Green building digital management method and system based on artificial intelligence

Abstract

The application discloses a kind of green building digital management method and system based on artificial intelligence.It first arranges the time series of the collected building environment data according to time and sample dimension, then inputs the obtained temperature time series vector and light intensity time series vector into a nonlinear response correlation coding module to obtain a temperature-light intensity time series correlation interaction vector, then extracts the local time series correlation pattern features of the temperature-light intensity time series correlation interaction vector to obtain a sequence of environment data local time series pattern feature vectors, then performs building energy consumption reasoning based on environment local time series pattern prior information on the sequence of environment data local time series pattern feature vectors to obtain a reasoning building energy consumption time series representation feature, and finally determines whether the air conditioner power at the current time point should be increased, decreased or unchanged based on the reasoning building energy consumption time series representation feature.

Description

Technical Field

[0001] This disclosure relates to the field of digital management, specifically to a digital management method and system for green buildings based on artificial intelligence. Background Technology

[0002] Green buildings are buildings that incorporate energy conservation, environmental protection, and sustainable development during their design, construction, and operation. In the construction industry, green buildings are becoming increasingly common because they reduce energy consumption, improve indoor air quality, and enhance occupant comfort. Effective management of building energy consumption is essential to achieving the goals of green buildings.

[0003] However, traditional green building management solutions typically rely on manual intervention and operation, resulting in high management costs and low efficiency. Furthermore, manual intervention is susceptible to subjective human factors, making it difficult to achieve precise control and optimization of building energy systems. In addition, traditional green building management methods often depend on static planning and design concepts, lacking flexibility and intelligence to adapt to dynamic environmental changes. In other words, traditional green building management solutions lack intelligence and adaptability, failing to intelligently adjust and optimize based on real-time environmental data and changing demands, leading to low building energy efficiency and an inability to effectively reduce energy waste and environmental impact.

[0004] Therefore, a digital management solution for green buildings based on artificial intelligence is desired. Summary of the Invention

[0005] This disclosure was made in consideration of the above issues. One object of this disclosure is to provide a digital management method and system for green buildings based on artificial intelligence.

[0006] Embodiments of this disclosure provide an artificial intelligence-based digital management method for green buildings, comprising:

[0007] Acquire time series of building environment data collected by a sensor array deployed on the monitored building, wherein the building environment data includes temperature and light intensity;

[0008] The time series of the building environment data is arranged according to time and sample dimensions to obtain the temperature time series input vector and the light intensity time series input vector;

[0009] The temperature time-series input vector and the light intensity time-series input vector are passed through a nonlinear response correlation coding module to obtain a temperature-light intensity time-series correlation interaction vector.

[0010] Local temporal correlation pattern feature extraction is performed on the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors of environmental data;

[0011] The sequence of local temporal pattern feature vectors of the environmental data is used to perform building energy consumption inference based on prior information of the local temporal pattern of the environment to obtain the temporal representation features of inferred building energy consumption; and

[0012] Based on the time-series characteristics of the inferred building energy consumption, it is determined whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

[0013] For example, according to an embodiment of the present disclosure, the AI-based digital management method for green buildings includes obtaining a temperature-light intensity time-series correlation interaction vector by passing the temperature time-series input vector and the light intensity time-series input vector through a nonlinear response correlation encoding module.

[0014] Calculate the positional response between the time-series input vector of light intensity and the time-series input vector of temperature to obtain the temperature-light intensity positional response feature vector;

[0015] Construct the first nonlinear factor at each position in the location-by-location response feature vector of the temperature-light intensity;

[0016] Construct a second nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector; and

[0017] The division between the first nonlinear factor and the second nonlinear factor is calculated to obtain the temperature-light intensity position-by-position nonlinear response correlation correction value, and the multiple temperature-light intensity position-by-position nonlinear response correlation correction values ​​are vectorized and normalized to obtain the temperature-light intensity time-series correlation interaction vector.

[0018] For example, according to an embodiment of the present disclosure, the method for digital management of green buildings based on artificial intelligence includes constructing a first nonlinear factor for each location in the temperature-light intensity location-by-location response feature vector, comprising:

[0019] The square of each location feature value in the temperature-light intensity location-by-location response feature vector is calculated and multiplied by the first predetermined adjustment hyperparameter to obtain the first nonlinear factor quadratic term adjustment parameter;

[0020] Calculate the product of the positional eigenvalues ​​in the temperature-light intensity positional response feature vector and the second predetermined adjustment hyperparameter to obtain the first nonlinear factor linear term adjustment parameter; and

[0021] The first nonlinear factor is obtained by summing the adjustment parameters of the quadratic term and the linear term of the first nonlinear factor.

[0022] For example, according to an embodiment of the present disclosure, the method for digital management of green buildings based on artificial intelligence includes constructing a second nonlinear factor for each location in the temperature-light intensity location-by-location response feature vector, comprising:

[0023] The square of each location feature value in the temperature-light intensity position-by-position response feature vector is calculated and multiplied by the third predetermined adjustment hyperparameter to obtain the second nonlinear factor quadratic term adjustment parameter;

[0024] Calculate the product of the eigenvalues ​​at each location in the temperature-light intensity position-by-location response feature vector and the fourth predetermined adjustment hyperparameter to obtain the adjustment parameter of the first-order term of the second nonlinear factor; and

[0025] The second nonlinear factor is obtained by summing the adjustment parameters of the first nonlinear factor's quadratic term and the first nonlinear factor's linear term, and then modulating the constant term by adding one.

[0026] For example, according to an embodiment of the present disclosure, the method for digital management of green buildings based on artificial intelligence includes extracting local temporal correlation pattern features from the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors for environmental data, comprising:

[0027] The temperature-light intensity temporal correlation interaction vector is divided according to a predetermined time scale based on temperature to obtain a sequence of local temperature-light intensity temporal correlation interaction vectors; and

[0028] The sequence of temperature-light intensity local temporal correlation interaction vectors is processed by an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model to obtain the sequence of environmental data local temporal pattern feature vectors.

[0029] For example, according to an embodiment of the present disclosure, a digital management method for green buildings based on artificial intelligence includes performing building energy consumption inference based on prior information of the local temporal pattern feature vectors of the environmental data to obtain inferred building energy consumption temporal representation features, including:

[0030] The sequence of local temporal pattern feature vectors of the environmental data is passed through a building energy consumption inference engine based on environmental temporal prior information to obtain the inference building energy consumption temporal representation feature vector, which is used as the inference building energy consumption temporal representation feature.

[0031] For example, according to an embodiment of the present disclosure, a digital management method for green buildings based on artificial intelligence includes a method for obtaining a time-series representation feature vector of inferred building energy consumption by passing a sequence of local temporal pattern feature vectors of environmental data through a building energy consumption inference engine based on prior environmental temporal information. This inference is used as the time-series representation feature of inferred building energy consumption. The method includes:

[0032] The mean feature vector of multiple local temporal pattern feature vectors of environmental data in the sequence of local temporal pattern feature vectors of environmental data is calculated as the positional difference between the local temporal pattern feature vector of environmental data at the last position in the sequence of local temporal pattern feature vectors of environmental data to obtain the average distribution information difference feature vector;

[0033] The positional difference between the local temporal pattern feature vector of the environmental data at a random position in the sequence of the local temporal pattern feature vectors of the environmental data and the local temporal pattern feature vector of the environmental data at the last position in the sequence of the local temporal pattern feature vectors of the environmental data is calculated to obtain the random distribution information difference feature vector.

[0034] The average distribution information difference feature vector is multiplied by the first predetermined weight hyperparameter and the first Gaussian distribution random number function value to obtain the first building energy consumption inference feature vector;

[0035] The random distribution information difference feature vector is multiplied by the second predetermined weight hyperparameter and the second Gaussian distribution random number function value to obtain the second building energy consumption inference feature vector;

[0036] Calculate the positional addition between the first building energy consumption inference feature vector and the second building energy consumption inference feature vector to obtain the inference building energy consumption time series characterization feature vector;

[0037] The first Gaussian random number function value and the second Gaussian random number function value are generated by a Gaussian random number function with a mean of 0 and a variance of 1.

[0038] For example, according to the AI-based digital management method for green buildings according to embodiments of this disclosure, determining whether the air conditioning power should increase, decrease, or remain unchanged at the current time point based on the inferred building energy consumption time-series characterization features includes:

[0039] The time-series characteristic vector of the inferred building energy consumption is passed through a classifier-based air conditioning controller to obtain the control result, which is used to indicate whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

[0040] For example, according to an embodiment of the present disclosure, the AI-based digital management method for green buildings further includes a training step: training the nonlinear response association encoding module, the environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the air conditioning controller based on a classifier.

[0041] The training steps include:

[0042] Acquire training data, which includes a time series of training building environment data collected by a sensor array deployed on the monitored building, wherein the training building environment data includes training temperature and training light intensity;

[0043] The time series of the training building environment data is arranged according to time and sample dimensions to obtain the training temperature time series input vector and the training light intensity time series input vector;

[0044] The training temperature time-series input vector and the training light intensity time-series input vector are passed through the nonlinear response correlation coding module to obtain the training temperature-light intensity time-series correlation interaction vector.

[0045] The training temperature-light intensity temporal correlation interaction vector is divided according to a predetermined time scale based on temperature to obtain a sequence of local temporal correlation interaction vectors of training temperature-light intensity.

[0046] The sequence of the training temperature-light intensity local temporal correlation interaction vector is passed through the environmental data temporal pattern feature extractor based on the one-dimensional convolutional neural network model to obtain the sequence of training environmental data local temporal pattern feature vectors.

[0047] The sequence of local temporal pattern feature vectors of the training environment data is passed through the building energy consumption inference engine based on environmental temporal prior information to obtain the training inference building energy consumption temporal representation feature vector.

[0048] The time-series feature vector representing the trained inference building energy consumption is passed through the classifier-based air conditioning controller to obtain a classification loss function value; and

[0049] The nonlinear response association coding module, the environmental data temporal pattern feature extractor based on the one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the air conditioning controller based on the classifier are trained based on the classification loss function value. In each iteration of the training, the trained inference building energy consumption temporal representation feature vector is corrected.

[0050] Embodiments of this disclosure also provide an artificial intelligence-based digital management system for green buildings, comprising:

[0051] An environmental data acquisition module is used to acquire time series of building environment data collected by a sensor array deployed on the monitored building, wherein the building environment data includes temperature and light intensity;

[0052] The vectorization module is used to arrange the time series of the building environment data according to time and sample dimensions to obtain the temperature time series input vector and the light intensity time series input vector;

[0053] The correlation interaction module is used to pass the temperature time-series input vector and the light intensity time-series input vector through the nonlinear response correlation encoding module to obtain the temperature-light intensity time-series correlation interaction vector.

[0054] The local correlation feature extraction module is used to extract local temporal correlation pattern features from the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors of environmental data;

[0055] The energy consumption inference module is used to perform building energy consumption inference based on prior information of the local temporal pattern feature vectors of the environmental data to obtain the inferred building energy consumption temporal representation features; and

[0056] The power control module is used to determine whether the air conditioning power should increase, decrease, or remain unchanged at the current time point based on the time-series characteristics of the inferred building energy consumption.

[0057] The artificial intelligence-based digital management method and system for green buildings according to embodiments of this disclosure collects and monitors building environment data in real time, including temperature and light intensity, using a sensor array. In the backend, it utilizes data processing and analysis algorithms based on artificial intelligence and deep learning to perform temporal collaborative correlation analysis of this building environment data. This allows for the capture and characterization of building energy consumption time-series patterns, thereby predicting future building energy consumption and developing reasonable energy-saving plans to optimize energy utilization efficiency. This enables intelligent management of building energy consumption, thereby improving building energy utilization efficiency and reducing energy waste. Attached Figure Description

[0058] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Obviously, the drawings described below only relate to some embodiments of this disclosure and are not intended to limit this disclosure.

[0059] Figure 1 A schematic diagram of the application architecture of the AI-based digital management method for green buildings is shown in an embodiment of this disclosure.

[0060] Figure 2 A flowchart of an artificial intelligence-based digital management method for green buildings, as illustrated in an embodiment of this disclosure, is shown.

[0061] Figure 3 A flowchart of sub-step S530 of the AI-based digital management method for green buildings in this embodiment is shown.

[0062] Figure 4 A flowchart of sub-step S540 of the AI-based digital management method for green buildings in this embodiment is shown.

[0063] Figure 5 A schematic diagram of the structure of an AI-based digital management system for green buildings is shown in an embodiment of this disclosure;

[0064] Figure 6 This disclosure illustrates an application scenario diagram of an AI-based digital management method for green buildings; and

[0065] Figure 7 A schematic diagram of a storage medium according to an embodiment of the present disclosure is shown. Detailed Implementation

[0066] The technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, and not all embodiments. Based on the embodiments of this disclosure, all other embodiments obtained by those skilled in the art without creative effort are also within the scope of protection of this disclosure.

[0067] The terminology used in this specification is that which is currently widely used in the art in consideration of the functionality of this disclosure; however, these terms may vary depending on the intent, precedent, or new technology of those skilled in the art. Furthermore, specific terms may be chosen by the applicant, and in such cases, their detailed meanings will be described in the detailed description of this disclosure. Therefore, the terminology used in this specification should not be construed as simple names, but rather based on the meaning of the terms and the overall description of this disclosure.

[0068] While this disclosure makes various references to certain modules of systems according to embodiments of this disclosure, any number of different modules may be used and run on user terminals and / or servers. The modules described are merely illustrative, and different aspects of the systems and methods may use different modules.

[0069] This disclosure uses flowcharts to illustrate the operations performed by a system according to embodiments of this disclosure. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, various steps can be processed in reverse order or simultaneously, as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0070] Figure 1 The diagram illustrates the application architecture of an AI-based digital management method for green buildings in this embodiment, including a server 100 and a terminal device 200.

[0071] Terminal device 200 and server 100 can be connected via the Internet to enable communication between them. Optionally, the Internet described above uses standard communication technologies and / or protocols. The Internet is typically the Internet, but can also be any network, including but not limited to any combination of Local Area Network (LAN), Metropolitan Area Network (MAN), Wide Area Network (WAN), mobile, wired or wireless networks, private networks or virtual private networks. In some embodiments, technologies and / or formats including Hyper Text Markup Language (HTML), Extensible Markup Language (XML), etc., are used to represent data exchanged over the network. Furthermore, conventional encryption technologies such as Secure Socket Layer (SSL), Transport Layer Security (TLS), Virtual Private Network (VPN), and Internet Protocol Security (IPsec) can be used to encrypt all or some links. In other embodiments, customized and / or dedicated data communication technologies can be used to replace or supplement the aforementioned data communication technologies.

[0072] Server 100 can provide various network services to terminal device 200. Server 100 can be a single server, a server cluster consisting of several servers, or a cloud computing center. Specifically, server 100 may include processor 110 (Center Processing Unit, CPU), memory 120, input device 130, and output device 140, etc. Input device 130 may include keyboard, mouse, touch screen, etc., and output device 140 may include display device, such as liquid crystal display (LCD), cathode ray tube (CRT), etc.

[0073] The memory 120 may include a read-only memory (ROM) and a random access memory (RAM), and provides the processor 110 with program instructions and data stored in the memory 120. In this embodiment of the present disclosure, the memory 120 may be used to store the program of the artificial intelligence-based digital management method for green buildings in this embodiment of the present disclosure.

[0074] The processor 110 executes the steps of any of the artificial intelligence-based digital management methods for green buildings in this disclosure according to the program instructions stored in the memory 120.

[0075] Furthermore, the application architecture diagrams in this disclosure are for the purpose of more clearly illustrating the technical solutions in this disclosure and do not constitute a limitation on the technical solutions provided in this disclosure. Of course, the technical solutions provided in this disclosure are also applicable to similar problems for other application architectures and business applications.

[0076] The following examples or embodiments illustrate, in a non-limiting manner, the artificial intelligence-based digital management method for green buildings provided according to at least one embodiment of the present disclosure. As described below, different features in these specific examples or embodiments can be combined with each other without conflict, thereby obtaining new examples or embodiments, which are also within the scope of protection of the present disclosure.

[0077] To address the aforementioned technical issues, this application proposes a digital management method for green buildings based on artificial intelligence. This method can monitor the building's environmental data in real time by installing sensor arrays in the building and analyze the collected data using artificial intelligence technology to identify energy consumption patterns and potential energy-saving opportunities.

[0078] Specifically, the technical concept of this application involves real-time monitoring and collection of building environment data, including temperature and light intensity, using a sensor array. At the backend, data processing and analysis algorithms based on artificial intelligence and deep learning are employed to perform temporal collaborative correlation analysis of this building environment data. This allows for the capture and characterization of building energy consumption time-series patterns, thereby predicting future building energy consumption and developing reasonable energy-saving plans. For example, automatically adjusting the building's air conditioning power to optimize energy utilization efficiency. This enables intelligent management of building energy consumption, thereby improving building energy utilization efficiency and reducing energy waste.

[0079] Figure 2 A flowchart of an AI-based digital management method for green buildings, as illustrated in an embodiment of this disclosure, is provided. For example, this AI-based digital management method for green buildings can be executed by a server, which may be... Figure 1 Server 100 is shown in the image. (As shown in the image...) Figure 2 As shown, the AI-based digital management method for green buildings according to an embodiment of this disclosure includes the following steps: S510, acquiring a time series of building environment data collected by a sensor group deployed on the monitored building, wherein the building environment data includes temperature and light intensity; S520, arranging the time series of the building environment data according to time and sample dimensions to obtain a temperature time series input vector and a light intensity time series input vector; S530, passing the temperature time series input vector and the light intensity time series input vector through a nonlinear response association coding module to obtain a temperature-light intensity time series association interaction vector; S540, extracting local time series association pattern features from the temperature-light intensity time series association interaction vector to obtain a sequence of local time series pattern feature vectors of environmental data; S550, performing building energy consumption inference based on prior information of local time series patterns of the environment on the sequence of local time series pattern feature vectors of environmental data to obtain inferred building energy consumption time series characterization features; and S560, determining, based on the inferred building energy consumption time series characterization features, whether the air conditioning power at the current time point should increase, decrease, or remain unchanged.

[0080] More specifically, in the technical solution of this application, firstly, the time series of building environment data collected by a sensor array deployed on the monitored building is obtained, wherein the building environment data includes temperature and light intensity. Next, considering that the building environment data includes temperature data and light intensity data, and that temperature and light intensity change continuously over time, meaning that both temperature and light intensity exhibit dynamic temporal variations, the technical solution of this application requires arranging the time series of the building environment data according to time and sample dimensions to obtain a temperature time series input vector and a light intensity time series input vector. By arranging the temperature and light intensity in the building environment data according to the time dimension, the vectors of these building environment time series data can retain their temporal order, facilitating subsequent analysis and modeling of the temporal correlation of the building environment data.

[0081] Then, considering that changes in light intensity directly affect the temperature of the building environment, especially under sufficient sunlight, light will heat the interior space of the building. Furthermore, considering that the relationship between light intensity and temperature is usually non-linear, simple linear models are insufficient to accurately capture their complex temporal interactions. Therefore, in order to better understand the mechanism by which light affects temperature and provide important reference for building energy management and regulation, the technical solution of this application requires that the temperature temporal input vector and the light intensity temporal input vector be processed by a non-linear response correlation encoding module to obtain a temperature-light intensity temporal correlation interaction vector. Through the processing of the non-linear response correlation encoding module, the complex non-linear relationship between light intensity and temperature can be learned and captured, thereby better understanding their temporal interactions and providing a basis for subsequent building energy consumption reasoning.

[0082] Accordingly, such as Figure 3 As shown, in step S530, the temperature time-series input vector and the light intensity time-series input vector are processed by a nonlinear response correlation encoding module to obtain a temperature-light intensity time-series correlation interaction vector, including: S531, calculating the positional response between the light intensity time-series input vector and the temperature time-series input vector to obtain a temperature-light intensity positional response feature vector; S532, constructing a first nonlinear factor at each position in the temperature-light intensity positional response feature vector; S533, constructing a second nonlinear factor at each position in the temperature-light intensity positional response feature vector; and S534, calculating the division between the first nonlinear factor and the second nonlinear factor to obtain a temperature-light intensity positional nonlinear response correlation correction value, and vectorizing and normalizing the multiple temperature-light intensity positional nonlinear response correlation correction values ​​to obtain the temperature-light intensity time-series correlation interaction vector.

[0083] In step S532, constructing the first nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector includes: calculating the product of the square of the eigenvalue at each position in the temperature-light intensity position-by-position response feature vector and a first predetermined adjustment hyperparameter to obtain the quadratic term adjustment parameter of the first nonlinear factor; calculating the product of the eigenvalue at each position in the temperature-light intensity position-by-position response feature vector and a second predetermined adjustment hyperparameter to obtain the linear term adjustment parameter of the first nonlinear factor; and calculating the sum between the quadratic term adjustment parameter and the linear term adjustment parameter to obtain the first nonlinear factor.

[0084] In step S533, constructing the second nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector includes: calculating the product of the square of the eigenvalue at each position in the temperature-light intensity position-by-position response feature vector and a third predetermined adjustment hyperparameter to obtain the second nonlinear factor quadratic term adjustment parameter; calculating the product of the eigenvalue at each position in the temperature-light intensity position-by-position response feature vector and a fourth predetermined adjustment hyperparameter to obtain the second nonlinear factor linear term adjustment parameter; and calculating the sum of the first nonlinear factor quadratic term adjustment parameter and the first nonlinear factor linear term adjustment parameter, followed by constant term addition modulation to obtain the second nonlinear factor.

[0085] Specifically, in a specific example of this application, obtaining a temperature-light intensity time-series correlation interaction vector by passing the temperature time-series input vector and the light intensity time-series input vector through a nonlinear response correlation encoding module includes: processing the temperature time-series input vector and the light intensity time-series input vector through the nonlinear response correlation encoding module using the following nonlinear response formula to obtain the temperature-light intensity time-series correlation interaction vector; wherein, the nonlinear response formula is:

[0086]

[0087] in, This refers to the feature value at the i-th position in the time-series input vector of illumination intensity. Let v be the feature value at the i-th position in the temperature time-series input vector, where A, B, C, and D are adjustment hyperparameters. i Let be the feature value at the i-th position in the time-series correlation interaction vector of temperature and light intensity.

[0088] Furthermore, the complex nonlinear relationship between light intensity and temperature exists not only in the entire time domain but also in various local time periods. Moreover, considering that capturing and characterizing the temporal correlation features of light intensity and temperature data across the entire time domain would reduce the ability to capture implicit nonlinear correlations and detailed features, thus affecting the accuracy of understanding and capturing the complex nonlinear relationship between light intensity and temperature, this application's technical solution divides the temperature-light intensity temporal correlation interaction vector according to a predetermined time scale based on temperature to obtain a sequence of local temperature-light intensity temporal correlation interaction vectors. This division based on a predetermined time scale based on temperature facilitates the subsequent capture of nonlinear local temporal correlation relationships and pattern features between temperature and light intensity in different local time periods, further improving data representation and model performance, enabling the model to better understand and predict temporal changes in building energy consumption.

[0089] Subsequently, the sequence of temperature-light intensity local temporal correlation interaction vectors is then subjected to feature mining using an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model. This process extracts the local temporal patterns and correlation interaction features between temperature and light intensity in each local time period, thereby obtaining a sequence of environmental data local temporal pattern feature vectors. This provides richer and more effective feature representations for subsequent building energy consumption inference and building energy management.

[0090] Accordingly, such as Figure 4 As shown, in step S540, local temporal correlation pattern feature extraction is performed on the temperature-light intensity temporal correlation interaction vector to obtain a sequence of environmental data local temporal pattern feature vectors, including: S541, dividing the temperature-light intensity temporal correlation interaction vector according to a predetermined time scale based on temperature to obtain a sequence of temperature-light intensity local temporal correlation interaction vectors; and S542, passing the sequence of temperature-light intensity local temporal correlation interaction vectors through an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model to obtain a sequence of environmental data local temporal pattern feature vectors.

[0091] It should be understood that building energy consumption is influenced by various factors, including ambient temperature and light intensity. Furthermore, the local temporal correlation characteristics of temperature and light intensity in each local time period reflect the local temporal patterns and trends of building energy consumption, respectively. To utilize these local temporal characteristics of building energy consumption for future energy consumption inference, the technical solution of this application further uses a building energy consumption inference engine based on environmental temporal prior information to obtain a temporal representation feature vector for inferring building energy consumption. It should be understood that through the processing of the building energy consumption inference engine based on environmental temporal prior information, the local temporal nonlinear responsiveness correlation characteristics of ambient temperature and light intensity in each local time period can be used as prior knowledge for short-term inference and temporal representation of building energy consumption, helping building managers better understand and control energy consumption.

[0092] Accordingly, in step S550, the sequence of the local temporal pattern feature vector of the environmental data is subjected to building energy consumption inference based on the prior information of the local temporal pattern of the environment to obtain the inferred building energy consumption temporal representation feature, including: passing the sequence of the local temporal pattern feature vector of the environmental data through a building energy consumption inferencer based on the prior information of the environment temporal pattern to obtain the inferred building energy consumption temporal representation feature vector as the inferred building energy consumption temporal representation feature.

[0093] Specifically, the sequence of local temporal pattern feature vectors of the environmental data is processed by a building energy consumption inferencer based on environmental temporal prior information to obtain the inferred building energy consumption temporal representation feature vector, which serves as the inferred building energy consumption temporal representation feature. This includes: calculating the positional difference between the mean feature vector of multiple local temporal pattern feature vectors in the sequence of local temporal pattern feature vectors of the environmental data and the environmental data local temporal pattern feature vector at the last position in the sequence of local temporal pattern feature vectors of the environmental data to obtain the average distribution information difference feature vector; calculating the loop of the environmental data local temporal pattern feature vector at a random position in the sequence of local temporal pattern feature vectors of the environmental data and the loop of the environmental data local temporal pattern feature vector at the last position in the sequence of local temporal pattern feature vectors of the environmental data. The location difference between the feature vectors of the local time-series pattern of the environmental data is used to obtain a random distribution information difference feature vector; the average distribution information difference feature vector is multiplied by a first predetermined weight hyperparameter and a first Gaussian random number function value to obtain a first building energy consumption inference feature vector; the random distribution information difference feature vector is multiplied by a second predetermined weight hyperparameter and a second Gaussian random number function value to obtain a second building energy consumption inference feature vector; the location addition between the first building energy consumption inference feature vector and the second building energy consumption inference feature vector is calculated to obtain the inference building energy consumption time-series characterization feature vector; wherein, the first Gaussian random number function value and the second Gaussian random number function value are generated by a Gaussian random number function with a mean of 0 and a variance of 1.

[0094] Specifically, in a specific example of this application, performing building energy consumption inference based on prior information of the local temporal pattern feature vector of the environmental data to obtain inferred building energy consumption temporal representation features includes: processing the sequence of the local temporal pattern feature vector of the environmental data through the building energy consumption inference engine based on prior information of the environmental temporal data using the following energy consumption inference formula to obtain the inferred building energy consumption temporal representation feature vector; wherein, the energy consumption inference formula is:

[0095]

[0096] Wherein, V1 is the environmental data local temporal pattern feature vector at the last position in the sequence of environmental data local temporal pattern feature vectors, V rand V is the local temporal pattern feature vector of the environmental data at a random position in the sequence of local temporal pattern feature vectors of the environmental data. * is the mean feature vector of the sequence of local temporal pattern feature vectors of the environmental data, p1 and p2 are weight hyperparameters, and f1 and f2 are hyperparameters that use a Gaussian distribution function to generate random numbers with a mean of 0 and a variance of 1 as the coefficients of the Gaussian distribution function. and These are vector subtraction and vector addition, respectively, and V is the time-series characteristic vector representing the energy consumption of the inferred building.

[0097] Subsequently, the inferred building energy consumption time-series characteristic vector is passed through a classifier-based air conditioning controller to obtain control results. These control results indicate whether the air conditioning power should increase, decrease, or remain unchanged at the current time point. In other words, the short-term time-series characteristic features of building energy consumption are used for classification processing to formulate reasonable energy-saving plans. For example, the air conditioning power at the current time point can be automatically adjusted to optimize energy utilization efficiency. This enables intelligent management of building energy consumption, thereby improving building energy utilization efficiency and reducing energy waste.

[0098] Accordingly, in step S560, determining whether the air conditioning power should increase, decrease, or remain unchanged at the current time point based on the inferred building energy consumption time series characterization features includes: passing the inferred building energy consumption time series characterization feature vector through a classifier-based air conditioning controller to obtain a control result, wherein the control result is used to indicate whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

[0099] As you can understand, the role of a classifier is to learn classification rules and classifiers using given categories and known training data, and then classify (or predict) unknown data. Logistic regression and SVM are commonly used to solve binary classification problems. For multi-class classification problems, logistic regression or SVM can also be used, but multiple binary classifications are needed to form the multi-class classification. However, this is prone to errors and inefficient. A commonly used multi-class classification method is the Softmax classification function.

[0100] Furthermore, in the technical solution disclosed herein, the artificial intelligence-based digital management method for green buildings further includes a training step: training the nonlinear response association encoding module, the environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the air conditioning controller based on a classifier.

[0101] The training step includes: acquiring training data, which includes a time series of training building environment data collected by a sensor array deployed on the monitored building, wherein the training building environment data includes training temperature and training light intensity; arranging the time series of the training building environment data according to time and sample dimensions to obtain a training temperature time series input vector and a training light intensity time series input vector; passing the training temperature time series input vector and the training light intensity time series input vector through the nonlinear response association coding module to obtain a training temperature-light intensity time series association interaction vector; dividing the training temperature-light intensity time series association interaction vector according to a predetermined time scale based on temperature to obtain a sequence of training temperature-light intensity local time series association interaction vectors; and passing the sequence of training temperature-light intensity local time series association interaction vectors through the nonlinear response association coding module. The method describes an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model to obtain a sequence of local temporal pattern feature vectors for training environmental data; passing the sequence of local temporal pattern feature vectors for training environmental data through a building energy consumption inferencer based on environmental temporal prior information to obtain a training inference building energy consumption temporal representation feature vector; passing the training inference building energy consumption temporal representation feature vector through a classifier-based air conditioning controller to obtain a classification loss function value; and training the nonlinear response association coding module, the environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the classifier-based air conditioning controller based on the classification loss function value, wherein the training inference building energy consumption temporal representation feature vector is corrected in each iteration of the training.

[0102] In the technical solution described above, the sequence of the training temperature-light intensity local temporal correlation interaction vector expresses the source temporal nonlinear response correlation of training temperature-light intensity in the local time domain. Therefore, after passing the sequence of the training temperature-light intensity local temporal correlation interaction vector through an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model, and then passing the sequence of the training environmental data local temporal pattern feature vector through a building energy consumption inferencer based on environmental temporal prior information, the training inference building energy consumption temporal representation feature vector will also have temporal feature expression complexity corresponding to the local temporal prior feature distribution of temporal correlation features in each local time domain within the high-dimensional temporal aggregation space. That is, the training inference building energy consumption temporal representation feature vector will have rich geometric shape details in the high-dimensional feature space, resulting in class representation distortion when classified by a classifier to map to the class probability domain. This affects the class representation of the training inference building energy consumption temporal representation feature vector for the classifier's predetermined class, thereby reducing the accuracy of the classification results.

[0103] Based on this, the applicant of this application corrects the training inference building energy consumption time series representation feature vector each time the training inference building energy consumption time series representation feature vector is classified and iterated by the classifier-based air conditioning controller.

[0104] Accordingly, in one example, in each iteration of the training, the trained inference building energy consumption time-series representation feature vector is corrected using the following correction formula to obtain the corrected trained inference building energy consumption time-series representation feature vector; wherein, the correction formula is:

[0105]

[0106] Where V is the time-series representation feature vector of the training and inference building energy consumption, p is the class probability value obtained by the classifier through the time-series representation feature vector V of the training and inference building energy consumption, and v i It is the i-th eigenvalue of the time-series characteristic vector V representing the energy consumption of the building during training and inference. V is the mean of all eigenvalues ​​of the training inference building energy consumption time-series representation feature vector V, ||V||1 represents the first norm of the training inference building energy consumption time-series representation feature vector V, and α is a weight hyperparameter. exp(·) represents the exponential operation of the numerical value, which means calculating the value of the natural exponential function raised to the power of the numerical value. ′i It is the i-th eigenvalue of the time-series characteristic vector of the corrected training inference building energy consumption.

[0107] Specifically, while preserving the geometric details of the high-dimensional feature manifold under the class probability mapping of the training inference building energy consumption time-series representation feature vector V, the shape attributes of the feature values ​​themselves relative to the feature set of the training inference building energy consumption time-series representation feature vector V are edited through a class probability distortion consultation. This transforms the class probability distortion mapping of the training inference building energy consumption time-series representation feature vector V onto the latent class space feature representation. Then, consultation fusion is performed by supplementing with a basic low-rank constrained latent inversion representation to bridge the gap between the edited consultation representation and the original geometric details. This improves the class representation of the high-dimensional features of the training inference building energy consumption time-series representation feature vector V, thereby improving the accuracy of the classification results. In this way, future building energy consumption can be predicted more accurately, enabling the formulation of reasonable energy-saving plans, achieving intelligent management of building energy consumption, improving building energy utilization efficiency, and reducing energy waste.

[0108] Based on the above embodiments, see Figure 5The diagram shown is a structural schematic of an AI-based green building digital management system 800 according to an embodiment of this disclosure. The AI-based green building digital management system 800 includes: an environmental data acquisition module 810, used to acquire time series of building environment data collected by a sensor array deployed on the monitored building, wherein the building environment data includes temperature and light intensity; a vectorization module 820, used to arrange the time series of the building environment data according to time and sample dimensions to obtain a temperature time series input vector and a light intensity time series input vector; and an association interaction module 830, used to pass the temperature time series input vector and the light intensity time series input vector through a nonlinear response association encoding module to obtain the temperature... - Light intensity temporal correlation interaction vector; Local correlation feature extraction module 840, used to extract local temporal correlation pattern features from the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors of environmental data; Energy consumption inference module 850, used to perform building energy consumption inference based on prior information of local temporal patterns of the environment on the sequence of local temporal pattern feature vectors of environmental data to obtain inferred building energy consumption temporal characterization features; and Power control module 860, used to determine whether the air conditioning power should increase, decrease, or remain unchanged at the current time point based on the inferred building energy consumption temporal characterization features.

[0109] Here, those skilled in the art will understand that the specific functions and operations of each module in the aforementioned AI-based green building digital management system 800 have been referenced above. Figures 2 to 4 The description of the AI-based digital management method for green buildings is detailed here, and therefore, its repeated description will be omitted.

[0110] Figure 6 This diagram illustrates an application scenario of the AI-based digital management method for green buildings according to embodiments of this disclosure. Figure 6 As shown, in this application scenario, firstly, time series data of the building environment (e.g., data collected by a sensor array deployed in the monitored building) is acquired. Figure 6 As shown in D), the building environment data includes temperature and light intensity, and then the time series of the building environment data is input to a server deployed with an AI-based green building digital management algorithm (e.g., Figure 6 In the S shown, the server is able to use the AI-based green building digital management algorithm to process the time series of the building environment data to obtain a control result indicating whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

[0111] Based on the above embodiments, this disclosure also provides an electronic device according to another exemplary embodiment. In some possible implementations, the electronic device in this disclosure may include a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, can implement the steps of the artificial intelligence-based digital management method for green buildings described above.

[0112] For example, disclosure based on electronic devices Figure 1 Taking server 100 as an example, the processor in the electronic device is the processor 110 in server 100, and the memory in the electronic device is the memory 120 in server 100.

[0113] Embodiments of this disclosure also provide a computer-readable storage medium. Figure 7 A schematic diagram of a computer-readable storage medium 1000 according to an embodiment of the present disclosure is shown. Figure 7 As shown, the computer-readable storage medium 1000 stores computer-executable instructions 1001. When the computer-executable instructions 1001 are executed by a processor, the artificial intelligence-based digital management method for green buildings according to embodiments of this disclosure, as described with reference to the above figures, can be executed. The computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. The volatile memory may include, for example, random access memory (RAM) and / or cache memory. The non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc.

[0114] Embodiments of this disclosure also provide a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform an artificial intelligence-based digital management method for green buildings according to embodiments of this disclosure.

[0115] Those skilled in the art will understand that the contents disclosed herein can be varied and modified in many ways. For example, the various devices or components described above can be implemented in hardware, or in software, firmware, or a combination of some or all of the three.

[0116] Furthermore, while this disclosure makes various references to certain units in systems according to embodiments of this disclosure, any number of different units may be used and run on clients and / or servers. The units described are merely illustrative, and different aspects of the systems and methods may use different units.

[0117] Those skilled in the art will understand that all or part of the steps in the above methods can be implemented by a program instructing related hardware, and the program can be stored in a computer-readable storage medium, such as a read-only memory, a disk, or an optical disk. Optionally, all or part of the steps in the above embodiments can also be implemented using one or more integrated circuits. Accordingly, each module / unit in the above embodiments can be implemented in hardware or as a software functional module. This disclosure is not limited to any particular combination of hardware and software.

[0118] Unless otherwise defined, all terms used herein (including technical and scientific terms) shall have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure pertains. It should also be understood that terms such as those defined in a common dictionary shall be interpreted as having a meaning consistent with their meaning in the context of the relevant art, and not as having an idealized or highly formalized meaning, unless expressly defined herein.

[0119] The foregoing description is a description of the present disclosure and should not be construed as limiting it. Although several exemplary embodiments of the present disclosure have been described, those skilled in the art will readily understand that many modifications can be made to the exemplary embodiments without departing from the novel teachings and advantages of the present disclosure. Therefore, all such modifications are intended to be included within the scope of the present disclosure as defined by the claims. It should be understood that the foregoing description is a description of the present disclosure and should not be construed as limiting it to the specific embodiments disclosed, and modifications to the disclosed embodiments and other embodiments are intended to be included within the scope of the appended claims. The present disclosure is defined by the claims and their equivalents.

Claims

1. A digital management method for green buildings based on artificial intelligence, characterized in that, include: Acquire time series of building environment data collected by a sensor array deployed on the monitored building, wherein the building environment data includes temperature and light intensity; The time series of the building environment data is arranged according to time and sample dimensions to obtain the temperature time series input vector and the light intensity time series input vector; The temperature time-series input vector and the light intensity time-series input vector are passed through a nonlinear response correlation coding module to obtain a temperature-light intensity time-series correlation interaction vector. Local temporal correlation pattern feature extraction is performed on the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors of environmental data; The sequence of local temporal pattern feature vectors of the environmental data is used to perform building energy consumption inference based on prior information of the local temporal pattern of the environment to obtain the temporal representation features of inferred building energy consumption; and Based on the time-series characteristics of the inferred building energy consumption, it is determined whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

2. The artificial intelligence-based digital management method for green buildings according to claim 1, characterized in that, The temperature time-series input vector and the light intensity time-series input vector are processed through a nonlinear response correlation encoding module to obtain a temperature-light intensity time-series correlation interaction vector, including: Calculate the positional response between the time-series input vector of light intensity and the time-series input vector of temperature to obtain the temperature-light intensity positional response feature vector; Construct the first nonlinear factor at each position in the location-by-location response feature vector of the temperature-light intensity; Construct a second nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector; and The division between the first nonlinear factor and the second nonlinear factor is calculated to obtain the temperature-light intensity position-by-position nonlinear response correlation correction value, and the multiple temperature-light intensity position-by-position nonlinear response correlation correction values ​​are vectorized and normalized to obtain the temperature-light intensity time-series correlation interaction vector.

3. The artificial intelligence-based digital management method for green buildings according to claim 2, characterized in that, Constructing the first nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector includes: The square of each location feature value in the temperature-light intensity location-by-location response feature vector is calculated and multiplied by the first predetermined adjustment hyperparameter to obtain the first nonlinear factor quadratic term adjustment parameter; Calculate the product of the positional eigenvalues ​​in the temperature-light intensity positional response feature vector and the second predetermined adjustment hyperparameter to obtain the first nonlinear factor linear term adjustment parameter; and The first nonlinear factor is obtained by summing the adjustment parameters of the quadratic term and the linear term of the first nonlinear factor.

4. The artificial intelligence-based digital management method for green buildings according to claim 3, characterized in that, Constructing the second nonlinear factor at each position in the temperature-light intensity position-by-position response feature vector includes: The square of each location feature value in the temperature-light intensity position-by-position response feature vector is calculated and multiplied by the third predetermined adjustment hyperparameter to obtain the second nonlinear factor quadratic term adjustment parameter; Calculate the product of the eigenvalues ​​at each location in the temperature-light intensity position-by-location response feature vector and the fourth predetermined adjustment hyperparameter to obtain the adjustment parameter of the first-order term of the second nonlinear factor; and The second nonlinear factor is obtained by summing the adjustment parameters of the first nonlinear factor's quadratic term and the first nonlinear factor's linear term, and then modulating the constant term by adding one.

5. The artificial intelligence-based digital management method for green buildings according to claim 4, characterized in that, Local temporal correlation pattern feature extraction is performed on the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors for environmental data, including: The temperature-light intensity temporal correlation interaction vector is divided according to a predetermined time scale based on temperature to obtain a sequence of local temperature-light intensity temporal correlation interaction vectors; and The sequence of temperature-light intensity local temporal correlation interaction vectors is processed by an environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model to obtain the sequence of environmental data local temporal pattern feature vectors.

6. The artificial intelligence-based digital management method for green buildings according to claim 5, characterized in that, The sequence of local temporal pattern feature vectors of the environmental data is used to perform building energy consumption inference based on prior information of the local temporal pattern of the environment to obtain the temporal representation features of inferred building energy consumption, including: The sequence of local temporal pattern feature vectors of the environmental data is passed through a building energy consumption inference engine based on environmental temporal prior information to obtain the inference building energy consumption temporal representation feature vector, which is used as the inference building energy consumption temporal representation feature.

7. The artificial intelligence-based digital management method for green buildings according to claim 6, characterized in that, The sequence of local temporal pattern feature vectors of the environmental data is passed through a building energy consumption inference engine based on environmental temporal prior information to obtain the inferred building energy consumption temporal representation feature vector, which is used as the inferred building energy consumption temporal representation feature, including: The mean feature vector of multiple local temporal pattern feature vectors of environmental data in the sequence of local temporal pattern feature vectors of environmental data is calculated as the positional difference between the local temporal pattern feature vector of environmental data at the last position in the sequence of local temporal pattern feature vectors of environmental data to obtain the average distribution information difference feature vector; The positional difference between the local temporal pattern feature vector of the environmental data at a random position in the sequence of the local temporal pattern feature vectors of the environmental data and the local temporal pattern feature vector of the environmental data at the last position in the sequence of the local temporal pattern feature vectors of the environmental data is calculated to obtain the random distribution information difference feature vector. The average distribution information difference feature vector is multiplied by the first predetermined weight hyperparameter and the first Gaussian distribution random number function value to obtain the first building energy consumption inference feature vector; The random distribution information difference feature vector is multiplied by the second predetermined weight hyperparameter and the second Gaussian distribution random number function value to obtain the second building energy consumption inference feature vector; Calculate the positional addition between the first building energy consumption inference feature vector and the second building energy consumption inference feature vector to obtain the inference building energy consumption time series characterization feature vector; The first Gaussian random number function value and the second Gaussian random number function value are generated by a Gaussian random number function with a mean of 0 and a variance of 1.

8. The artificial intelligence-based digital management method for green buildings according to claim 7, characterized in that, Based on the aforementioned time-series characteristics of building energy consumption, determining whether the air conditioning power should increase, decrease, or remain unchanged at the current time point includes: The time-series characteristic vector of the inferred building energy consumption is passed through a classifier-based air conditioning controller to obtain the control result, which is used to indicate whether the air conditioning power should increase, decrease, or remain unchanged at the current time point.

9. The artificial intelligence-based digital management method for green buildings according to claim 8, characterized in that, It also includes a training step: for training the nonlinear response association coding module, the environmental data temporal pattern feature extractor based on a one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the air conditioning controller based on a classifier; The training steps include: Acquire training data, which includes a time series of training building environment data collected by a sensor array deployed on the monitored building, wherein the training building environment data includes training temperature and training light intensity; The time series of the training building environment data is arranged according to time and sample dimensions to obtain the training temperature time series input vector and the training light intensity time series input vector; The training temperature time-series input vector and the training light intensity time-series input vector are passed through the nonlinear response correlation coding module to obtain the training temperature-light intensity time-series correlation interaction vector. The training temperature-light intensity temporal correlation interaction vector is divided according to a predetermined time scale based on temperature to obtain a sequence of local temporal correlation interaction vectors of training temperature-light intensity. The sequence of the training temperature-light intensity local temporal correlation interaction vector is passed through the environmental data temporal pattern feature extractor based on the one-dimensional convolutional neural network model to obtain the sequence of training environmental data local temporal pattern feature vectors. The sequence of local temporal pattern feature vectors of the training environment data is passed through the building energy consumption inference engine based on environmental temporal prior information to obtain the training inference building energy consumption temporal representation feature vector. The time-series feature vector representing the trained inference building energy consumption is passed through the classifier-based air conditioning controller to obtain a classification loss function value; and The nonlinear response association coding module, the environmental data temporal pattern feature extractor based on the one-dimensional convolutional neural network model, the building energy consumption inferencer based on environmental temporal prior information, and the air conditioning controller based on the classifier are trained based on the classification loss function value. In each iteration of the training, the trained inference building energy consumption temporal representation feature vector is corrected.

10. A green building digital management system based on artificial intelligence, characterized in that, include: An environmental data acquisition module is used to acquire time series of building environment data collected by a sensor array deployed on the monitored building, wherein the building environment data includes temperature and light intensity; The vectorization module is used to arrange the time series of the building environment data according to time and sample dimensions to obtain the temperature time series input vector and the light intensity time series input vector; The correlation interaction module is used to pass the temperature time-series input vector and the light intensity time-series input vector through the nonlinear response correlation encoding module to obtain the temperature-light intensity time-series correlation interaction vector. The local correlation feature extraction module is used to extract local temporal correlation pattern features from the temperature-light intensity temporal correlation interaction vector to obtain a sequence of local temporal pattern feature vectors of environmental data; The energy consumption inference module is used to perform building energy consumption inference based on prior information of the local temporal pattern feature vectors of the environmental data to obtain the inferred building energy consumption temporal representation features; and The power control module is used to determine whether the air conditioning power should increase, decrease, or remain unchanged at the current time point based on the time-series characteristics of the inferred building energy consumption.

Images