Building energy consumption saving optimization control method based on reinforcement learning

Abstract

The present application relates to the technical field of building management data analysis, and particularly relates to a building energy consumption saving optimization control method based on reinforcement learning. The method comprises the following steps: performing building envelope thermal dynamic response characteristic analysis based on building multi-source heterogeneous energy consumption data and building envelope basic data to generate building envelope thermal dynamic response characteristic data; analyzing building envelope multi-scheme simulation energy consumption characteristic data based on the building envelope thermal dynamic response characteristic data, and analyzing building envelope multi-scheme energy consumption trend data; establishing a building envelope energy saving control learning model based on the building envelope multi-scheme energy consumption trend data; transmitting building operation and maintenance target requirement data to the building envelope energy saving control learning model to perform building energy consumption saving collaborative control relationship analysis, so as to design a building energy consumption saving optimization control strategy. The present application realizes intelligent control of building energy consumption saving optimization.

Description

Technical Field

[0001] This invention relates to the field of building management data analysis technology, and in particular to a building energy consumption optimization and control method based on reinforcement learning. Background Technology

[0002] Thermal conduction loss in building envelopes is one of the main sources of energy waste in buildings, directly affecting energy efficiency and operation and maintenance costs. Optimizing building energy consumption requires meeting comfort and safety goals during operation and maintenance while minimizing energy consumption to achieve a synergistic balance between energy saving and operational objectives. However, existing building energy consumption control methods largely focus on adjusting building equipment operation, lacking a systematic analysis of the building envelope's own thermal dynamics and material thermal conductivity. They fail to fully consider the intrinsic relationship between the building envelope and building energy consumption. While some energy consumption optimization methods attempt to introduce simulation technology for analysis, they are mostly limited to simulating the energy consumption of a single envelope scheme, lacking systematic simulation comparison analysis of multiple envelope material schemes. This makes it difficult to accurately capture the differences in energy consumption trends under different material combinations and ratios, and thus difficult to select the optimal envelope scheme. Summary of the Invention

[0003] Based on this, the present invention provides a building energy consumption optimization and control method based on reinforcement learning to solve at least one of the above-mentioned technical problems.

[0004] To achieve the above objectives, a building energy consumption optimization and control method based on reinforcement learning includes the following steps: Step S1: Obtain building multi-source heterogeneous energy consumption data and building envelope basic data; perform thermal dynamic response characteristic analysis of building envelope based on building multi-source heterogeneous energy consumption data and building envelope basic data, and generate building envelope thermal dynamic response characteristic data. Step S2: Based on the basic data of the building envelope and the thermal dynamic response characteristic data of the building envelope, perform material thermal conductivity characteristic analysis of the building envelope to generate material thermal conductivity characteristic data of the building envelope. Step S3: Based on the thermal conductivity characteristics data of the building envelope materials, perform simulation energy consumption characteristics analysis of multiple schemes of building envelope materials to generate simulation energy consumption characteristic data of multiple schemes of building envelope materials; Step S4: Perform multi-scheme energy consumption trend analysis on the multi-scheme simulation energy consumption characteristic data of the building envelope, and generate multi-scheme energy consumption trend data of the building envelope. Step S5: Obtain building operation and maintenance target demand data; establish a building envelope energy-saving control learning model based on the energy consumption trend data of multiple schemes of building envelope; transmit the building operation and maintenance target demand data to the building envelope energy-saving control learning model to conduct collaborative control relationship analysis of building energy consumption and energy saving, so as to design building energy consumption and energy saving optimization control strategy.

[0005] Furthermore, step S1 includes the following steps: Step S11: Obtain building multi-source heterogeneous energy consumption data and basic data of building envelope; Step S12: Perform time-series alignment and spatial mapping of building multi-source heterogeneous energy consumption data and building envelope basic data to analyze building multi-source characterization attributes and generate building multi-source characterization attribute data; Step S13: Design the building equivalent thermal network envelope based on the building multi-source characterization attribute data, and generate the building equivalent thermal network envelope data; Step S14: Perform thermal parameter identification and solution processing on the building equivalent thermal network envelope data to generate building identification thermal parameter data; Step S15: Analyze the thermal dynamic response characteristics of the building envelope based on the building equivalent thermal network envelope data and the building identified thermal parameter data, and generate thermal dynamic response characteristic data of the building envelope.

[0006] Furthermore, the multi-source heterogeneous energy consumption data of the building mentioned in step S11 includes building environment data, building equipment operation data, and building equipment energy consumption data, and the basic data of the building envelope includes building envelope geometric data and building envelope concrete mix proportion data.

[0007] Furthermore, step S13 includes the following steps: Step S131: Analyze the thermal state and boundary characteristics of the building envelope based on the multi-source characterization attribute data of the building, and generate thermal state boundary characteristic data of the building envelope; Step S132: Based on the multi-source characterization attribute data of the building, perform structural characteristic classification processing of the building envelope to generate building envelope classification data; Step S133: Design the equivalent thermal network envelope structure by using the thermal state boundary feature data of the building envelope structure and the building envelope structure division data, and generate the equivalent thermal network envelope structure data.

[0008] Furthermore, step S2 includes the following steps: Step S21: Extract the concrete mix proportion area of ​​the building envelope structure based on the basic data of the building envelope structure, and generate concrete mix proportion area data of the building envelope structure. Step S22: Perform thermal conductivity analysis of the concrete mix proportion area based on the concrete mix proportion area data of the building envelope structure, and generate thermal conductivity data of the concrete mix proportion area based on thermal property mapping. Step S23: Analyze the neighborhood thermal conductivity influence characteristics of the thermal dynamic response of the building envelope on the thermal conductivity data of the concrete mix proportion area mapped by the thermal property data, and generate neighborhood thermal conductivity influence characteristic data of the thermal dynamic response. Step S24: Based on the thermal dynamic response neighborhood thermal conductivity influence characteristic data and the thermal property mapping concrete mix proportion region thermal conductivity data, perform material thermal conductivity characteristic analysis of the building envelope structure to generate material thermal conductivity characteristic data of the building envelope structure.

[0009] Furthermore, step S3 includes the following steps: Step S31: Discretize the concrete mix proportion data of the building envelope to generate discrete concrete mix proportion data. Step S32: Based on the discretized data of concrete mix proportions, design simulation conditions for multiple schemes of building envelope materials to obtain simulation condition data for multiple schemes of building envelope materials; Step S33: Analyze the boundary conditions for building energy consumption simulation based on the building's multi-source characterization attribute data, and generate building energy consumption simulation boundary condition data; Step S34: Based on the multi-scheme simulation condition data of building envelope materials and the boundary condition data of building energy consumption simulation, perform multi-scheme simulation thermal balance processing of building envelope to generate multi-scheme simulation thermal balance data of building envelope. Step S35: Perform multi-scheme simulation energy consumption characteristic analysis on the multi-scheme simulation thermal balance data of the building envelope, and generate multi-scheme simulation energy consumption characteristic data of the building envelope.

[0010] Furthermore, step S32 includes the following steps: Step S321: Perform material-specific combination characteristic analysis based on the discretized data of concrete mix proportions to generate material-specific combination characteristic data; Step S322: Perform thermal conductivity characteristic mapping on the material-specific combination characteristic data using the thermal conductivity characteristic data of the building envelope material to generate material combination thermal conductivity characteristic mapping data; Step S323: Design simulation conditions for multiple schemes of building envelope materials based on the thermal conductivity characteristic mapping data of material combination, and obtain simulation condition data for multiple schemes of building envelope materials.

[0011] Furthermore, step S4 includes the following steps: Step S41: Analyze the time-series changes in the thermal state of the building envelope using the thermal dynamic response characteristic data of the building envelope and the simulation energy consumption characteristic data of the building envelope under multiple schemes, and generate time-series change characteristic data of the thermal state of the building envelope under multiple schemes. Step S42: Establish an optimized trend prediction model for the thermal state of the multi-scheme building envelope based on the time-series variation characteristic data of the thermal state of the multi-scheme building envelope; Step S43: By optimizing the multi-scheme building envelope thermal state trend prediction model, perform multi-scheme building envelope simulation energy consumption trend analysis on the multi-scheme building envelope energy consumption characteristic data, and generate multi-scheme building envelope energy consumption trend data.

[0012] Furthermore, step S42 includes the following steps: A model for the trend change of the thermal state of a multi-scheme enclosure structure is established by using a pre-defined long short-term memory neural network algorithm and time-series change characteristic data of the thermal state of the multi-scheme enclosure structure. This model is used to obtain a prediction model for the thermal state of the multi-scheme enclosure structure. The prediction model is then subjected to particle swarm optimization of the trend prediction relationship of the multi-scheme enclosure structure to obtain an optimized prediction model for the thermal state of the multi-scheme enclosure structure.

[0013] Furthermore, step S5 includes the following steps: Step S51: Obtain building operation and maintenance target requirements data; Step S52: Based on the energy consumption trend data of multiple schemes for the building envelope, perform state representation spatial design of the energy consumption trend of multiple schemes to obtain the state representation spatial data of the energy consumption trend of multiple schemes. Step S53: Extract the influencing factors of the multi-scheme energy consumption trend of the building envelope based on the multi-scheme energy consumption trend data, and generate multi-scheme energy consumption trend influencing factor data of the building envelope. Step S54: Design the action space of the energy consumption trend of multiple schemes by using the data of the factors affecting the energy consumption trend of the building envelope, and obtain the action space data of the energy consumption trend of multiple schemes. Step S55: Analyze the collaborative relationship of the influence on the action space based on the energy consumption trend of multiple schemes, and generate collaborative relationship data of the influence on the action space; Step S56: Based on the preset reinforcement learning algorithm, establish the energy-saving control learning relationship of the building envelope structure by using the spatial data of the energy consumption trend state representation of multiple schemes and the spatial data of the energy consumption trend influence of multiple schemes, and generate the energy-saving control learning model of the building envelope structure. Step S57: Transmit the building operation and maintenance target requirement data to the building envelope energy-saving control learning model to analyze the collaborative control relationship of building energy consumption and energy saving, so as to design a building energy consumption and energy saving optimization control strategy, wherein the building energy consumption and energy saving optimization control strategy includes building envelope concrete mix control data and building equipment operation control data.

[0014] The beneficial effects of this application are that it acquires multi-source heterogeneous energy consumption data of buildings and basic data of building envelopes, and then, through a series of processes such as time-series alignment and spatial mapping of multi-source building characterization attribute analysis, building equivalent thermal network envelope design, identification and solution of building envelope thermal parameters, and analysis of building envelope thermal dynamic response characteristics, it clarifies the thermal dynamic response characteristics of building envelopes, overcomes the limitations of single and one-sided detection of building envelope characteristics, and makes the analysis of building envelope thermal characteristics more comprehensive and accurate; subsequently, based on the basic data of building envelopes and thermal dynamic response characteristics... The data undergoes processing including extraction of concrete mix proportions for building envelope structures, thermal conductivity analysis of concrete mix proportions through thermophysical mapping, analysis of the influence of neighborhood thermal conductivity on thermal dynamic response, and comprehensive analysis of the thermal conductivity characteristics of building envelope materials. This addresses the problem of inaccurately capturing the thermal conductivity laws of building envelope materials, achieving precise transfer and quantitative analysis from the basic characteristics of the building envelope to the thermal conductivity characteristics of the materials. This avoids subsequent simulation deviations caused by inaccurate material thermal conductivity analysis, ensuring that the generated thermal conductivity characteristic data of building envelope materials accurately reflects their thermal conductivity laws. Furthermore, through precise processing of the building equivalent thermal network design and the identification and solution of building envelope thermal parameters, the analysis accuracy of thermal dynamic response characteristics is further optimized, ensuring that the thermal dynamic response characteristic data accurately reflects the essential characteristics and core influencing factors of heat transfer in the building envelope. Based on the thermal conductivity characteristics of building envelope materials, this study performed discretization of concrete mix design, design of simulation conditions for multiple building envelope material schemes, analysis of boundary conditions for building energy consumption simulation, thermal balance simulation of multiple building envelope schemes, and energy consumption characteristic analysis of multiple building envelope schemes. This clearly outlined the energy consumption trends and differences under different building envelope material schemes, providing intuitive and accurate simulation data support for subsequent energy consumption trend analysis and energy-saving control strategy design for multiple building envelope schemes. Based on the thermal dynamic response characteristics of building envelopes and the energy consumption characteristics of multiple schemes, this study analyzed the temporal changes in the thermal state of building envelopes under multiple schemes, optimized the construction of a predictive model for the thermal state of building envelopes under multiple schemes, and conducted energy consumption trend analysis of multiple building envelope schemes. This process gradually deconstructed the evolution of building energy consumption under different schemes, accurately identified the energy consumption differences and energy-saving potential of each scheme, clarified the dynamic mechanism of energy consumption trends, and enabled early prediction of building energy consumption changes under different schemes. This avoids the problem of unreasonable energy-saving control strategy design due to inaccurate energy consumption trend prediction.Based on multi-scheme energy consumption trend data and obtained building operation and maintenance target demand data, this study conducts state representation space design for multi-scheme energy consumption trends, extraction of influencing factors of multi-scheme energy consumption trends in the building envelope, design of the influence action space for multi-scheme energy consumption trends, analysis of the collaborative relationship of the influence action space, establishment of a learning model for energy-saving control of the building envelope, and design of energy-saving optimization control strategies for building energy consumption. This provides a foundation for intelligent control of building energy consumption. Subsequently, the control strategy is optimized through reinforcement learning algorithms and multi-dimensional analysis results. The evolution law and interrelationship of each control parameter are clearly identified, achieving optimal collaborative energy saving between the building envelope and building equipment. This avoids the problem of control strategies lagging behind changes in building operating status, accurately optimizes the control method of building energy consumption, ensures efficient energy saving of building energy consumption, and can specifically prevent and control the problem of building energy waste, while taking into account the stability of building operation and maintenance target requirements.

[0015] Therefore, the reinforcement learning-based building energy-saving optimization control method of this invention takes the characteristics of the building envelope as its core starting point, overcoming the limitations of existing technologies that mainly focus on the operation and adjustment of building equipment. It achieves a systematic analysis of the building envelope's own thermal dynamic characteristics and material thermal conductivity, fully considering the intrinsic relationship between the building envelope and building energy consumption. Simultaneously, through systematic simulation and comparative analysis of multiple material schemes for the building envelope, it can accurately capture the differences in energy consumption trends under different material combinations and ratios, effectively solving the problem of simulation being limited to a single scheme and making it difficult to select the optimal building envelope scheme. Furthermore, by combining reinforcement learning algorithms to establish a building envelope energy-saving control learning model, it can achieve coordinated control of building energy saving and operation and maintenance goals based on building operation and maintenance objectives, adapting to complex and ever-changing building operating environments and operation and maintenance needs, thereby improving the effectiveness and intelligence level of building energy-saving optimization control. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the steps of a reinforcement learning-based building energy consumption optimization and control method according to the present invention. Figure 2 for Figure 1 A detailed flowchart illustrating the implementation steps of step S3. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0018] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. The same reference numerals in the drawings denote the same or similar parts, and therefore repeated descriptions of them will be omitted. Some block diagrams shown in the drawings are functional entities and do not necessarily correspond to physically or logically independent entities. Functional entities may be implemented in software, or in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods. The term "and / or" as used herein includes any and all combinations of one or more of the associated items listed.

[0019] To achieve the above objectives, please refer to Figures 1 to 2 This invention provides a building energy consumption optimization and control method based on reinforcement learning. In the embodiments of this invention, please refer to... Figure 1 The diagram shown is a flowchart illustrating the steps of a reinforcement learning-based building energy conservation optimization and control method according to the present invention. The reinforcement learning-based building energy conservation optimization and control method includes the following steps: Step S1: Obtain building multi-source heterogeneous energy consumption data and building envelope basic data; perform thermal dynamic response characteristic analysis of building envelope based on building multi-source heterogeneous energy consumption data and building envelope basic data, and generate building envelope thermal dynamic response characteristic data. In this embodiment of the invention, multi-source heterogeneous energy consumption data of buildings and basic data of building envelope structures are collected. Both types of data are obtained from the actual operation of the building and the structure itself, ensuring that the data reflects the building's energy consumption and basic characteristics of the building envelope structure. After the data collection is completed, the thermal dynamic response characteristics of the building envelope structure are analyzed based on the two types of data. The analysis process focuses on the heat conduction law, temperature change response law, and heat accumulation characteristics of the building envelope structure under different environmental conditions. Through correlation analysis of the two types of data, the heat transfer path, thermal response delay characteristics, and thermal balance state of each part of the building envelope structure are clarified, accurately capturing the intrinsic correlation between the thermal dynamic changes of the building envelope structure and energy consumption data, and generating thermal dynamic response characteristic data of the building envelope structure.

[0020] Step S2: Based on the basic data of the building envelope and the thermal dynamic response characteristic data of the building envelope, perform material thermal conductivity characteristic analysis of the building envelope to generate material thermal conductivity characteristic data of the building envelope. In this embodiment of the invention, the thermal conductivity characteristics of the building envelope are analyzed based on both basic building envelope data and thermal dynamic response characteristic data, clarifying the thermal conductivity performance patterns and influencing factors of the building envelope materials. Combining the structural geometry and material composition information contained in the basic building envelope data with characteristics such as heat transfer efficiency and temperature change rate in the thermal dynamic response characteristic data, the analysis focuses on the correlation between material thermal conductivity and thermal dynamic response, exploring the influence of different material compositions and structural configurations on thermal conductivity, and identifying the core influencing factors and mechanisms of material thermal conductivity. Through systematic analysis of the changes in material thermal conductivity with thermal dynamic response, core characteristic parameters of material thermal conductivity are extracted, generating thermal conductivity characteristic data of the building envelope materials, providing a clear basis for subsequent multi-scheme energy consumption simulation.

[0021] Step S3: Based on the thermal conductivity characteristics data of the building envelope materials, perform simulation energy consumption characteristics analysis of multiple schemes of building envelope materials to generate simulation energy consumption characteristic data of multiple schemes of building envelope materials; In this embodiment of the invention, the thermal conductivity characteristics of building envelope materials are used as the core basis for simulation analysis of energy consumption characteristics under multiple schemes of building envelope materials. The core is to compare the differences in building energy consumption under different material schemes through multi-scheme simulation, providing data support for subsequent energy consumption trend analysis. Based on the thermal conductivity characteristics of materials, multiple building envelope material schemes are designed. Each scheme has clear differences in material composition and structural construction, ensuring that the energy consumption characteristics of the schemes are comparable. Through simulation, the heat exchange process and energy consumption generation process of the building envelope under different material schemes are simulated, capturing the energy consumption change law, energy consumption distribution characteristics, and core information such as energy consumption peak and steady-state energy consumption under each scheme, comprehensively analyzing the impact of different material schemes on building energy consumption. After the simulation analysis is completed, the simulation data of each scheme are integrated, the energy consumption characteristic parameters of each scheme are extracted, and multi-scheme simulated energy consumption characteristic data of building envelope is generated.

[0022] Step S4: Perform multi-scheme energy consumption trend analysis on the multi-scheme simulation energy consumption characteristic data of the building envelope, and generate multi-scheme energy consumption trend data of the building envelope. In this embodiment of the invention, the core analysis object is the simulated energy consumption characteristic data of multiple schemes for building envelopes. The analysis focuses on the energy consumption trend of building envelopes under different material schemes, aiming to uncover the changing patterns and future trends of energy consumption, providing trend data support for the subsequent construction of energy-saving control models. During the analysis, the thermal dynamic response characteristic data of the building envelopes are correlated, and the simulated energy consumption characteristics of each scheme are combined. The analysis focuses on the changing patterns of energy consumption over time, the differences in energy consumption distribution among different envelope units, and the comparative patterns of energy consumption between different schemes. The system analyzes the correlation between energy consumption changes and thermal dynamic response characteristics, capturing the core influencing factors of energy consumption trends. Through time-series analysis and trend extraction of energy consumption data for each scheme, the rising and falling patterns and steady-state trends of energy consumption under different schemes are clarified. The intrinsic correlation between energy consumption trends and material thermal conductivity and thermal dynamic response characteristics is analyzed, generating multi-scheme energy consumption trend data for building envelopes.

[0023] Step S5: Obtain building operation and maintenance target demand data; establish a building envelope energy-saving control learning model based on the energy consumption trend data of multiple schemes of building envelope; transmit the building operation and maintenance target demand data to the building envelope energy-saving control learning model to conduct collaborative control relationship analysis of building energy consumption and energy saving, so as to design building energy consumption and energy saving optimization control strategy.

[0024] In this embodiment of the invention, building operation and maintenance target demand data is collected. This data is set around three core objectives: building energy consumption optimization, indoor thermal comfort assurance, and operation and maintenance cost control. It clarifies the energy consumption control baseline, thermal comfort standards, and cost constraints, providing clear target guidance for subsequent energy-saving control strategy design. Subsequently, based on the energy consumption trend data of multiple schemes for the building envelope, an energy-saving control learning model for the building envelope is established. This model is constructed based on reinforcement learning logic. By associating energy consumption trend states with adjustable actions, it discovers the optimal control actions under different energy consumption states, clarifies the correlation between states, actions, and energy-saving effects, and achieves autonomous learning of energy-saving control logic. After the model is established, the building operation and maintenance target demand data is incorporated into the analysis to conduct a collaborative control relationship analysis of building energy consumption. Combined with operation and maintenance target constraints, the optimal combination of energy-saving control actions that meets the target requirements is selected, and a building energy consumption optimization control strategy is designed. This strategy includes two core components: concrete mix ratio control of the building envelope and building equipment operation control, ensuring that the strategy can achieve multiple objectives of energy consumption optimization, thermal comfort standards, and controllable operation and maintenance costs.

[0025] Furthermore, step S1 includes the following steps: Step S11: Obtain building multi-source heterogeneous energy consumption data and basic data of building envelope; In this embodiment of the invention, multi-source heterogeneous energy consumption data and basic data of the building envelope of the target building are collected. The multi-source heterogeneous energy consumption data includes building environment data, building equipment operation data, and building equipment energy consumption data. The building environment data covers the real-time temperature, humidity, and light intensity of each floor and functional area inside the building, and the real-time ambient temperature, wind speed, and solar radiation intensity outside the building. The collection cycle is set to once every 15 minutes, and the data is collected continuously for 30 days. The building equipment operation data covers the operating status, start-up and shutdown time, and operating load of the air conditioning system, lighting system, and ventilation system. The building equipment energy consumption data covers the hourly power consumption and water consumption of each device. The basic data of the building envelope includes the geometric data of the building envelope and the concrete mix proportion data of the building envelope. The geometric data covers the wall thickness, door and window size, roof area, exterior wall area, building floor height, and the layout of each envelope component. The concrete mix proportion data covers the mixing ratio of cement, sand, gravel, and water in the concrete used for the exterior walls, roof, and floor slabs, as well as the type and amount of admixtures.

[0026] Step S12: Perform time-series alignment and spatial mapping of building multi-source heterogeneous energy consumption data and building envelope basic data to analyze building multi-source characterization attributes and generate building multi-source characterization attribute data; In this embodiment of the invention, multi-source characterization attribute analysis is performed on the collected multi-source heterogeneous energy consumption data of buildings and the basic data of building envelope structures through time-series alignment and spatial mapping. Time-series alignment uses 15-minute time granularity to adjust equipment operation data and environmental data from different collection periods to the same time dimension, ensuring the correspondence of various types of data at the same time node. For example, the hourly operating load data of the air conditioning system is split into four corresponding 15-minute time periods to accurately match the ambient temperature data of the same period. Spatial mapping uses each functional area of ​​the building as the basic unit to associate the geometric data of the building envelope structure with the energy consumption data and environmental data of the corresponding area, clarifying the energy consumption impact range of each envelope structure component. For example, the geometric data of the exterior wall is associated with the indoor and outdoor temperature data and air conditioning energy consumption data of the corresponding area, and the geometric data of the roof is associated with the solar radiation intensity data and lighting energy consumption data of the roof area. Through time-series alignment and spatial mapping, the characterization attributes of various types of data are extracted to generate multi-source characterization attribute data of buildings. This data includes the correlation characteristics of energy consumption, environment, and building envelope structure at each time node and in each spatial area.

[0027] Step S13: Design the building equivalent thermal network envelope based on the building multi-source characterization attribute data, and generate the building equivalent thermal network envelope data; In this embodiment of the invention, the equivalent thermal network envelope structure of a building is designed based on multi-source characterization attribute data of the building. First, based on the geometric data and environmental data of the envelope structure in the multi-source characterization attribute data of the building, the thermal state and boundary characteristics of the building envelope structure are analyzed to clarify the heat exchange path between the interior and exterior of the building. The building envelope structure is divided into four independent heat conduction units: exterior walls, roof, floors, and doors and windows. Each unit serves as a core node of the equivalent thermal network. Then, based on the geometric dimensions and material properties of each unit, the thermal resistance and heat capacity parameters of each node are set. The thermal resistance parameter is calculated based on the thickness of the envelope structure and the thermal conductivity of the material, while the heat capacity parameter is calculated based on the volume of the envelope structure, the density of the material, and the specific heat capacity. An equivalent thermal network composed of thermal nodes, thermal resistance, and heat capacity is constructed. Each thermal node corresponds to an envelope structure unit, and the nodes are connected by thermal resistance to simulate the heat conduction process between the units. Heat exchange paths are established between the nodes and the interior and exterior environments of the building to simulate indoor and outdoor heat exchange, generating equivalent thermal network envelope structure data of the building. This data includes the number of nodes in the equivalent thermal network, the parameters of each node, the connection relationship between nodes, and the heat exchange path parameters.

[0028] Step S14: Perform thermal parameter identification and solution processing on the building equivalent thermal network envelope data to generate building identification thermal parameter data; In this embodiment of the invention, the thermal parameters of the building envelope are identified and solved using the equivalent thermal network envelope data. The least squares method is used as the thermal parameter identification algorithm, and a thermal parameter identification objective function is constructed. The objective function is to minimize the sum of squares of the deviations between the simulated temperature of the building envelope and the actual collected temperature of the building envelope. Its mathematical expression is: ,in The objective function value, This represents the total number of data collections. Let K be the temperature of the building envelope in the k-th equivalent thermal network simulation. The temperature of the building envelope is the actual temperature collected in the kth iteration. The thermal parameters to be identified include the thermal resistance and heat capacity of each equivalent thermal network node. The temperature data in the multi-source characterization attribute data of the building are substituted into the objective function. The thermal resistance and heat capacity parameter values ​​that minimize the objective function are solved through iterative calculation. During the iteration process, the thermal parameter values ​​are adjusted each time, and the objective function value is recalculated until the objective function value converges. The convergence condition is that the difference between the objective function values ​​of two adjacent iterations is less than 0.01. The building identification thermal parameter data is generated. This data contains the accurate thermal resistance and heat capacity parameters of each equivalent thermal network node, ensuring that the equivalent thermal network can reflect the thermal conduction characteristics of the building envelope.

[0029] Step S15: Analyze the thermal dynamic response characteristics of the building envelope based on the building equivalent thermal network envelope data and the building identified thermal parameter data, and generate thermal dynamic response characteristic data of the building envelope.

[0030] In this embodiment of the invention, the thermal dynamic response characteristics of the building envelope are analyzed based on the building equivalent thermal network envelope data and the building identification thermal parameter data. The thermal resistance and heat capacity parameters from the building identification thermal parameter data are substituted into the building equivalent thermal network to construct the thermal dynamic response equation of the envelope. The thermal dynamic response equation is described by a first-order linear differential equation, and its mathematical expression is: Where C is the node heat capacity, R is the node thermal resistance, T is the node temperature, t is time, and q is the node heat flux density; by substituting indoor and outdoor temperatures, solar radiation intensity, and other thermal excitation data from the building's multi-source characterization attribute data into the thermal dynamic response equation, and solving the differential equation, the temperature change curves of each node of the building envelope under different thermal excitations are obtained. The rise rate, fall rate, and steady-state value of the temperature change curves are analyzed to extract the thermal response law of the building envelope under different environmental conditions. For example, the temperature response delay time, temperature peak value, and time to reach the peak value of the exterior wall nodes are analyzed when the solar radiation intensity changes, and the heat conduction rate of the roof nodes is analyzed when the indoor and outdoor temperature difference changes. The thermal dynamic response characteristic data of the building envelope is generated, which includes characteristic parameters such as the thermal response delay time, temperature change rate, steady-state temperature, and heat conduction efficiency of each building envelope unit.

[0031] Furthermore, the multi-source heterogeneous energy consumption data of the building mentioned in step S11 includes building environment data, building equipment operation data, and building equipment energy consumption data, and the basic data of the building envelope includes building envelope geometric data and building envelope concrete mix proportion data.

[0032] Furthermore, step S13 includes the following steps: Step S131: Analyze the thermal state and boundary characteristics of the building envelope based on the multi-source characterization attribute data of the building, and generate thermal state boundary characteristic data of the building envelope; In this embodiment of the invention, the thermal state and boundary characteristics of the building envelope are analyzed based on multi-source characterization attribute data. This data includes the correlation characteristics of energy consumption, environment, and building envelope at various time points and spatial regions. Core information such as the geometric dimensions of the building envelope, indoor and outdoor temperatures, solar radiation intensity, and initial heat flux data are extracted to analyze the real-time thermal state and boundary characteristics of the building envelope. Thermal state analysis focuses on the temperature distribution, heat conduction direction, and heat accumulation degree of various parts of the building envelope. By comparing the temperature changes of the same part of the building envelope at different time points, the pattern of thermal state changes with the environment is clarified. Boundary characteristic analysis revolves around the heat exchange interface between the building envelope and the indoor and outdoor environments. The scope, type, and heat exchange mode of the heat exchange boundary are determined, distinguishing the areas of convective heat transfer, radiative heat transfer, and conductive heat transfer. The difference in heat flux density at each boundary is calculated to clarify the main paths and intensities of indoor and outdoor heat exchange. Based on the above analysis, thermal state boundary characteristic data of the building envelope is generated. This data includes real-time temperature, heat flux density, heat conduction direction, boundary heat exchange type, and heat exchange intensity parameters for various parts of the building envelope.

[0033] Step S132: Based on the multi-source characterization attribute data of the building, perform structural characteristic classification processing of the building envelope to generate building envelope classification data; In this embodiment of the invention, the structural characteristics of the building envelope are classified based on multi-source characterization attribute data. The geometric data and material parameters of the building envelope in the multi-source characterization attribute data serve as the basis for classification. The building envelope is classified according to its structural form, material composition, and thermal conductivity characteristics, with clear classification standards and boundaries. The building envelope is divided into four primary units: exterior wall unit, roof unit, floor slab unit, and door and window unit. The exterior wall unit is further divided into four secondary units according to orientation: south wall, north wall, east wall, and west wall. Exterior wall units with different orientations have different thermal conductivity characteristics due to different solar radiation reception. The roof unit is divided into two secondary units according to the insulation layer construction: flat roof and pitched roof. The insulation layer thickness and laying method of each unit are clearly defined. The floor slab unit is divided into three secondary units according to the number of floors: ground floor slab, intermediate floor slab, and top floor slab. The differences in heat exchange between each floor slab and the soil, as well as between adjacent floors, are distinguished. The door and window unit is divided into two secondary units according to the material: glass doors and windows and aluminum alloy doors and windows. The material thickness and sealing performance of each unit are clearly defined. Each partitioned unit corresponds to unique geometric parameters, material parameters, and spatial location information, generating building envelope partitioning data. This data includes the level, number, geometric dimensions, material composition, and spatial location of the partitioned units. Through precise partitioning, targeted thermal analysis of different building envelope units can be achieved, providing a clear unit basis for the subsequent design of equivalent thermal network nodes.

[0034] Step S133: Design the equivalent thermal network envelope structure by using the thermal state boundary feature data of the building envelope structure and the building envelope structure division data, and generate the equivalent thermal network envelope structure data.

[0035] In this embodiment of the invention, the equivalent thermal network of the building envelope is designed using thermal state boundary feature data and building envelope segmentation data. Each segmentation unit is treated as an independent node in the equivalent thermal network, with node numbers corresponding one-to-one with the segmentation units, ensuring that each node accurately corresponds to a specific area of ​​the building envelope. Based on the heat flux density and heat conduction direction in the thermal state boundary feature data of the building envelope, combined with the geometric dimensions and material parameters of the segmentation units, the thermal resistance and heat capacity parameters of each node are calculated. The thermal resistance calculation model is as follows: ,in To divide the thickness of the unit, To determine the thermal conductivity of the unit material; the heat capacity calculation model is as follows: ,in To divide the volume of the unit, To classify the density of unit materials, To classify the specific heat capacity of the unit materials, the connection relationships between nodes are designed based on the heat exchange type and intensity in the thermal state boundary characteristic data of the building envelope. Nodes corresponding to the same level of unit division are connected by thermal resistance to simulate the heat conduction process between units. Heat exchange paths are established between the exterior wall, roof, and door / window nodes and the outdoor environment, while heat exchange paths are established between the floor slab nodes and adjacent floor nodes. All nodes establish heat exchange paths with the indoor environment, and the thermal resistance parameters of the heat exchange paths are determined based on the boundary heat exchange intensity. Through the above design, a complete equivalent building thermal network is constructed, generating equivalent building thermal network envelope data. This data includes network node numbers, thermal resistance and heat capacity parameters of each node, connection relationships between nodes, and heat exchange path parameters, enabling accurate simulation of the heat conduction process of the building envelope.

[0036] Furthermore, step S2 includes the following steps: Step S21: Extract the concrete mix proportion area of ​​the building envelope structure based on the basic data of the building envelope structure, and generate concrete mix proportion area data of the building envelope structure. In this embodiment of the invention, the concrete mix design regions of the building envelope are extracted based on the basic data of the building envelope. The basic data includes geometric data and concrete mix design data. The geometric data specifies the spatial location, size range, and boundary division of each part of the exterior walls, roof, and floor slabs. The concrete mix design data specifies the mixing ratio of cement, sand, gravel, and water, as well as the admixture parameters, for each part of the concrete. Based on the unit boundaries in the building envelope division data, the building envelope is divided into regions according to the differences in concrete mix design. Regions with the same mix design parameter are divided into an independent mix design region, and regions with different mix design parameters are divided into different mix design regions. For exterior wall units, multiple mix design regions corresponding to the four orientations (south, north, east, west) and concrete mix design differences are extracted. For roof units, mix design regions corresponding to flat roofs and pitched roofs are extracted based on the insulation layer structure and concrete mix design differences. For floor slab units, mix design regions corresponding to the ground floor, middle floors, and top floor are extracted based on the floor level and concrete mix design differences. Each mix designation area clearly corresponds to the building envelope structural unit, spatial location, geometric dimensions, and complete concrete mix design parameters, generating concrete mix designation area data for the building envelope structure.

[0037] Step S22: Perform thermal conductivity analysis of the concrete mix proportion area based on the concrete mix proportion area data of the building envelope structure, and generate thermal conductivity data of the concrete mix proportion area based on thermal property mapping. In this embodiment of the invention, a thermal conductivity analysis of the concrete mix proportion region is performed based on the thermal property mapping of the concrete mix proportion region data of the building envelope. The concrete mix proportion parameters and regional geometric dimensions in the concrete mix proportion region data of the building envelope are the core analysis basis. A thermal property mapping model between concrete mix proportion and thermal conductivity is constructed. The model structure is a linear correlation model, for example... ,in The thermal conductivity of concrete, This refers to the proportion of cement by weight. The percentage of sand and gravel by weight. For water quality percentage, This refers to the mass percentage of the admixture. These are model coefficients, determined through calibration using experimental data on the thermal properties of concrete materials. Training data comprises concrete mix proportions and their measured thermal conductivity data for each mix proportion region. The thermal conductivity of concrete with different mix proportions is measured experimentally, and the mix proportions and measured thermal conductivity are substituted into the model to obtain fixed coefficients through calibration. After determining the model parameters, the concrete mix proportion parameters of each mix proportion region are substituted into the model to calculate the thermal conductivity of each mix proportion region. At the same time, the uniformity of thermal conductivity distribution within the region is analyzed to clarify the variation law of thermal conductivity within the same mix proportion region, and thermal property mapping of concrete mix proportion region thermal conductivity data is generated.

[0038] Step S23: Analyze the neighborhood thermal conductivity influence characteristics of the thermal dynamic response of the building envelope on the thermal conductivity data of the concrete mix proportion area mapped by the thermal property data, and generate neighborhood thermal conductivity influence characteristic data of the thermal dynamic response. In this embodiment of the invention, the thermal dynamic response characteristic data of the building envelope is used to analyze the neighborhood thermal conductivity influence characteristics of the thermal property mapping concrete mix design data. The building envelope thermal dynamic response characteristic data includes parameters such as the thermal response delay time, temperature change rate, and heat transfer efficiency of each envelope unit. The thermal property mapping concrete mix design data includes the thermal conductivity and distribution characteristics of each mix design region. A neighborhood division standard is determined: taking each mix design region as the core, mix design regions adjacent to this region and closely related in heat exchange are defined as its neighborhood. The scope and correlation strength of the neighborhood are clarified. Adjacent mix design regions within the same envelope unit are direct neighbors, while mix design regions in different envelope units but connected by heat transfer paths are indirect neighbors. A neighborhood thermal conductivity influence model is constructed, with the mathematical expression as follows: ,in Let represent the change in thermal conductivity in the i-th proportioning region. The influence coefficient, The number of neighboring regions. Let J be the temperature change in the j-th neighborhood region. Let be the distance between the i-th region and the j-th neighboring region. Let be the thermal conductivity of the j-th neighboring region. By substituting parameters such as the rate of temperature change and heat transfer efficiency from the thermal dynamic response characteristic data of the building envelope into the model, the change in thermal conductivity of each mix area under the influence of the neighboring thermal dynamic response is calculated. The influence of changes in the thermal state of the neighboring region on the thermal conductivity performance of the core area is analyzed, the degree of influence and the conduction path are clarified, and characteristic data of the thermal dynamic response neighboring thermal conductivity influence are generated. This data includes parameters such as the change in thermal conductivity of each mix area, the intensity of the neighboring influence, and the influence path, thus improving the comprehensiveness of the thermal conductivity characteristic analysis.

[0039] Step S24: Based on the thermal dynamic response neighborhood thermal conductivity influence characteristic data and the thermal property mapping concrete mix proportion region thermal conductivity data, perform material thermal conductivity characteristic analysis of the building envelope structure to generate material thermal conductivity characteristic data of the building envelope structure.

[0040] In this embodiment of the invention, the thermal conductivity characteristics of building envelope structures are analyzed based on the thermal dynamic response neighborhood thermal conductivity influence characteristic data and the thermal conductivity data of concrete mix proportion regions mapped by thermal property mapping. The core parameters of both types of data are integrated, and each mix proportion region is used as an analysis unit. Combining the unit classification in the building envelope structure division data, the thermal conductivity characteristics of different envelope structure units and different mix proportion regions are comprehensively analyzed. The basic thermal conductivity obtained from thermal property mapping is superimposed with the change in thermal conductivity caused by neighborhood influence to obtain the actual effective thermal conductivity of each mix proportion region. A material thermal conductivity characteristic evaluation index system is constructed, including three core indicators: effective thermal conductivity, thermal conductivity stability, and neighborhood influence sensitivity. Thermal conductivity stability is calculated by the change in effective thermal conductivity at different time points, and neighborhood influence sensitivity is calculated by the ratio of the change in thermal conductivity to the change in neighborhood temperature. By analyzing the differences in effective thermal conductivity across different mix proportions, the correlation between concrete mix proportions and thermal conductivity characteristics is clarified. Through analysis of thermal conductivity stability, the changing trends of thermal conductivity performance under thermal dynamic response in different mix proportions are identified. Furthermore, by analyzing neighborhood influence sensitivity, the degree to which the thermal conductivity performance of different mix proportions is affected by the thermal state of the surrounding area is determined. Based on these analyses, the overall thermal conductivity characteristics of each building envelope unit are extracted, generating thermal conductivity characteristic data for building envelope materials. This data includes parameters such as effective thermal conductivity, thermal conductivity stability, and neighborhood influence sensitivity for each mix proportion region and each building envelope unit.

[0041] Furthermore, as an embodiment of the present invention, reference is made to... Figure 2 As shown, Figure 1 A detailed flowchart illustrating the implementation steps of step S3 is provided in this embodiment. Step S3 includes: Step S31: Discretize the concrete mix proportion data of the building envelope to generate discrete concrete mix proportion data. In this embodiment of the invention, the concrete mix proportion data of the building envelope structure is discretized. This data includes the unit affiliation, spatial location, geometric dimensions, and complete concrete mix proportion parameters for each mix proportion region. The discretization process is based on each mix proportion region, dividing each region into several uniform discrete units according to a spatial grid division standard. The size of each discrete unit is determined based on the geometric dimensions of the mix proportion region, ensuring that the concrete mix proportion parameters within each discrete unit are uniform and without significant differences. The discretization uses an orthogonal grid division method, with the grid direction parallel to the surface of the building envelope structure. The discrete units are rectangular in shape, and each unit corresponds to unique spatial coordinates, geometric dimensions, and mix proportion parameters. The correspondence between the discrete units and the original mix proportion regions is clearly defined, and the mix proportion region number and the type of the building envelope structure unit to which each discrete unit belongs are labeled. After discretization, the information of all discrete units is integrated to generate discretized data of the concrete mix proportion region. This data includes the number of the discrete unit, spatial coordinates, geometric dimensions, concrete mix proportion parameters, the mix proportion region to which it belongs, and information of the retaining structure unit. Discretization enables subsequent multi-scheme simulations to accurately locate each tiny region, improving simulation accuracy.

[0042] Step S32: Based on the discretized data of concrete mix proportions, design simulation conditions for multiple schemes of building envelope materials to obtain simulation condition data for multiple schemes of building envelope materials; In this embodiment of the invention, multi-scheme simulation conditions for building envelope materials are designed based on discretized data of concrete mix proportions. The discrete unit mix proportion parameters, geometric dimensions, and spatial locations in the discretized data serve as the design basis. Multiple simulation schemes are constructed around the adjustment of concrete mix proportion parameters, clarifying the parameter differences and design standards for each scheme. The simulation scheme design focuses on energy-saving optimization. Combining the thermal conductivity characteristics data of building envelope materials, the mass proportions of cement, sand, water, and admixtures in the concrete mix proportion parameters are adjusted, constructing 10 independent simulation schemes. Each scheme corresponds to a different concrete mix proportion combination. One scheme uses the original concrete mix proportion parameters as the baseline scheme, while the remaining 9 schemes adjust the proportion of a single mix proportion parameter, keeping other parameters unchanged. This ensures the uniqueness of the differences between each scheme, facilitating subsequent comparative analysis of the impact of different mix proportion parameters on energy consumption. Simultaneously, the simulation scope and duration for each scheme are clearly defined. The simulation scope covers all discrete units and corresponding envelope units, and the simulation duration is consistent with the data acquisition period mentioned earlier, set at 30 days, with each day's simulation occurring at 15-minute intervals. Based on the above design, simulation condition data for multiple schemes of building envelope materials is generated. This data includes the number of each simulation scheme, concrete mix proportion parameters, simulation range, simulation duration, and the corresponding relationship of discrete elements.

[0043] Step S33: Analyze the boundary conditions for building energy consumption simulation based on the building's multi-source characterization attribute data, and generate building energy consumption simulation boundary condition data; In this embodiment of the invention, boundary condition analysis for building energy consumption simulation is performed based on building multi-source characterization attribute data. This data includes the correlation features of energy consumption, environment, and building envelope at various time points and spatial regions. Core parameters related to energy consumption simulation, such as indoor and outdoor temperature, solar radiation intensity, wind speed, and heat flux density, are extracted from this data to clarify the boundary types and parameter values ​​for energy consumption simulation. The building energy consumption simulation boundaries are divided into indoor environment boundaries, outdoor environment boundaries, and internal boundaries of the building envelope. Outdoor environment boundary parameters directly utilize real-time outdoor environmental data from the building multi-source characterization attribute data, extracting corresponding parameters at 15-minute time points as input parameters for the outdoor environment during simulation. Indoor environment boundary parameters are set to fixed values, and an indoor reference temperature is set based on routine building operation and maintenance requirements to ensure the stability of the indoor environment boundary. Internal boundary parameters of the building envelope are determined based on building envelope segmentation data and thermal state boundary characteristic data, clarifying the heat exchange boundaries between discrete units and setting boundary thermal conductivity coefficients to match the thermal conductivity parameters of the discrete units. Based on the above analysis, boundary condition data for building energy consumption simulation is generated. This data includes boundary type, values ​​of each boundary parameter and corresponding time nodes, boundary heat exchange method, and clarifies the constraints of each boundary during the simulation process.

[0044] Step S34: Based on the multi-scheme simulation condition data of building envelope materials and the boundary condition data of building energy consumption simulation, perform multi-scheme simulation thermal balance processing of building envelope to generate multi-scheme simulation thermal balance data of building envelope. In this embodiment of the invention, multi-scheme simulation thermal balance processing of the building envelope structure is performed based on simulation condition data of building envelope materials and boundary condition data of building energy consumption simulation. The thermal balance method is used to conduct the simulation, and a thermal balance equation for the building envelope structure is constructed. The expression of the thermal balance equation is as follows: ,in The total heat flow received by the building envelope. The total heat flux dissipated by the building envelope. Heat flow stored in the building envelope This represents the heat loss of the building envelope. This mainly includes solar radiation heat flow and indoor heat conduction heat flow, which are calculated from the solar radiation intensity and indoor temperature parameters in the building energy consumption simulation boundary condition data; It mainly refers to the heat flow emitted from the building envelope to the outside, which is calculated from the thermal conductivity of the discrete element, the temperature difference between the indoor and outdoor elements, and the area of ​​the discrete element. It is calculated from the heat capacity and temperature change of the discrete unit. The heat loss during heat conduction is calculated at a fixed ratio. For each simulation scheme, the corresponding concrete mix proportions are substituted into the heat balance equation. Combined with the parameters in the building energy consumption simulation boundary condition data, the heat flow and temperature change of each discrete unit are calculated sequentially at 15-minute time nodes. This simulates the heat accumulation and heat conduction process of the building envelope under different schemes, and records the heat balance state parameters of each discrete unit at each time node. Based on the above simulation processing, multi-scheme simulation heat balance data of the building envelope is generated. This data includes parameters such as heat flow, temperature, heat storage, and heat loss for each simulation scheme, each time node, and each discrete unit, fully presenting the heat balance change law of the building envelope under different material schemes.

[0045] Step S35: Perform multi-scheme simulation energy consumption characteristic analysis on the multi-scheme simulation thermal balance data of the building envelope, and generate multi-scheme simulation energy consumption characteristic data of the building envelope.

[0046] In this embodiment of the invention, the energy consumption characteristics of a building envelope are analyzed based on the simulated thermal balance data of multiple simulation schemes. The thermal balance data of each simulation scheme are integrated, and based on discrete elements and combined with the building envelope segmentation data, the thermal balance parameters of each envelope element are summarized to analyze the energy consumption characteristics under different schemes. An energy consumption characteristic analysis model is constructed, and the model expression is as follows: ,in The total energy consumption of the building envelope. This represents the total number of simulation time points. The total number of discrete units. Let be the heat loss of the i-th discrete unit at time t. Let be the area of ​​the i-th discrete unit. The thermal efficiency is determined by the heat transfer efficiency parameter from the building's multi-source characterization attribute data. This model calculates the total energy consumption of the building envelope for each simulation scheme, analyzes the energy consumption trend over time for each scheme, and examines the differences in energy consumption distribution among various building envelope units. Core features such as peak energy consumption, average energy consumption, and energy consumption fluctuation amplitude are extracted. The energy consumption differences between different simulation schemes are compared to clarify the impact of concrete mix proportion adjustments on the building envelope's energy consumption. For example, the impact of changes in cement ratio on the energy consumption of the south-facing exterior wall unit and changes in admixture ratio on the energy consumption of the roof unit are analyzed. Based on the above analysis, multi-scheme simulation energy consumption characteristic data for the building envelope is generated. This data includes the total energy consumption of each simulation scheme, the energy consumption trend over time, the energy consumption distribution of each building envelope unit, energy consumption characteristic parameters, and differences between the schemes.

[0047] Furthermore, step S32 includes the following steps: Step S321: Perform material-specific combination characteristic analysis based on the discretized data of concrete mix proportions to generate material-specific combination characteristic data; In this embodiment of the invention, material-specific combination feature analysis is performed based on the discretized data of the concrete mix proportion region. The discretized data includes the discrete unit number, spatial coordinates, geometric dimensions, concrete mix proportion parameters, the mix proportion region to which it belongs, and information on the enclosing structure unit. The material-specific combination feature analysis focuses on the combination differences of concrete mix proportion parameters and the spatial combination rules of discrete units. The core is to extract the specific features of different mix proportion combinations, providing a basis for subsequent thermal conductivity mapping and simulation scheme design. Based on discrete units, they are classified according to the type of enclosing structure unit (exterior wall, roof, floor slab, doors and windows). Discrete units within the same enclosing structure unit are associated according to their spatial location. The differences in concrete mix proportion parameter combinations of different discrete units are analyzed, clarifying the combination types of the mass proportions of cement, sand, gravel, water, and admixtures. The number, spatial distribution range, and total geometric dimensions of discrete units for each combination type are statistically analyzed. A material-specific combination feature evaluation model is constructed, with the model expression being: ,in For specific combination feature values, This is the sum of parameter differences between this combination and the benchmark combination within the same type of enclosure structure unit. This represents the total surface area of ​​the discrete units corresponding to this combination. This represents the total volume of the discrete units corresponding to this combination. The feature weights are determined by the correlation between the discrete unit proportioning parameters and spatial distribution. The training data uses the proportioning parameters and spatial dimensions of the discrete units. By calculating the feature values ​​of different combinations, the specificity of different combinations is distinguished. Based on the above analysis, material-specific combination feature data is generated. This data includes the type, parameter details, discrete unit correlation information, spatial distribution characteristics, and specific combination feature values ​​for each proportioning combination, clarifying the differentiated characteristics of different proportioning combinations.

[0048] Step S322: Perform thermal conductivity characteristic mapping on the material-specific combination characteristic data using the thermal conductivity characteristic data of the building envelope material to generate material combination thermal conductivity characteristic mapping data; In this embodiment of the invention, thermal conductivity characteristic data of building envelope materials are used to map the thermal conductivity performance of material-specific combination characteristic data to the thermal conductivity performance of the material combinations. The thermal conductivity characteristic data of building envelope materials includes parameters such as effective thermal conductivity, thermal stability, and neighborhood influence sensitivity for each mix ratio region and each envelope unit. The material-specific combination characteristic data contains core information for each mix ratio combination. The core of thermal conductivity performance feature mapping is to establish the correspondence between mix ratio combination characteristics and thermal conductivity parameters, clarifying the differences in thermal conductivity performance among different mix ratio combinations. A mapping model between material combinations and thermal conductivity performance is constructed. The model structure is a multiple linear regression model, and the mathematical expression is: ,in The effective thermal conductivity of the material combination, These are material-specific combination characteristic values. These represent the mass percentages of cement, sand, gravel, and admixtures, respectively. These are the model coefficients. The training data uses the proportioning parameters and specific eigenvalues ​​from the material-specific combination feature data, as well as the corresponding effective thermal conductivity data from the thermal conductivity feature data of building envelope materials. The training data is substituted into the model, and the model coefficients are calibrated using the least squares method. During the calibration process, the deviation between the model's predicted value and the actual effective thermal conductivity is calculated. The coefficients are iteratively adjusted until the deviation converges. After determining the fixed coefficients, the feature parameters of each proportioning combination are substituted into the model to calculate the predicted effective thermal conductivity value for the corresponding combination. At the same time, combined with the thermal conductivity stability parameter, the change law of thermal conductivity performance of different proportioning combinations over time is analyzed to generate material combination thermal conductivity performance feature mapping data. This data includes the predicted effective thermal conductivity value, thermal conductivity stability, specific eigenvalues, and parameter correlations for each proportioning combination, fully presenting the correspondence between proportioning combinations and thermal conductivity performance, and providing core thermal conductivity performance basis for subsequent simulation condition design.

[0049] Step S323: Design simulation conditions for multiple schemes of building envelope materials based on the thermal conductivity characteristic mapping data of material combination, and obtain simulation condition data for multiple schemes of building envelope materials.

[0050] In this embodiment of the invention, simulation conditions for multiple schemes of building envelope materials are designed based on the thermal conductivity characteristic mapping data of material combinations. The core is to adjust the material ratio combination to construct multiple simulation schemes with significant differences in thermal conductivity, providing diverse material conditions for subsequent multi-scheme energy consumption simulation and reinforcement learning energy-saving control model construction. Based on the thermal conductivity characteristic mapping data of material combinations, 10 representative ratio combinations are selected. One group adopts the original benchmark ratio combination (corresponding to the benchmark thermal conductivity), and the remaining 9 groups are selected with different specific characteristic values ​​and significant differences in effective thermal conductivity. Each scheme corresponds to a unique concrete mix ratio parameter combination. The details of the mix ratio parameters, the predicted value of the effective thermal conductivity, and the corresponding discrete unit range of each scheme are clearly defined to ensure that the thermal conductivity of each scheme has significant differences, which facilitates subsequent comparative analysis of the impact of different thermal conductivity on building energy consumption. Simultaneously, the simulation constraints for each scheme were clearly defined. The simulation scope covered all discrete elements and associated building envelope elements corresponding to the mix proportion. The simulation duration was consistent with the data acquisition and discretization cycle described earlier, set at 30 days, with each day's simulation lasting 15 minutes. During the simulation, other conditions (such as environmental parameters and boundary conditions) remained consistent, with only the concrete mix proportion and corresponding thermal conductivity parameters being changed. Based on the above design, simulation condition data for multiple schemes of building envelope materials was generated. This data included the number of each simulation scheme, the concrete mix proportion parameter combination, the predicted effective thermal conductivity, the corresponding discrete element range, the simulation duration, and the simulation constraints. The core conditions of each simulation scheme were clearly defined, ensuring that subsequent multi-scheme simulations could accurately compare the energy consumption differences of different material combinations.

[0051] Furthermore, step S4 includes the following steps: Step S41: Analyze the time-series changes in the thermal state of the building envelope using the thermal dynamic response characteristic data of the building envelope and the simulation energy consumption characteristic data of the building envelope under multiple schemes, and generate time-series change characteristic data of the thermal state of the building envelope under multiple schemes. In this embodiment of the invention, the thermal dynamic response characteristic data of the building envelope is used to analyze the temporal variation characteristics of the thermal state of the building envelope under multiple simulation schemes based on the energy consumption characteristic data of the building envelope. The thermal dynamic response characteristic data of the building envelope includes parameters such as the thermal response delay time, temperature change rate, steady-state temperature, and heat conduction efficiency of each building envelope unit. The multi-scheme simulation energy consumption characteristic data of the building envelope includes parameters such as the total energy consumption of each simulation scheme, the energy consumption time change trend, and the energy consumption distribution of each building envelope unit. The core of the analysis is to correlate the thermal dynamic response characteristics with the temporal variation of energy consumption and extract the change law of the thermal state of the building envelope over time under different simulation schemes. Taking each simulation scheme as an independent analysis unit, the thermal dynamic response parameters and energy consumption data are correlated at 15-minute time nodes to analyze the correlation between the temperature change rate and the energy consumption change rate of each building envelope unit, clarifying the influence of the thermal response delay time on the temporal variation of energy consumption. For example, when the thermal response delay time of the south wall unit is long, the difference between the time of the peak energy consumption and the time of the peak solar radiation occurs is observed. A thermal state temporal variation characteristic analysis model is constructed, and the model expression is as follows: ,in Let be the change in energy consumption at time point t. For lag Temperature change at each time point ( (thermal response delay time) For effective thermal conductivity, The correlation coefficient is determined by calibration using thermal dynamic response characteristic data and energy consumption characteristic data. This model calculates the correlation between energy consumption change and temperature change at each time point, extracting the temporal variation patterns of the thermal state for each scheme, including energy consumption change characteristics during temperature rise / fall phases, stable energy consumption values ​​corresponding to steady-state temperatures, and energy consumption fluctuation patterns caused by thermal response delays. This generates temporal variation characteristic data of the thermal state of the building envelope for multiple schemes. This data includes the temperature change rate, energy consumption change, correlation parameters, and temporal variation patterns for each time point in each scheme.

[0052] Step S42: Establish an optimized trend prediction model for the thermal state of the multi-scheme building envelope based on the time-series variation characteristic data of the thermal state of the multi-scheme building envelope; In this embodiment of the invention, an optimized predictive model for the thermal state trend of a multi-scheme building envelope is established based on the time-series variation characteristic data of the thermal state of the building envelope. The predictive model is based on a long short-term memory neural network and uses a particle swarm optimization algorithm to optimize the model parameters, ensuring the model's prediction accuracy. The model structure, training data, and algorithm logic are fully disclosed. The model structure includes an input layer, a hidden layer, and an output layer. The input layer nodes are the temperature change rate, energy consumption change, correlation parameters, and time node information from the time-series variation characteristic data of the thermal state of the multi-scheme building envelope. The number of input layer nodes is the same as the number of feature parameters. The hidden layer has three layers, each with twice the number of nodes as the input layer, and uses the tanh activation function to mine deep correlation features of the time-series data. The output layer nodes are the thermal state parameters (temperature, heat conduction rate) at future time nodes, and the output layer uses a linear activation function. The training data uses the first 80% of the time node data from the time-series variation characteristic data of the thermal state of the multi-scheme building envelope, divided into training sets according to the time series. The remaining 20% ​​of the data is used as the test set. The training objective is to minimize the deviation between the predicted thermal state parameters and the actual thermal state parameters. The algorithm logic is as follows: First, initialize the weights and bias parameters of the Long Short-Term Memory Neural Network (LSTM). Input the training data into the network and calculate the predicted value through forward propagation. Calculate the mean square error between the predicted and actual values ​​as the loss function and adjust the network parameters through backpropagation. Then, use the Particle Swarm Optimization (PSO) algorithm to optimize the network weights and bias parameters. The objective function of PSO is consistent with the network loss function, and the particle dimension is consistent with the number of network parameters. Iterate and update the particle position and velocity to find the optimal parameter combination. The iteration terminates when the loss function value converges. After convergence, determine the optimized network parameters to obtain an optimized multi-scheme thermal state trend prediction model for the building envelope. This model can accurately predict the future trend of thermal state changes of the building envelope under different simulation schemes.

[0053] Step S43: By optimizing the multi-scheme building envelope thermal state trend prediction model, perform multi-scheme building envelope simulation energy consumption trend analysis on the multi-scheme building envelope energy consumption characteristic data, and generate multi-scheme building envelope energy consumption trend data.

[0054] In this embodiment of the invention, an optimized multi-scheme building envelope thermal state trend prediction model is used to analyze the multi-scheme building envelope energy consumption trend of simulated energy consumption characteristic data. The core is to use the thermal state trend data obtained from the prediction model, correlate it with energy consumption characteristics, analyze the future change trend of energy consumption under different schemes, and clarify the energy consumption change law. Historical energy consumption data and thermal state time series data from the multi-scheme building envelope simulation energy consumption characteristic data are input into the optimized thermal state trend prediction model to predict the thermal state parameters (temperature, heat conduction rate) of the building envelope at each time point within the next 30 days. Combined with the energy consumption characteristic analysis model constructed above, the predicted thermal state parameters are substituted into the model to calculate the predicted energy consumption values ​​at each time point in the future. Analyze the energy consumption prediction trends of each simulation scheme, extract energy consumption trend characteristics, including energy consumption increase / decrease trends, future energy consumption peaks and their occurrence times, steady-state energy consumption trends, and differences in energy consumption trends among different building envelope units. Compare the differences in energy consumption trends among different schemes to clarify the influence of concrete mix design and thermal dynamic response characteristics on future energy consumption changes. For example, schemes with lower effective thermal conductivity show a more significant future energy consumption decrease trend and lower energy consumption peaks. Construct an energy consumption trend evaluation model, the model expression of which is... ,in This represents the energy consumption trend fluctuation value. Let be the energy consumption value at the t-th prediction time point. To predict the average energy consumption over the period, To predict the number of time points, this model quantifies the stability of energy consumption trends for each scheme; the smaller the fluctuation value, the more stable the energy consumption trend. Based on the above analysis, energy consumption trend data for multiple schemes of building envelope is generated. This data includes the future energy consumption prediction value, energy consumption change trend, energy consumption trend fluctuation value, energy consumption trend of each envelope unit, and differences between schemes for each simulation scheme.

[0055] Furthermore, step S42 includes the following steps: A model for the trend change of the thermal state of a multi-scheme enclosure structure is established by using a pre-defined long short-term memory neural network algorithm and time-series change characteristic data of the thermal state of the multi-scheme enclosure structure. This model is used to obtain a prediction model for the thermal state of the multi-scheme enclosure structure. The prediction model is then subjected to particle swarm optimization of the trend prediction relationship of the multi-scheme enclosure structure to obtain an optimized prediction model for the thermal state of the multi-scheme enclosure structure.

[0056] In this embodiment of the invention, a model of the thermal state trend of a multi-scheme building envelope is established using a preset long short-term memory neural network algorithm and time-series change feature data of the thermal state of the building envelope. This model is then used to obtain a prediction model of the thermal state trend of the multi-scheme building envelope. The model is further optimized using particle swarm optimization to obtain an optimized prediction model of the thermal state trend of the multi-scheme building envelope, providing accurate support for subsequent energy consumption trend analysis and energy-saving control model construction. The time-series change feature data of the thermal state of the multi-scheme building envelope includes the temperature change rate, energy consumption change, correlation parameters, and time-series change patterns of each scheme at each time point. This data serves as the core training and modeling basis for the long short-term memory neural network. The model structure of the long short-term memory neural network is fixed as a three-layer structure consisting of an input layer, a hidden layer, and an output layer. The number of nodes in the input layer is fixed at four, corresponding to the temperature change rate, energy consumption change, correlation parameters, and time point identifiers in the time-series change feature data of the thermal state of the multi-scheme building envelope. Each input node corresponds to a type of feature parameter, ensuring accurate matching between the input data and the model input layer. The hidden layer has three layers, with each layer having twice the number of nodes as the input layer. The tanh activation function is used, and its expression is: This method is used to uncover deep correlations in time-series data, avoid gradient vanishing, and ensure that the model can accurately capture long-term patterns in thermal state changes over time. The output layer has a fixed number of two nodes, corresponding to the predicted building envelope temperature and predicted heat transfer rate at future time points, respectively. A linear activation function is used, expressed as follows: To ensure consistency between the output results and the actual thermal state parameters, the training data comprises all time-series data from the multi-scheme building envelope thermal state temporal change feature data, divided in chronological order without additional data processing. This data is directly used as the input and label data for model training. The input data includes the temperature change rate, energy consumption change, and correlation parameters of the preceding time nodes, while the label data consists of the thermal state parameters (temperature, heat conduction rate) of the corresponding subsequent time nodes. The training objective is to minimize the deviation between the predicted and actual thermal state parameters. The weights and bias parameters of the Long Short-Term Memory (LSTM) neural network are initialized. The training data is input into the network in chronological order, passed through the input layer to the hidden layer. The hidden layer uses the tanh activation function to perform feature transformation and deep mining on the input data. The processed feature data is then passed to the output layer to output the predicted thermal state values. The mean squared error between the predicted and actual label values ​​is calculated as the loss function, expressed as follows: ,in The loss value. The total amount of training data, These are the actual thermal state parameter values. To predict thermal state parameters, the network weights and bias parameters are adjusted using a backpropagation algorithm, iteratively updated until the loss function converges. The convergence condition is that the difference between the loss function values ​​of two adjacent iterations is less than a fixed threshold, at which point a multi-scheme thermal state trend prediction model for the building envelope is obtained. The multi-scheme thermal state trend prediction model is then subjected to particle swarm optimization (PSO) of the multi-scheme trend prediction relationship. The core logic of the PSO algorithm is to find the optimal parameter combination of the model by simulating the motion trajectory of the particle swarm. The optimization objects are the weights and bias parameters of the long short-term memory neural network. The parameters of the PSO optimization are fixed, the particle dimension is consistent with the number of parameters to be optimized, each particle corresponds to a set of model parameter combinations, and the initial values ​​of the particle position and velocity are set within a fixed range. The fitness function is consistent with the loss function of the long short-term memory neural network, both being the mean square error between the predicted and actual values, used to evaluate the quality of the model parameter combinations corresponding to the particles. The position and velocity of the particle swarm are initialized, the fitness value of each particle is calculated, and the optimal particle position in the swarm and the optimal particle position of the individual are recorded. The position and velocity of each particle are iteratively updated using the following formulas: , ,in Let be the velocity of the i-th particle in the j-th dimension during the (t+1)-th iteration. For inertial weights, As a learning factor, It is a random constant. Let be the optimal position for the i-th particle. The optimal position for the group. Let be the position of the i-th particle in the j-th dimension during the t-th iteration. After each iteration, update the individual optimal and the group optimal positions until the number of iterations reaches a fixed value. At this point, the model parameter combination corresponding to the group optimal position is the optimal parameter. Substitute this parameter into the original long short-term memory neural network to obtain an optimized multi-scheme enclosure structure thermal state trend prediction model. This model can accurately capture the future change trend of the enclosure structure thermal state under different simulation schemes.

[0057] Furthermore, step S5 includes the following steps: Step S51: Obtain building operation and maintenance target requirements data; In this embodiment of the invention, building operation and maintenance target demand data is obtained, and combined with the actual building operation and maintenance scenario, the energy consumption control baseline, thermal comfort standard, and operation and maintenance cost constraints during the operation and maintenance process are clarified. The building operation and maintenance target demand data includes three types of core parameters: energy consumption control target parameters, thermal comfort target parameters, and operation and maintenance cost constraint parameters. Among them, the energy consumption control target parameters clarify the upper limit of energy consumption per unit time of the building envelope and the upper limit of total energy consumption throughout the entire cycle; the thermal comfort target parameters clarify the reference temperature range and humidity range of each functional area of ​​the building interior to ensure that the indoor environment meets the usage requirements during the operation and maintenance process; the operation and maintenance cost constraint parameters clarify the cost upper limit of concrete mix adjustment and equipment operation adjustment. All parameters are determined based on the actual building operation and maintenance scenario requirements.

[0058] Step S52: Based on the energy consumption trend data of multiple schemes for the building envelope, perform state representation spatial design of the energy consumption trend of multiple schemes to obtain the state representation spatial data of the energy consumption trend of multiple schemes. In this embodiment of the invention, a state representation space for the energy consumption trends of multiple schemes of building envelope is designed based on the energy consumption trend data of multiple schemes. The core is to transform the energy consumption trend data into state features that can be recognized by reinforcement learning algorithms, establish the correspondence between states and energy consumption trends, and provide standardized input states for reinforcement learning models. The energy consumption trend data of multiple schemes of building envelope includes the future energy consumption prediction value, energy consumption change trend, energy consumption trend fluctuation value, energy consumption trend of each envelope unit, and scheme comparison differences for each scheme. Based on this, the core state features of each scheme are extracted, including the average energy consumption, peak energy consumption, energy consumption decrease rate, and the correlation between thermal state and energy consumption. These feature parameters are quantified and integrated to construct the state representation space. The state representation space is constructed in vector form, with the vector dimension matching the number of extracted state features. Each vector component corresponds to a state feature parameter, and the vector index corresponds one-to-one with the simulation scheme number and the enclosure structure unit type. This clarifies the energy consumption trend scenario corresponding to each state vector and generates multi-scheme energy consumption trend state representation space data. This data includes state vectors, the meaning of vector components, scheme correlations, and unit correlation information, achieving a standardized representation of energy consumption trend states and providing accurate support for the state input of reinforcement learning models.

[0059] Step S53: Extract the influencing factors of the multi-scheme energy consumption trend of the building envelope based on the multi-scheme energy consumption trend data, and generate multi-scheme energy consumption trend influencing factor data of the building envelope. In this embodiment of the invention, the influencing factors of the multi-scheme energy consumption trend of the building envelope are extracted based on the energy consumption trend data of the building envelope. The core is to clarify the influence weight and law of each factor on the energy consumption trend, so as to provide a basis for subsequent action space design. From the multi-scheme energy consumption trend data of the building envelope, the energy consumption trend parameters of each simulation scheme are correlated with the relevant data obtained above, including the thermal conductivity characteristics of the building envelope materials, concrete mix proportion parameters, and thermal dynamic response characteristics data, to extract the core factors affecting the energy consumption trend, which are specifically divided into three categories: material factors (concrete mix proportion parameters, effective thermal conductivity), thermal state factors (thermal response delay time, temperature change rate), and environmental factors (indoor and outdoor temperature difference, solar radiation intensity). An influencing factor extraction model is constructed, and the model expression is as follows: ,in Let i be the weight of the i-th type of influencing factor. This is the energy consumption value. Let be the parameter value of the i-th type of influencing factor. The model calculates the weights of various influencing factors to obtain the average energy consumption value, clarifies the influence intensity of different factors on the energy consumption trend, and generates data on the influencing factors of energy consumption trends for multiple building envelope schemes. This data includes the parameters, weight values, influence patterns, and correlations with energy consumption trends of various influencing factors.

[0060] Step S54: Design the action space of the energy consumption trend of multiple schemes by using the data of the factors affecting the energy consumption trend of the building envelope, and obtain the action space data of the energy consumption trend of multiple schemes. In this embodiment of the invention, the impact action space design of the energy consumption trend of multiple schemes is carried out by using the data of influencing factors of the energy consumption trend of the building envelope. The core is to design adjustable and executable actions based on the influencing factors to achieve precise intervention in the energy consumption trend and provide action input for the energy-saving control model. Based on the three types of influencing factors extracted in step S53, three types of actions are designed: material adjustment actions, thermal state adjustment actions, and environmental adaptation actions. Each type of action corresponds to specific adjustment parameters and adjustment ranges. The material adjustment action corresponds to the adjustment of concrete mix proportion parameters, clarifying the adjustment step size and adjustment range of each mix proportion parameter, and maintaining consistency with the mix proportion differences of the previous simulation scheme; the thermal state adjustment action corresponds to the adjustment of the heat conduction process of the building envelope, and is related to the adjustment methods of thermal resistance and heat capacity parameters; the environmental adaptation action corresponds to the adjustment of indoor and outdoor heat exchange, and is related to the adjustment of ventilation, shading and other related parameters. Each action corresponds to a unique action identifier, adjustment parameter, and adjustment method. The actions are clearly distinguished from each other and there is no overlap or redundancy. The multi-scheme energy consumption trend impact action space data is generated. This data includes action identifier, action type, adjustment parameter, adjustment range and corresponding influencing factors.

[0061] Step S55: Analyze the collaborative relationship of the influence on the action space based on the energy consumption trend of multiple schemes, and generate collaborative relationship data of the influence on the action space; In this embodiment of the invention, the analysis of the synergistic relationships in the action space based on the influence of multi-scheme energy consumption trends on action space data is conducted. The core objective is to clarify the synergistic effects between different actions, avoid action conflicts, improve energy-saving control effects, and provide a basis for the synergistic control of reinforcement learning models. Based on the action types and adjustment parameters in the action space data, the correlation between different actions is analyzed, and an action synergistic relationship analysis model is constructed. The model expression is as follows: ,in Let be the coordination coefficient between action type i and action type j. This represents the change in energy consumption when two types of actions work together. This represents the change in energy consumption when two types of actions act alone. The model calculates the synergy coefficient between any two types of actions; a positive coefficient indicates synergistic effect, while a negative coefficient indicates conflict. Based on the calculation results, the synergistic combination method of actions is determined, synergistic actions are distinguished from conflicting actions, the synergistic priority of each type of action is clarified, and data on the synergistic relationship affecting the action space is generated. This data includes the synergistic coefficient, synergistic combination method, action priority, and conflicting action identifier.

[0062] Step S56: Based on the preset reinforcement learning algorithm, establish the energy-saving control learning relationship of the building envelope structure by using the spatial data of the energy consumption trend state representation of multiple schemes and the spatial data of the energy consumption trend influence of multiple schemes, and generate the energy-saving control learning model of the building envelope structure. In this embodiment of the invention, an energy-saving control learning model for a building envelope is established based on a preset reinforcement learning algorithm, using spatial data representing the state of multiple energy consumption trends and spatial data representing the actions influenced by these trends. This model employs the Q-learning reinforcement learning algorithm and comprises four layers: a state input layer, an action decision layer, a reward calculation layer, and a parameter update layer. The state input layer receives the state vector from the spatial data representing the state of multiple energy consumption trends. The action decision layer outputs the actions from the spatial data representing the actions influenced by these trends. The reward calculation layer calculates the reward value based on the energy consumption change after the action is executed. The parameter update layer updates the model parameters. The training data consists of spatial data representing the state of multiple energy consumption trends, action space data, and corresponding energy consumption change data. The training objective is to maximize the cumulative reward value, and the reward function expression is as follows: ,in As a reward value, As the baseline energy consumption value, This represents the energy consumption value after the action is performed. For the cost of action execution, The reward weights are determined by the energy consumption control target and cost constraints. The algorithm logic is as follows: initialize the Q-value matrix, input the state vector and action vector into the model, calculate the initial Q-value, execute the corresponding action and acquire energy consumption change data, calculate the reward value through the reward function, and use the Q-value update formula. Update the Q value, where For learning rate, As a discount factor, Current state For the current action, This is the next state after the action is executed. The optimal action for the next state is iteratively updated until the Q value converges. At this point, a learning model for energy-saving control of the building envelope is generated. This model can accurately establish the correlation between state, action and reward, providing core model support for the design of subsequent energy-saving control strategies.

[0063] Step S57: Transmit the building operation and maintenance target requirement data to the building envelope energy-saving control learning model to analyze the collaborative control relationship of building energy consumption and energy saving, so as to design a building energy consumption and energy saving optimization control strategy, wherein the building energy consumption and energy saving optimization control strategy includes building envelope concrete mix control data and building equipment operation control data.

[0064] In this embodiment of the invention, building operation and maintenance target demand data is transmitted to the building envelope energy-saving control learning model for collaborative control relationship analysis of building energy consumption and energy saving. An optimized control strategy for building energy consumption and energy saving is designed. The core is to combine operation and maintenance target constraints and output the optimal action combination through the model to achieve a balance between energy consumption control and operation and maintenance needs. The energy consumption control target, thermal comfort target, and cost constraint parameters from the building operation and maintenance target demand data are input into the energy-saving control learning model. Based on the established state-action-Q-value correlation, the model analyzes the fit between the energy consumption results corresponding to different action combinations and the operation and maintenance targets, and selects the optimal action combination that meets the operation and maintenance target constraints. The optimal action combination includes concrete mix proportion control actions for the building envelope and building equipment operation control actions. The concrete mix proportion control actions correspond to the material adjustment actions in step S54, specifying the specific control values ​​of each mix proportion parameter to match the mix proportion parameters in the simulation scheme. The building equipment operation control actions correspond to the operation parameter control of the air conditioning, lighting, and ventilation systems, specifying the equipment start-up and shutdown times and operating load control values ​​to adapt to the energy consumption trend of the building envelope. Based on the above collaborative control analysis, a building energy consumption optimization control strategy is generated. This strategy includes concrete mix ratio control data for the building envelope and building equipment operation control data. It clarifies the execution time, adjustment parameters, and control range of each control action, ensuring that the control strategy can accurately meet the operation and maintenance target requirements, achieve building energy consumption optimization, and simultaneously meet indoor thermal comfort standards and operation and maintenance cost constraints.

[0065] Therefore, the embodiments should be considered as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of the equivalents of the application are intended to be included within the invention.

[0066] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.

Claims

1. A building energy consumption optimization and control method based on reinforcement learning, characterized in that, Includes the following steps: Step S1: Obtain building multi-source heterogeneous energy consumption data and basic data of building envelope; Based on the building's multi-source heterogeneous energy consumption data and the building envelope's basic data, the thermal dynamic response characteristics of the building envelope are analyzed, and thermal dynamic response characteristic data of the building envelope are generated. Step S2: Based on the basic data of the building envelope and the thermal dynamic response characteristic data of the building envelope, perform material thermal conductivity characteristic analysis of the building envelope to generate material thermal conductivity characteristic data of the building envelope. Step S3: Based on the thermal conductivity characteristics data of the building envelope materials, perform simulation energy consumption characteristics analysis of multiple schemes of building envelope materials to generate simulation energy consumption characteristic data of multiple schemes of building envelope materials; Step S4: Perform multi-scheme energy consumption trend analysis on the multi-scheme simulation energy consumption characteristic data of the building envelope, and generate multi-scheme energy consumption trend data of the building envelope. Step S5: Obtain building operation and maintenance target requirements data; A learning model for energy-saving control of building envelope was established based on energy consumption trend data of multiple schemes for building envelope. The building operation and maintenance target requirements data are transmitted to the building envelope energy-saving control learning model to analyze the collaborative control relationship of building energy consumption and energy saving, so as to design building energy consumption and energy saving optimization control strategies.

2. The building energy consumption optimization and control method based on reinforcement learning according to claim 1, characterized in that, Step S1 includes the following steps: Step S11: Obtain building multi-source heterogeneous energy consumption data and basic data of building envelope; Step S12: Perform time-series alignment and spatial mapping of building multi-source heterogeneous energy consumption data and building envelope basic data to analyze building multi-source characterization attributes and generate building multi-source characterization attribute data; Step S13: Design the building equivalent thermal network envelope based on the building multi-source characterization attribute data, and generate the building equivalent thermal network envelope data; Step S14: Perform thermal parameter identification and solution processing on the building equivalent thermal network envelope data to generate building identification thermal parameter data; Step S15: Analyze the thermal dynamic response characteristics of the building envelope based on the building equivalent thermal network envelope data and the building identified thermal parameter data, and generate thermal dynamic response characteristic data of the building envelope.

3. The building energy consumption optimization and control method based on reinforcement learning according to claim 2, characterized in that, The building multi-source heterogeneous energy consumption data mentioned in step S11 includes building environment data, building equipment operation data, and building equipment energy consumption data. The building envelope basic data includes building envelope geometric data and building envelope concrete mix proportion data.

4. The building energy consumption optimization and control method based on reinforcement learning according to claim 2, characterized in that, Step S13 includes the following steps: Step S131: Analyze the thermal state and boundary characteristics of the building envelope based on the multi-source characterization attribute data of the building, and generate thermal state boundary characteristic data of the building envelope; Step S132: Based on the multi-source characterization attribute data of the building, perform structural characteristic classification processing of the building envelope to generate building envelope classification data; Step S133: Design the equivalent thermal network envelope structure by using the thermal state boundary feature data of the building envelope structure and the building envelope structure division data, and generate the equivalent thermal network envelope structure data.

5. The building energy consumption optimization and control method based on reinforcement learning according to claim 1, characterized in that, Step S2 includes the following steps: Step S21: Extract the concrete mix proportion area of ​​the building envelope structure based on the basic data of the building envelope structure, and generate concrete mix proportion area data of the building envelope structure. Step S22: Perform thermal conductivity analysis of the concrete mix proportion area based on the concrete mix proportion area data of the building envelope structure, and generate thermal conductivity data of the concrete mix proportion area based on thermal property mapping. Step S23: Analyze the neighborhood thermal conductivity influence characteristics of the thermal dynamic response of the building envelope on the thermal conductivity data of the concrete mix proportion area mapped by the thermal property data, and generate neighborhood thermal conductivity influence characteristic data of the thermal dynamic response. Step S24: Based on the thermal dynamic response neighborhood thermal conductivity influence characteristic data and the thermal property mapping concrete mix proportion region thermal conductivity data, perform material thermal conductivity characteristic analysis of the building envelope structure to generate material thermal conductivity characteristic data of the building envelope structure.

6. The building energy consumption optimization and control method based on reinforcement learning according to claim 5, characterized in that, Step S3 includes the following steps: Step S31: Discretize the concrete mix proportion data of the building envelope to generate discrete concrete mix proportion data. Step S32: Based on the discretized data of concrete mix proportions, design simulation conditions for multiple schemes of building envelope materials to obtain simulation condition data for multiple schemes of building envelope materials; Step S33: Analyze the boundary conditions for building energy consumption simulation based on the building's multi-source characterization attribute data, and generate building energy consumption simulation boundary condition data; Step S34: Based on the multi-scheme simulation condition data of building envelope materials and the boundary condition data of building energy consumption simulation, perform multi-scheme simulation thermal balance processing of building envelope to generate multi-scheme simulation thermal balance data of building envelope. Step S35: Perform multi-scheme simulation energy consumption characteristic analysis on the multi-scheme simulation thermal balance data of the building envelope, and generate multi-scheme simulation energy consumption characteristic data of the building envelope.

7. The building energy consumption optimization and control method based on reinforcement learning according to claim 6, characterized in that, Step S32 includes the following steps: Step S321: Perform material-specific combination characteristic analysis based on the discretized data of concrete mix proportions to generate material-specific combination characteristic data; Step S322: Perform thermal conductivity characteristic mapping on the material-specific combination characteristic data using the thermal conductivity characteristic data of the building envelope material to generate material combination thermal conductivity characteristic mapping data; Step S323: Design simulation conditions for multiple schemes of building envelope materials based on the thermal conductivity characteristic mapping data of material combination, and obtain simulation condition data for multiple schemes of building envelope materials.

8. The building energy consumption optimization and control method based on reinforcement learning according to claim 1, characterized in that, Step S4 includes the following steps: Step S41: Analyze the time-series changes in the thermal state of the building envelope using the thermal dynamic response characteristic data of the building envelope and the simulation energy consumption characteristic data of the building envelope under multiple schemes, and generate time-series change characteristic data of the thermal state of the building envelope under multiple schemes. Step S42: Establish an optimized trend prediction model for the thermal state of the multi-scheme building envelope based on the time-series variation characteristic data of the thermal state of the multi-scheme building envelope; Step S43: By optimizing the multi-scheme building envelope thermal state trend prediction model, perform multi-scheme building envelope simulation energy consumption trend analysis on the multi-scheme building envelope energy consumption characteristic data, and generate multi-scheme building envelope energy consumption trend data.

9. The building energy consumption optimization and control method based on reinforcement learning according to claim 8, characterized in that, Step S42 includes the following steps: A model for the trend change of the thermal state of a multi-scheme enclosure structure is established by using a pre-defined long short-term memory neural network algorithm and time-series change characteristic data of the thermal state of the multi-scheme enclosure structure. This model is used to obtain a prediction model for the thermal state of the multi-scheme enclosure structure. The prediction model is then subjected to particle swarm optimization of the trend prediction relationship of the multi-scheme enclosure structure to obtain an optimized prediction model for the thermal state of the multi-scheme enclosure structure.

10. The building energy consumption optimization and control method based on reinforcement learning according to claim 1, characterized in that, Step S5 includes the following steps: Step S51: Obtain building operation and maintenance target requirements data; Step S52: Based on the energy consumption trend data of multiple schemes for the building envelope, perform state representation spatial design of the energy consumption trend of multiple schemes to obtain the state representation spatial data of the energy consumption trend of multiple schemes. Step S53: Extract the influencing factors of the multi-scheme energy consumption trend of the building envelope based on the multi-scheme energy consumption trend data, and generate multi-scheme energy consumption trend influencing factor data of the building envelope. Step S54: Design the action space of the energy consumption trend of multiple schemes by using the data of the factors affecting the energy consumption trend of the building envelope, and obtain the action space data of the energy consumption trend of multiple schemes. Step S55: Analyze the collaborative relationship of the influence on the action space based on the energy consumption trend of multiple schemes, and generate collaborative relationship data of the influence on the action space; Step S56: Based on the preset reinforcement learning algorithm, establish the energy-saving control learning relationship of the building envelope structure by using the spatial data of the energy consumption trend state representation of multiple schemes and the spatial data of the energy consumption trend influence of multiple schemes, and generate the energy-saving control learning model of the building envelope structure. Step S57: Transmit the building operation and maintenance target requirement data to the building envelope energy-saving control learning model to analyze the collaborative control relationship of building energy consumption and energy saving, so as to design a building energy consumption and energy saving optimization control strategy, wherein the building energy consumption and energy saving optimization control strategy includes building envelope concrete mix control data and building equipment operation control data.

Images