Building energy consumption management method, device, equipment and medium

Abstract

The application discloses a kind of building energy consumption management method, device, equipment and medium, it is related to building energy consumption management technical field, comprising: collection and preprocessing building multi-source heterogeneous data, provide input for energy consumption prediction model based on machine learning, can realize high-precision prediction to building energy consumption;Based on this prediction result, building equipment control strategy is obtained using reinforcement learning algorithm, and energy-saving strategy is generated using generative adversarial network, which can provide intelligent energy consumption optimization decision support for building managers and break away from the dependence on traditional experience;Intelligent collaborative control of multiple devices is realized using building equipment control strategy, and the operation mode of equipment is optimized by combining energy-saving strategy, which can avoid energy waste and conflict of single-device independent operation.The whole process forms a dynamic closed loop, which can respond to changes in building operation in real time, promote building energy consumption management to realize dynamic self-adaptive adjustment, and ensure that the building is stably in energy-saving operation state.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of building energy consumption management, and in particular to a building energy consumption management method, device, equipment and medium. BACKGROUND

[0002] Under the background of accelerating urbanization and increasingly complex building functions, the use of buildings presents significant dynamic characteristics. At the same time, the external environment of the building is also in a state of continuous change, which affects the energy consumption demand of the building air conditioning, lighting and other systems. However, the traditional building energy consumption management strategy is mostly based on fixed threshold or empirical formula, for example, starting and stopping the equipment according to the preset period, ignoring the actual use demand difference of different areas; setting energy consumption benchmark according to historical data, which is difficult to respond to the load fluctuation caused by real-time weather changes; relying on manual inspection to adjust the operation parameters, which not only has strong lag, but also cannot accurately match the dynamic energy consumption characteristics of the building. This mode is easy to lead to the building system in a state of over-supply or under-supply energy for a long time, and cannot adapt to the management needs of the building under the use scene and external environment. SUMMARY

[0003] The purpose of the present application is to provide a building energy consumption management method, device, equipment and medium, which can promote the dynamic adaptive adjustment of building energy consumption management and ensure that the building is in a stable energy-saving operation state.

[0004] In order to solve the above technical problems, the present application provides a building energy consumption management method, comprising:

[0005] Collecting building multi-source heterogeneous data and preprocessing the building multi-source heterogeneous data;

[0006] Inputting the processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model to output a building energy consumption prediction result;

[0007] According to the building energy consumption prediction result, a reinforcement learning algorithm is used to obtain a building equipment control strategy, and a generative adversarial network is used to generate an energy-saving strategy;

[0008] The building equipment control strategy is used for equipment collaborative control, and the energy-saving strategy is used for optimizing the equipment operation mode, so that the building is in an energy-saving operation state.

[0009] In the first aspect, in the building energy consumption management method provided by the present application, before the processed building multi-source heterogeneous data is input into the machine learning-based energy consumption prediction model, it further comprises:

[0010] The energy consumption prediction model is constructed by using a convolutional neural network combined with a recurrent neural network; wherein the convolutional neural network is used for processing image data, and the recurrent neural network is used for analyzing time series data and capturing the change trend of building energy consumption at different time scales.

[0011] A historical building multi-source heterogeneous data sample set is obtained.

[0012] The energy consumption prediction model is trained by using the historical building multi-source heterogeneous data sample set; during the training process, the energy consumption prediction model is updated and optimized by using online learning.

[0013] On the other hand, in the above building energy consumption management method provided by the application, according to the building energy consumption prediction result, a reinforcement learning algorithm is used to obtain a building equipment control strategy, which includes:

[0014] The building equipment control system is defined as an agent, and the internal and external environment of the building is regarded as an environment interacting with the agent.

[0015] During the interaction between the agent and the environment, according to different control actions combined with the building energy consumption prediction result, a corresponding reward feedback is obtained.

[0016] Through continuous action trial and error, the building equipment control strategy is learned and optimized to obtain an optimal building equipment control strategy that maximizes the reward.

[0017] Correspondingly, the building equipment control strategy is used for equipment collaborative control, which includes adjusting the strength of the air conditioning system, adjusting the ventilation frequency of the ventilation system and adjusting the brightness of the lighting system by using the optimal building equipment control strategy, so as to realize the collaborative control between the equipment.

[0018] On the other hand, in the above building energy consumption management method provided by the application, according to different control actions combined with the building energy consumption prediction result, a corresponding reward feedback is obtained, which includes:

[0019] When the agent performs a control action, the actual energy consumption corresponding to the control action is associated and analyzed with the building energy consumption prediction result.

[0020] If the actual energy consumption is lower than the building energy consumption prediction result, a first reward feedback is obtained.

[0021] If the actual energy consumption is equal to the building energy consumption prediction result, a basic reward feedback is obtained; the basic reward feedback is less than the first reward feedback.

[0022] If the actual energy consumption is higher than the building energy consumption prediction result, a second reward feedback is obtained; the second reward feedback is less than the basic reward feedback.

[0023] In another aspect, in the building energy consumption management method provided by the present application, the energy saving strategy is generated by using a generative adversarial network, which comprises:

[0024] According to the characteristics and energy saving goals of the building, a generator is used to generate an energy saving strategy;

[0025] According to the target indicators including the building energy consumption prediction results, cost analysis, and technical feasibility, a discriminator is used to evaluate and screen the energy saving strategy;

[0026] The energy saving strategy is optimized through the adversarial training between the generator and the discriminator to obtain an optimal energy saving strategy;

[0027] Correspondingly, the device operation mode is optimized in combination with the energy saving strategy optimization device, which comprises optimizing the device operation mode in combination with the optimal energy saving strategy.

[0028] In another aspect, in the building energy consumption management method provided by the present application, according to the target indicators including the building energy consumption prediction results, cost analysis, and technical feasibility, a discriminator is used to evaluate and screen the energy saving strategy, which comprises:

[0029] The energy consumption data corresponding to the energy saving strategy is compared with the building energy consumption prediction results by using the discriminator, and according to the deviation degree of the energy consumption data and the building energy consumption prediction results, it is judged whether the energy saving strategy is within a first preset range in terms of energy consumption control; if not, the energy saving strategy is excluded;

[0030] At the same time, the cost analysis is carried out in combination with the cost required for implementing the energy saving strategy, the key indicators including the cost benefit ratio in the energy saving strategy are calculated, and it is evaluated whether the energy saving strategy is within a second preset range in terms of cost; if not, the energy saving strategy is excluded;

[0031] According to the device performance, system compatibility, and operation feasibility required by the energy saving strategy, it is judged whether the energy saving strategy is within a third preset range in terms of technical feasibility; if not, the energy saving strategy is excluded.

[0032] In another aspect, in the building energy consumption management method provided by the present application, it further comprises:

[0033] Receiving instructions for a building; the instructions include query instructions or control instructions;

[0034] Carrying out semantic understanding on the instructions;

[0035] According to the content understood by the semantic understanding, relevant information is extracted from a knowledge base to generate answer content and solutions, and is responded and output in the form of natural language.

[0036] Meanwhile, unstructured data related to building energy consumption is processed, and the processing result is uploaded to the knowledge base.

[0037] To solve the above technical problems, the application further provides a building energy consumption management device, comprising:

[0038] a data acquisition module, configured to acquire building multi-source heterogeneous data and pre-process the building multi-source heterogeneous data;

[0039] a model inference module, configured to input the pre-processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model and output a building energy consumption prediction result;

[0040] a strategy generation module, configured to obtain a building equipment control strategy by using a reinforcement learning algorithm according to the building energy consumption prediction result, and generate an energy-saving strategy by using a generative adversarial network;

[0041] an optimization decision module, configured to perform equipment collaborative control by using the building equipment control strategy, and optimize equipment operation mode in combination with the energy-saving strategy, so that the building is in an energy-saving operation state.

[0042] To solve the above technical problems, the application further provides an electronic device, comprising:

[0043] a memory, configured to store a computer program;

[0044] a processor, configured to implement the steps of the building energy consumption management method when the computer program is executed.

[0045] To solve the above technical problems, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the building energy consumption management method.

[0046] As can be seen from the above technical solutions, the building energy consumption management method provided by the application comprises the following steps: acquiring building multi-source heterogeneous data and pre-processing the building multi-source heterogeneous data; inputting the pre-processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model and outputting a building energy consumption prediction result; obtaining a building equipment control strategy by using a reinforcement learning algorithm according to the building energy consumption prediction result, and generating an energy-saving strategy by using a generative adversarial network; performing equipment collaborative control by using the building equipment control strategy, and optimizing equipment operation mode in combination with the energy-saving strategy, so that the building is in an energy-saving operation state.

[0047] The building energy consumption management method provided by the application can realize high-precision prediction of building energy consumption, accurately capture short-term fluctuations and long-term trends, and provide an important basis for energy consumption optimization management.

[0048] In addition, the application also provides a building energy consumption management device, an electronic device and a computer readable storage medium for building energy consumption management, which have the same or corresponding technical features as the building energy consumption management method mentioned above, and the effects are the same as above. BRIEF DESCRIPTION OF DRAWINGS

[0049] In order to more clearly illustrate the embodiments of the application, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative labor on the basis of these drawings.

[0050] Figure 1 The flow chart of the building energy consumption management method provided by the embodiments of the application is shown in the figure.

[0051] Figure 2 The structural schematic diagram of the building energy consumption management device provided by the embodiments of the application is shown in the figure. DETAILED DESCRIPTION

[0052] The technical solutions in the embodiments of the application will be described clearly and completely in combination with the drawings in the embodiments of the application. Obviously, the described embodiments are only some embodiments of the application, not all embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0053] It should be noted that in the description of the present application, the terms "comprising", "containing" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or equipment including a series of elements not only includes those elements, but also includes other elements not explicitly listed or inherent to such a process, method, article or equipment. The terms "first", "second" and the like in the present application are used to distinguish similar objects, not to describe a specific order or sequence.

[0054] In order to enable those skilled in the art to better understand the present application, the present application will be further described in detail below in combination with the drawings and specific embodiments.

[0055] In combination with the specific application environment architecture or specific hardware architecture on which the building energy consumption management method is executed, the specific application environment architecture or specific hardware architecture is described herein.

[0056] The embodiment of the present application provides a building energy consumption management method, and the method is described in detail in combination with the execution process of the building energy consumption management method. Figure 1 The flowchart of the building energy consumption management method provided by the embodiment of the present application is shown in Figure 1 The method comprises the following steps.

[0057] S101, collecting building multi-source heterogeneous data and pre-processing the building multi-source heterogeneous data.

[0058] In implementation, the building multi-source heterogeneous data of the present application can include building energy consumption data (such as energy consumption of each device, total energy consumption, etc.), indoor and outdoor environment parameters (such as temperature and humidity, illumination, carbon dioxide concentration, etc.), device operation data (such as device start-stop state, operation parameter, fault information, etc.), personnel activity data (such as personnel quantity, distribution, stay duration, etc.). The present application can collect building energy consumption, indoor and outdoor environment parameters, device operation data, personnel activity data and other multi-source heterogeneous data through a data collection system.

[0059] In order to meet the subsequent processing requirements of the building multi-source heterogeneous data, in the process of pre-processing the building multi-source heterogeneous data, the building multi-source heterogeneous data can be cleaned, noise can be removed, missing values can be filled, abnormal data can be corrected, key features can be extracted, and finally integrated to form a unified available data set.

[0060] S102, inputting the processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model to output a building energy consumption prediction result.

[0061] In implementation, this invention can utilize machine learning algorithms such as regression analysis, neural networks, and random forests to construct a high-precision energy consumption prediction model. This model can automatically adjust based on real-time changes in internal and external conditions. When processed multi-source heterogeneous building data is input into this energy consumption prediction model, this cleaned, feature-extracted, and integrated high-quality data precisely matches the model's input requirements. During model operation, machine learning algorithms perform in-depth mining and analysis of the input data. By learning the potential correlations between energy consumption and various influencing factors in historical data, the model can continuously optimize its parameters and structure, constructing a mapping relationship between energy consumption and multi-source data features. Based on the constructed mapping relationship, the corresponding building energy consumption prediction results can be output.

[0062] S103. Based on the building energy consumption prediction results, a reinforcement learning algorithm is used to obtain the building equipment control strategy, and a generative adversarial network is used to generate an energy-saving strategy.

[0063] In implementation, this invention can, based on the building energy consumption prediction output in step S102, first apply reinforcement learning algorithms to obtain suitable building equipment control strategies, such as rationally adjusting air conditioning temperature, lighting brightness, and elevator operation modes; simultaneously, it utilizes a Generative Adversarial Network (GAN) to allow the generator and discriminator to compete and co-evolve, generating feasible energy-saving strategies. These strategies not only respond to the trends in energy consumption prediction but also find a balance between building functional requirements and energy-saving goals, ultimately providing comprehensive and powerful strategic support for efficient energy-saving management of buildings. The building equipment control strategies and energy-saving strategies of this invention can also be automatically adjusted according to real-time changes in internal and external conditions.

[0064] S104. Adopt building equipment control strategies for coordinated equipment control, and combine them with energy-saving strategies to optimize equipment operation modes so that the building is in an energy-saving operation state.

[0065] In practice, this invention can employ building equipment control strategies to conduct coordinated equipment control. Based on the functional characteristics and operational needs of different equipment, it enables equipment such as air conditioning, lighting, elevators, and ventilation systems to form a linkage mechanism. For example, during peak hours, it can automatically increase the cooling efficiency of air conditioning and simultaneously brighten the area lighting. When there are few people, it can collaboratively reduce equipment power to avoid energy waste caused by the independent operation of individual equipment. At the same time, it can optimize equipment operation modes by combining energy-saving strategies. By adjusting equipment start-up and shutdown times and operating parameters, such as allowing air conditioning to store cold and heat in advance during off-peak hours to avoid peak energy consumption, and the lighting system to automatically adjust brightness according to the intensity of natural light, all types of equipment can operate in a low-energy-consumption and high-efficiency state while meeting the normal use needs of the building, thereby keeping the entire building in a stable energy-saving operation mode.

[0066] It should be noted that the above building energy consumption management method can be understood as a method of intelligent prediction and optimization management of building energy consumption based on generative artificial intelligence. The method realizes intelligent prediction of building energy consumption, and automatically optimizes the equipment operation mode in combination with energy-saving strategies, thereby realizing intelligent management and efficient energy saving of building energy consumption.

[0067] In the above building energy consumption management method provided by the embodiment of the application, through the integration and preprocessing of building multi-source heterogeneous data, high-quality input is provided for the energy consumption prediction model based on machine learning, high-precision prediction of building energy consumption can be realized, short-term fluctuations and long-term trends can be accurately captured, and important basis is provided for energy consumption optimization management; based on the prediction result, the building equipment control strategy is obtained by using the reinforcement learning algorithm, and the energy-saving strategy is generated by using the generative adversarial network, intelligent energy consumption optimization decision support is provided for building managers, and dependence on traditional experience is eliminated; the intelligent collaborative control of multiple devices is realized by using the equipment control strategy, and the equipment operation mode is optimized in combination with the energy-saving strategy, thereby avoiding energy consumption waste and conflicts of single device independent operation. The whole process forms a dynamic closed loop of data acquisition, prediction analysis, strategy generation and control execution, can respond to changes in use and external environment in building operation in real time, promotes dynamic self-adaptive adjustment of building energy consumption management, ensures that the building is stably in an energy-saving operation state, and greatly improves the flexibility and effectiveness of building energy consumption management.

[0068] Further, in specific implementation, in the above building energy consumption management method provided by the embodiment of the application, before the step S102 of inputting the processed building multi-source heterogeneous data into the energy consumption prediction model based on machine learning, the following steps can also be included: a convolutional neural network (CNN) is used in combination with a recurrent neural network (RNN) to construct an energy consumption prediction model; the convolutional neural network is used to process image data, and the recurrent neural network is used to analyze time series data and capture the change trend of building energy consumption at different time scales; a historical building multi-source heterogeneous data sample set is obtained; the energy consumption prediction model is trained by using the historical building multi-source heterogeneous data sample set; and in the training process, the energy consumption prediction model is updated and optimized by using online learning.

[0069] In implementation, the present application can adopt a convolutional neural network to process image data such as building thermal imaging images, combine the ability of a recurrent neural network to analyze time series data, establish a hybrid prediction model that fuses the features of multi-source data, effectively capture the change trend of building energy consumption at different time scales, and accurately predict the energy consumption in the future hours to months. The recurrent neural network can be a long short-term memory (LSTM). At the same time, by using online learning, the model can continuously update and optimize according to new real-time data, adapt to dynamic changes in energy consumption, thereby improving the accuracy and reliability of the prediction, and providing an important basis for building managers to understand energy demand in advance, reasonably arrange energy supply and equipment operation plans, and realize energy consumption optimization management.

[0070] Further, in the specific implementation, in the building energy consumption management method provided by the embodiment of the present application, step S103 obtains the building equipment control strategy by using a reinforcement learning algorithm according to the building energy consumption prediction result, which can specifically include: defining the building equipment control system as an agent, and the internal and external environment of the building as the environment interacting with the agent; in the interaction process between the agent and the environment, the corresponding reward feedback is obtained according to different control actions combined with the building energy consumption prediction result; through continuous action trial and error, the building equipment control strategy is learned and optimized to obtain the optimal building equipment control strategy that maximizes the reward.

[0071] Correspondingly, step S104 adopts the building equipment control strategy to perform equipment collaborative control, which can specifically include: adjusting the strength of the air conditioning system, adjusting the ventilation frequency of the ventilation system, and adjusting the brightness of the lighting system by using the optimal building equipment control strategy, so as to realize the collaborative control between the equipment.

[0072] In implementation, the agent refers to an entity that learns the optimal behavior strategy by interacting with the environment in a specific environment. The present application can regard the building equipment control system as an agent, and the internal and external environment of the building as the interactive environment of the agent. In the interaction with the environment, the agent of the present application can perceive the state of the environment, select and execute control actions such as equipment start-stop and parameter adjustment according to its own decision mechanism, obtain reward feedback such as energy consumption reduction degree and indoor comfort degree satisfaction from the environment, and learn the optimal control strategy that maximizes the long-term reward by continuously trying different control strategies. For example, in the elevator group control system, the reinforcement learning algorithm is used to optimize the scheduling strategy of the elevator according to real-time floor call information, elevator running state and other factors, reduce the number of empty trips and waiting time of the elevator, and reduce the energy consumption of the elevator system.

[0073] In addition, since the coordinated operation of many equipment systems such as heating, ventilation, air conditioning, lighting and the like in the building has a significant impact on energy consumption, the building energy consumption management method provided by the embodiment of the present application can monitor the operation state and energy consumption data of each equipment in real time, and learn the best coordination mode between the equipment by using a reinforcement learning algorithm. Under the premise of meeting the indoor comfort standard, the air conditioning refrigeration / heating intensity, ventilation frequency, lighting brightness and the like are automatically adjusted to realize efficient coordination and avoid excessive operation or repeated energy consumption. For example, when the number of personnel in a certain area is detected to be reduced, the air conditioning load and lighting brightness in the area are automatically reduced, and the operation of the equipment in other areas is coordinated, so that the overall energy consumption of the building is reduced to the greatest extent without affecting the overall comfort.

[0074] Further, in the building energy consumption management method provided by the embodiment of the present application, the step of obtaining the corresponding reward feedback according to the different control actions in combination with the building energy consumption prediction result can specifically include: when the agent executes each control action, the actual energy consumption corresponding to the control action is associated with the building energy consumption prediction result for analysis; if the actual energy consumption is lower than the building energy consumption prediction result, a first reward feedback is obtained; if the actual energy consumption is equal to the building energy consumption prediction result, a basic reward feedback is obtained; the basic reward feedback is less than the first reward feedback; if the actual energy consumption is higher than the building energy consumption prediction result, a second reward feedback is obtained; and the second reward feedback is less than the basic reward feedback.

[0075] In the implementation, when the agent executes each control action, the agent can be guided to optimize the control strategy through differentiated reward settings. When the actual energy consumption is lower than the building energy consumption prediction result, the highest first reward is given, which can encourage the agent to actively explore better control actions, such as further adjusting the equipment operation parameters, optimizing the start and stop time and the like, so as to continuously improve the strategy in the direction of reducing energy consumption; and when the actual energy consumption is out of limit, the lowest second reward is given, which can prompt the agent to avoid inefficient or unreasonable control behaviors. In this way, the association analysis of the actual energy consumption and the prediction result as the reward basis can make the agent pay more attention to the matching of the prediction trend in the learning process.

[0076] Further, in the building energy consumption management method provided by the embodiment of the present application, the step S103 of generating the energy saving strategy by using the generative adversarial network can specifically include: generating the energy saving strategy by using the generator according to the characteristics and energy saving target of the building; evaluating and screening the energy saving strategy by using the discriminator according to the target indexes including the building energy consumption prediction result, cost analysis and technical feasibility; and optimizing the energy saving strategy by the adversarial training between the generator and the discriminator to obtain the optimal energy saving strategy.

[0077] Correspondingly, the step S104 of optimizing the equipment operation mode in combination with the energy saving strategy can specifically include: optimizing the equipment operation mode in combination with the optimal energy saving strategy.

[0078] In implementation, the generator can generate various potential energy-saving design schemes, equipment operation schemes or energy management strategies according to the characteristics of the building and the energy-saving target; the discriminator evaluates and screens the generated schemes according to the energy consumption prediction results, economic cost analysis, technical feasibility and other indicators. Through the adversarial training of the two, the quality of the schemes generated by the generator is continuously optimized, and finally an efficient energy-saving scheme that meets the actual demand is obtained. For example, in a building renovation project, the generative adversarial network can generate different enclosure renovation schemes, energy system upgrade schemes, etc., and evaluate the energy-saving effect and investment return rate of each scheme, thereby providing a scientific basis for decision-making for building managers.

[0079] Further, in the specific implementation, in the building energy consumption management method provided by the embodiment of the present application, the step of evaluating and screening the energy-saving strategy by the discriminator according to the target indicators including the building energy consumption prediction results, cost analysis, and technical feasibility can specifically include: comparing the energy consumption data corresponding to the energy-saving strategy with the building energy consumption prediction results by the discriminator, and judging whether the energy-saving strategy is within the first preset range in terms of energy consumption control according to the deviation degree of the energy consumption data and the building energy consumption prediction results; if not within the first preset range, the energy-saving strategy is excluded; at the same time, the cost analysis is combined with the cost required for implementing the energy-saving strategy, the key indicators including the cost benefit ratio in the energy-saving strategy are calculated, and whether the energy-saving strategy is within the second preset range in terms of cost is evaluated; if not within the second preset range, the energy-saving strategy is excluded; whether the energy-saving strategy is within the third preset range in terms of technical feasibility is judged according to the equipment performance, system compatibility and operation feasibility required by the energy-saving strategy; if not within the third preset range, the energy-saving strategy is excluded.

[0080] In implementation, in order to realize the energy consumption changes under the conditions of simulating different equipment operation parameter adjustment, energy distribution scheme change, building use mode change, etc., evaluate the energy saving effect and economic benefit of each scheme, and combine the real-time energy price information to provide energy procurement and use optimization suggestions for building managers to reduce energy cost, according to the target indicators including building energy consumption prediction results, cost analysis, technical feasibility, discriminators can be used to evaluate and screen energy saving strategies. Specifically, the discriminator will first compare the energy consumption data corresponding to the energy saving strategy with the building energy consumption prediction results, and judge whether the strategy is within the first preset range in terms of energy consumption control according to the deviation degree of the two. If it is not within this range, it means that it is difficult to achieve the expected energy consumption control effect and will be eliminated. At the same time, combined with the cost required for the implementation of the energy saving strategy, the cost analysis is carried out, the key indicators including cost benefit ratio are calculated, and it is evaluated whether the strategy is within the second preset range in terms of cost. If it is not within this range, it means that its economic benefit is not good and it will also be eliminated. In addition, according to the equipment performance, system compatibility and operation feasibility required by the energy saving strategy, it is judged whether it is within the third preset range in terms of technical feasibility. If it is not satisfied, it will be eliminated due to the difficulty in realizing the technical level. Through such multi-dimensional evaluation and screening, the remaining energy saving strategies can better adapt to the actual needs of the building and provide reliable support for the subsequent implementation of energy saving schemes and energy cost optimization.

[0081] Further, in the above building energy consumption management method provided by the embodiments of the present application, the following steps can also be included: receiving instructions for the building; the instructions include query instructions or control instructions; performing semantic understanding on the instructions; extracting relevant information from the knowledge base according to the content understood by the semantic understanding, generating answer content and solutions, and responding and outputting in natural language form; at the same time, processing unstructured data related to building energy consumption, and uploading the processing results to the knowledge base.

[0082] In implementation, the present application adopts natural language processing to provide a convenient human-computer interaction method for building energy consumption management. Users can initiate various query requests in the form of natural language to the system, such as "What is the reason for the increase in building energy consumption this week compared to last week?" and "How to reduce the energy consumption of the conference room?", etc. The present application can use semantic understanding, information retrieval and other technologies to understand the user's questions, quickly extract relevant information from a large amount of data and knowledge base for accurate answers, and provide corresponding solutions and optimization suggestions, so that users can easily obtain the required information. At the same time, users can also issue control instructions through natural language to realize remote operation and management of air conditioning, lighting, elevators and other equipment in the building, without relying on complex operation interfaces, greatly simplifying the management process. In addition, natural language processing can also be used to analyze and process unstructured data such as building energy consumption-related documents and reports, extracting key information such as historical energy consumption patterns and the correlation between equipment failure and energy consumption, and continuously enriching the system's knowledge reserves, providing more comprehensive data support for the optimization of energy consumption prediction models and the formulation of management strategies. This improves the convenience of human-computer interaction, reduces the user's threshold, and enables non-professionals to easily participate in energy consumption management; accelerates the efficiency of information acquisition and instruction execution, helping managers to quickly identify problems and take measures; and by mining the value of unstructured data, enhances the intelligence level and decision support capability of the system, laying a solid foundation for the fine and efficient management of building energy consumption.

[0083] It should be noted that, as the use of buildings and external environment are constantly changing, traditional energy consumption management strategies often fail to adapt to such dynamic changes. The building energy consumption management method of the present application has dynamic adaptive capability, by collecting building historical energy consumption data, environmental parameters, equipment operating status and other information, using generative artificial intelligence learning algorithm to establish energy consumption prediction model, predicting building energy consumption trend in advance, and combining with energy saving strategy to automatically optimize equipment operation mode, realizing intelligent management and efficient energy saving of building energy consumption. The present application can automatically adjust the energy consumption prediction model according to the real-time changes of internal and external conditions, optimize the management of building equipment control strategy and energy saving strategy. For example, when encountering sudden extreme weather changes or changes in building internal activity patterns, the present application can quickly perceive and update the prediction model, re-evaluate the current energy consumption status, and timely adjust the equipment operation scheme and energy distribution strategy to ensure that the building is always in energy-saving operation state. This dynamic adaptive adjustment capability greatly improves the flexibility and effectiveness of building energy consumption management.

[0084] Generative artificial intelligence plays an irreplaceable important role in the building energy consumption management method of the present application, realizing accurate prediction, intelligent decision-making, collaborative control and dynamic management of building energy consumption, i.e. realizing the whole-process optimization from data processing to intelligent management, providing stronger technical support for energy saving and carbon reduction and sustainable development of the building industry.

[0085] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be realized by means of software and a general hardware platform as required, and of course, it can also be realized by hardware, but in many cases, the former is a better embodiment.

[0086] The embodiment of the present application also provides a building energy consumption management device. Figure 2 The structural schematic diagram of the building energy consumption management device provided by the embodiment of the present application is shown in the figure. The embodiment is based on the angle of functional modules, and as shown in the figure, the device comprises: Figure 2

[0087] A data acquisition module 10 is configured to acquire building multi-source heterogeneous data and pre-process the building multi-source heterogeneous data.

[0088] A model inference module 11 is configured to input the processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model and output a building energy consumption prediction result.

[0089] A strategy generation module 12 is configured to obtain a building equipment control strategy by using a reinforcement learning algorithm according to the building energy consumption prediction result, and generate an energy-saving strategy by using a generative adversarial network.

[0090] An optimization decision module 13 is configured to perform equipment collaborative control by using the building equipment control strategy, and optimize the equipment operation mode in combination with the energy-saving strategy, so that the building is in an energy-saving operation state.

[0091] In the building energy consumption management device provided by the embodiment of the present application, the integration and preprocessing of the building multi-source heterogeneous data can provide high-quality input for the machine learning-based energy consumption prediction model, high-precision prediction of building energy consumption can be realized, short-term fluctuations and long-term trends can be accurately captured, and important basis for energy consumption optimization management can be provided. Based on the prediction result, a reinforcement learning algorithm is used to obtain a building equipment control strategy, and a generative adversarial network is used to generate an energy-saving strategy, intelligent energy consumption optimization decision support is provided for building managers, and the dependence on traditional experience is eliminated. The intelligent collaborative control of multiple devices is realized by using the equipment control strategy, and the energy consumption waste and conflicts of single-device independent operation are avoided by optimizing the equipment operation mode in combination with the energy-saving strategy. In this way, real-time response to the use condition and external environment changes in building operation can be realized, dynamic self-adaptive adjustment of building energy consumption management can be promoted, the building can be ensured to be in an energy-saving operation state, and the flexibility and effectiveness of building energy consumption management can be greatly improved.

[0092] ​Since the embodiments of the building energy consumption management device part correspond to the embodiments of the building energy consumption management method part, the description of the features in the embodiments of the building energy consumption management device part can refer to the relevant description of the embodiments of the building energy consumption management method part, which will not be repeated here. And has the same beneficial effects as the above-mentioned building energy consumption management method.

[0093] Further, in specific implementation, in the above-mentioned building energy consumption management device provided by the embodiments of the present application, an instruction processing module can also be included, which is configured to receive instructions for the building; the instructions include query instructions or control instructions; perform semantic understanding on the instructions; extract relevant information from the knowledge base according to the semantic understanding, generate answer content and solutions, and output responses in natural language form; at the same time, process unstructured data related to building energy consumption, and upload the processing results to the knowledge base.

[0094] The embodiments of the present application also provide an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor is configured to run the computer program to execute the steps in any of the above-mentioned building energy consumption management method embodiments.

[0095] The embodiments of the present application also provide a computer readable storage medium, which stores a computer program, wherein the computer program is configured to execute the steps in any of the above-mentioned building energy consumption management method embodiments when running.

[0096] In an exemplary embodiment, the above-mentioned computer readable storage medium can include but is not limited to: a U disk, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and various media that can store computer programs.

[0097] The embodiments of the present application also provide a computer program product, which comprises a computer program, and the computer program is executed by a processor to implement the steps in any of the above-mentioned building energy consumption management method embodiments.

[0098] The embodiments of the present application also provide another computer program product, which comprises a non-volatile computer readable storage medium, the non-volatile computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps in any of the above-mentioned building energy consumption management method embodiments.

[0099] Those skilled in the art will further realize that the mere concepts, teachings, and embodiments described herein are merely meant to provide an enabling description of the claimed invention and are not intended to limit the scope of the claimed invention to these embodiments. Accordingly, modifications and variations of the embodiments described herein are possible in light of the above teachings. It is therefore to be understood that within the scope of the claims and their equivalents, the claimed invention can be practiced otherwise than as specifically described in the description above. For the reader's convenience, the following terms are defined with that same or similar meaning throughout the present specification, unless it clearly appears to the contrary, or is contradicted by context.

[0100] The above describes in detail the building energy consumption management method, device, equipment and medium provided by the present application. The principles and implementation modes of the present application are described by applying specific examples in this paper. The above description of the embodiments is only used to help understand the method of the present application and its core idea. It should be pointed out that, for those skilled in the art, without departing from the principles of the present application, the present application can be improved and modified in many ways. These improvements and modifications also fall within the protection scope of the present application.

Claims

1. A building energy consumption management method, characterized by, The method comprises the following steps: Collecting building multi-source heterogeneous data and preprocessing the building multi-source heterogeneous data; The building multi-source heterogeneous data comprises building energy consumption data, indoor and outdoor environmental parameters, equipment operation data, and personnel activity data; Inputting the processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model to output a building energy consumption prediction result; the energy consumption prediction model is automatically adjusted according to real-time changes in internal and external conditions; the energy consumption prediction model comprises a mapping relationship between energy consumption and multi-source data features that has been constructed; According to the building energy consumption prediction result, a building equipment control strategy is obtained by using a reinforcement learning algorithm, comprising: defining a building equipment control system as an agent, and the internal and external environment of the building as an environment interacting with the agent; in the interaction process between the agent and the environment, according to different control actions, the corresponding reward feedback is obtained in combination with the building energy consumption prediction result; through continuous action trial and error, the building equipment control strategy is learned and optimized to obtain an optimal building equipment control strategy that maximizes the reward; Generating an energy-saving strategy by using a generative adversarial network, comprising: generating an energy-saving strategy by using a generator according to the characteristics and energy-saving goals of the building; comparing the energy consumption data corresponding to the energy-saving strategy with the building energy consumption prediction result by using a discriminator, and judging whether the energy-saving strategy is within a first preset range in terms of energy consumption control according to the deviation degree of the energy consumption data and the building energy consumption prediction result; if not, the energy-saving strategy is excluded; at the same time, a cost analysis is performed in combination with the cost required for implementing the energy-saving strategy, a key indicator containing a cost-benefit ratio in the energy-saving strategy is calculated, and whether the energy-saving strategy is within a second preset range in terms of cost is evaluated; if not, the energy-saving strategy is excluded; whether the energy-saving strategy is within a third preset range in terms of technical feasibility is judged according to the required equipment performance, system compatibility and operation feasibility of the energy-saving strategy; if not, the energy-saving strategy is excluded; the energy-saving strategy is optimized through the adversarial training between the generator and the discriminator to obtain an optimal energy-saving strategy; the building equipment control strategy and the energy-saving strategy are automatically adjusted according to real-time changes in internal and external conditions; The optimal building equipment control strategy is used for equipment collaborative control; the optimal building equipment control strategy comprises adjusting the strength of an air conditioning system, adjusting the ventilation frequency of a ventilation system, and adjusting the brightness of a lighting system; at the same time, the optimal energy-saving strategy is used to optimize the equipment operation mode, so that the building is in an energy-saving operation state.

2. The building energy consumption management method according to claim 1, wherein, Before inputting the processed building multi-source heterogeneous data into the machine learning-based energy consumption prediction model, the following steps are further included: The energy consumption prediction model is constructed by using a convolutional neural network combined with a recurrent neural network; wherein the convolutional neural network is used for processing image data, and the recurrent neural network is used for analyzing time series data to capture the change trend of building energy consumption at different time scales; A historical building multi-source heterogeneous data sample set is obtained; The historical building multi-source heterogeneous data sample set is used to train the energy consumption prediction model. In the training process, the energy consumption prediction model is updated and optimized by using online learning.

3. The building energy consumption management method according to claim 1, wherein, According to different control actions combined with the building energy consumption prediction results, the corresponding reward feedback is obtained, including: When the agent executes each control action, the actual energy consumption corresponding to the control action is associated with the building energy consumption prediction results for analysis. If the actual energy consumption is lower than the building energy consumption prediction results, a first reward feedback is obtained. If the actual energy consumption is equal to the building energy consumption prediction results, a basic reward feedback is obtained; the basic reward feedback is less than the first reward feedback. If the actual energy consumption is higher than the building energy consumption prediction results, a second reward feedback is obtained; the second reward feedback is less than the basic reward feedback.

4. The building energy consumption management method according to claim 1, wherein, Also including: Receiving instructions for the building; the instructions include query instructions or control instructions; Performing semantic understanding on the instructions; According to the content understood by the semantic understanding, relevant information is extracted from the knowledge base to generate answer content and solutions, and is responded to in natural language form; At the same time, process the unstructured data related to building energy consumption, and upload the processing results to the knowledge base.

5. A building energy consumption management device, characterized by, Including: A data acquisition module for collecting building multi-source heterogeneous data and preprocessing the building multi-source heterogeneous data; The building multi-source heterogeneous data includes building energy consumption data, indoor and outdoor environmental parameters, equipment operation data, and personnel activity data; A model inference module for inputting the processed building multi-source heterogeneous data into a machine learning-based energy consumption prediction model to output building energy consumption prediction results; the energy consumption prediction model is automatically adjusted according to real-time changes in internal and external conditions; the energy consumption prediction model includes a mapping relationship between the energy consumption and the characteristics of the multi-source data that has been constructed; The strategy generation module is configured to obtain a building equipment control strategy by using a reinforcement learning algorithm according to the building energy consumption prediction result, including: defining the building equipment control system as an agent, and defining the internal and external environment of the building as an environment interacting with the agent; in the interaction process between the agent and the environment, combining the building energy consumption prediction result according to different control actions to obtain corresponding reward feedback; through continuous action trial and error, learning and optimizing the building equipment control strategy to obtain an optimal building equipment control strategy maximizing the reward; and generating an energy-saving strategy by using a generative adversarial network, including: generating an energy-saving strategy by using a generator according to the characteristics and energy-saving goals of the building; comparing the energy consumption data corresponding to the energy-saving strategy with the building energy consumption prediction result by using a discriminator, and judging whether the energy-saving strategy is within a first preset range in terms of energy consumption control according to the deviation degree of the energy consumption data and the building energy consumption prediction result; if not, the energy-saving strategy is eliminated; at the same time, a cost analysis is performed in combination with the cost required for implementing the energy-saving strategy, a key indicator including a cost benefit ratio in the energy-saving strategy is calculated, and whether the energy-saving strategy is within a second preset range in terms of cost is evaluated; if not, the energy-saving strategy is eliminated; whether the energy-saving strategy is within a third preset range in terms of technical feasibility is judged according to the device performance, system compatibility and operation feasibility required by the energy-saving strategy; if not, the energy-saving strategy is eliminated; the energy-saving strategy is optimized through the adversarial training between the generator and the discriminator to obtain an optimal energy-saving strategy; and the building equipment control strategy and the energy-saving strategy are automatically adjusted according to the real-time changes of internal and external conditions. The optimization decision module is configured to perform equipment collaborative control by using the optimal building equipment control strategy; the optimal building equipment control strategy includes adjusting the strength of an air conditioning system, adjusting the ventilation frequency of a ventilation system and adjusting the brightness of a lighting system; and the optimal energy-saving strategy is combined to optimize the equipment operation mode, so that the building is in an energy-saving operation state.

6. An electronic device, comprising: The building energy consumption management method comprises the following steps: a memory for storing a computer program; a processor for executing the computer program to implement the steps of the building energy consumption management method according to any one of claims 1 to 4.

7. A computer readable storage medium characterized by The computer readable storage medium stores a computer program, and the computer program is executed by the processor to implement the steps of the building energy consumption management method according to any one of claims 1 to 4.

Images