Method and device for active control of power equipment according to remaining power capacity

Abstract

The present invention relates to an active control method for power facilities based on remaining power capacity, performed by a server, comprising: (a) selecting power consuming products capable of adjusting load usage in stages among power consuming products within a building and generating a total load usage pattern when all power consuming products are used; (b) simulating a current or future total load usage pattern while adjusting the load usage of each of the selected power consuming products; (c) predicting the remaining battery capacity of the energy consuming device by time period, considering the power generation amount of the energy consuming device over time and the total load usage pattern; (d) setting a minimum threshold value for the remaining battery capacity for stable operation of the building, deriving a plurality of remaining battery capacity patterns according to the simulated plurality of total load usage patterns, and detecting whether there exists a remaining battery capacity pattern among the remaining battery capacity patterns that temporarily deviates from the minimum threshold value. The present invention relates to an active control method for power equipment based on remaining power capacity, comprising: (e) a step of automatically controlling the power consumption of the power consuming product according to a battery remaining capacity pattern defined in a state where the battery remaining capacity is always above the minimum threshold value, wherein the total load usage pattern refers to load usage over time.

Description

Technology Field

[0001] The present invention relates to a method and apparatus for active control of power facilities based on remaining power capacity, and more specifically, to a technology for precisely analyzing the power consumption pattern of a plurality of power-consuming products located inside a building, and based on this, predicting the power generation amount and battery status of an energy-generating device that supplies power to the power-consuming products and actively controlling them. Background Technology

[0002] Conventional power facility operation methods have largely relied on static approaches. In other words, power-consuming devices were controlled through pre-set schedules or simple conditional statements, and the structure was unable to flexibly respond to real-time changes.

[0003] This control method lacks consideration for various external variables such as weather, user behavior, and equipment status, so problems such as power waste or under-supply frequently occur in actual operation.

[0004] In particular, it has been pointed out that when utilizing recently popular renewable energy, it is difficult to provide an efficient and stable power supply with existing systems because power generation fluctuates rapidly depending on time and weather conditions.

[0005] In addition, conventional technology lacked consideration for the characteristics and controllability of power-consuming devices. For example, although some devices could adjust load usage in stages, they were simply controlled by an On / Off method or treated identically, which resulted in missing the opportunity to finely control power consumption and an increased risk of battery over-discharge.

[0006] Therefore, technology is needed to solve such problems. The problem to be solved

[0007] The present invention aims to solve the problems of the aforementioned prior art and provides a method and device for active control of power facilities based on remaining power capacity.

[0008] In addition, through this, the purpose is to control power-consuming devices by segmenting them, perform simulations based on time-based load patterns, and actively adjust the load according to the predicted remaining battery capacity. means of solving the problem

[0009] As a technical means for achieving the technical problem described above, an active control method for power facilities based on remaining power capacity, performed by a server according to an embodiment of the present invention, comprises: (a) selecting power consuming products capable of adjusting load usage in stages among power consuming products within a building and generating a total load usage pattern when all power consuming products are used; (b) simulating a current or future total load usage pattern while adjusting the load usage of each of the selected power consuming products; (c) predicting the remaining battery capacity of the energy consuming device by time period by considering the power generation amount of the energy consuming device over time and the total load usage pattern; (d) setting a minimum threshold value for the remaining battery capacity for stable operation of the building, deriving a plurality of remaining battery capacity patterns according to the simulated plurality of total load usage patterns, and detecting whether there exists a remaining battery capacity pattern among the remaining battery capacity patterns that temporarily deviates from the minimum threshold value. and (e) a step of automatically controlling the power consumption of the power consuming product so as to follow a battery remaining capacity pattern defined in a state where the battery remaining capacity is always above the minimum threshold value; wherein the total load usage pattern may refer to load usage over time.

[0010] Additionally, the above step (a) comprises: (a-1) a step of subdividing the load usage for each power consumption product into multiple stages; (a-2) a step of extracting the maximum load usage and minimum load usage of the power consumption product from the load usage of the multiple stages; (a-3) a step of separately generating a load usage pattern for each power consumption product according to the usage stage for each power consumption product; (a-4) a step of distinguishing and generating load usage patterns for weekdays and weekends for each power consumption product; and (a-5) a step of generating a load usage pattern for a specific time interval for each power consumption product and summing them to generate an overall load usage pattern; wherein, if the power consumption product is an on / off control load, the multiple power consumption products are grouped into one group and the load usage is subdivided according to the number of power consumption products that are turned on / off within the group, and if the power consumption product is an energy usage control load, the load usage is subdivided according to the variation of its own set value.

[0011] Additionally, the above step (b) may involve simulating multiple load usage patterns for each power consuming product according to each step among multiple steps regarding load usage, and generating an overall load usage pattern for multiple cases by summing the load usage patterns for each power consuming product according to various cases.

[0012] Additionally, the above step (d) may be to set the threshold value to either the minimum power value that the SoC must maintain for a system containing multiple power-consuming products to operate, or the over-discharge upper limit value of the batteries included in the multiple power-consuming products, and to classify the power-consuming products into any one of the following: an ON / OFF control load capable of only turning on / off; an essential load that must operate regardless of user settings; and an energy consumption control load capable of adjusting the load usage amount in steps, and to control the output of the energy consumption control load.

[0013] Additionally, the above step (d) includes the step of measuring the current remaining battery capacity and predicting the remaining battery capacity for future time periods to generate a battery remaining capacity pattern; wherein the step of predicting the remaining battery capacity may be a prediction made by considering the maximum power generation of the photovoltaic power generation device predicted based on solar irradiance, the battery facility capacity, and the load usage of the power consuming product.

[0014] In addition, the step of generating the battery remaining capacity pattern may include: measuring the current battery remaining capacity; predicting the future battery remaining capacity based on the following regression equation; generating the battery remaining capacity pattern based on SoCt+1 = maximum power generation t + (battery facility capacity Х SoCt) - load usage t (where t is a natural number greater than or equal to 1) and the current and future battery remaining capacities.

[0015] In addition, (f) the power generation amount based on solar irradiance is re-predicted at preset time intervals, and the load usage pattern according to the change in time is updated at preset time intervals, thereby re-predicting the battery remaining capacity at preset time intervals. If the re-predicted battery remaining capacity is higher than the battery remaining capacity predicted in the previous time interval, the load usage of a predetermined type and number of power consuming products may be increased compared to the previous time interval. Effects of the invention

[0016] The present invention can reduce unnecessary energy waste by subdividing the load of power consuming devices and automatically adjusting it as needed through a method and device for active control of power facilities based on remaining power capacity.

[0017] Furthermore, it can contribute not only to reducing electricity costs but also to lowering the overall operating costs of buildings or facilities. In particular, by reducing unnecessary loads during peak hours and enabling strategic power distribution that considers battery status, it allows for the maintenance of optimal performance even with limited energy sources. Brief explanation of the drawing

[0018] FIG. 1 is an exemplary diagram of an active power equipment control system according to residual power capacity, according to one embodiment of the present invention. FIG. 2 is a block diagram showing the internal configuration of a main server according to one embodiment of the present invention. FIG. 3 is a graph showing the average power consumption on weekdays according to one embodiment of the present invention. FIG. 4 is a graph showing the average weekend power consumption according to one embodiment of the present invention. FIG. 5 is an exemplary diagram showing a graph of remaining battery capacity and a section exceeding the lowest threshold value, according to an embodiment of the present invention. FIG. 6 is an exemplary diagram of a GUI provided to a user terminal according to an embodiment of the present invention. FIG. 7 is an exemplary diagram of actual power generation and load power consumption provided to a user terminal according to an embodiment of the present invention. Figure 8 is an example diagram of the predicted power generation and predicted SOC provided to the user terminal. FIG. 9 is a flowchart showing the execution sequence of a power facility active control method according to residual power capacity, according to one embodiment of the present invention. Specific details for implementing the invention

[0019] Embodiments of the present invention are described below with reference to the attached drawings so that those skilled in the art can easily implement the invention. However, the present invention may be embodied in various different forms and is not limited to the embodiments described herein. Furthermore, in order to clearly explain the present invention in the drawings, parts unrelated to the explanation have been omitted, and similar parts throughout the specification are denoted by similar reference numerals.

[0020] Throughout the specification, when a part is described as being "connected" to another part, this includes not only cases where they are "directly connected," but also cases where they are "electrically connected" with other components interposed between them. Furthermore, 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.

[0021] In this specification, the term "part" includes a unit realized by hardware, a unit realized by software, and a unit realized using both. Additionally, one unit may be realized using two or more pieces of hardware, and two or more units may be realized by one piece of hardware. Meanwhile, "part" is not limited to software or hardware, and "part" may be configured to reside in an addressable storage medium or configured to run on one or more processors. Accordingly, as an example, "part" includes components such as software components, object-oriented software components, class components, and task components, as well as processes, functions, attributes, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuits, data, databases, data structures, tables, arrays, and variables. The functions provided within the components and "parts" may be combined into a smaller number of components and "parts" or further separated into additional components and "parts." In addition, the components and '~parts' may be implemented to play one or more CPUs within the device or secure multimedia card.

[0022] The "user terminal" mentioned below may be implemented as a computer or portable terminal capable of connecting to a server or other terminal via a network. Here, the computer may include, for example, a laptop, desktop, or notebook equipped with a web browser, and VR HMDs (e.g., HTC VIVE, Oculus Rift, GearVR, DayDream, PSVR, etc.). Here, the VR HMD includes PC models (e.g., HTC VIVE, Oculus Rift, FOVE, Deepon, etc.), mobile models (e.g., GearVR, DayDream, Storm Mirror, Google Cardboard, etc.), console models (PSVR), and stand-alone models implemented independently (e.g., Deepon, PICO, etc.). Portable terminals are wireless communication devices that ensure portability and mobility, and may include, for example, smartphones, tablet PCs, and wearable devices, as well as various devices equipped with communication modules such as Bluetooth (BLE, Bluetooth Low Energy), NFC, RFID, Ultrasonic, Infrared, WiFi, and LiFi. Additionally, "network" refers to a connection structure capable of exchanging information between each node, such as terminals and servers, and includes Local Area Networks (LAN), Wide Area Networks (WAN), the World Wide Web (WWW), wired and wireless data communication networks, telephone networks, wired and wireless television communication networks, etc.Examples of wireless data communication networks include, but are not limited to, 3G, 4G, 5G, 3GPP (3rd Generation Partnership Project), LTE (Long Term Evolution), WIMAX (World Interoperability for Microwave Access), Wi-Fi, Bluetooth communication, infrared communication, ultrasonic communication, Visible Light Communication (VLC), and LiFi.

[0023] The present invention relates to a method for actively controlling power facilities based on remaining power capacity and a device for providing the same. Specifically, it relates to a technology that precisely analyzes the power consumption pattern of a plurality of power-consuming products located inside a building, and based on this, predicts the power generation amount and battery status of an energy-generating device supplying power to the power-consuming products and actively controls them.

[0024] Referring to FIG. 1, for this purpose, an active power facility control system according to a residual power capacity according to one embodiment of the present invention may be composed of a main server (100), a power plant server (200), a user terminal (300), a battery (400), and a plurality of loads (500).

[0025] Here, the main server (100) is a device that has a program or application installed to perform an active control method for power facilities according to the remaining power capacity, and may include a processor and memory for executing the program or application as shown in FIG. 2, a database (DB) in which data for executing the program or application is stored, and a communication module for communicating via wired or wireless connection with a power plant server (200) or a user terminal (300) to be described later.

[0026] Next, the power plant server (200) provides the generated power to a battery (400) installed in a specific building or a collection of one or more such buildings in which a plurality of power consuming products are installed, and includes an energy generation device for generating power, which may be implemented in different forms of energy generation devices according to various embodiments. Typically, the energy generation device may be implemented in the form of a thermal, wind, nuclear, or hydroelectric power generation device, and may also be implemented as a power generation device using new and renewable energy.

[0027] Meanwhile, the user terminal (300) may include a standard smartphone, tablet PC, laptop, and desktop, etc., having a program or application already installed that performs the active control method of power facilities according to the remaining power capacity. In this case, the user terminal (300) may be a terminal owned by the manager of a specific building that is the target of the active control method of power facilities according to the remaining power capacity of the present invention, and in an additional embodiment of the present invention, a procedure for the user terminal (300) to be authenticated as a manager terminal by the main server (100) may be further included. In the case of such a procedure, the main server (100) may request an image of a manager appointment letter or business card for the building to be recognized as a manager from the user terminal (300), and when the user terminal (300) transmits this, the user terminal (300) may be set as a manager terminal.

[0028] Returning to the topic, the battery (400) receives and stores power produced by the energy generation device of the power plant server (200) and may be used as part of an Energy Storage System (ESS). Typically, the battery (400) of the present invention may be implemented in the form of a lithium-ion battery, a lithium iron phosphate battery (LiFePO₄), a lead-acid battery, a sodium-sulfur battery (NaS), and a flow battery, but is not necessarily limited thereto.

[0029] Next, according to one embodiment of the present invention, the load (500) refers to the amount of power consumed by each power-consuming product, and may be a subject that changes over time and can be adjusted according to the situation. As illustrated, there may be multiple loads (500) (N), and this number may change according to the building conditions or the user's intention.

[0030] Below, we will explain an active control method for power facilities based on remaining power capacity, performed through the components and system described above.

[0031] A method for active control of power facilities according to residual power capacity according to one embodiment of the present invention may be disclosed by a main server (100) selecting power consuming products among power consuming products in a building that can adjust the load usage amount in stages.

[0032] The power consumption products selected by the main server (100) may consist of power consumption products capable of only turning on / off (ON / OFF control load), power consumption products that must operate regardless of user settings (essential load), and power consumption products capable of adjusting load usage in stages (energy usage control load).

[0033] Here, the ON / OFF control load is a load that can only be turned on and off, such as a light that cannot be adjusted in brightness or a computer case, and may be a power-consuming product that always consumes a constant amount of power. The essential load is a load that requires constant operation, such as a refrigerator, a security device, or a fire alarm device, and may be a power-consuming product that must operate regardless of the user's settings. Meanwhile, the energy usage control load may be a load (500) that can flexibly adjust power usage, such as a cooling unit or a heating unit.

[0034] Meanwhile, a load controller may be further included between the load (500) and the main server (100). A load controller according to one embodiment of the present invention controls the usage (power consumption) of a plurality of loads and can be implemented in various forms. In a representative embodiment, the load controller may include a relay and contactor-based controller, an SCR / Thyristor-controlled controller, and an inverter (VFD, Variable Frequency Drive)-controlled controller. The load controller may be implemented in a mixed form of local control and central control, which may be installed near each load or in a building or space where a plurality of loads are installed, and which is either synchronized with the main server (100) after independent control or controlled in real time by the control of the main server.

[0035] Next, the main server (100) generates a total load usage pattern when selected power consumption products are used.

[0036] Referring to FIG. 3, the total load usage pattern according to one embodiment of the present invention can be implemented in the form of a graph that visually represents the power consumption pattern, i.e., the load usage pattern, which changes over time due to the use of all power-consuming products within a building. Here, the horizontal axis represents the flow of time over a day, and the vertical axis represents the amount of power consumed during that time period. Through this, the load usage pattern of the present invention can show the total power usage at a specific point in time and may include various load patterns that appear when multiple power-consuming products operate simultaneously or operate at different output levels. That is, the total load usage pattern according to one embodiment of the present invention may refer to the load usage over time.

[0037] The overall load usage pattern may have distinct characteristics depending on the time of day. For example, power usage may be low in the morning, then increase sharply after 9:00 AM when work begins, decrease slightly during lunchtime, increase again in the afternoon, and decrease in the evening after work. This is similar to the actual flow of power usage in a building, and based on this, the main server (100) can derive an optimal control scenario by predicting energy demand through the process described below.

[0038] To generate this load usage pattern, the main server (100) subdivides the load usage for each power-consuming product into multiple stages. For example, assuming the power-consuming product is an air conditioner, if the cooling temperature of the air conditioner ranges from 16 degrees to 30 degrees, the main server (100) can subdivide the power consumed by each cooling temperature of the air conditioner into multiple sections. If it is subdivided into 10 stages, the section with the highest power consumption is the section operating at 16 to 18 degrees, so this section becomes the 10th stage, and conversely, the section with the lowest power consumption may be subdivided into the 28 to 30 degree range.

[0039] Additionally, the main server (100) can perform subdivision in different ways depending on the type of power consumption product. When the power consumption product is an on / off control load, the main server (100) groups multiple power consumption products into one group and subdivides the load usage according to the number of power consumption products that are turned on / off within the group. On the other hand, when the power consumption product is an energy usage control load, the load usage may be subdivided according to changes in its own setting value.

[0040] Next, the main server (100) extracts the maximum load usage and minimum load usage of the power consumption product among the load usages of multiple stages. At this time, depending on the embodiment, the maximum and minimum load usage values ​​may be received directly from the power consumption product; however, according to another embodiment of the present invention, the main server (100) may receive the product name from the power consumption product and extract the maximum load usage and minimum load usage by crawling the internet to collect output information of the product. This can be applied in the same way to the segmentation process described above, and in this case, in addition to the output information of the product, information such as mode information (modes with different power consumption, such as cooling or dehumidification in the case of an air conditioner) and temperature setting values ​​may be additionally collected as a result of the internet crawling.

[0041] The main server (100) separately generates a load usage pattern for each power consumption product according to the usage level of each power consumption product extracted.

[0042] Referring to Fig. 4, load usage patterns for weekdays and weekends can be generated separately for each power-consuming product. Additionally, load usage patterns for specific time intervals are generated for each power-consuming product and aggregated to generate an overall load usage pattern.

[0043] Next, the main server (100) performs a process of simulating the current or future total load usage pattern while adjusting the load usage of each of the selected power consumption products.

[0044] The simulation can be performed by predicting the State of Charge (SOC, remaining battery capacity), which will be explained later, over time.

[0045] In addition, the main server (100) simulates multiple load usage patterns for each power consumption product according to each stage among multiple stages of load usage, and generates an overall load usage pattern for multiple cases by summing the load usage patterns for each power consumption product according to various cases. That is, when multiple power consumption products are installed, the load usage pattern for each power consumption product is simulated and the results are combined, or when even a single power consumption product has multiple outputs (e.g., a fan's low wind, medium wind, and high wind), the load usage pattern for each output is output separately and combined to generate an overall load usage pattern.

[0046] According to an additional embodiment of the present invention, a load usage pattern may be generated using a linear regression model. In such an embodiment, the linear regression model is completed by setting past data regarding the load usage pattern generated through the method described above as training data for the linear regression model and training it, and by inputting a present or future point in time into the completed linear regression model, the load usage pattern from the past to the present or from the present to a future point in time can be predicted and output.

[0047] At this time, the load usage pattern used as training data may include at least one of the power usage for the last 24 hours for each power product, the daily average power usage for the last 7 days, the maximum and minimum values ​​of power usage, and the daily standard deviation of power usage, and as previously explained, it may be received directly from the power consuming product or collected by crawling the internet by the main server (100).

[0048] Next, the main server (100) can perform an operation of predicting the remaining capacity of the battery (400) of the energy generation device by time period, taking into account the amount of energy generated by the energy generation device over time and the overall load usage pattern.

[0049] The main server (100) receives the capacity that was charged in the battery (400) before performing the above operation, and can estimate it through the value obtained by subtracting the sum of the previously predicted load usage patterns from the received charge capacity.

[0050] Afterward, the main server (100) sets a minimum threshold value for the remaining capacity of the battery (400) for the stable operation of the building, and then derives a pattern of remaining capacity of the multiple batteries (400) according to a simulated pattern of multiple total load usage.

[0051] At this time, the minimum threshold value may be set as either the minimum power value that the SoC must maintain for a system containing multiple power-consuming products to operate, or the over-discharge upper limit value of each of the multiple power-consuming products, and may be set in advance prior to the implementation of the present invention.

[0052] Meanwhile, according to an additional embodiment of the present invention, the main server (100) can also predict the remaining battery capacity at a future point in time.

[0053] In the corresponding embodiment, the main server (100) first measures the remaining capacity of the current battery (400), predicts the remaining capacity of the battery at different time intervals in the future, and generates a battery remaining capacity pattern.

[0054] Here, the future remaining battery capacity can be calculated based on the following regression equation ([Equation 1]).

[0055] [Formula 1]

[0056] SoCt+1 = Maximum Power Generation t + (Battery Capacity × SoCt) - Load Usage t (t is a natural number greater than or equal to 1)

[0057] Through such a formula, the main server (100) can generate a battery remaining capacity pattern based on the current and future battery remaining capacity, and the battery remaining capacity may be predicted by considering the maximum power generation amount of the solar power generation device, i.e., the power plant server (200), the battery facility capacity, and the load usage amount of the power consuming product based on solar irradiance.

[0058] Referring to FIG. 5, the main server (100) detects whether there is a section (600) that temporarily exceeds the previously set minimum threshold value among the battery remaining capacity patterns derived through the above process.

[0059] According to an additional embodiment of the present invention, when a section (600) that deviates from the minimum threshold value is detected by the main server (100), if the section is a past point in time, the main server (100) may provide a message to the user terminal (300) indicating that a deviation from the threshold value occurred in the section. On the other hand, if it is a future point in time, that is, if it is determined that a deviation from the threshold value will occur in the future, the power consumption of the power consuming product may be automatically controlled to follow a battery remaining capacity pattern defined in a state where the battery remaining capacity is always above the minimum threshold value.

[0060] For example, as illustrated in FIG. 5, if a section (600) that deviates from the lowest threshold value is detected between 03:00 on May 19 and 06:00 on May 19, since the threshold deviation occurs when the air conditioner usage is 100% and 90%, the main server (100) can control the output of the corresponding power consumption product (air conditioner) to be 80%.

[0061] Referring to FIG. 6, when such control operations occur or monitoring by the main server (100) continues, a GUI (Graphic User Interface) as illustrated can be provided to a user terminal (300).

[0062] A GUI according to one embodiment of the present invention may include, as illustrated, buttons for power currently being consumed, current status, monitoring, restart, stop, and reboot, and a power button capable of shutting down each power-consuming product.

[0063] Additionally, threshold values ​​or operating modes may be provided, and if the operating mode is a recommended mode, a recommendation to control a specific power-consuming device (air conditioner) (turning it off) may be displayed as illustrated. In this case, the reason for the recommendation (risk of low battery) may also be displayed along with the recommendation.

[0064] Meanwhile, in the additional embodiment described above, the main server (100) can re-predict the amount of power generated based on the amount of solar radiation received from the power plant server (200) at preset time intervals, and update the load usage pattern according to the change in time at preset time intervals.

[0065] Through this, the main server (100) can re-predict the remaining battery capacity at preset time intervals, and if the re-predicted remaining battery capacity is higher than the remaining battery capacity predicted in the previous time interval, it can immediately reflect the load usage of a predetermined type and number of power consuming products at a higher level than in the previous time interval.

[0066] At this time, a screen as illustrated in FIGS. 7 and 8 may be provided to a user terminal (300). The main server (100) may provide the actual power generation amount received from the power plant server (200) and the load power usage directly predicted or measured as a graph as in FIG. 7 to the user terminal (300).

[0067] Likewise, the power generation amount predicted by the main server (100) and the predicted SOC (remaining battery capacity) can also be configured as a graph as in FIG. 8 and displayed on the user terminal (300).

[0068] Hereinafter, with reference to FIG. 9, an active control method for power equipment according to residual power capacity according to one embodiment of the present invention will be summarized once again.

[0069] A method for active control of power equipment according to a remaining power capacity according to one embodiment of the present invention begins by first transmitting the load usage amount (S100) from the load side to the main server (100) and transmitting the remaining battery capacity (101) from the battery side to the main server (100).

[0070] The main server (100) selects power consumption products (loads) that can adjust the load usage in stages (S102).

[0071] The main server (100) generates a total load usage pattern (S103) when all power-consuming products are used.

[0072] In addition, the main server (100) generates (S104) a load usage pattern for each power-consuming product, distinguishing between weekdays and weekends.

[0073] The main server (100) simulates the load usage pattern (S105) while adjusting the load usage of each selected power-consuming product, and predicts the remaining battery capacity by time period (S106) by considering the power generation amount and the load usage pattern.

[0074] Next, the main server (100) predicts the remaining battery capacity by time period (S107) by considering the power generation amount and load usage pattern.

[0075] Afterwards, the main server (100) sets a minimum threshold value for the remaining battery capacity and detects (S108) whether there is a remaining capacity pattern that exceeds the minimum threshold value.

[0076] The main server (100) automatically controls the power usage (S109) of the power-consuming product, i.e., the load (500), according to the result detected through the process described above.

[0077] Afterwards, the main server (100) provides control information, etc. to the user terminal (300) through a status notification UI (S110).

[0078] Meanwhile, as an additional embodiment, the present invention may optimize load control based on reinforcement learning.

[0079] In the corresponding embodiment, the main server (100) receives state data including battery remaining capacity, renewable energy generation amount prediction value, load usage prediction value, power unit price information, time information (day of the week, time zone, season, etc.), and power facility status information from the weather station server, power facility, and energy facility.

[0080] The main server determines an action to adjust the load level of power-consuming devices within the building based on the status data. Here, the action may include a charge / discharge control command for a power storage system (ESS).

[0081] In addition, the action decision performed by the main server (100) is performed by a reinforcement learning algorithm, wherein the reinforcement learning algorithm may perform learning by setting reward elements such as reducing power costs, minimizing user inconvenience, maintaining battery life, and improving the self-consumption rate of renewable energy, and setting constraints such as failing to meet the minimum threshold value of the remaining battery capacity, exceeding the power facility capacity, and exceeding the comfort standard.

[0082] Here, reinforcement learning algorithms that support continuous or mixed action spaces (such as PPO, TD3, DDPG, etc.) may be applied, and may further include LSTM (Long Short-Term Memory) or Transformer-based time series encoders to learn long-term time series characteristics within the state data.

[0083] In addition, the above reinforcement learning policy can be trained (offline training) using a digital twin model created using past Energy Management System (EMS) log data, and then periodically fine-tuned online in a real-time operating environment. For example, during digital twin-based online training, the main server (100) can perform fine-tuning by gradually reducing the search rate and prioritizing search in a state where the probability of violating safety constraints is low.

[0084] Meanwhile, the power generation prediction and load usage prediction in the embodiment may be updated at intervals of 5 or 15 minutes, and may be configured to immediately switch to rule-based control logic to ensure safety when abnormal events such as sudden load fluctuations or a decrease in power generation are detected.

[0085] As another additional embodiment, the present invention may further include a dynamic threshold adjustment algorithm.

[0086] That is, in one embodiment of the present invention, the main server (100) can dynamically adjust the minimum threshold value of the remaining battery capacity, rather than a fixed value, by considering changes in time, predicted power generation load, power unit price, weather conditions, operating goals, etc.

[0087] In the corresponding embodiment, the minimum threshold value is set based on chance-constrained optimization. That is, the main server (100) can be set such that, considering the uncertainty of the predicted net load, the probability that the remaining battery capacity will fall below the minimum threshold value during a future period of time is less than or equal to a predefined tolerance. In this case, the tolerance may be set prior to the implementation of the present invention or may be input from a user or administrator terminal.

[0088] The above-described minimum threshold calculation process can be performed in the following order. First, the main server (100) applies quantile forecasting to the predicted net load (i.e., total power consumption of loads within the building - power generation) to generate a conservative scenario that prevents underestimation of the prediction. For example, using the 90% quantile forecast value among all possible cases, a scenario is generated in which the power generation is set low and the load is set high, and the total power consumption and power generation of the loads corresponding to that scenario are determined.

[0089] Next, for the above scenario, the main server searches for a minimum threshold value that minimizes power costs, battery degradation, and user inconvenience while ensuring that the remaining battery capacity does not drop below a threshold. This search operation may be performed using a Genetic Algorithm or Bayesian Optimization. When using Bayesian Optimization, the objective function can be defined as the 'total cost, which is the sum of power purchase costs and degradation costs due to battery charging and discharging when the battery is operated at minimum SOC.' Subsequently, the objective function is evaluated at an initial sample point, the surrogate model is updated, and the next evaluation point (a candidate value for the minimum threshold) is selected using the acquisition function. This process is then repeated. When a predefined number of iterations is reached, the search is terminated, and the minimum threshold value that minimizes the objective function value is finally determined. The finally determined minimum threshold value is applied in real time, and a dynamic minimum threshold value can be generated by periodically re-executing the search operation in response to changes in weather conditions, fluctuations in electricity rates, etc.

[0090] The dynamically calculated minimum threshold value is applied as a safety constraint for reinforcement learning-based load control and is also reflected in the reward function of the reinforcement learning policy to prevent unnecessary fluctuations near the threshold value. Through this, the present invention can simultaneously achieve power cost reduction, battery protection, and operational stability.

[0091] One embodiment of the present invention may also be implemented in the form of a recording medium comprising computer-executable instructions, such as program modules executed by a computer. A computer-readable medium may be any available medium accessible by a computer and includes both volatile and non-volatile media, and both removable and non-removable media. Additionally, a computer-readable medium may include all computer storage media. A computer storage medium includes both volatile and non-volatile, removable and non-removable media implemented by any method or technique for storing information, such as computer-readable instructions, data structures, program modules, or other data.

[0092] Although the method and system of the present invention have been described in relation to specific embodiments, some or all of their components or operations may be implemented using a computer system having a general-purpose hardware architecture.

[0093] The foregoing description of the present invention is for illustrative purposes only, and those skilled in the art will understand that other specific forms can be easily modified without altering the technical spirit or essential features of the present invention. Therefore, the embodiments described above should be understood as illustrative in all respects and not restrictive. For example, each component described as a single unit may be implemented in a distributed manner, and components described as distributed may likewise be implemented in a combined form.

[0094] The scope of the present invention is defined by the claims set forth below rather than by the detailed description above, and all modifications or variations derived from the meaning and scope of the claims and equivalent concepts thereof should be interpreted as being included within the scope of the present invention. Explanation of the symbols

[0095] 100: Main Server 200: Power Plant Server 300: User terminal 400: Battery 500: Load 600: Section exceeding the minimum threshold

Claims

Claim 1 A method for active control of power facilities based on remaining power capacity, performed by a main server, comprising: (a) selecting power consuming products within a building that can adjust load usage in stages and generating an overall load usage pattern when all power consuming products are in use; (b) simulating a current or future overall load usage pattern while adjusting the load usage of each of the selected power consuming products; (c) predicting the remaining battery capacity of the energy generating device by time period, considering the power generation amount of the energy generating device over time and the overall load usage pattern; (d) setting either the minimum power value that the SoC must maintain for the operation of a plurality of power consuming products included in the building or the over-discharge upper limit value of the batteries included in the plurality of power consuming products as the minimum threshold value for the remaining battery capacity for the stable operation of the building, and then classifying the power consuming products into one of an ON / OFF control load capable of only turning on / off, an essential load that must operate regardless of user settings, and an energy usage control load capable of adjusting load usage in stages, and (e) A step of collecting output information and mode information of a power consumption product, and controlling the output while controlling the mode of the power consumption product to measure the current remaining battery capacity and predict the future remaining battery capacity by time period, wherein, considering the maximum power generation of the photovoltaic power generation device predicted based on solar irradiance, the battery facility capacity, and the load usage of the power consumption product, a plurality of battery remaining capacity patterns are derived according to the simulated plurality of total load usage patterns, and a step of detecting whether there exists a battery remaining capacity pattern among the battery remaining capacity patterns that temporarily deviates from the minimum threshold value; (e) A step of automatically controlling the power usage of the power consumption product so as to follow the battery remaining capacity pattern defined in a state where the battery remaining capacity is always above the minimum threshold value;and (f) a step of re-predicting the power generation amount based on solar irradiance at preset time intervals and updating the load usage pattern according to changes in time at preset time intervals, thereby re-predicting the battery remaining capacity at preset time intervals, and if the re-predicted battery remaining capacity is higher than the battery remaining capacity predicted in the previous time interval, increasing the load usage of a predetermined type and number of power consuming products compared to the previous time interval, wherein the total load usage pattern refers to load usage according to time, and the step (e) receives state data including the battery remaining capacity, the predicted value of renewable energy generation amount, the predicted value of load usage, power unit price information, time information, and power equipment status information, and determines an action for adjusting the load level of power consuming products within the building based on the state data by a reinforcement learning algorithm, wherein the action includes a charge / discharge control command of the power storage device, and the reinforcement learning algorithm sets power cost reduction, battery life maintenance, and improvement of renewable energy self-consumption rate as compensation factors, and the battery remaining capacity falls below the minimum threshold value, exceeds power equipment capacity and Active control method for power equipment based on remaining power capacity, wherein learning is performed by setting exceeding comfort standards as a constraint. Claim 2 In claim 1, the above step (a) comprises: (a-1) a step of subdividing the load usage for each power consumption product into multiple stages; (a-2) a step of extracting the maximum load usage and minimum load usage of the power consumption product from the load usage of the multiple stages; (a-3) a step of separately generating a load usage pattern for each power consumption product according to the usage stage for each power consumption product; (a-4) a step of distinguishing and generating load usage patterns for weekdays and weekends for each power consumption product; and (a-5) a step of generating a load usage pattern for each power consumption product for a specific time interval and summing them to generate an overall load usage pattern; wherein the step (a-1) is, if the power consumption product is an on / off control load, grouping a plurality of power consumption products into one group and subdividing the load usage according to the number of power consumption products that are on / off within the group, and if the power consumption product is an energy usage control load, subdividing the load usage according to the variation of its own set value, a method for active control of power equipment according to remaining power capacity. Claim 3 A method for active control of power equipment according to residual power capacity, wherein step (b) simulates multiple load usage patterns for each power consuming product according to each step among multiple steps regarding load usage, and generates an overall load usage pattern for multiple cases by summing the load usage patterns for each power consuming product according to various cases. Claim 4 delete Claim 5 delete Claim 6 A method for active control of power equipment according to residual power capacity, wherein the step of generating the battery remaining capacity pattern comprises: measuring the current battery remaining capacity; predicting the future battery remaining capacity based on the following regression equation; generating the battery remaining capacity pattern based on SoCt+1 = maximum power generation amount t + (battery facility capacity Х SoCt) - load usage t (where t is a natural number greater than or equal to 1) and the current and future battery remaining capacities. Claim 7 delete

Images