Integrated Control Apparatus for Installation, Design, and Operation Optimization of AI-Based Energy Storage System

Abstract

The present invention discloses a control device for integrally optimizing the installation, design, and operation stages of an energy storage system (ESS) by utilizing artificial intelligence (AI). The present invention has the technical objective of achieving maximum economic effects with minimum investment by simultaneously realizing reductions in initial investment costs during the installation phase and reductions in operating costs during the operation phase. To this end, the invention comprises a data collection unit that collects various operational data including contracted power, tariff plans, load power, renewable energy generation, battery status, and external environmental information; a prediction unit that predicts changes in load and generation based on said data; a planning unit that generates multiple operation scenarios and evaluates economic feasibility by reflecting the prediction results, tariff systems, and operating conditions; a search processing unit that searches for and selects a scenario where overall efficiency is maximized by considering constraints; a control unit that outputs charge / discharge control signals according to the selected scenario; and a replanning unit that detects fluctuations or prediction errors occurring during operation and updates the operation plan in real time. By combining an artificial intelligence (AI) prediction model and a multi-objective optimization algorithm to integrate and control the entire process of installation, design, and operation as a single closed-loop, the invention possesses the technical effect of simultaneously achieving electricity bill reduction, investment cost minimization, battery life extension, and long-term profitability improvement.

Description

Technology Field

[0001] The present invention belongs to the field of artificial intelligence (AI) and energy management system (EMS) technology, and in particular relates to the installation, design, and operation optimization of energy storage systems (ESS). More specifically, it relates to a technology for an AI-based integrated control device designed to simultaneously achieve reductions in initial investment costs during the installation phase and operating costs during the operation phase by analyzing power consumption patterns, tariff plans, and renewable energy generation amounts of manufacturing facilities or smart factories. Background Technology

[0002] Recently, energy storage systems (ESS) have been attracting attention as a key infrastructure to complement the intermittency of renewable energy and to realize load peak reduction and electricity bill savings. However, existing ESS control technologies are mainly focused on hardware-based charge and discharge control and single battery state management, and have not yet reached the level of integrally optimizing the installation, design, and operation stages of the entire system.

[0003] For example, prior art 1 (US 2023 / 0396067 A1, Honeywell) presents a structure for managing energy flow in each container unit with a central controller in a distributed BESS and microgrid environment, but this is limited to physical control and communication management and does not include a function to optimize operational efficiency from an economic perspective by analyzing load patterns, tariffs, and renewable energy generation. In other words, it has limitations in that it cannot perform data-based predictive control or cost-saving scheduling.

[0004] Prior Art 2 (US 2024 / 0266856 A1, SparkCognition) is a battery cell-level performance prediction and control technology using artificial intelligence. While it can improve charge / discharge efficiency and lifespan degradation, this is limited to management at the individual battery level and is difficult to apply to system optimization that comprehensively considers multiple ESS facilities and external operating factors. Furthermore, since it does not perform economic analysis such as capacity estimation or return on investment (ROI) during the design phase, it lacks decision support functions before installation.

[0005] Prior art 3 (Registered Patent No. 10-2663507) is a KPI-based artificial intelligence optimization system that implements data-based indicator improvement and cloud distributed learning technology; however, it is limited to general management indicators or process efficiency improvement and fails to reflect the real-time operation and constraints (e.g., SOC, SOH, tariff schemes, power quality, etc.) of physical energy systems, such as ESS design and operation control. Therefore, there are structural limitations in directly applying KPI-based optimization techniques to the operation planning, charge / discharge control, and installation cost analysis of energy systems.

[0006] Ultimately, the current level of technology is classified into (1) hardware-centric BESS control technology, (2) battery-unit predictive control technology, and (3) non-physical KPI optimization systems. Due to their respective limitations, these technologies have not yet evolved into an integrated optimization system capable of simultaneously achieving reductions in capital expenditure (CapEx) during the installation phase and operating expenditure (OpEx) during the operation phase.

[0007] Therefore, there is a need for technology capable of comprehensively analyzing various data, such as load power, tariff plans, renewable energy generation, weather information, and industrial operation schedules, and using artificial intelligence (AI) to optimize ESS installation capacity, charging and discharging strategies, and economic indicators (ROI, payback period, etc.). Furthermore, there is an urgent need for AI-based integrated control technology that can continuously guarantee system stability and profitability by re-executing operation plans in real-time in response to prediction errors or changes in the external environment. Prior art literature

[0008] US 2024 / 0266856 A1 (Published: August 8, 2024) US 2023 / 0396067 A1 (Published: December 7, 2023) KR 10-2663507 (Registered: April 30, 2024) The problem to be solved

[0009] The present invention aims to solve the aforementioned problems and relates to an integrated control device for simultaneously optimizing the capital expenditure (CapEx) and operating expenditure (Opex) of an artificial intelligence (AI)-based energy storage system. Its purpose is to analyze and predict power consumption patterns based on data in various manufacturing facility environments, such as industrial plants, smart factories, and commercial buildings, thereby maximizing the economic efficiency and operational efficiency of the facility.

[0010] In addition, the present invention aims to maximize the efficiency of ESS design by integrating, collecting, and analyzing heterogeneous data such as power load patterns, tariff plans, and renewable energy generation amounts of manufacturing facilities in real time, and by automatically calculating customer-tailored ESS capacity and charging / discharging strategies.

[0011] In addition, the present invention aims to minimize capital expenditure (CapEx) during the installation phase through an artificial intelligence (AI)-based constraint optimization algorithm. That is, by comprehensively considering variables such as the unit price of the ESS, battery type, capacity, and installation environment, it aims to minimize the customer's initial investment burden by deriving the most economical design plan while achieving the same power saving effect.

[0012] In addition, the present invention aims to support customers in making quick and rational investment decisions by setting the investment payback period as a key decision-making indicator and automatically predicting and visualizing profitability and payback periods for various scenarios during the pre-installation stage.

[0013] In addition, the present invention aims to minimize operating costs (Opex) by analyzing real-time power usage data during the operation phase and automatically optimizing the charging and discharging schedule by integrating and considering changes in electricity rates, peak demand, renewable energy generation, weather information, market prices, etc.

[0014] Furthermore, the present invention aims to maximize the client's total revenue and achieve a sustainable operational structure by performing Multi-Objective Optimization that simultaneously considers multiple KPIs, such as electricity bill reduction, power trading revenue, and carbon reduction. Through this, it evolves from a simple energy saving system into a revenue-generating energy management platform.

[0015] In addition, the present invention aims to support the latest optimal control at all times without separate intervention by the operator by enabling an artificial intelligence (AI) control algorithm to autonomously update ESS operation strategies based on predictions by learning market price fluctuations, peak rates, weather forecasts, etc., in advance.

[0016] Furthermore, the present invention aims to reduce not only short-term costs but also long-term maintenance costs by having an artificial intelligence (AI) control algorithm analyze the charge and discharge cycles of a battery and perform battery life protection control. This extends the lifespan of the equipment and minimizes the Total Cost of Ownership (TCO).

[0017] Furthermore, the present invention aims to implement a flexible system applicable to various operating scenarios, such as industrial, commercial, and residential use, by providing an integrated structure that supports both self-consumption (Behind the Meter) and power trading (Front of the Meter) operations.

[0018] In addition, the present invention aims to secure faster proposal capabilities than competitors during the bidding and proposal stages and improve sales and order success rates by providing a function that can automatically generate fast and quantitative ESS estimates. Through this, it is possible to secure data-based design and proposal competitiveness and create differentiated added value in the EPC business.

[0019] In addition, the present invention aims to contribute to customers' continuous reduction of electricity costs, ESG response, carbon reduction, and improvement of energy independence by implementing an autonomous ESS management structure that continuously learns operating data and gradually improves system efficiency over time through a KPI-based automatic update function over the long term. means of solving the problem

[0020] To achieve the above objective, an integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system according to one aspect of the present invention comprises: a data collection unit that collects a plurality of operation-related data including contract power information, tariff plan information, load power data, renewable energy generation data, battery status data, and external environment data provided by a power supplier or consumer;

[0021] A prediction unit that predicts future changes in load and power generation using the collected data; and a planning unit that automatically generates multiple operating scenarios and evaluates costs and revenues for each scenario by reflecting the prediction results, electricity rate system, battery status, and operating conditions;

[0022] A search processing unit that searches for and selects an optimal operating scenario from among a plurality of scenarios generated by the planning unit to maximize overall operating efficiency while satisfying set constraints; and a control unit that outputs a control signal to control the charging and discharging operation of an energy storage device based on the optimal operating scenario selected by the search processing unit;

[0023] It is characterized by including a replanning unit that detects load fluctuations, renewable energy generation fluctuations, or prediction errors occurring during operation, updates the operation plans of the planning unit and the search processing unit, and ensures that these plans are reflected in real time through the control unit.

[0024] In addition, according to the present invention, the planning unit is characterized by calculating an operation scenario to maximize the economic efficiency of the entire system by comprehensively considering a plurality of evaluation indicators including operating costs, energy charges, factors that degrade the performance of the device, and revenue-related factors.

[0025] In addition, according to the present invention, the search processing unit analyzes the interconnected energy flow targeting energy storage devices, renewable energy generation facilities, and load facilities distributed across a plurality of points or resources, and performs integrated optimization to maximize the operating efficiency of the entire system by comprehensively considering power demand, surplus power generation, transmission constraints, and economic feasibility between each point.

[0026] In addition, according to the present invention, the replanning unit updates the operation plan in real time according to a predefined threshold condition when a difference between the prediction result and actual operation data, a sudden fluctuation in renewable energy output, an abnormal pattern of load, or a change in the external environment is detected, and immediately updates the control command by repeatedly performing the optimization procedure of the planning unit and the search processing unit in a rolling horizon or adaptive manner.

[0027] In addition, according to the present invention, the control unit controls the timing and amount of charging and discharging in conjunction with an energy management device including a power conversion device (PCS), and is characterized by automatically generating and outputting a control signal corresponding to the detection of the system's operating status, contracted power limit, power system constraints, or emergency situation, thereby stably maintaining the overall energy flow.

[0028] In addition, according to the present invention, the data collection unit is configured to be connected to at least one of a water electrolysis facility, a combined heat and power (CHP) facility, and a pumped storage hydropower (PSH) facility, and includes an energy flow data collection function for collecting electric, heat, and hydropower energy flow data from said facilities.

[0029] The above control unit is characterized by being configured to adjust the power input of the water electrolysis facility, the power and heat production of the combined heat and power generation facility, and the pumping and power generation operation mode of the pumped storage hydroelectric facility by taking into account the charging and discharging status of the energy storage system (ESS) together with the energy flow data.

[0030] In addition, according to the present invention, the device further comprises an installation and design optimization unit that receives operation-related input data, including load characteristics, storage capacity, power converter capacity, and renewable energy facility capacity, provided by a user or system during the installation or design phase, automatically calculates economic indicators including initial investment costs, estimated operating costs, return on investment, or payback period based on said input data, and automatically presents an installation plan or design configuration that maximizes cost efficiency and operational efficiency by searching for and evaluating a plurality of candidate installation plans or design plans within predefined constraints.

[0031] In addition, according to the present invention, the device further comprises a visualization output unit that analyzes calculated economic indicators, design results, or driving performance data, converts them into a visual form for presentation, and provides them in the form of a graph, dashboard, or report through an interface of a user terminal or management system so that they can be utilized for decision-making.

[0032] An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system according to another aspect of the present invention comprises: a data collection unit that collects operation-related input data, including load characteristics, storage capacity, power conversion system (PCS) capacity, renewable energy facility capacity, tariff plan information, and operation constraints, from a user or an external system;

[0033] An economic evaluation unit that automatically calculates economic indicators including initial investment costs, estimated operating costs, return on investment, payback period, and energy saving effects based on the above-mentioned collected input data; and a scenario generation unit that automatically generates multiple installation or design scenarios by considering the results of the economic evaluation together with input constraints, and calculates priorities by evaluating cost efficiency and operational efficiency for each scenario;

[0034] An installation and design optimization unit that analyzes the evaluation results of the above-mentioned scenario generation unit to select an installation or design plan that maximizes cost-effectiveness, and automatically calculates system configuration specifications, equipment capacity, and estimated investment and operating costs based on the selected installation or design plan;

[0035] It is characterized by including a result providing unit that provides the optimal result derived above in a visualized form to a user terminal or management system interface, and automatically recalculates and displays the latest result in real-time when input conditions are changed or design variables are updated.

[0036] In addition, according to the present invention, the economic evaluation unit is characterized by using an artificial intelligence prediction model to predict changes in electricity pricing schemes, load growth rates, battery performance degradation, and fluctuations in renewable energy generation as long-term time series, and dynamically calculating future investment costs and payback periods based on the prediction results.

[0037] In addition, according to the present invention, the result providing unit is characterized by visualizing the calculated optimal installation or design results in the form of a three-dimensional model or a digital twin, and being configured so that economic indicators and design results are automatically updated in real time when a user changes each design variable. Effects of the invention

[0038] Through the above-mentioned means of solution, the present invention provides the following effects.

[0039] First, according to the present invention, the user does not have to directly perform complex ESS operation control or scheduling, and artificial intelligence automatically establishes an optimal charging and discharging plan, thereby improving the convenience of management.

[0040] Second, users can make quick and accurate decisions based on the optimal capacity and investment cost information automatically calculated during the installation phase, thereby saving time and effort in the design and quotation processes.

[0041] Third, through real-time forecasting and replanning functions, stable power operation is ensured even amidst unexpected load fluctuations or weather changes, reducing user operational anxiety and management burden.

[0042] Fourth, since the system automatically reduces electricity costs, users can achieve economical operation effects without any specialized knowledge.

[0043] Fifth, the integrated visualization interface allows for intuitive checking of driving status and economic indicators, enhancing user understanding and accessibility.

[0044] Sixth, since consistent management from installation to operation is possible through AI-based integrated control, users can simplify the complex procedures required for maintenance and operational efficiency management.

[0045] 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 configuration of the invention described in the detailed description or claims of the present invention. Brief explanation of the drawing

[0046] FIG. 1 is a drawing showing the entire system according to the present invention. FIG. 2 illustrates a block diagram of an integrated control device for optimizing the installation, design, and operation of an AI-based energy storage system according to an embodiment of the present invention. FIG. 3 illustrates a block diagram of an integrated control device for optimizing the installation, design, and operation of an AI-based energy storage system according to another embodiment of the present invention. FIG. 4 is a conceptual diagram illustrating the peak shaving effect of an energy storage system according to the present invention. FIG. 5 is a conceptual diagram illustrating the peak shaving and load leveling effects through charge and discharge control of an artificial intelligence (AI)-based energy storage system (ESS) according to the present invention. FIG. 6 is a flowchart illustrating the AI-based automatic estimation calculation procedure of the AI-based energy storage system (ESS) installation and design optimization device of the present invention. Specific details for implementing the invention

[0047] Hereinafter, embodiments of the present invention are described in detail with reference to the attached drawings so that those skilled in the art can easily implement the present invention. The present invention may be embodied in various different forms and is not limited to the embodiments described herein. In the drawings, parts unrelated to the explanation have been omitted to clearly explain the present invention, and the same reference numerals have been used for identical or similar components throughout the specification.

[0048] The words and terms used in this specification and claims are not limited to their ordinary or dictionary meanings, but should be interpreted in a meaning and concept consistent with the technical spirit of the invention in accordance with the principles by which the inventor defines terms and concepts to best describe his invention.

[0049] Therefore, the embodiments described in this specification and the configurations illustrated in the drawings correspond to preferred embodiments of the present invention and do not represent all technical concepts of the present invention; thus, various equivalents and modifications that may replace such configurations may exist at the time of filing the present invention.

[0050] In this specification, terms such as "comprising" or "having" are intended to describe the existence of the features, numbers, steps, actions, components, parts, or combinations thereof described in the specification, and should be understood as not precluding the existence or addition of one or more other features, numbers, steps, actions, components, parts, or combinations thereof. Furthermore, the fact that a component is "connected" to another component includes, unless there are special circumstances, cases where they are directly connected as well as indirectly connected.

[0051] In this specification, singular expressions include plural expressions unless the context clearly specifies them as singular. Additionally, plural expressions include singular expressions unless the context clearly specifies them as plural. Throughout the specification, when a part is described as including a certain component, this means that, unless specifically stated otherwise, it does not exclude other components but may include additional components.

[0052] Hereinafter, preferred embodiments of the present invention will be described in detail in conjunction with the attached drawings.

[0053] FIG. 1 is a diagram showing the entire system according to the present invention. The system is composed of a power company (11), a meteorological agency (12), a power exchange (13), a user terminal (20), a network (30), and an artificial intelligence-based integrated control device (200), and each component enables data exchange and the execution of control commands through mutual communication.

[0054] The information provided by the power company (11) includes past hourly power usage, contracted power, and rate plan, which are details of the power contract subscribed to by the consumer. This information is used to learn future power usage and electricity rate prediction models for companies and households that install an energy storage system (ESS). Past power usage data can be obtained by the installer of the energy storage system (ESS) requesting it from the power company in an online or offline manner. For example, in Korea, it is provided in 15-minute intervals, and in Japan, in 30-minute intervals. Since this data corresponds to the user's power usage history, it is mandatory to provide it upon request in accordance with relevant laws. The installer of the energy storage system (ESS) uploads the acquired data to a web-based service on a one-time basis, and subsequently, in the real-time control stage, real-time power usage information is automatically collected through the power meter installed at the installation location.

[0055] In addition, the electricity rate information provided by the power company (11) includes detailed items such as time-of-use rates, seasonal rate changes, electricity market trading prices, and peak rate ranges. This information is used as key input data when automatically calculating the charging and discharging schedule during the operation phase. For example, the system switches to charging mode in the low rate range and performs discharging in the high rate range to achieve the maximum electricity cost reduction effect. By reflecting rate information in real time in this way, economic efficiency is significantly improved compared to the existing manual operation method.

[0056] Time-based load power data is data that records the power usage patterns of each consumer's facility in a time series, and is used as a primary input for learning and demand forecasting models. The integrated control device (200) of the present invention performs short-term load forecasting (1 minute to 24 hours) and medium-to-long-term load trend analysis based on this data, and thereby learns an AI model that determines the optimal energy storage system (ESS) capacity and charging / discharging timing. This AI model performs the function of comprehensively analyzing power consumption, renewable energy generation, and the battery's state of charge (SoC) input in real time during the operation phase, and immediately generating control commands regarding the charging or discharging of the energy storage system (ESS) according to the results.

[0057] The Korea Meteorological Administration (12) or private weather information providers can provide weather forecast data as well as past and real-time weather data. The system calculates solar power generation in different ways depending on the type of weather data provided. For example, if only past and real-time weather information is provided, a solar radiation prediction model is trained using this data, and the amount of power generated by the solar panel is calculated through the trained solar radiation prediction model. If a weather service that provides a solar radiation prediction model with higher accuracy is used, the amount of power generated can be calculated by directly utilizing the solar radiation prediction value. In addition, if a service that provides a high-precision power generation prediction model that directly predicts the amount of power generated is used, the prediction value can be used as input for the operation plan of the energy storage system (ESS).

[0058] The Meteorological Agency (12) provides external environmental data necessary for the efficient operation and design optimization of energy storage systems. The Meteorological Agency (12) periodically transmits weather data including solar radiation, ambient temperature, atmospheric pressure, wind speed, precipitation, humidity, and short-term, medium-term, and long-term forecast information. This data is essential for predicting the variability of renewable energy generation, such as solar and wind power, and in particular, solar radiation and cloud cover information serve as key factors in determining the accuracy of solar power generation output prediction. For example, based on real-time solar radiation data and forecast information provided by the Meteorological Agency (12), an artificial intelligence prediction model can predict the potential solar power generation for the next day or the next week and automatically adjust charging and discharging strategies accordingly. This maintains a balance between energy production and storage, and minimizes losses due to overgeneration or power shortages.

[0059] Data on ambient temperature, wind speed, atmospheric pressure, and precipitation are directly utilized not only for renewable energy generation but also for load forecasting and facility efficiency analysis. Changes in ambient temperature have a significant impact on fluctuations in cooling and heating loads, and wind speed is a determining factor for wind power generation output. The integrated control unit (200) utilizes these environmental data as input variables for the AI-based prediction unit to dynamically analyze fluctuations in energy demand and power generation according to weather conditions. In addition, forecast information from the Korea Meteorological Administration is used for establishing long-term operation scenarios and economic evaluations, enabling the optimization of energy storage system (ESS) operation strategies in advance under specific seasons or weather conditions. For example, if prolonged cloudy conditions are expected, the charging strategy is strengthened, or the discharge schedule is automatically reset to reflect the increase in cooling and heating loads when a heatwave or cold wave is forecast.

[0060] The data from the Korea Meteorological Administration (12) is not merely reference information, but acts as a key input element that directly contributes to improving the prediction accuracy and economic efficiency of AI-based energy management. By collecting and analyzing weather data in real time, the integrated control device (200) of the present invention can move away from static control methods centered on past data and implement a proactive energy operation system that responds to future weather changes. Through this, the energy storage system (ESS) can autonomously update its operation plan according to environmental changes, maximize the utilization rate of renewable energy, and simultaneously reduce energy costs and system inefficiencies.

[0061] The Power Exchange (13) is an external organization responsible for operating the power market and provides market-based data such as real-time power market (SMP, real-time price), settlement unit price, system congestion information, demand response (DR) signals and power supply and demand plans to an integrated control system.

[0062] Market price information and grid status information collected from the power exchange (13) are used as essential input values ​​to establish operation strategies for energy storage systems (ESS), hydroelectric power plants (HEF), combined heat and power (CHP) plants, and pumped storage hydroelectric power (PSH) plants in the integrated control unit (200). In particular, the predicted SMP, time-based rate changes, grid congestion, or reserve capacity indicators provided by the power exchange (13) affect not only the charging and discharging decisions of the energy storage system (ESS) but also the switching of pumped storage and power generation modes of pumped storage hydroelectric power (PSH) plants, power input scheduling of hydroelectric power plants (HEF), and the adjustment of the power and heat production ratio of combined heat and power (CHP) plants.

[0063] That is, the power exchange (13) functions as a central energy market information provider that enables economic assessment based on grid conditions and market-linked energy operation, thereby supporting the integrated control system to maximize the operational efficiency and cost efficiency of all energy resources.

[0064] The network (30) is a communication infrastructure for data transmission between a power company (11), a meteorological agency (12), a power exchange (13), a user terminal (20), and an integrated control device (200), and can be configured based on wired / wireless internet, a cloud server, or a private network. The network (30) of the present invention includes a high-speed communication protocol and a security encryption module to ensure real-time data, and stably transmits large-capacity operational information without data bottlenecks through a multi-channel structure.

[0065] The user terminal (20) can be implemented in the form of a web-based or dedicated application, and visualizes and displays data received from the power company (11) and the meteorological agency (12), and transmits user inputs such as operating conditions, constraint settings, and investment plans to the integrated control device (200). The integrated control device (200) integrates and analyzes the received operational data and external environment data to automatically perform the calculation of the installation capacity of the energy storage device (ESS), the design of the facility configuration, the planning of charging and discharging operations, and the economic evaluation.

[0066] In addition, the integrated control device (200) utilizes an artificial intelligence prediction model and a multi-objective optimization algorithm to derive a control strategy that achieves reduction in capital expenditure (CapEx) during the installation phase, reduction in operating expenditure (OpEx) during the operation phase, reduction in electricity costs, extension of battery life, and maximization of long-term profitability. Through such an integrated control structure, the present invention reflects data from external organizations, such as power companies and meteorological agencies, in real time, and updates plans according to changes in the user's operating environment, thereby continuously improving the installation and operation efficiency of the energy storage system (ESS).

[0067] Self-consumption energy storage systems (ESS) are primarily designed to reduce electricity costs, and artificial intelligence prediction models operate to calculate optimal charging and discharging times based on real-time changes in power usage collected from businesses or households where the ESS is installed. Accordingly, the ESS is controlled to operate in a manner that achieves the greatest cost savings by matching the user's load patterns.

[0068] On the other hand, the revenue-generating energy storage system (ESS) receives real-time power price fluctuations provided by the power exchange (13) as input, and an artificial intelligence prediction model establishes an optimal trading strategy, and the charging and discharging of the energy storage system (ESS) is controlled to be performed automatically according to that strategy. In the domestic and Japanese power markets, participation in the power exchange is limited to a few licensed companies, but in some parts of the United States, an open power trading market is operated in which all operators can participate.

[0069] FIG. 2 illustrates a block diagram of an integrated control device (200) for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system according to an embodiment of the present invention. The integrated control device (200) is configured to include a data collection unit (210), a prediction unit (220), a planning unit (230), a search processing unit (240), a replanning unit (260), an installation and design optimization unit (270), a visualization output unit (280), and a control unit (250) that comprehensively controls the operation and signal flow of all of these.

[0070] The data collection unit (210) collects and manages all basic data for the installation, design, and operation optimization of the energy storage system (ESS). The data collection unit (210) is connected to power companies, meteorological agencies, and user terminals via a communication network and collects various types of data such as contracted power, rate plans, load power data by time of day, renewable energy generation amount, battery status (e.g., SOC, SOH, RTE, etc.), and weather conditions (e.g., solar radiation, temperature, wind speed, etc.). Since each data has a different collection cycle and format, the data collection unit (210) includes a communication protocol conversion module and a data matching module to convert all data into a standardized data format (e.g., JSON, CSV, MQTT message, etc.) and then records it in an internal storage. This allows for the integrated processing of data transmitted from different devices and agencies, and enables the maintenance of consistency in input data during future prediction and optimization calculations.

[0071] The data collection unit (210) adopts a real-time data streaming structure to collect information flowing in simultaneously from multiple data channels in parallel, and includes buffering and synchronization technologies to minimize data loss or delay. This structure ensures stable data processing performance even in large-scale ESS installation environments or multi-site operation structures. The collected data undergoes procedures such as filtering, outlier removal, missing value compensation, and unit normalization during the preprocessing stage. This preprocessing process is an essential step for increasing data reliability and prediction accuracy, and prevents judgment errors caused by noise or sensor errors.

[0072] The data collection unit (210) is linked with internal facilities such as a power management system (PMS), an energy management system (EMS), a power conversion device (PCS), and a battery management system (BMS), and can collect real-time data such as current, voltage, and temperature measured by field sensors. This information is used not only for operation control but also for analyzing the battery's performance degradation status or factors causing efficiency reduction. In addition, API-based data provided by external organizations (e.g., the Korea Meteorological Administration, the Korea Power Exchange, demand response programs, etc.) is automatically collected and periodically updated by a scheduler. Through this, the data collection unit (210) of the present invention integrates and normalizes heterogeneous data from multiple sources to secure a high-quality operational data set in real time, thereby fundamentally improving the accuracy and reliability of AI-based prediction and optimization calculations.

[0073] The prediction unit (220) is an artificial intelligence-based analysis module for quantitatively predicting future changes in load power and renewable energy generation based on various operational data transmitted from the data collection unit (210). The prediction unit (220) plays a key role in establishing the operation plan for the energy storage system and predicts power consumption and generation in minutes, hours, or days through time-series-based data pattern learning. Input data includes load power data, renewable energy generation, weather data (solar radiation, temperature, wind speed, humidity, etc.), equipment operation schedules, weekday / weekend patterns, and holiday information. The prediction unit (220) is designed to learn non-linear patterns and inter-correlation by considering these complex factors and to estimate the variability of power demand and generation with minimal error.

[0074] The prediction unit (220) utilizes machine learning and statistical models to achieve optimal prediction accuracy by combining various artificial intelligence algorithms. Additionally, the prediction unit corrects errors caused by a single factor by simultaneously learning variables such as power consumption, solar radiation, temperature, and industrial facility operating rate through a multi-input structure. This composite prediction model improves prediction accuracy by precisely reflecting the effects of power demand patterns by time period, power generation efficiency, and weather changes, and the prediction results are used as key input values ​​for the charging and discharging scheduling and economic evaluation of the planning unit (230).

[0075] The prediction unit (220) is designed to enable continuous learning and correction based on a data stream collected in real time. When a deviation occurs between the prediction result and the actual operation data, the prediction unit (220) automatically updates model parameters through a relearning module to respond to environmental changes or equipment aging over time. Through this, the prediction unit (220) secures adaptability to various external factors such as seasonal changes, industrial load fluctuations, and weather forecasting errors, and implements a self-learning prediction system based on real-time feedback rather than a simple prediction based on past data. As a result, the prediction unit (220) serves as a foundation for achieving efficient operation and economic optimization of the ESS, and significantly improves the intelligent control precision of the entire system.

[0076] The planning unit (230) is a core module that automatically generates multiple operation scenarios by comprehensively considering the power rate system, battery status (e.g., SOC, SOH, RTE, etc.), facility capacity, and operation constraints, based on the load power and renewable power generation prediction results transmitted from the prediction unit (220). The planning unit (230) determines the hourly operation schedule between the charging, discharging, and standby modes of the ESS, and calculates various evaluation indicators for each scenario, such as the expected power rate, savings, energy efficiency, battery degradation costs, and return on investment (ROI). In addition, the planning unit (230) can establish a realistic operation plan by simultaneously reflecting various rate systems, such as the time-of-use tariff, contracted power limit, peak rate range, and demand response (DR) participation conditions. For example, by automatically generating an optimized schedule in which charging operation is performed during the low-rate hours at night and discharging is performed during the peak rate range, economic operation that minimizes total power costs is possible.

[0077] The planning unit (230) is also configured to perform multi-objective optimization beyond simple scheduling. In this case, the objective function may consist of multiple items such as reducing total power consumption, reducing peak demand, minimizing operating costs, maximizing energy efficiency, extending battery life, reducing greenhouse gases, and increasing the utilization rate of renewable energy. The planning unit (230) defines an objective function in the form of a polynomial with weights assigned to each item and generates various operating scenarios that satisfy it. The generated scenarios are transmitted to the search processing unit (240), and the optimal scenario is selected after evaluating the economic feasibility and operating efficiency of each candidate. Through this structure, the planning unit (230) can establish a balanced operating plan that simultaneously satisfies economic feasibility and technical stability, rather than single-objective optimization that considers only a specific purpose.

[0078] The planning unit (230) ensures the realism of scenario generation by explicitly considering operating constraints. Constraints include the upper and lower limits of the battery's SOC, charging / discharging power limits, PCS capacity, inverter efficiency, and power quality maintenance conditions (allowable voltage and frequency ranges). The planning unit (230) models these constraints in the form of mathematical constraint equations and combines them with an objective function to convert them into a Constrained Optimization Problem. Consequently, the planning unit (230) functions not as a simple simulation generator, but as an AI-based automatic planning generation engine applicable in actual operating environments, and can derive an optimized operating plan in real time in response to various environmental variables and changes in the rate system without user intervention. Through this, the present invention moves away from the existing empirical scheduling-based operating method and realizes a data-based intelligent energy operation planning system.

[0079] The search processing unit (240) is a core computational module for deriving an optimal scenario that maximizes the operating efficiency and economic viability of the entire system while satisfying given constraints among a plurality of operating scenarios generated by the planning unit (230). The search processing unit (240) is configured to apply various optimization algorithms, including Mixed-Integer Linear Programming (MILP), Genetic Algorithm (GA), Reinforcement Learning, or metaheuristic search techniques. Each operating scenario is represented by multidimensional data such as hourly charge / discharge patterns, power unit price, load prediction results, and battery state information (e.g., SOC, SOH), and the search processing unit (240) inputs this into an objective function to search for an operating solution that simultaneously achieves cost minimization and efficiency maximization.

[0080] The search processing unit (240) does not stop at simple mathematical optimization but includes an adaptive optimization structure that applies an artificial intelligence-based learning search technique to automatically improve the control policy according to environmental changes. In the case of a reinforcement learning-based search technique, the operating result (reward value) is used as a feedback signal to learn a policy in which economic efficiency and stability are simultaneously improved over time. This structure has a higher responsiveness to prediction errors or disturbances (weather changes, sudden load surges, etc.) compared to existing fixed algorithms. In addition, by utilizing the evolutionary search characteristics of genetic algorithms, it is possible to derive a set of optimal solutions that simultaneously consider multiple goals, such as cost, efficiency, lifespan, and greenhouse gas reduction effects, in multi-objective optimization problems. Among the multiple solutions calculated in this way, the search processing unit (240) can select the scenario with the highest weight according to the system policy or dynamically select it according to real-time operating conditions.

[0081] The search processing unit (240) simultaneously considers constraints such as preventing contracted power exceeding, upper and lower SOC limits, charging and discharging efficiency, PV generation prediction error, and inverter capacity limit during the search process. Each constraint is internalized into the objective function in the form of an inequality equation or a penalty function, and the algorithm adopts only solutions satisfying these conditions as valid scenarios. Through this, a stable solution reflecting a realistic operating environment can be derived. The finally selected optimal operating scenario is transmitted to the control unit (250) and converted into an actual charging and discharging control signal for the energy storage system (ESS), and the search results are continuously updated in conjunction with the replanning unit (260) even during real-time operation. In this way, the search processing unit (240) performs high-dimensional optimization search based on candidate solutions generated by the planning unit (230), and functions as the core artificial intelligence engine of the present invention to perform intelligent operation decisions that consider economic feasibility, efficiency, stability, and adaptability.

[0082] The replanning unit (260) is a core control module that maintains the stability and efficiency of the system by detecting fluctuations in external variables or prediction errors occurring in real time during the operation of the AI-based energy storage system and immediately updating the operation plan accordingly. The replanning unit (260) operates in a rolling horizon manner that periodically re-executes the calculation procedures of the planning unit (230) and the search processing unit (240), and ensures that the operation strategy does not remain in a static state by continuously reflecting real-time data. When the replanning unit (260) detects abnormal situations such as a sudden surge in load power, a rapid decrease in renewable power generation, a discrepancy with the weather forecast, or reaching the SOC limit of the battery, it immediately performs a replanning process. In this process, it may operate by updating only a part of the existing plan or by recalculating the entire operation plan, and the updated plan is transmitted to the control unit (250) in real time to enable immediate response control.

[0083] The replanning method based on moving windows moves through fixed time intervals (e.g., 15 minutes, 1 hour, 24 hours) and repeatedly performs optimization by reflecting new data at each point in time. Unlike the existing static scheduling method that relies only on past prediction results, this structure can actively adapt to environmental changes by continuously updating the plan over time. For example, if solar power generation is lower than predicted, the replanning unit (260) receives real-time power generation data, reduces the charging schedule, and adjusts the discharging plan to compensate for the remaining load. Conversely, if power generation is higher than predicted, the plan is changed to store surplus power, thereby increasing the self-consumption rate without power generation loss. At this time, the replanning unit (260) simultaneously considers constraints (contracted power limit, SOC upper and lower limits, charge / discharge efficiency, etc.) to ensure that the new plan does not violate system constraints.

[0084] The replanning unit (260) includes an artificial intelligence learning-based error correction algorithm to gradually improve the model accuracy of the prediction unit (220). When the error between the prediction result and the actual measurement value continuously accumulates in a specific direction, the replanning unit (260) corrects the parameters of the prediction model or activates an auxiliary model to minimize the error. Through this, the system maintains stable operating performance even in environments that change moment by moment due to external factors such as weather changes, load uncertainty, and market price fluctuations. Furthermore, the replanning unit (260) is closely linked with the control unit (250), and when the plan update is completed, it immediately reflects the new control signal to update the charging and discharging status of the ESS, thereby continuously ensuring the energy efficiency of the entire system and the effect of reducing electricity costs. As a result, the replanning unit (260) serves as a real-time adaptive control engine that connects prediction, planning, search, and control in the integrated control device of the present invention, enabling the energy storage system to actively respond to changes in the external environment.

[0085] The installation and design optimization unit (270) operates as an artificial intelligence (AI)-based design decision engine to automatically calculate the optimal system configuration during the new installation, expansion, or remodeling phases of an energy storage system (ESS) within the integrated control device (200) of the present invention. This module generates customized design scenarios according to the user's operating purpose (e.g., electricity bill reduction type, peak reduction type, power trading profit type, etc.), and the input values ​​include load power profiles, target self-consumption rate, battery capacity (kWh), power conversion device (PCS) capacity (kW), renewable energy facility capacity, physical constraints of the installation site, expected operating patterns, regional solar irradiance, and power unit price fluctuation data. The input data is standardized through a predefined parameter normalization process and linked with an internal database (DB) to form a basic data set for design calculations. Through this, the system can perform economic evaluation and design optimization based on consistent standards even in various installation environments.

[0086] The installation and design optimization unit (270) goes beyond simple facility capacity calculation and automatically calculates economic indicators such as initial investment costs (CapEx), expected operating costs (OpEx), return on investment (ROI), internal rate of return (IRR), net present value (NPV), and payback period (PBP). To this end, the installation and design optimization unit (270) includes a multi-layered database containing information on electricity tariffs, electricity trading prices, facility unit costs, efficiency of batteries and PCS, operating loss rates, maintenance costs, system depreciation rates, and government subsidies or REC (Renewable Energy Certificate) incentives. This database is updated in real-time according to changes in region, time, and tariff policy, and quantitatively analyzes the financial feasibility for each design scenario. For example, if different ESS capacities and PCS capacities are entered for two business sites with the same load pattern, this module automatically calculates the investment amount, electricity bill savings, and payback period for each case and presents an optimal design plan.

[0087] The AI-based search algorithm is designed to select a design among these multiple design scenarios that simultaneously maximizes cost efficiency and operational efficiency. In this invention, multi-objective optimization can be performed using genetic algorithms, Bayesian optimization, reinforcement learning techniques, and equivalent algorithms or mathematical methods. The algorithm generates a Pareto Front that considers the trade-off between investment costs and efficiency, and derives the optimal solution that satisfies the best economic-performance balance among them. For example, excessively increasing battery capacity may increase peak reduction effects but cause investment costs to skyrocket, potentially worsening ROI; conversely, insufficient capacity may result in negligible reduction effects. This module minimizes the user's investment risk by automatically adjusting these conflicting factors to present the optimal Energy Storage System (ESS) capacity and PCS specifications.

[0088] The installation and design optimization unit (270) derives a realistic and feasible design plan by considering the physical constraints of the actual installation environment and regional operating regulations. Constraints include site area, grid connection capacity, power quality maintenance standards, energy storage system (ESS) safety standards (e.g., fire safety, insulation resistance, etc.), and climate characteristics (e.g., high / low temperature, humidity, etc.). The calculated design results may be provided as a list of multiple candidate design plans sorted in order of cost-effectiveness, in addition to a single optimal plan. The final results are displayed on the user terminal (20) in the form of graphs, economic reports, 3D digital twin simulations, etc., through the visualization output unit (280). Through this, the user can intuitively compare and review the design results and check the economic effects of each scenario in real time. Therefore, the installation and design optimization unit (270) functions as a highly intelligent design system that goes beyond simple design support functions by combining data-based economic evaluation and AI optimization algorithms to automate energy storage system (ESS) investment decision-making and ensure the maximization of profitability from the design stage.

[0089] The control unit (250) is a core module that comprehensively controls the data flow and computation procedures between each component, generates operation commands for the Power Conditioning System (PCS) and battery system according to the planned charge / discharge schedule, and reflects updated control commands from the replanning unit (260) in real time. The control unit (250) comprehensively manages the operational stability, energy efficiency, and economic feasibility of the entire system, and integrates and controls all procedures from the installation stage to the operation stage into a single closed-loop structure, thereby realizing continuous efficiency improvement and maximization of profitability.

[0090] The data collection unit (210) is configured to be connected to at least one of a water electrolysis facility, a combined heat and power (CHP) facility, and a pumped water storage hydropower (PSH) facility, and may include an energy flow data collection function for collecting power, heat, and hydropower energy flow data from the facilities.

[0091] In this case, the control unit (250) can simultaneously analyze energy flow data and the charging and discharging status of the ESS to adjust the power input of the water electrolysis facility, control the power and heat production of the combined heat and power (CHP) facility, and operate to appropriately switch the pumping and power generation operation modes of the pumped storage hydropower (PSH) facility. Through this, a control signal is generated so that each facility can be operated in accordance with current energy demand and supply conditions.

[0092] In addition, the control unit (250) can optimize the operating conditions of the entire system by comprehensively considering the energy exchange structure between the electrolysis, combined heat and power (CHP), and pumped storage hydropower (PSH) facilities, including the energy storage system (ESS), and the operating constraints of individual facilities. Through such integrated control, the cost efficiency and energy conversion efficiency of the entire energy system are improved, and operational stability is also greatly enhanced.

[0093] FIG. 3 illustrates a block diagram of an integrated control device (300) for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system according to another embodiment of the present invention. The device (300) includes a data collection unit (310), an economic evaluation unit (320), a scenario generation unit (330), an installation and design optimization unit (340), a visualization output unit (350), and a control unit (360) that comprehensively controls the computation and signal flow between these modules. The control unit (360) orchestrates the execution order of each module and automatically re-executes dependent modules upon input update to maintain consistency and up-to-dateness of results.

[0094] The data collection unit (310) collects various operational data required for ESS design and operation from users or external systems. Input data includes hourly consumption characteristics of load power, battery storage capacity (kWh), power conversion device (PCS) capacity (kW), renewable energy facility capacity such as solar and wind power, power tariffs, and operating constraints. Since the collection cycle, data format, and unit system of these data vary by source, the data collection unit (310) includes a schema mapping structure to automatically identify and integrate all inputs. This prevents data loss due to format mismatch even when linked with external systems (API-based power company servers, weather databases, facility sensor networks, etc.).

[0095] The collected data undergoes a unit normalization process before being converted into an internal common format. For example, inconsistencies in power units (kW, MW), energy units (kWh, MWh), rate units (won / kWh, $ / MWh), and time units (second, minute, hour) are automatically detected and converted into a unified unit system. Unit conversion is performed based on a predefined standard conversion table, and each data item is tagged with metadata and managed after conversion. In addition, when the same physical quantity is collected from multiple data sources (e.g., load power data collected simultaneously from a power company API and a user-side meter), a reliability-based weighted integration method is applied to calculate a final representative value. This standardization and unit normalization process ensures input consistency for the subsequent economic evaluation unit (320) and scenario generation unit (330) and prevents calculation errors that may occur during the modeling process.

[0096] The data collection unit (310) adopts a streaming-based data collection structure to ensure both real-time performance and reliability. This is implemented as an event-driven structure, so that updates and buffering are performed automatically whenever new data occurs. In the streaming path, a timestamp synchronization function is applied so that data collected at different periods are aligned on the same time axis. In addition, if a missing value occurs, the missing value can be automatically corrected by applying, for example, linear interpolation of an adjacent time point or a prediction-based imputation algorithm. Outliers caused by sensor noise or measurement errors are identified and removed through anomaly detection models such as statistical deviation analysis (Z-score), IQR (Interquartile Range), or Isolation Forest. This process maintains the consistency and quality of the data, thereby improving the accuracy of subsequent artificial intelligence prediction and economic evaluation.

[0097] The data collection unit (310) includes encryption and authentication procedures to ensure security and integrity. All data streams collected from external systems are transmitted via encrypted communication based on TLS (Transport Layer Security), and when linking with APIs, unauthorized access can be blocked by using authentication tokens such as OAuth 2.0, for example. Collected data is checked for tampering through a blockchain-based hash verification procedure, and only verified data can be stored in the internal database. In addition, the data collection unit (310) can periodically evaluate the data quality index and calculate the reliability level of the data to transmit it to each module within the system. Through this, the data collection unit (310) of the present invention functions as a highly reliable AI data pipeline that automates the integration, normalization, correction, and verification of heterogeneous data, going beyond a simple information input function, and fundamentally guarantees the accuracy and stability of the entire ESS design and operation optimization process.

[0098] The economic evaluation unit (320) quantitatively calculates the financial feasibility and long-term economic feasibility required for the installation and operation of an energy storage system (ESS) based on the operational input values ​​provided by the data collection unit (310). The economic evaluation unit (320) automatically calculates key economic indicators such as initial investment cost (CapEx), expected operating cost (OpEx), return on investment (ROI), payback period (PBP), net present value (NPV), and energy saving effect, and each item is dynamically calculated according to changes in time, tariff plan, load pattern, equipment efficiency, etc. Initial investment cost consists of batteries, power conversion system (PCS), inverters, grid connection facilities, installation costs, and site costs, while operating costs include maintenance costs, power loss costs, module replacement costs, and control system operating costs. The economic evaluation unit (320) is configured to realistically estimate actual installation costs and long-term operating costs at a specific point in time by breaking down each item into detailed element units and reflecting unit prices, depreciation rates, and exchange rate fluctuation rates stored in the database in real time.

[0099] The economic evaluation unit (320) applies a dynamic economic simulation structure based on long-term time series prediction. The economic evaluation unit (320) calls an artificial intelligence prediction model and receives long-term time series data such as changes in the power rate system at each stage (e.g., changes in Time-of-Use, demand response incentives, fluctuations in market unit price, etc.), load growth rate, battery performance degradation rate (SOH degradation curve), and variability in renewable energy generation. For example, in a scenario assuming operation for more than 10 years, the actual cash flow is calculated by considering the annual battery capacity degradation (e.g., 1.5% / year), the inflation-reflected rate of increase in unit price (e.g., 2% / year), and the rate of increase in maintenance costs. Through this, financial indicators such as return on investment (ROI), payback period (PBP), and net present value (NPV) are calculated in the form of dynamic functions that reflect variability at each point in time, rather than static values ​​at a single point in time.

[0100] The calculated economic indicators are used as inputs for the objective function and constraints in the subsequent stages, the scenario generation unit (330) and the installation and design optimization unit (340). For example, a multi-objective optimization problem is constructed to maximize the return on investment (ROI) while ensuring that the initial investment cost (CapEx) does not exceed a certain budget limit, and the function is used as a criterion for evaluating the cost efficiency and profitability of the entire system. The economic evaluation unit (320) outputs the indicators in a structured form (e.g., JSON-based Key-Value Pair or Matrix form) during this function construction stage so that the search algorithm can directly reference them. In addition, constraints such as site limitations, grid connection capacity limits, and renewable energy output uncertainty are attached in the form of metadata to support the automatic reflection of realistic conditions during optimization calculations.

[0101] As a result, the economic evaluation unit (320) functions as a highly intelligent economic simulation engine that goes beyond a simple calculator and combines AI-based long-term prediction, probabilistic risk analysis, and data structuring for multi-purpose optimization. Through this, users can predict investment performance in advance due to the passage of time, policy changes, and technological advancements, and make optimal financial decisions during the ESS installation and design phases.

[0102] The scenario generation unit (330) is a core component that automatically generates multiple installation and design combinations by comprehensively considering various financial indicators, operating conditions, and external constraints calculated by the economic evaluation unit (320). This module automatically generates complex scenarios that include not only ESS facility configurations (battery capacity, PCS capacity, inverter specifications, renewable energy connection ratio) but also operating policies (charge / discharge schedule, operating mode by tariff plan, peak reduction strategy, etc.). At this time, the scenario generation process consists of a two-stage structure consisting of a rule-based initialization stage and an artificial intelligence (AI)-based combination search stage. In the initialization stage, the range of basic design parameters is defined based on goals set by the user (e.g., minimizing investment costs, maximizing profits, shortening the payback period, etc.), and thousands of candidate scenarios are generated within that range. Subsequently, in the AI-based combination search stage, inefficient combinations are automatically removed through genetic algorithms, reinforcement learning, or Bayesian optimization techniques, and a group of candidate scenarios that simultaneously satisfy economic efficiency and technical efficiency are derived.

[0103] Each generated scenario is evaluated not merely as a combination of design configurations, but as an independent execution unit capable of quantitative comparison from an economic and operational perspective. For example, indicators such as cost efficiency, operation efficiency, self-consumption rate, peak shaving effect, and CO₂ reduction are calculated for each scenario. These evaluation indicators are not evaluated based on a single standard, but rather their relative importance is reflected according to the user's target weighting factor. For example, industrial consumers may set high weights for electricity bill reduction and payback period reduction, while public institutions or eco-friendly certification projects may assign higher weights to carbon reduction effects and self-consumption rate. Based on these multidimensional indicators, the scenario generation unit (330) constructs each scenario into a multi-objective evaluation vector and comprehensively compares and analyzes economic performance and efficiency.

[0104] Scenario priority assignment is performed using a weighted sum or Pareto optimization method. In the weighted sum method, a total score is calculated by applying the weight assigned to each evaluation item, and scenarios are ranked according to the result. On the other hand, in the Pareto-based method, multiple conflicting objective functions (e.g., cost minimization vs. efficiency maximization) are considered simultaneously to explore scenarios on the Pareto Front where improving one objective does not worsen another. Through this process, not only a single “optimal solution” but also multiple “non-dominated solutions” that allow the user to make strategic decisions are derived. This Pareto-based approach is highly effective in the field of energy system design requiring a balance between multiple objectives, and the scenario generation unit (330) of the present invention implements this with high computational efficiency by combining it with an AI search algorithm.

[0105] The scenario generation unit (330) transmits a scenario with a higher priority to the installation and design optimization unit (340) through linkage with the control unit (360). The control unit (360) verifies the detailed parameters (battery capacity, PCS ratio, predicted power generation amount, facility investment cost, etc.) of the transmitted higher-priority scenarios, evaluates their feasibility, and determines the final execution candidate. Through this mutual linkage structure, the system can derive a realistic optimal plan that considers not only economic efficiency but also actual operational feasibility, such as the installation environment, operating conditions, and grid constraints. Therefore, the scenario generation unit (330) of the present invention functions as an AI-based decision support engine that comprehensively reflects design, economic feasibility, and operation strategy, and significantly improves the quality of ESS installation and operation plans and design productivity.

[0106] The installation and design optimization unit (340) is an artificial intelligence (AI)-based search and optimization engine for automatically selecting the optimal installation plan that maximizes cost-performance from a plurality of installation and design candidate plans derived from the scenario generation unit (330). This unit mathematically models a multi-objective optimization problem and simultaneously considers economic indicators (e.g., initial investment cost (CapEx), operating cost (OpEx), return on investment (ROI), payback period (PBP), net present value (NPV)) and technical indicators (e.g., charge / discharge efficiency, self-consumption rate, battery life, power quality maintenance performance, etc.).

[0107] To this end, the installation and design optimization unit (340) converts the variables of each scenario (e.g., battery capacity, PCS capacity, installation location, renewable energy ratio, constraints, etc.) into inputs for an objective function and derives an optimal solution by repeatedly performing a search algorithm. In the search process, computational techniques such as genetic algorithm, Bayesian optimization, or mixed integer linear programming (MILP) are applied.

[0108] The search process consists of a structure of iterative evaluation and updating. At each stage, multiple candidate proposals are compared to select the best-performing ones, and new combinations are generated based on these superior candidates to find progressively better solutions. In this process, the search technique applies the concept of Pareto optimization to derive a multi-equilibrium solution where improving one objective does not worsen another. For example, while increasing battery capacity enhances peak reduction, it also raises initial investment costs; therefore, the optimization unit calculates the balance point between these two factors to propose an optimal capacity that satisfies both economic efficiency and effectiveness above a certain level. Furthermore, the Bayesian optimization algorithm improves search efficiency by learning from results obtained in previous iterations, enabling it to find accurate optimal values ​​while minimizing computational costs.

[0109] The selected optimal installation plan or design plan is automatically calculated in the form of system configuration specifications (e.g., number of battery modules, PCS capacity, inverter rating, facility layout), estimated initial investment cost (CapEx) / operating cost (OpEx), and return on investment (ROI). In addition, the installation and design optimization unit (340) performs sensitivity analysis to quantitatively evaluate the impact of key design variables (e.g., battery capacity, installation location, change in electricity rate scheme, etc.) on economic feasibility and operational efficiency. Through this, the user can visually identify which variables have the greatest impact on profitability and secure the ability to respond to decision-making in response to policy changes or market fluctuations. As such, the installation and design optimization unit of the present invention is an Intelligent Design Decision Engine that combines an AI search algorithm, multi-objective optimization, and sensitivity analysis, thereby minimizing investment risk from the initial stage of ESS installation and simultaneously securing economic efficiency and technical reliability.

[0110] The visualization output unit (350) is an interface module that converts and provides optimal results derived from the installation and design optimization unit (340) into various forms of visual information so that the user can intuitively verify and analyze them. This unit includes a 2D and 3D rendering engine and expresses data such as configuration specifications, layout structure, facility capacity, power flow, and economic indicators of the ESS facility in a visualized form. The user can grasp the results at a glance not only through simple table-format data but also through graphs, gauges, heatmaps, and energy flow diagrams. In addition, the visualization output unit (350) enables simulation verification in the pre-installation stage by implementing design results in a virtual space identical to the actual installation environment based on Digital Twin technology. For example, the user can virtually place battery modules, PCS, inverters, grid connection lines, etc., on a 3D model and simulate environmental changes or operating conditions to check expected power flow, energy saving effects, and facility response characteristics in real time.

[0111] The visualization output unit (350) is configured so that economic indicators and system configurations are updated in real time according to changes in the user's design variables. When the user adjusts parameters such as battery capacity, installation scale, electricity rate scheme, and load conditions on the interface, the system immediately performs recalculation and updates and displays key indicators such as IRR, NPV, PBP, energy saving rate, investment cost (CapEx), and operating cost (OpEx). Unlike existing static report-type systems, this real-time bidirectional interaction structure significantly improves the speed and accuracy of the user's decision-making by immediately visualizing the impact of design changes. In addition, the changed results are dynamically reflected through a graphic engine, so that, for example, if the ESS installation capacity is increased, the trend of rising electricity rate savings or increasing investment cost is immediately expressed in the form of a curve on the graph. The visualization output unit of the present invention supports efficient decision-making and equipment optimization at the design stage simultaneously by implementing a real-time visualization feedback loop that integrates data simulation, economic analysis, and design modeling.

[0112] The visualization output unit (350) is linked to a user terminal or management system and provides key analysis results in the form of a dashboard. Through this, the manager can monitor key performance indicators (KPIs) such as project-specific economic feasibility, installation efficiency, and system utilization rate, and can compare and analyze design change history and results. The visualization output unit (350) is designed to operate in a web-based or cloud-based environment, providing the same 3D visualization interface even when accessed remotely. In addition, digital twin data is used as feedback data during the actual operation phase, enabling continuous model improvement by comparing the difference between prediction and actual operation performance. Therefore, the visualization output unit (350) goes beyond a simple result display function and serves as an end-user interface that connects installation, design, and operation optimization in a single closed loop, realizing integrated management and visual-based decision support throughout the entire lifecycle of the energy storage system.

[0113] Figure 4 is a conceptual diagram illustrating the peak shaving effect of an energy storage system according to the present invention. The graph on the left of the figure represents a state where peak control is not performed, and the graph on the right represents a state where load distribution is achieved using an energy storage system. In both cases, the total power consumption for the day (S₁= S₂) is the same, but the phenomenon is shown where the electricity bill varies significantly depending on the difference in the maximum demand at the time of peak power occurrence.

[0114] In the state on the left, the load increases sharply around 5 PM, causing the peak demand to rise to the level of 100 units. This is a typical example where the demand charge skyrockets due to a short-term peak load, and even with the same energy consumption, the power charge during the peak period has a significant impact on the overall electricity bill. On the other hand, the state on the right shows the result of load leveling the load curve through AI-based ESS operation control.

[0115] In the state shown on the right, the ESS discharges the energy it has charged prior to the peak hours to suppress a rapid increase in load and reduce the maximum power demand to approximately 50 units. Consequently, the maximum value of the power demand curve is reduced by half, resulting in a significant reduction in peak charges. This diagram quantitatively demonstrates that electricity rates can vary significantly depending on the load distribution pattern, even when consuming the same total energy (S₁= S₂), and visually explains the effect of ESS peak reduction control on reducing demand charges. Therefore, the AI-based ESS operation optimization technology of the present invention provides the technical effect of simultaneously realizing electricity cost reduction and equipment efficiency improvement by combining load forecasting and peak control to mitigate unnecessary peak power usage in advance.

[0116] Figure 5 is a conceptual diagram illustrating the peak shaving and load leveling effects through charge-discharge control of an artificial intelligence (AI)-based energy storage system (ESS) according to the present invention. The graph on the left shows the load pattern when only photovoltaic (PV) power generation is installed, and the graph on the right shows the load pattern when charge-discharge control is performed with the additional installation of an ESS.

[0117] P1 in the graph on the left represents the power load curve before the installation of the solar power generation facility, where solar power generation (PV1) occurs during the daytime (approx. 9:00 to 17:00) when solar irradiance is high, offsetting some of the load. Accordingly, the power load curve after the application of solar power generation is represented as P2, where total power consumption decreases, but peak load still occurs during the evening hours (after 18:00). In other words, solar power generation alone cannot completely suppress the surge in power demand during periods of low solar irradiance, so the effect of reducing peak charges is limited.

[0118] In contrast, the graph on the right shows the case where AI-based charge and discharge control is performed by additionally installing an ESS. During the day, the ESS stores energy by charging the surplus power generated by solar power generation, and distributes the load by discharging the stored energy during the evening peak hours. As a result, the existing maximum load point P2 shifts to a lower P3 after ESS operation, and the peak height of the overall load curve decreases significantly.

[0119] The ESS control algorithm of the present invention performs predictive control rather than simple time-based control. That is, by simultaneously considering load forecasting, solar power generation forecasting, tariffs, contracted power, and the battery's charge-discharge efficiency (RTE), it automatically determines an efficient charging time during the day and initiates discharge just before the peak occurs in the evening. This prevents unnecessary energy circulation, minimizes battery degradation, and maximizes the effect of reducing electricity costs. Furthermore, by predictively determining the charging and discharging times, the present invention can simultaneously achieve an extension of battery life and improved long-term economic efficiency.

[0120] FIG. 6 is a flowchart illustrating the AI-based automatic estimation calculation procedure of the AI-based energy storage system (ESS) installation and design optimization device of the present invention. This figure illustrates, step-by-step, a process in which a user can derive installation estimates and economic analysis results in real-time (<1 minute) with minimal input information. The entire process consists of three stages: input of electricity usage data (S710), input of installation address (S720), and input of electricity bill reduction target (S730). Each stage is interconnected to ultimately perform automatic estimation calculation by an AI model.

[0121] First, in step S710, the user's electricity usage data input step is performed. The user can upload time-based power usage data provided by the power company or the Energy Management System (EMS), or automatically input it through API integration. This data includes load patterns, peak occurrence times, and time-based consumption characteristics, and is used as basic training data for the AI ​​model. The input load data is immediately preprocessed to remove outliers, correct missing values, and perform unit normalization, after which it is converted into a structure suitable for annual (365-day) simulation.

[0122] In the next step, S720, the installation address input step is performed. The user enters the address of the installation site, and based on this, the system automatically links regional weather data (solar radiation, temperature, humidity, wind speed, etc.) and power system information (regional tariffs, peak hours, transmission loss coefficients, etc.). In this step, the optimal facility configuration plan can be derived by considering the photovoltaic power generation potential (PV Potential), appropriate ESS capacity, and site constraints according to the installation location. In other words, by simply entering the installation address, it is possible to design a system optimized for each region by reflecting real-time data from the Korea Meteorological Administration and the Korea Power Exchange.

[0123] In the final step S730, the input of electricity cost reduction targets and automatic estimate calculation are performed. The user sets criteria such as target reduction rates (e.g., 30%, 50%), investment limits, and payback periods, and the AI ​​model uses these as the objective function to automatically calculate economic indicators such as installation costs (CapEx), estimated operating costs (OpEx), ROI, IRR, and NPV. In this invention, the AI ​​algorithm comprehensively analyzes input load data, weather conditions, tariff plans, and reduction targets to present the optimal ESS and PV capacities, system specifications, and estimated investment costs in real time. The calculation results are visualized in the form of graphs, tables, and simulation curves and provided to the user, allowing for comparative analysis of power consumption, cost reduction effects, and peak suppression performance before and after installation.

[0124] Accordingly, FIG. 6 is a diagram visually illustrating the integrated automatic estimation and simulation process performed by the AI-based installation and design optimization device of the present invention. The system receives only three input values ​​from the user and can derive an “AI real-time installation estimate” including actual installation costs and economic feasibility predictions within one minute, providing a processing speed more than 95% faster and high accuracy compared to conventional manual estimation. Furthermore, by automatically generating pre- and post-installation performance comparison reports based on 365-day simulations, it provides a strong competitive advantage in the ESS investment feasibility review and bidding / proposal stages.

[0125] Although embodiments of the present invention have been described, the spirit of the present invention is not limited by the embodiments presented in this specification. Those skilled in the art who understand the spirit of the present invention may easily propose other embodiments within the scope of the same spirit by adding, changing, deleting, or adding components, and such are also to be considered to be included within the spirit of the present invention. Explanation of the symbols

[0126] 11: Power company 12: Korea Meteorological Administration 13: Power Exchange 20: User terminal 30: Network 200: Integrated control unit for installation, design, and operation optimization of AI-based energy storage systems

Claims

Claim 1 An integrated control device for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system comprises: a data collection unit that collects user basic input data, including electricity usage data, installation location information, and operation target information, from a user or external system, and operation-related data, including contracted power information, tariff plan information, load power data, renewable energy generation data, battery status data, and external environment data provided by a power supplier or consumer, and aligns the data collected from different sources into a common time axis and a common data format; a prediction unit that predicts future load changes and renewable energy generation changes using the data aligned by the data collection unit; a planning unit that generates a plurality of operation scenarios by reflecting the prediction results of the prediction unit, the tariff plan information, the contracted power information, the battery status data, and operation constraints, and calculates operation evaluation indicators for each operation scenario, including at least one of cost, revenue, energy saving effect, and battery degradation factor; and an economic evaluation unit that generates a plurality of candidate installation plans or candidate design plans based on the user basic input data and the operation-related data, and for each candidate installation plan or candidate design plan, includes at least one of initial investment cost, estimated operating cost, return on investment, and payback period. An installation and design optimization unit that calculates system configuration specifications including at least one of an indicator, storage device capacity, power converter capacity, and renewable energy facility capacity; and a search processing unit that compares and evaluates a plurality of operation scenarios generated by the planning unit and a plurality of candidate installation plans or candidate design plans generated by the installation and design optimization unit within the same constraint system to produce an integrated optimization result that reflects both the economic indicator of the installation and design phase and the operation evaluation indicator of the operation phase, and selects an optimal installation plan or optimal design plan and an optimal operation scenario based on the integrated optimization result.An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system, comprising: a control unit that outputs a control signal to control the charging or discharging operation of an energy storage device according to the optimal operation scenario based on system configuration specifications corresponding to the optimal installation plan or optimal design plan selected by the search processing unit; and a replanning unit that detects a difference between actual load data, actual renewable energy generation data, or actual battery status data occurring during operation and the prediction result of the prediction unit, and if the difference satisfies a predefined threshold condition, updates at least one operation among the planning unit, the installation and design optimization unit, and the search processing unit so that the updated optimal operation scenario is reflected in the control signal of the control unit, wherein the system configuration specifications calculated by the installation and design optimization unit are used as the charging / discharging control range of the control unit and the update target constraint of the replanning unit, and the operation result updated by the replanning unit is reflected in the recalculation of the economic indicator and the operation evaluation indicator, so that the installation and design optimization and the operation optimization are linked in a single closed-loop structure. Claim 2 An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the data collection unit collects data having different collection cycles, data formats, or unit systems, and performs at least one of time synchronization, unit normalization, missing value correction, and outlier removal on the collected data to generate an operational data set that can be commonly used by the prediction unit, the planning unit, and the installation and design optimization unit. Claim 3 An integrated control device for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the planning unit sets a plurality of evaluation indicators including at least two of operating costs, energy rates, power trading revenue, demand response revenue, carbon reduction effects, and battery performance degradation factors as an objective function, and generates the plurality of operating scenarios by reflecting the objective function together with at least one constraint among a contracted power limit, a charge state limit, a discharge power limit, and a power converter capacity limit. Claim 4 An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the installation and design optimization unit sets at least one of battery capacity, power converter capacity, renewable energy facility capacity, installation location, installation site conditions, and operation target information as a design variable, and calculates cost efficiency and operation efficiency for each combination of the design variables to generate the plurality of candidate installation plans or candidate design plans. Claim 5 An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the search processing unit generates a plurality of integrated candidate plans by combining the plurality of candidate installation plans or candidate design plans with the plurality of operation scenarios, and for each integrated candidate plan, evaluates at least one of initial investment cost, estimated operating cost, return on investment, payback period, peak reduction effect, energy saving effect, and battery degradation cost, and selects the integrated candidate plan with the maximum cost-effectiveness as the integrated optimization result. Claim 6 An integrated control device for the installation, design, and operation optimization of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the replanning unit updates control commands by repeatedly performing optimization procedures of the planning unit and the search processing unit in a moving window method or an adaptive method when a difference between the prediction result and actual operation data, a sudden fluctuation in renewable energy output, an abnormal pattern of load, a change in tariff plan, a fluctuation in power market prices, or a change in battery status is detected. Claim 7 An integrated control device for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the control unit controls the timing and amount of charging and discharging in conjunction with an energy management device including a power converter, and automatically generates a control signal corresponding to the detection of exceeding contracted power, violation of power system constraints, or an emergency situation to maintain the overall energy flow. Claim 8 An integrated control device for optimizing the installation, design, and operation of an artificial intelligence (AI)-based energy storage system, wherein, in claim 1, the integrated control device provides at least one of the optimal installation plan or optimal design plan calculated by the installation and design optimization unit, the optimal operation scenario selected by the search processing unit, the economic indicator, and the operation evaluation indicator in the form of a graph, dashboard, report, 3D model, or digital twin to a user terminal or management system interface, and further comprises a visualization output unit that causes the economic indicator and the operation evaluation indicator to be recalculated and updated and displayed when the user changes design variables or operation goal information. Claim 9 delete Claim 10 delete Claim 11 delete

Images