulation result analysis model, and an optimal operation process generation model by training a deep learning neural network with a training data set.

4. In the first paragraph, the optimal operating platform providing server; An optimal operation platform providing server characterized by classifying monitoring data collected from a smart energy community according to data providing objects and generating model input data by time-synchronizing monitoring data collected from other objects.

5. In the first paragraph, the optimal operating platform providing server; A server providing an optimal operating platform characterized by modeling and simulating monitoring data collected from a smart energy community to build a digital twin model for the smart energy community.

6. In paragraph 5, the operating platform providing server; A server that provides an optimal operating platform that analyzes data collected through digital twins to identify patterns and provides feedback on the analysis results to the actual target, the smart energy community.

7. In paragraph 1, the operating platform providing server; A server providing an optimal operating platform that features a digital twin utilization solution that visualizes collected monitoring data, model input data, and simulation result data in 3D, AR, and VR.

8. In the third paragraph, the optimal operating platform providing server; An optimal operation platform providing server characterized by performing fine-tuning of a pre-trained model through feedback or performing fine-tuning of a pre-trained model through supervised learning on multiple training data sets including training input data and training label data.

9. In the first paragraph, the optimal operating platform providing server; An optimal operation platform providing server characterized by collecting abnormal situation data related to an abnormal situation from among collected monitoring data, analyzing the abnormal situation data, and extracting an optimal process for the abnormal situation.

10. In paragraph 9, the above abnormal situation data; An optimal operation platform providing server characterized by including an object in which an abnormal situation has occurred, a type of abnormal situation, and monitoring data of the object for a certain period of time before the occurrence of the abnormal situation.

11. Optimal operation method of smart energy community using digital twin by electronic device (A) A step of creating a digital twin that simulates the smart energy community using smart energy community monitoring data; (B) A step of performing a simulation by inputting data identical to the operating status of a smart energy community into the digital twin generated above; (C) a step of analyzing the simulation results; and (D) A method for optimal operation of a smart energy community using a digital twin, including a step of extracting an optimal operation process of a smart energy community based on the above analysis results.

12. In the 11th paragraph, the step (A); A method for optimally operating a smart energy community using a digital twin, comprising: a step of modeling and simulating monitoring data collected from a smart energy community to build a digital twin model for the smart energy community;

Images