KR20230101526A - Apparatus and method for saving building energy based on artificial intelligence - Google Patents

Abstract

The present invention relates to an artificial intelligence-based building energy saving device and method. In the artificial intelligence-based building energy saving device according to an embodiment of the present invention, a communication module; a memory for storing at least one process for performing building energy saving; and a control module that performs an operation for performing the building energy saving based on the at least one process, wherein the control module includes environmental data for a target building and the target building based on a predetermined cycle. Operation data for at least one target facility is collected, and major indoor environmental factors are predicted through a pre-learned first machine learning model using environmental data and operation data accumulated and collected for a predetermined period of time, and the It is possible to generate control signals for each target facility according to the predicted major indoor environmental factors.

Description

AI-based building energy saving device and method {APPARATUS AND METHOD FOR SAVING BUILDING ENERGY BASED ON ARTIFICIAL INTELLIGENCE}

The present invention relates to an artificial intelligence-based building energy saving device and method, and more particularly, to an artificial intelligence-based building energy saving device and method for controlling target facilities to save building energy using environmental data and operation data. It is about.

Recently, various autonomous driving models that use appropriate driving strategies for each hour, season, and situation are being studied to reduce energy consumption, and a new baseline model for verifying the energy consumption reduction of these controlled facilities is needed.

According to the energy usage analysis methodology of IPMVP (Inter-national Performance Measurement & Verification Protocol, 2002) and ASHRAE Handbook Fundamentals (ASHRAE Handbook Fundamentals, 2013), M&V (Measurement and Verification) is mainly used as a data-based regression analysis model. and the multiple regression M&V model was mainly used to predict monthly energy use patterns.

However, since the control signals for each facility are determined in units of hours or minutes depending on the operation strategy, prediction of energy use patterns based on monthly data has limitations in M&V work for each operation strategy.

As an alternative to solving this problem, machine learning algorithm models can be used, and general machine learning algorithm-based models for predicting building energy in time units include KNN (K-nearest Neighbor), Random Forest (RF), and artificial neural networks. (Artificial Neural Network, ANN).

Through the development of such a machine learning algorithm-based model, even if a highly accurate hourly or minutely predicted model exists, it is necessary to predict the energy consumption of facilities that take different driving strategies for each situation according to the change in operating strategy. Since strategy-related information is not reflected in the model, it is difficult to analyze specific and accurate performance by driving strategy.

Therefore, it is necessary to develop a technology that can reduce building energy based on artificial intelligence by considering not only indoor and outdoor environmental data, but also operation strategies and control information for each facility.

Korean Registered Patent Publication No. 10-2085956 (registration date: March 02, 2020)

The present invention has been proposed to solve the above problems, and does not rely on the subjective judgment of the manager, but uses objective data such as indoor and outdoor environmental data and facility-related data by time, season, and situation for experience buildings. It is to provide an artificial intelligence-based building energy saving device and method that can improve economic efficiency through saving building energy by automatically changing or establishing an operating strategy for a building based on learning.

The problems to be solved by the present invention are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

An artificial intelligence-based building energy saving device according to an embodiment of the present invention for solving the above problems includes a communication module; a memory for storing at least one process for performing building energy saving; and a control module performing an operation for performing the building energy saving based on the at least one process, wherein the control module includes environmental data for a target building and the target building based on a predetermined cycle. Operation data for at least one target facility is collected, and major indoor environmental factors are predicted through a pre-learned first machine learning model using environmental data and operation data accumulated and collected for a predetermined period of time, and the It is possible to generate control signals for each target facility according to the predicted major indoor environmental factors.

On the other hand, the artificial intelligence-based building energy saving method according to an embodiment of the present invention collects environmental data for a target building and operation data for at least one target facility included in the target building based on a preset cycle, respectively. step; predicting key indoor environmental factors through a pre-learned first machine learning model using environmental data and driving data accumulated and collected for a predetermined period of time; generating control signals for each target facility according to the predicted major indoor environmental factors; calculating predicted power consumption through a pre-learned second machine learning model using the generated control signal for each target facility; and calculating a real-time energy saving amount by comparing actual power consumption after the generated control signal for each target facility is applied with the predicted power consumption.

Other specific details of the invention are included in the detailed description and drawings.

According to the present invention, without relying on the manager's subjective judgment, using objective data such as indoor/outdoor environmental data and facility-related data for each time, season, and situation for the experienced building, the operation strategy for the building is automatically set up based on machine learning. By changing or establishing, it is possible to improve economic efficiency through saving building energy.

The effects of the present invention are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a diagram showing the network structure of an artificial intelligence-based building energy saving system according to an embodiment of the present invention.
2 is a block diagram showing the configuration of an artificial intelligence-based building energy saving device according to an embodiment of the present invention.
3 is a diagram showing an example of a method of pre-processing environmental data in an artificial intelligence-based building energy saving device according to an embodiment of the present invention.
4 is a diagram showing the structure of a first machine learning model according to an embodiment of the present invention.
5 is a diagram showing the structure of a second machine learning model according to an embodiment of the present invention.
6 is a diagram showing the structure of a baseline model for saving building energy based on a first machine learning model and a second machine learning model according to an embodiment of the present invention.
7 is a graph showing energy saving efficiency after a control signal generated by a building energy saving device according to an embodiment of the present invention is applied to each target facility.
8 is a flowchart illustrating an operation for generating a control signal in a building energy saving method according to an embodiment of the present invention.
9 is a flowchart illustrating an operation for calculating an energy saving amount in a building energy saving method according to an embodiment of the present invention.

Advantages and features of the present invention, and methods of achieving them, will become clear with reference to the detailed description of the following embodiments taken in conjunction with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various different forms, only these embodiments are intended to complete the disclosure of the present invention, and are common in the art to which the present invention belongs. It is provided to fully inform the person skilled in the art of the scope of the invention, and the invention is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present invention. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements. Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present invention.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

The spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe a component's correlation with other components. Spatially relative terms should be understood as including different orientations of elements in use or operation in addition to the orientations shown in the drawings. For example, if you flip a component that is shown in a drawing, a component described as "below" or "beneath" another component will be placed "above" the other component. can Thus, the exemplary term “below” may include directions of both below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

The term "unit" or "module" used in the specification means a hardware component such as software, FPGA or ASIC, and "unit" or "module" performs certain roles. However, "unit" or "module" is not meant to be limited to software or hardware. A “unit” or “module” may be configured to reside in an addressable storage medium and may be configured to reproduce one or more processors. Thus, as an example, a “unit” or “module” may refer to components such as software components, object-oriented software components, class components, and task components, processes, functions, properties, procedures, subroutines, segments of program code, drivers, firmware, microcode, circuitry, data, databases, data structures, tables, arrays and variables. Functions provided within components and “units” or “modules” may be combined into smaller numbers of components and “units” or “modules” or may be combined into additional components and “units” or “modules”. can be further separated.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which the present invention belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

In general, building equipment systems are operated through automatic or optimal control, but in most cases, absolute authority is given to the subjective judgment of the manager (operator) for the final comprehensive decision on the operation of the building. At this time, the manager actively establishes an operation strategy based on his or her experience and intuition, but also passively controls the operation of equipment systems according to indoor and outdoor environmental conditions occurring in the building. However, it is difficult to see that such subjective operation is always accurately diagnosing and responding to various situations occurring in the building.

The present invention is to reduce building energy by automatically generating control signals based on indoor and outdoor environmental data using at least one machine learning model based on artificial intelligence and applying them to the operation of a target building.

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

Prior to explaining the present invention, the concept of 'facility' described in this specification includes air conditioning equipment, sanitary equipment, power equipment, lighting equipment, and disaster prevention equipment provided in the building, and each provided for maintenance and operation of the building. It is a concept that collectively refers to species facilities, and should not necessarily be interpreted as a concept limited to the facilities described in this specification.

1 is a diagram showing the network structure of an artificial intelligence-based building energy saving system according to an embodiment of the present invention.

Referring to FIG. 1 , a building energy saving system 10 according to an embodiment of the present invention may include a building energy saving device 100, at least one facility 200, and a manager terminal 300.

At least one facility 200 is a variety of facilities that provide energy (building energy) necessary for the building, and is located (installed) inside or outside the building, and is controlled by a control signal generated by the building energy saving device 100. do.

When a plurality (200-1 to 200-n) of the at least one facility 200 is provided, it may be classified into groups and managed according to the installation location, function, and manager in charge of each facility.

The building energy saving device 100 collects indoor/outdoor environmental data and operation data for at least one facility 200 for a preset period of time, and learns based on the accumulated environmental data and operation data for a preset period of time. Predict major indoor environmental factors through the first machine learning model. Thereafter, the building energy saving device 100 generates a control signal for each target facility according to the predicted major indoor environmental factors.

On the other hand, the building energy saving device 100 calculates the predicted power consumption through the pre-learned second machine learning model using the previously generated control signal for each target facility, and after the generated control signal for each target facility is applied, That is, the real-time energy saving amount is calculated by comparing the actual power consumption after operation with the control signal corresponding to each target facility with the previously calculated predicted power consumption.

In addition, the building energy saving device 100 transmits various data or notifications set (selected) by the manager as well as various collected data, generated control signals, and calculated energy savings to the previously registered manager terminal 300. You can report by doing this. At this time, there may be at least one manager terminal 300, and since data or notifications to be reported or received may be set differently for each manager, all information included in data or notifications transmitted to each manager terminal 300 is different. can do.

When various data or notifications are received from the building energy saving device 100, the manager terminal 300 displays them so that the manager can confirm them, and if necessary, when the manager inputs a control signal, the input control signal is transmitted and at least one The facility 200 of may be manually controlled. At this time, the manager terminal 300 may transmit the control signal input by the manager to the building energy saving device 100 to control the corresponding facility through the building energy saving device 100, or transmit it directly to the corresponding facility You can also control the equipment.

As described above, at least one manager terminal 300 may also be provided, and may have access to control as it is registered in the building energy saving device 100 and/or at least one facility 200. In addition, when a plurality of manager terminals 300 (300-1 to 300-n) are provided, they may be classified into groups and managed according to departments, positions, facilities in charge, setting authority, authority level, etc. of each manager.

On the other hand, the administrator terminal 300 is a computer, UMPC (Ultra Mobile PC), workstation, net-book, PDA (Personal Digital Assistants), portable computers, web tablets, wireless phones, mobile phones, smart phones, pads, smart watches, Wearable terminals, e-books, portable multimedia players (PMPs), portable game consoles, navigation devices, black boxes or digital cameras, other mobile communication terminals, etc. can

Meanwhile, FIG. 1 corresponds to an embodiment of the present invention, and at least one facility 200 may be controlled by a separate control system (not shown), in which case the building energy saving device 100 controls the control system. It can also act as part of a system.

2 is a block diagram showing the configuration of an artificial intelligence-based building energy saving device according to an embodiment of the present invention.

2, the artificial intelligence-based building energy saving device (hereinafter referred to as 'building energy saving device') 100 according to an embodiment of the present invention includes a communication module 110, a storage module 130 and a control module (150).

The communication module 110 performs wired or wireless communication with at least one external device (such as a server). In particular, in performing wireless communication, a wireless signal is transmitted and received in a communication network according to wireless Internet technologies.

Wireless Internet technologies include, for example, WLAN (Wireless LAN), Wi-Fi (Wireless-Fidelity), Wi-Fi (Wireless Fidelity) Direct, DLNA (Digital Living Network Alliance), WiBro (Wireless Broadband), WiMAX (World Interoperability for Microwave Access), HSDPA (High Speed Downlink Packet Access), HSUPA (High Speed Uplink Packet Access), LTE (Long Term Evolution), LTE-A (Long Term Evolution-Advanced), etc. 100) transmits and receives data according to at least one wireless Internet technology within a range including Internet technologies not listed above.

As for short range communication, Bluetooth™, RFID (Radio Frequency Identification), Infrared Data Association (IrDA), UWB (Ultra Wideband), ZigBee, NFC (Near Field Communication), Wi -Local communication may be supported using at least one of wireless-fidelity (Fi), Wi-Fi Direct, and wireless universal serial bus (USB) technologies. Such a local area wireless communication network (Wireless Area Networks) is between the building energy saving device 100 and at least one facility 200, between the building energy saving device 100 and the manager terminal 300, at least one facility 200 And wireless communication between the manager terminal 300 may be supported. At this time, the local area wireless communication network may be a local area wireless personal area network (Wireless Personal Area Networks).

The storage module 130 stores at least one data (information) and at least one process necessary for building energy saving, as well as various other data (information). For example, a program for saving building energy, at least one facility information, and at least one manager information may be stored as data. In addition, the storage module 130 stores various commands, algorithms, etc. for executing the building energy saving method.

To this end, the storage module 130 may include a memory, and the memory may be of a flash memory type, a hard disk type, a multimedia card micro type, or a card type. Memory (eg SD or XD memory, etc.), RAM (Random Access Memory, RAM), SRAM (Static Random Access Memory), ROM (Read-Only Memory, ROM), EEPROM (Electrically Erasable Programmable Read-Only Memory), It may include at least one type of storage medium among a programmable read-only memory (PROM) magnetic memory, a magnetic disk, and an optical disk. In addition, the memory may store information temporarily, permanently or semi-permanently, and may be provided in a built-in or removable type.

The control module 150 is for controlling components in the building energy saving device 100 or for processing and calculating various information, and controls to save building energy based on at least one process stored in the storage module 130.

Specifically, the control module 150 collects indoor and/or outdoor environment data and operation data for at least one facility 200, and collects environment data and operation data accumulated over a predetermined period of time based on the collected environment data and operation data. A major indoor environmental factor is predicted through the pre-learned first machine learning model.

Here, the environment data includes first environment data including indoor environment factors and second environment data including outdoor environment factors, and the first environment data and the second environment data may be individually collected through respective sensors. there is. For example, the first environmental data and the second environmental data include at least one of indoor temperature, outdoor temperature, humidity, relative humidity, CO2 concentration, noise, electromagnetic waves, asbestos, pressure, rainfall, light pollution, illuminance, and fine dust concentration. can include

In addition, the operation data includes operation and specification-related data for each facility, and for example, at least one of power consumption, On/Off status, operation mode status, air volume, set temperature, upper limit temperature and lower limit temperature for each operation mode. can include

The environment data and operation data are collected in units of preset time, and the collected environment data and operation data are accumulated and stored in the storage module 130 . That is, time-series data of a preset time unit is accumulated. Thereafter, environmental data and driving data accumulated and stored during a predetermined period among the stored environmental data and driving data may be used to predict major indoor environmental factors.

On the other hand, the control module 150 performs preprocessing to predict key indoor environmental factors predicted using environmental data and driving data, and then converts the preprocessed data to the pre-learned first machine learning model as input data. input to generate control signals for each target facility.

In addition, the control module 150 calculates the predicted power consumption through the pre-learned second machine learning model using the previously generated control signal for each target facility, and after the generated control signal for each target facility is applied, that is, Real-time energy saving is calculated by comparing the actual power consumption after operation with the control signal corresponding to each target facility with the predicted power consumption previously calculated.

In addition, the control module 150 transmits not only collected various data, generated control signals and calculated energy savings, but also various other data or notifications set (selected) by the manager to the previously registered manager terminal 300. report by doing

3 is a diagram showing an example of a method of pre-processing environmental data in an artificial intelligence-based building energy saving device according to an embodiment of the present invention.

The building energy saving device 100 may use at least one of an Inter Quantile Range (IQR) outlier removal method and a MinMaxScaling method to perform preprocessing on environmental data. IQR outlier elimination is a method of dividing the numerical range of data into quartiles to remove data outside a specific range, and MinMaxScaling, as shown in FIG. 3, is normalization that facilitates model learning by rescaling the range of data. (Normalization) method.

In other words, based on the minimum and maximum values of the data specified in the specifications of the sensor that detects the environmental data, data outside the range is substituted with 0, and the input data is adjusted using the Min-Max Scaler used for calibration learning with linear interpolation. normalize it.

Thus, the environmental data and driving data can be used as one input data input to the first machine learning model.

4 is a diagram showing the structure of a first machine learning model according to an embodiment of the present invention.

Referring to FIG. 4 , an Indoor Environment Prediction Model (IEPM) may be applied as a first machine learning model for predicting major indoor environmental factors.

This IEPM is designed to use the first environment data, the second environment data, and driving data as input data. In the case of future indoor temperature and humidity to be predicted, the outdoor temperature and humidity pattern accumulated up to the previous point in time, and the indoor temperature and humidity pattern and the control patterns of the air conditioning facilities may be comprehensively influenced. Therefore, there is a need to process and utilize time series data accumulated over a certain period of time as one input data.

Accordingly, the input data set of IEPM accumulates and utilizes 1-minute unit data for a certain period of time (Window Size) for all types of data. In addition, the IEPM for each major indoor environment factor can be designed with a 1D Convolution Network and GRU, which have strengths in learning time series patterns, as the basic components of the model, and the hyperparameters of each IEPM are model evaluation. The results can be optimized.

IEPM learns the non-linear correlation between operating data for each facility and environmental data including indoor and outdoor environmental factors based on learning data operated with a specific driving strategy to predict the value of key indoor environmental factors at the next point in time. designed for the purpose

Referring to FIG. 4 , during data preprocessing for learning the IEPM, data outside the corresponding range is removed based on the minimum and maximum values of data specified in the sensor specification. At this time, based on the inter quantile range (IQR), for example, data out of the standard Q1 ± 1.5 * IQR range is removed, and the data of the removed time step is corrected by linear interpolation. Then, based on the minimum and maximum values of the data specified in the sensor specifications, the input data is normalized using the Min-Max Scaler. At this time, time-series data of all sensor data is accumulated and used as one input data. For example, when using 1-hour time-series data, if the collection period is set to 1 minute, a total of 60 data is obtained.

On the other hand, in the data input layer, indoor/outdoor environment data (in/outdoor temp., RH etc.) and operation data (on/off, cooling set point etc.) are collected, concatenated, and input to the predictive neural network.

This first machine learning model is created by learning about changes in major indoor environmental factors according to cross-influence relationships between target facilities and comprehensive control situations (states) of control target facility groups. In this case, the hidden layer depth is designed to be 3 or more, and at least one algorithm of Conv1D, Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) may be applied to reflect the time-series pattern.

5 is a diagram showing the structure of a second machine learning model according to an embodiment of the present invention.

Referring to FIG. 5 , an Electricity Load Prediction Model (ELPM) may be applied as a second machine learning model for predicting a power load for each facility of a target building.

Unlike the IEPM, this ELPM uses only the main indoor environmental factors that belong to the prediction target of the IEPM as environmental data, excluding outdoor temperature and humidity data, which are outdoor environmental data. This is because the operating conditions and control strategies of air conditioning equipment depend mostly on the indoor environment. In addition, unlike the IEPM that utilizes the operation and specification data of all facilities as input, the ELPM for each facility can be designed as a model focused on predicting the load of the facility itself by using the operation and specification data of the facility as input data. Meanwhile, the ELPM for each facility can be designed based on the LGBM, and the hyperparameters of each LGBM can be optimized through model evaluation results.

Referring to FIG. 5 , during data pre-processing for learning the ELPM, data outside the corresponding range is removed based on the minimum and maximum values of data specified in the sensor specifications. At this time, based on the inter quantile range (IQR), for example, data out of the standard Q1 ± 1.5 * IQR range is removed, and the data of the removed time step is corrected by linear interpolation. Then, based on the minimum and maximum values of the data specified in the sensor specifications, the input data is normalized using the Min-Max Scaler.

On the other hand, key indoor environmental factors (indoor temp., RH, CO ₂ etc.) data and operation data (on/off, Cooling set point etc.) are input in the data input layer, and environmental data (in the data combination layer for each facility) Common data) and operation data (operation data) of each facility are combined and input.

This second machine learning model is generated by learning the correlation between environmental data and driving data and energy consumption. At this time, the depth of the hidden layers (L1 to L3) is designed to be 3 or more, and at least one algorithm of Conv1D, Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) may be applied to the neural network series for each layer.

Accordingly, energy consumption power (or amount of power) for each facility is output through the output layer based on indoor and outdoor environment data and operation data.

6 is a diagram showing the structure of a baseline model for saving building energy based on a first machine learning model and a second machine learning model according to an embodiment of the present invention.

The present invention has a two-step model structure in which the IEPM (first machine learning model) and the ELPM (second machine learning model) are linked, and the indoor environment factor prediction model in which time-series pattern learning is important is a CNN and/or RNN-based network. can be designed as a model basic component.

Referring to FIG. 6, the IEPM receives operational data for each facility and environmental data including indoor and outdoor environmental factors, predicts major indoor environmental factors of the next time step, and conditions the existing control pattern (strategy) of the target building. By using as , the operation data of the next time step according to the existing control pattern is derived. That is, a control signal for each facility to be applied is generated. In addition, the ELPM predicts the power load for each facility by receiving the predicted indoor environment factor and the derived operation data (control signal) of the next time step.

7 is a graph showing energy saving efficiency after a control signal generated by a building energy saving device according to an embodiment of the present invention is applied to each target facility.

7 is an example of a graph showing the power load for each facility predicted through the procedure of FIG. 6, through which energy savings can be confirmed.

8 is a flowchart illustrating an operation for generating a control signal in a building energy saving method according to an embodiment of the present invention.

Referring to FIG. 8, the building energy saving device 100 collects environmental data for a target building based on a preset cycle (S301), and operates at least one target facility included in (related to) the target building. Data is collected (S303). In this case, although step S301 is illustrated as being performed before step S303 in FIG. 8 , step S303 may be performed first or steps S301 and S303 may be performed simultaneously, and the order in which they are performed is not limited.

Next, the building energy saving device 100 generates one input data by performing pre-processing on the environmental data and driving data accumulated and collected for a predetermined time (S305), and converts the generated input data into a first Input to the machine learning model to predict major indoor environmental factors (S307).

Thereafter, the building energy saving device 100 generates a control signal for each target facility according to the major indoor environmental factors predicted in step S307 (S309).

9 is a flowchart illustrating an operation for calculating an energy saving amount in a building energy saving method according to an embodiment of the present invention.

Referring to FIG. 9, after the control signal for each target facility is generated, the building energy saving device 100 calculates predicted power consumption through a second machine learning model using the control signal for each target facility (S311), After the control signal for each target facility is applied, the actual power consumption is checked (S313).

Next, the building energy saving device 100 compares the predicted power consumption calculated in step S311 and the actual power consumption confirmed in step S313 to calculate real-time energy savings (S315).

On the other hand, although not shown in FIGS. 8 and 9, the building energy saving device 100 transmits the information on the control signal generated in step S309 and/or the information on the amount of real-time energy savings calculated in step S315 according to a preset manager. It can be transmitted to the terminal 300.

As described above, the present invention can calculate more accurate and detailed data by upscaling measurement data of a specific area using other measurement data that includes at least a part of the specific area or is measured in an adjacent area.

Meanwhile, steps of a method or algorithm described in relation to an embodiment of the present invention may be directly implemented as hardware, implemented as a software module executed by hardware, or implemented by a combination thereof. A software module may include random access memory (RAM), read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), flash memory, hard disk, removable disk, CD-ROM, or It may reside in any form of computer readable recording medium well known in the art to which the present invention pertains.

Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art to which the present invention pertains can be implemented in other specific forms without changing the technical spirit or essential features of the present invention. you will be able to understand Therefore, it should be understood that the embodiments described above are illustrative in all respects and not restrictive.

10: building energy saving system 100: building energy saving device
200: at least one facility 300: manager terminal
110: communication module 130: storage module
150: control module

Claims (10)

communication module;
a memory for storing at least one process for performing building energy saving; and
A control module that performs an operation for performing the building energy saving based on the at least one process,
The control module,
Based on a predetermined period, environmental data for a target building and operation data for at least one target facility included in the target building are collected, respectively, and the environment data and operation data accumulated and collected during a predetermined period are used to Characterized in that predicting major indoor environmental factors through the learned first machine learning model and generating control signals for each target facility according to the predicted major indoor environmental factors,
AI-based building energy saving device.
According to claim 1,
The control module,
The predicted power consumption is calculated through the previously learned second machine learning model using the generated control signal for each target facility, and the actual power consumption after the generated control signal for each target facility is applied is compared with the predicted power consumption. Characterized in calculating real-time energy savings,
AI-based building energy saving device.
According to claim 1,
The control module,
When collecting the above environmental data,
Characterized in that first environmental data including indoor environmental factors and second environmental data including outdoor environmental factors are respectively collected.
AI-based building energy saving device.
According to claim 1,
The control module,
Characterized in that, in order to predict major indoor environmental factors through the pre-learned first machine learning model, preprocessing is performed on the accumulated and collected environmental data and driving data to be processed into one input data.
AI-based building energy saving device.
According to claim 4,
The pretreatment,
Characterized in that the cumulatively collected environmental data and operation data are substituted with 0 for data outside the preset range, corrected by linear interpolation, and data normalization is performed for the corrected environmental data,
AI-based building energy saving device.
According to claim 1,
The pre-learned first machine learning model,
It was created by learning about changes in major indoor environmental factors according to the cross-influence relationship between the target facilities and the control situation of the control target facility group,
Characterized in that the hidden layer depth is designed to be 3 or more, and at least one algorithm of Conv1D, Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) is applied to reflect the time-series pattern,
AI-based building energy saving device.
According to claim 2,
The pre-learned second machine learning model,
It is generated by learning about the correlation between the environmental data and the driving data and energy consumption,
Characterized in that the hidden layer depth is designed to be 3 or more, and the layer-by-layer neural network series is applied with at least one algorithm of Conv1D, LSTM (Long Short-Term Memory), and GRU (Gated Recurrent Unit),
AI-based building energy saving device.
In the artificial intelligence-based building energy saving method performed by the device,
Collecting environmental data for a target building and operation data for at least one target facility included in the target building, respectively, based on a predetermined cycle;
predicting key indoor environmental factors through a pre-learned first machine learning model using environmental data and driving data accumulated and collected for a predetermined period of time;
generating control signals for each target facility according to the predicted major indoor environmental factors;
calculating predicted power consumption through a pre-learned second machine learning model using the generated control signal for each target facility; and
Calculating real-time energy savings by comparing actual power consumption after the generated control signal for each target facility is applied with the predicted power consumption,
AI-based building energy saving method.
According to claim 8,
When predicting major indoor environmental factors,
In order to predict major indoor environmental factors through the pre-learned first machine learning model, preprocessing is performed on the accumulated and collected environmental data and driving data to process them into one input data;
The pretreatment,
Characterized in that the cumulatively collected environmental data and operation data are substituted with 0 for data outside the preset range, corrected by linear interpolation, and data normalization is performed for the corrected environmental data,
AI-based building energy saving method.
Combined with a computer, a computer program stored in a computer readable recording medium for executing the artificial intelligence-based building energy saving method according to any one of claims 8 to 9.

Images