Building energy saving optimization design system based on BIM technology

Abstract

This invention relates to the field of building energy-saving design technology, specifically a building energy-saving optimization design system based on BIM technology. This system includes modules for data integration, environmental simulation, building envelope optimization, model updating, energy consumption analysis, and scheme generation. It first imports an initial BIM model and extracts building information; then, it analyzes the model's geographical, geometric, and material properties, and generates an external micro-environment field by combining it with an external environment database; next, it inputs this field data and building information into an improved particle swarm optimization algorithm. This algorithm dynamically adjusts the particle interaction network based on the spatial topological relationships of building components, iteratively searching to obtain an optimized scheme for the building envelope's structural parameters; then, it applies this scheme to update the BIM model, and integrates internal dynamic load data to form an energy consumption analysis model; finally, it generates a comprehensive building energy-saving design scheme.

Description

Technical Field

[0001] This invention relates to the field of building energy-saving design technology, and in particular to a building energy-saving optimization design system based on BIM technology. Background Technology

[0002] In the field of architectural design, achieving building energy efficiency targets relies on precise performance analysis and optimized design. Current conventional practice typically applies Building Information Modeling (BIM) technology, building performance simulation, and optimization algorithms as independent processes. Designers first complete geometric modeling in BIM software, then import the simplified model into separate energy simulation software (such as EnergyPlus and IESVE), manually setting or inputting material properties of building components, geographical location and climate data, and internal load parameters. After performance simulation, if the results do not meet energy efficiency targets, designers must rely on experience to return to the BIM model to adjust the design scheme, such as changing the building envelope or window types, and then re-perform the simulation calculation. This iterative process is inefficient and highly dependent on the designer's experience. Another common approach is to use a parametric platform to couple BIM and the simulation engine, and introduce standard optimization algorithms (such as genetic algorithms and standard particle swarm optimization) to automatically optimize design parameters.

[0003] Existing technical solutions have significant drawbacks. First, there is a disconnect and information loss in data transfer between the BIM model and the performance simulation environment. The rich semantic, precise geometric, and material information of components contained in the BIM model is often simplified or requires extensive manual redefinition when converted to the simulation model. The loading of external environmental parameters is also often coarse-grained static data, which is difficult to reflect the local micro-environmental differences caused by the specific building form, affecting the simulation accuracy and the reliability of the optimization starting point. Second, in the optimization stage, conventional optimization algorithms (such as the standard particle swarm optimization algorithm) pre-set and fixed the information interaction network between particles when searching for energy-saving design parameters, without considering the physical connections and spatial topological relationships between the building's own components (such as adjacent walls, floors, and exterior walls). This "black box" optimization may lead to a mismatch between the algorithm's search direction and the actual thermal performance coupling relationship of the building, resulting in low optimization efficiency and parameter combinations that may be physically incompatible, making them impractical.

[0004] Therefore, the problems to be solved focus on two aspects: first, how to achieve lossless and automated data transfer from BIM design model to high-precision environmental simulation, especially generating external micro-environment fields that fit the specific form of the building, so as to provide more realistic boundary conditions for optimization; second, how to embed the spatial logic of the building into the optimization algorithm so that the parameter search process can intelligently reflect the topological relationship of building components, thereby improving the optimization efficiency and the effectiveness of the results. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by proposing a building energy-saving optimization design system based on BIM technology.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a building energy-saving optimization design system based on BIM technology, comprising:

[0007] The data integration module is used to import the initial BIM design model of the building and identify and extract the building envelope information, spatial division information and equipment system information in the initial BIM design model;

[0008] The environment simulation module is used to parse the geographical location, geometric attributes and material attributes of building components from the initial BIM design model, and combine them with a preset external environment parameter database to generate building external micro-environment field data.

[0009] The building envelope optimization module is used to input the building external microenvironment field data, building envelope structure information and spatial division information into an improved particle swarm optimization algorithm. The improved particle swarm optimization algorithm dynamically adjusts the interaction network of the particle swarm according to the spatial topological relationship of the building components, and iteratively searches to obtain an optimized scheme of building envelope construction parameters that meets the preset thermal performance constraints.

[0010] The model update module is used to update the initial BIM design model by applying the optimization scheme of the building envelope construction parameters, and generate a thermal performance optimization model.

[0011] The energy consumption analysis module is used to acquire preset hourly data on personnel activities and equipment operating loads inside the building, and integrate it with the thermal performance optimization model to form a building energy consumption analysis model that includes dynamic internal disturbances.

[0012] The scheme generation module generates a comprehensive energy-saving design scheme for buildings based on the building energy consumption analysis model.

[0013] As a further aspect of the present invention, the environmental simulation module generates building external microenvironment field data, including:

[0014] Based on the geographical location of the building components, typical annual meteorological data of the area where the building is located are obtained from a preset external environmental parameter database, including hourly dry-bulb temperature, wet-bulb temperature, direct solar radiation, diffuse radiation, wind speed and wind direction.

[0015] Using computational fluid dynamics principles and combining the geometric properties of the building components, the wind field distribution and solar radiation shading around the building are simulated.

[0016] The simulated wind field, temperature field, and radiation field data are coupled with the material properties of the building's outer surface to obtain the hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution of each outer surface of the building, thus forming the building's external microenvironment field data.

[0017] As a further aspect of the present invention, the environmental simulation module couples the simulated wind field, temperature field, and radiation field data with the material properties of the building's exterior surface to obtain the hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution of each exterior surface of the building, forming the building's external micro-environmental field data, including:

[0018] The hourly wind speed, air temperature, and total solar radiation irradiance at the location of the building's outer surface are obtained from the simulation results of the external microenvironment field data.

[0019] Based on the material properties of the building's exterior surface, determine its solar radiation absorptivity, long-wave emissivity, roughness, and the azimuth and tilt angles determined by the geometric properties of the building components.

[0020] Based on the total solar radiation illuminance, the azimuth and tilt angle of the building's outer surface, calculate the corrected direct solar radiation illuminance and diffuse solar radiation illuminance received by the building's outer surface.

[0021] The solar radiation gain is calculated by combining the solar radiation absorptivity of the building's exterior surface. The hourly solar heat gain coefficient is calculated by using the ratio of the solar radiation gain to the corrected total solar irradiance. The hourly convective heat transfer coefficient is calculated by combining the formula for calculating the convective heat transfer coefficient of the building's exterior surface with the hourly wind speed and the roughness of the building's exterior surface as input variables.

[0022] Using the surface heat balance equation, with the solar radiation heat gain, convective heat transfer, radiative heat transfer with the surrounding environment, and the heat conduction process of the building's outer surface as input, the hourly surface temperature distribution of each outer surface of the building is obtained by iterative solution.

[0023] The hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution are associated with the corresponding building exterior surface components and organized according to time series to generate the building exterior microenvironment field data.

[0024] As a further aspect of the present invention, the improved particle swarm optimization algorithm in the enclosure optimization module dynamically adjusts the interaction network of the particle swarm based on the spatial topological relationship of building components, including:

[0025] Define each particle in the particle swarm as a candidate combination of enclosure structure construction parameters;

[0026] Based on the initial BIM design model, the spatial adjacency and thermal bridge relationships between wall, roof, and window components in the building envelope are analyzed, and a building component adjacency diagram is constructed.

[0027] With reference to the structure of the adjacency graph of the building components, the interaction network of the particle swarm is initialized so that the particle parameters corresponding to the components that are adjacent in space or have thermal correlation have a higher connection probability in the interaction network.

[0028] During the particle swarm optimization process, the overall thermal performance of the parameter combination represented by each particle is evaluated, and its fitness is calculated.

[0029] The updating of particle velocity and position depends not only on the individual historical best solution and the group historical best solution, but also on the range of neighborhood best solutions for each particle, which is limited to the solutions of other particles directly connected to the current particle in the dynamic interaction network.

[0030] As the iteration proceeds, based on the convergence of particle fitness, the connection weights between different particles in the interaction network are dynamically sparsified or strengthened, guiding the search to achieve a balance between parameter space with superior local performance and global exploration, and finally outputting the optimized scheme for the construction parameters of the enclosure structure.

[0031] As a further aspect of the present invention, the scheme generation module generates a comprehensive building energy-saving design scheme based on the building energy consumption analysis model, including:

[0032] The building energy consumption analysis model is subjected to hourly dynamic energy consumption simulation throughout the year to calculate a detailed energy consumption sequence including cooling energy consumption, heating energy consumption and lighting energy consumption.

[0033] High-energy-consumption characteristic periods are extracted from the detailed energy consumption sequence, and the simulated boundary conditions corresponding to the high-energy-consumption characteristic periods are input into an ensemble learning model. The ensemble learning model is trained to predict the energy-saving potential of different equipment control strategies.

[0034] Based on the prediction results of the ensemble learning model, a set of optimized equipment operation strategies are selected using a multi-objective decision-making method.

[0035] The optimization scheme for the building envelope construction parameters is combined with the optimization scheme for the equipment operation strategy to generate the final comprehensive building energy-saving design scheme.

[0036] As a further aspect of the present invention, the building energy consumption analysis model is subjected to hourly dynamic energy consumption simulation throughout the year, including:

[0037] Set the simulation time step to hours and the simulation period to years;

[0038] Within each simulation time step, the external disturbance data, hourly personnel activity and equipment operating load data, and thermal parameters of the building envelope are taken as inputs from the external microenvironment field data of the building.

[0039] Solve the heat balance equations for each zone of the building to calculate the indoor air temperature, the inner surface temperature of each building envelope, and the real-time load of the heating and cooling system for each zone of the building at the current simulation time step.

[0040] The energy consumption of the lighting system, equipment system, heating system, and cooling system are accumulated and calculated to generate the detailed hourly energy consumption sequence.

[0041] As a further aspect of the present invention, extracting high-energy-consumption characteristic periods from the detailed energy consumption sequence includes:

[0042] A sliding window analysis is performed on the detailed energy consumption sequence to identify continuous time windows in which the total energy consumption exceeds a preset high energy consumption threshold;

[0043] Within each high-energy-consumption time window, further analyze the energy consumption composition to distinguish the time period types dominated by cooling energy consumption, heating energy consumption, or overall energy consumption.

[0044] Record the start and end times, outdoor weather conditions, building indoor load conditions, and thermal state of the building envelope for each high energy consumption characteristic period as the simulation boundary conditions.

[0045] As a further aspect of the present invention, the training and prediction of the ensemble learning model includes:

[0046] Construct a pool of prediction models that use multiple energy consumption simulation engines as base learners;

[0047] Using simulation data or actual operation data of historical building projects as training samples, each sample includes the simulated boundary conditions under specific working conditions, the equipment control strategy adopted under the corresponding working conditions, and the energy consumption change corresponding to the equipment control strategy.

[0048] The ensemble learning model is trained to learn the complex mapping relationship from simulated boundary conditions and candidate policies to energy-saving potential;

[0049] In application, the simulated boundary conditions of the high energy consumption period are input into the trained ensemble learning model, and a series of candidate equipment control strategies are enumerated or searched. The ensemble learning model outputs a predicted value of the energy-saving potential of each candidate strategy.

[0050] As a further aspect of the present invention, the step of using a multi-objective decision-making method to select a set of optimized equipment operation strategies includes:

[0051] Define multiple decision objectives, including maximizing energy efficiency, minimizing investment costs, and minimizing the impact on indoor thermal comfort;

[0052] The energy-saving potential predicted by the integrated learning model is transformed into energy-saving effect, and the performance of each candidate strategy is evaluated on various decision objectives by combining the preset strategy implementation cost data and the simulation data of the strategy's impact on the indoor thermal environment.

[0053] By applying the Pareto optimal solution selection algorithm, a set of non-dominated solutions is found from all candidate strategies, forming the Pareto optimal solution set;

[0054] Based on the preference weights preset during the design phase, a final solution is selected from the Pareto optimal solution set as the optimization scheme for the equipment operation strategy.

[0055] As a further aspect of the present invention, the optimization scheme for the building envelope construction parameters is combined with the optimization scheme for the equipment operation strategy to generate a final comprehensive building energy-saving design scheme, including:

[0056] The technical feasibility of the optimized structural parameters of the enclosure structure is verified to ensure that the selected structural parameters comply with local specifications, the materials are available, and the construction is feasible.

[0057] The system compatibility verification of the equipment operation strategy optimization scheme is performed to ensure that the strategy does not exceed the physical operating limits of the equipment system.

[0058] The two verified solutions are integrated in terms of time and measure type to generate an implementation list that includes passive technical measures and active control strategies.

[0059] In the implementation list, each measure is marked with its associated BIM components, systems, expected energy-saving contribution, estimated cost, and implementation stage, forming a structured document of the comprehensive building energy-saving design scheme.

[0060] Compared with the prior art, the advantages and positive effects of the present invention are as follows:

[0061] By automatically extracting the precise geographical location, 3D geometric attributes, and hierarchical material composition of building components from the BIM design model, and deeply integrating this data with a pre-set external environmental parameter database containing local hourly weather, solar radiation, and wind field information, high-resolution building external microenvironmental field data is directly generated. This method replaces the traditional approach of relying on manually simplified models, using uniform static climate files, or simple interpolation. Because the microenvironmental field accurately reflects the differences in solar radiation heat gain, shading, and wind pressure distribution caused by the building's shape, orientation, and local details, the thermal boundary conditions of the building envelope used for subsequent optimization calculations are more realistic and detailed. This fundamentally improves the accuracy of the input conditions for building energy consumption simulation, enabling the envelope optimization design based on this data to more accurately evaluate the actual performance of different construction schemes under real microclimates, thereby significantly enhancing the reliability of the optimization scheme and the predictive ability of actual energy-saving effects.

[0062] Based on the actual spatial adjacency, enclosure, and support topological relationships between building components, a dynamic information interaction and learning network among individual particles in the particle swarm optimization algorithm is constructed and adjusted. During the optimization iteration process, particles representing design parameters of different building components (such as east-facing exterior walls, west-facing windows, and roofs) do not exchange information randomly or in a completely connected manner. Instead, they form subgroups or specific network structures with priority connection rights based on the spatial relationships of their corresponding components in the BIM model (such as components sharing thermal bridges or components in the same facade's solar radiation sequence). This mechanism enables the optimization search process to more intelligently identify and coordinate component parameter combinations with closely related thermal performance, guiding the search direction to better align with building physics. Compared to the standard particle swarm optimization algorithm, which treats all building parameters as a flat set of independent or completely coupled components, this method significantly improves optimization efficiency and convergence speed in complex multi-parameter spaces, and makes the output optimized parameter scheme not only mathematically optimal but also more coordinated and constructible in terms of physical spatial relationships. Attached Figure Description

[0063] Figure 1 This is a sequence diagram of the BIM-based building energy-saving optimization design system described in this invention;

[0064] Figure 2 A flowchart illustrating the process of generating building external microenvironment field data through coupled calculations of the environmental simulation module;

[0065] Figure 3 A flowchart for generating a comprehensive building energy-saving design scheme for the scheme generation module. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0067] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0068] See Figure 1 This invention provides a building energy-saving optimization design system based on BIM technology, specifically including:

[0069] The initial BIM design model of the building is imported through the data integration module, which automatically identifies and extracts information on the building envelope, spatial division, and equipment systems. The environmental simulation module parses the geographical location, geometric attributes, and material properties of building components from this initial model. Combined with a pre-set external environmental parameter database, it generates micro-environmental field data describing the building's surrounding conditions, such as wind, heat, and light. The envelope optimization module receives the aforementioned micro-environmental field data, envelope information, and spatial division information, and inputs them into an improved particle swarm optimization algorithm. Based on the spatial topology of building components analyzed from the BIM model, this algorithm dynamically adjusts the interaction network structure of the particle swarm during the optimization process. Through iterative search, it outputs an optimized scheme for the envelope construction parameters that meets pre-set thermal performance constraints. The model update module applies this optimization scheme to update the relevant component parameters in the initial BIM design model, thereby generating a thermal performance optimization model reflecting the optimized physical characteristics. The energy consumption analysis module further acquires pre-set hourly data on personnel activity and equipment operating load within the building, integrates it with the thermal performance optimization model, and constructs a building energy consumption analysis model containing dynamic internal disturbances that can be used for fine-grained simulation. Based on this building energy consumption analysis model, the scheme generation module performs simulation analysis and strategy optimization to generate a comprehensive building energy-saving design scheme that includes passive envelope optimization and active equipment control.

[0070] In one embodiment of the present invention, see [reference] Figure 2The process of generating micro-environmental field data for the building's external environment using the environmental simulation module involves multiple steps. Based on the geographical location of building components analyzed from the initial BIM design model, the module retrieves typical annual meteorological data for the building's location from a pre-set external environmental parameter database. This data includes hourly dry-bulb temperature, wet-bulb temperature, direct solar radiation, diffuse radiation, wind speed, and wind direction. Utilizing computational fluid dynamics principles and combining the geometric properties of the building components, the module numerically simulates the wind field distribution around the building and the solar radiation shading caused by the building itself and surrounding objects. After acquiring the simulated wind field, temperature field, and radiation field data, the module couples these with the material properties of the building's external surfaces. Specifically, it obtains the hourly wind speed, air temperature, and total solar irradiance at the center point of each external surface from the simulation results. Based on the material properties of the building's external surfaces provided by the BIM model, it determines the solar radiation absorptivity, long-wave emissivity, surface roughness, and the azimuth and tilt angles determined by the component's geometric properties. Based on the total solar irradiance and the azimuth and tilt angle of the building's exterior surface, the actual direct and diffuse solar irradiance received by the surface after angle correction is calculated. Combined with the surface's solar radiation absorptivity, the solar radiative heat gain is calculated, and the hourly solar heat gain coefficient of the surface is obtained by using the ratio of solar radiative heat gain to the corrected total solar irradiance. Simultaneously, hourly wind speed and surface roughness are used as input variables and substituted into the empirical formula for calculating the convective heat transfer coefficient of the building's exterior surface to obtain the hourly convective heat transfer coefficient. The module utilizes the surface heat balance equation, which considers solar radiative heat gain, convective heat transfer based on the convective heat transfer coefficient and air temperature, long-wave radiation heat transfer with the sky and surrounding environmental surfaces, and heat conduction through the building envelope itself. By iteratively solving this equation, the hourly surface temperature distribution of each exterior surface of the building is obtained. Finally, the module associates the calculated hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution data with the building exterior surface component objects in the BIM model, and organizes them according to the time series to form structured building exterior microenvironment field data.

[0071] In its implementation, the environmental simulation module generates building external microenvironmental field data through a series of sequential calculation and data coupling steps. From the initial BIM design model provided by the "data integration module," the module parses the geographic coordinates, geometric dimensions, shape information, and material type identifiers of each building component instance. Based on the geographic coordinates, the module accesses a pre-set external environmental parameter database, which stores typical annual meteorological data for different geographical regions. The module extracts hourly meteorological parameters matching the building coordinates for the entire year, including dry-bulb temperature, wet-bulb temperature, direct solar radiation intensity, diffuse solar radiation intensity, wind speed, and wind direction. After obtaining the basic meteorological data, the environmental simulation module utilizes a computational fluid dynamics software kernel, combined with the 3D geometric mesh information of the building structure and surrounding major obstructions extracted from the BIM model, to perform unsteady-state numerical simulations. This simulation calculates the wind speed vector field, air temperature field, and hourly solar radiation shadow distribution formed by the mutual shading of the building shape and its surrounding environment under typical annual meteorological conditions.

[0072] In practical implementation, coupling the wind field, temperature field, and radiation field data obtained from computational fluid dynamics simulations with the physical properties of the building's external surface is a crucial step in generating microenvironmental field data for the building's external environment. For each building external surface component defined in the BIM model, the module locates and extracts the hourly wind speed, hourly air temperature, and hourly total solar irradiance (without angle correction) at the center point of the external surface from the computational fluid dynamics simulation results. Simultaneously, based on the material type assigned to the external surface in the BIM model, the module retrieves the corresponding thermophysical property parameters from the associated material library. These parameters include solar radiation absorptivity, long-wave emissivity, and surface roughness. The module also determines the azimuth and tilt angles by calculating the geometric normal vector of the external surface. It is understandable that calculating the actual solar radiation received by the building's external surface is a necessary step. Based on the solar cycle, the module uses extracted direct and diffuse solar radiation data, combined with the azimuth and tilt angles of the building's exterior surface, to calculate the hourly direct and diffuse solar radiation illuminance actually received by the surface using a solar angle correction model. The sum of these two is the corrected total solar irradiance. In some embodiments, the module then calculates the hourly solar heat gain coefficient and the hourly convective heat transfer coefficient. For the solar heat gain coefficient, the module multiplies the solar radiation absorptivity of the building's exterior surface by the corrected total solar irradiance to obtain the solar heat gain. The solar heat gain coefficient is defined as the ratio of solar heat gain to the corrected total solar irradiance. For the convective heat transfer coefficient, the module uses an empirical formula based on surface roughness and wind speed for calculation. The formula is as follows:

[0073]

[0074] in: The coefficient of convective heat transfer representing the exterior surface of a building. Hourly wind speed representing the center of the building's outer surface , , is an empirical coefficient related to the roughness of the building's exterior surface. The module substitutes the hourly wind speed value and the coefficient corresponding to the roughness obtained from the material properties into the above formula to calculate the hourly convective heat transfer coefficient of the building's exterior surface.

[0075] In practical implementation, solving for the temperature distribution on the building's exterior surface requires the use of a surface heat balance equation. The module establishes a heat balance equation for each building's exterior surface. The heat gain and loss terms considered in the equation include: solar radiation heat gain calculated based on the solar heat gain coefficient and total solar irradiance; convective heat transfer calculated based on the convective heat transfer coefficient, air temperature, and the surface temperature to be determined; far-infrared radiation heat transfer calculated based on the surface long-wave emissivity and the average radiant temperature of the surrounding environment; and heat conduction from the exterior surface to the interior surface, determined by the thermal conductivity of the building envelope material provided by the BIM model. Optionally, the module uses an iterative numerical solution method to solve this nonlinear equation. At each hourly step, the module sets an initial estimate of the surface temperature, then calculates the heat transfer terms, checks whether the heat balance equation is satisfied, and if not, adjusts the surface temperature estimate and recalculates until the equation residual is less than the convergence tolerance, thus obtaining the stable surface temperature of the building's exterior surface at the current moment. Understandably, this iterative process is repeated for all building exterior surfaces over all 8760 annual time steps, ultimately yielding the hourly surface temperature distribution for each exterior surface. In some embodiments, after completing all calculations, the module performs data organization and association. The module arranges the three types of data—hourly solar heat gain coefficient, hourly convective heat transfer coefficient, and hourly surface temperature distribution—according to a time series (e.g., from 00:00 on January 1st to 23:00 on December 31st). Optionally, the module uses unique identifiers to tightly associate and bind the data of each time series with the original building exterior surface component object in the BIM model that generated the data, ultimately forming a structured, queryable building external microenvironment field dataset. This dataset contains the boundary condition parameters for heat exchange between each building exterior surface and the external environment every hour throughout the year.

[0076] In one embodiment of the present invention, in the building envelope optimization module, the improved particle swarm optimization algorithm dynamically adjusts the interaction network of the particle swarm based on the spatial topological relationships of the building components. The algorithm defines each particle in the optimization problem as having a position vector representing a candidate combination of building envelope construction parameters, such as the thickness of insulation materials, heat transfer coefficient, and shading coefficient of walls, roofs, and windows facing different directions. Based on the initial BIM design model, the algorithm analyzes the spatial relationships between different components in the building envelope, identifies the spatial adjacency relationships between walls, between walls and windows, and between walls and roofs, and analyzes the structural associations that may form thermal bridges, thereby constructing a building component adjacency graph with building components as nodes and thermal connections as edges. During the algorithm initialization phase, using the structure of this building component adjacency graph as a reference, an initial interaction network for the particle swarm is established, so that the parameter dimensions corresponding to adjacent components or those with strong thermal connections in physical space have a higher connection probability in the particle interaction network, meaning that the particle dimensions representing these parameters can more easily exchange information. During the particle swarm optimization process, the overall thermal performance of the parameter combination represented by each particle is evaluated, and its fitness value is calculated. The updates to particle velocity and position depend not only on the individual's historical best solution and the historical best solution of the entire swarm, but also on the neighborhood best solution referenced by each particle during the update. The selection range is limited to the solution set corresponding to other particles directly connected to the current particle in the dynamic interaction network. As the iteration progresses, the algorithm dynamically sparsifies or strengthens the weights of connections between different particles in the interaction network based on the convergence and distribution of the particle swarm fitness. For example, it reduces connections between groups of particles with similar fitness to promote local development, or increases connections between groups of particles with significantly different fitness to promote global exploration. This guides the search process to achieve a balance between a parameter space with superior local performance and global exploration, ultimately outputting the optimal enclosure structure construction parameter optimization scheme with the best overall performance under the constraints.

[0077] In practical implementation, the improved particle swarm optimization algorithm in the building envelope optimization module dynamically adjusts the particle swarm interaction network based on the spatial topological relationships of building components. The process begins with defining the optimization problem. The improved particle swarm optimization algorithm defines each particle in the particle swarm as a multi-dimensional vector. This multi-dimensional vector corresponds to a candidate combination of building envelope construction parameters. Each dimension of the multi-dimensional vector represents a specific thermal parameter of a particular building envelope component, such as the insulation layer thickness of the south-facing exterior wall, the heat transfer coefficient of the east-facing exterior window, and the solar radiation absorptivity of the roof. In practical implementation, constructing a building component adjacency graph reflecting the relationships between components is a fundamental step in the improved particle swarm optimization algorithm. Based on the initial BIM design model provided by the "data integration module," the improved particle swarm optimization algorithm automatically analyzes the spatial adjacency relationships between components in the building envelope by parsing the spatial coordinates and geometric relationships of building components. For example, it identifies which windows a wall connects to, and which walls the roof intersects with. Simultaneously, the improved particle swarm optimization algorithm identifies component connections with thermal bridging risks based on a construction knowledge base, such as the locations where reinforced concrete beams and columns penetrate the insulation layer. The improved particle swarm optimization algorithm uses identified components such as walls, roofs, and windows as nodes, and spatial adjacency and thermal bridge relationships between components as edges, to construct an undirected adjacency graph of building components. It can be understood that the structure of the adjacency graph reflects the intensity of the mutual influence of the thermal performance of various components in the building physics model.

[0078] In its implementation, the improved particle swarm optimization algorithm initializes and dynamically adjusts the inter-particle interaction network based on the adjacency graph of building components. The improved algorithm uses the structure of the adjacency graph as a reference to initialize the connection probability matrix of the inter-particle interaction network. Two building components directly connected by edges in the adjacency graph have their corresponding parameter dimensions assigned a higher initial connection probability in the interaction network. This means that in subsequent optimization, the particle dimensions representing the parameters of these two components are more likely to be considered "neighbors" and exchange information. In some embodiments, the improved particle swarm optimization algorithm evaluates particle fitness and updates particle states in each iteration. In each iteration, the improved particle swarm optimization algorithm substitutes the combination of building envelope construction parameters represented by each particle into the energy consumption simulation model to calculate the overall building thermal performance indicators, such as the annual cumulative heating and cooling load, and uses this to calculate the particle fitness value. The updating of particle velocity and position depends not only on the particle's own historical best solution and the historical best solution of the entire particle swarm, but also, when updating its velocity and position, the neighborhood best solution used as a reference is strictly limited to the set of historical best solutions corresponding to other particle dimensions that are connected to the current particle dimension within the connection probability matrix of the dynamic interaction network. Optionally, the specific update strategy for particle velocity and position in the improved particle swarm optimization algorithm follows a specific formula. The velocity update formula for the particle dimension in the iteration is:

[0079]

[0080] in: Representative particles In the Dimensional parameters on Speed ​​during the next iteration; Representative particles In the Dimensional parameters on The position at the next iteration; It is inertial weight; , , It is a learning factor; , , It is a random number between 0 and 1; It is a particle In the Dimensional parameters up to The individual historical optimal solution in the next iteration; The entire particle swarm is at the th Dimensional parameters up to The global historical optimal solution for the next iteration; In dynamic interactive networks with particles In the Among the particle neighbors directly connected in terms of dimension parameter, up to The neighborhood historical best solution in the next iteration.

[0081] In practical implementation, dynamically adjusting the connection weights of the interaction network as the optimization process progresses is a core feature of the improved particle swarm optimization algorithm. The improved PSO algorithm monitors the convergence of the particle swarm's fitness, for example, by calculating the particle swarm fitness variance or the average distance between particles. In some embodiments, when premature convergence of the particle swarm's fitness and a decrease in diversity are detected, the improved PSO algorithm dynamically sparsifies the connection weights within high-fitness particle groups in the interaction network connection probability matrix, while selectively strengthening certain connection weights between high-fitness and low-fitness particle groups, thereby encouraging particles to escape local optima and explore globally. Conversely, when the particle swarm is in a divergent search state, the improved PSO algorithm strengthens the connection weights within particle groups with similar fitness, guiding particles to perform a refined local search within a parameter space with superior performance. Optionally, the adjustment of connection weights is based on a composite function of the difference in fitness between particles and the parameter distance. After a preset number of iterations, the improved particle swarm optimization algorithm outputs a multi-dimensional vector representing the particle with the best fitness. This multi-dimensional vector is the optimized scheme of the building envelope construction parameters that satisfies the preset thermal performance constraints.

[0082] In one embodiment of the present invention, see [reference] Figure 3The scheme generation module generates a comprehensive energy-saving design scheme for buildings based on a building energy consumption analysis model. This module first performs hourly dynamic energy consumption simulations on the building energy consumption analysis model, which integrates all boundary conditions, throughout the year, calculating a detailed energy consumption sequence including cooling, heating, lighting, and equipment energy consumption. The module extracts high-energy-consumption characteristic periods from this detailed energy consumption sequence and inputs the corresponding simulation boundary conditions—including meteorological parameters, indoor load conditions, and building envelope status—into a pre-trained ensemble learning model. This ensemble learning model is trained to predict the energy-saving potential of implementing various equipment control strategies under different simulation boundary conditions. Based on the prediction results of the ensemble learning model, the module uses a multi-objective decision-making method to select a set of optimized equipment operation strategies that balance energy-saving effects, economic costs, and comfort impacts from numerous candidate equipment control strategies. Finally, the module merges the optimized building envelope construction parameters output by the building envelope optimization module with the aforementioned optimized equipment operation strategies to generate the final comprehensive energy-saving design scheme for buildings. Prior to the merger, the technical feasibility of the optimized building envelope structural parameters was verified. This included checking whether the selected structural parameters met local building codes, whether the materials involved were available in the local market, and whether the construction methods were feasible. Simultaneously, the system compatibility of the optimized equipment operation strategy was verified. This involved checking whether the equipment operation parameters set in the strategy, such as the air conditioning temperature range and fresh air volume adjustment range, exceeded the physical operating limits of the actual equipment systems within the building. The two verified solutions were then integrated along the time and measure type dimensions to generate an implementation list containing passive technical measures and active control strategies. This implementation list included the associated BIM model components or systems for each measure, its expected energy-saving contribution, estimated incremental cost, and suggested implementation stage, ultimately forming a structured comprehensive building energy-saving design document.

[0083] In its implementation, the scheme generation module generates a comprehensive energy-saving building design scheme based on the building energy consumption analysis model, involving a series of logical steps. The module performs hourly dynamic energy consumption simulations on the building energy consumption analysis model for 8760 hours throughout the year, simulating detailed energy consumption sequences including cooling, heating, lighting, and equipment energy consumption. These detailed energy consumption sequences serve as the foundation for subsequent analysis. The module extracts high-energy-consumption characteristic periods from the detailed energy consumption sequences and inputs the simulation boundary conditions corresponding to these periods into a pre-trained ensemble learning model. This model is trained to predict the energy-saving potential of different equipment control strategies under various boundary conditions. Based on the predicted output of the ensemble learning model, the module employs a multi-objective decision-making method to select a set of optimized equipment operation strategies from a series of candidate strategies. These optimized strategies balance energy-saving effects, investment costs, and indoor thermal comfort. Finally, the module merges the optimized building envelope construction parameters output by the building envelope optimization module with the optimized equipment operation strategies to generate the final comprehensive energy-saving building design scheme. In practical implementation, verifying the technical feasibility of the optimized building envelope structural parameters is a necessary step before merging. The scheme generation module checks whether the structural parameters of each type of building envelope in the optimized scheme comply with the mandatory building energy-saving design code of the project location, verifies the feasibility of procuring the specified insulation materials and glass types in the local building materials market, and assesses the feasibility of special structural nodes in terms of construction technology. Simultaneously, the scheme generation module performs system compatibility verification on the optimized equipment operation strategy scheme, verifying whether the recommended equipment operation parameter settings in the optimized scheme, such as the set temperature adjustment range of the air conditioning system and the air volume adjustment ratio of the fresh air system, exceed the physical operating limits of the actual designed or installed equipment systems within the building, such as the rated cooling capacity range of the chiller unit and the minimum allowable air volume of the fan.

[0084] In practical implementation, the validated optimized schemes for building envelope structural parameters and equipment operation strategies are integrated and a structured document is generated. The scheme generation module categorizes the optimized schemes for building envelope structural parameters as long-term passive technical measures throughout the building's entire lifecycle, and the optimized schemes for equipment operation strategies as active control strategies during the building's operation phase. The module also merges and classifies the specific items of the two schemes based on the type of measure, generating a unified list of implementation schemes for comprehensive building energy conservation design. In some embodiments, each measure in the implementation list is associated with multiple attributes. The scheme generation module labels each measure in the implementation list with its associated BIM model component or equipment system number, estimates the percentage of energy-saving contribution of the measure to the building's total energy consumption on an annual scale, estimates the potential incremental or retrofit costs of implementing the measure, and suggests whether to implement the measure during the building design, construction, or operation phases. It is understandable that the scheme generation module ultimately outputs a structured comprehensive building energy conservation design scheme document as a list of implementation schemes with complete attribute annotations. This document can be in the form of tables, graphic reports, or data formats associated with the BIM model. The process of selecting equipment operation strategy optimization schemes using a multi-objective decision-making method in the scheme generation module involves explicit mathematical processing. The scheme generation module defines multiple decision objectives, including maximizing energy-saving effects, minimizing investment costs, and minimizing the impact on indoor thermal comfort. The module transforms the energy-saving potential predicted by the ensemble learning model into specific annual energy-saving data, and combines this with a pre-set equipment strategy implementation cost database and simulated data on the impact of strategies on the indoor predicted average voting index to evaluate the quantitative performance of each candidate equipment control strategy on the three decision objectives of energy-saving effects, investment costs, and thermal comfort impact. The scheme generation module applies the Pareto optimal solution selection algorithm to handle the multi-objective decision problem, finding a set of non-dominated solutions from all candidate equipment control strategies to form a Pareto optimal solution set. In some embodiments, each solution in the Pareto optimal solution set is not inferior to any other solution on all decision objectives, and is strictly superior to other solutions on at least one decision objective. The solution generation module makes a final selection from the Pareto optimal solution set based on the preference weight vector preset during the design phase. This preference weight vector reflects the decision-maker's relative emphasis on the three objectives of energy saving, investment cost, and thermal comfort impact. The solution generation module determines the final solution by calculating the weighted comprehensive evaluation value of each Pareto solution. The formula for calculating the weighted comprehensive evaluation value is as follows:

[0085]

[0086] in: Representing the The comprehensive evaluation value of each Pareto solution; Representing the Each decision objective has a preset preference weight, and the sum of the weights is 1; It is the first The Pareto solution is at the th... The normalized performance values ​​for each decision objective are used. The solution generation module selects the Pareto solution with the highest comprehensive evaluation value as the final selected equipment operation strategy optimization scheme. Optionally, see Table 1, which shows a simplified multi-objective decision evaluation segment.

[0087] Table 1: Multi-objective evaluation table of candidate equipment operation strategies

[0088] Strategy Number Predicted annual energy consumption (kWh) Estimated investment cost (RMB) Predicted PMV impact value Normalized energy saving Normalized cost Normalized PMV impact A1 15,000 8,000 0.12 0.85 0.60 0.70 A2 12,500 5,000 0.18 0.71 0.80 0.50 A3 17,000 12,000 0.10 1.00 0.40 0.80

[0089] Optionally, after calculating the weighted comprehensive evaluation value, assuming the preference weight vector is (energy saving: 0.5, cost: 0.3, comfort: 0.2), the comprehensive evaluation value of strategy A1 calculated by the scheme generation module is 0.5 * 0.85 + 0.3 * 0.60 + 0.2 * 0.70 = 0.735, and so on, selecting the strategy with the highest score. It can be understood that the multi-objective decision-making method enables the finally selected equipment operation strategy optimization scheme to systematically balance multiple often conflicting design objectives.

[0090] In one embodiment of the present invention, the process of performing hourly dynamic energy consumption simulation of the building energy consumption analysis model throughout the year in the scheme generation module follows specific steps. The simulation time step is set to 1 hour, with a complete calendar year as the simulation cycle. Within each simulation time step, the external disturbance data corresponding to this moment in the building's external microenvironment field data, including the temperature of each external surface, solar heat gain, convective heat transfer coefficient, etc., along with preset hourly internal disturbance data such as personnel activity density, equipment power density, and lighting switch schedule, as well as the updated thermal parameters of the building envelope, are used as input conditions. The core of the simulation is to solve the heat balance equations for each thermal zone of the building, calculating the indoor air temperature of each zone, the internal surface temperature of each building envelope, and the real-time load of the heating or cooling system required to maintain the set temperature at the current simulation time step. At each time step, the energy consumption of the lighting system, indoor equipment system, heating system, and cooling system are cumulatively calculated, thereby generating a detailed energy consumption sequence arranged in chronological order. When extracting high-energy-consumption characteristic periods from the detailed energy consumption sequence, a sliding window analysis is performed on the obtained hourly total energy consumption sequence for the entire year. The window length can be set to 6 hours or 12 hours to identify continuous time windows within which the average total energy consumption exceeds a preset high-energy-consumption threshold. Within each identified high-energy-consumption time window, the composition of energy consumption is further analyzed to distinguish whether the period is dominated by cooling energy consumption, heating energy consumption, or a combination of multiple energy consumption such as cooling, heating, lighting, and equipment. The start and end times, outdoor dry-bulb and wet-bulb temperatures, solar radiation, wind speed, and other meteorological conditions, indoor occupant and equipment load conditions, and thermal conditions such as the internal surface temperature and heat transfer characteristics of the building envelope are recorded for each high-energy-consumption characteristic period. This information together constitutes the set of simulation boundary conditions upon which subsequent analyses depend.

[0091] In practice, the scheme generation module follows a standardized process for performing hourly dynamic energy consumption simulations of the building energy consumption analysis model throughout the year. The simulation time step is set to 1 hour, with a complete calendar year (8760 hours) as the simulation period, starting at 00:00 on January 1st and ending at 23:00 on December 31st. Within each simulation time step, the module extracts corresponding external disturbance data from the building's external microenvironment data. This data includes hourly surface temperatures, solar heat gain coefficients, and convective heat transfer coefficients of each external surface. The module also extracts hourly occupant activity density, equipment power density, lighting on / off schedules, and dimming levels from a pre-set internal disturbance database. Simultaneously, the module acquires updated thermal parameters of the building envelope from the building energy consumption analysis model, including wall heat transfer coefficients and window shading coefficients. The scheme generation module uses this data as input to drive the core solver of the building energy consumption analysis model, solving the transient heat balance equations for each thermal zone of the building. These transient heat balance equations consider the effects of heat conduction through the building envelope, convective heat transfer between indoor and outdoor spaces, heat generation from indoor heat sources, and the heating and cooling supply of the air conditioning system. By solving these transient heat balance equations, the module calculates the indoor air temperature of each zone, the temperature of the inner surfaces of each building envelope, and the real-time load of the heating or cooling system required to maintain the set temperature of each zone at the current simulation time step. The scheme generation module then accumulates and calculates the energy consumption of the lighting system, indoor equipment system, heating system, and cooling system for each time step. Lighting and equipment system energy consumption are calculated based on power and operating time, while heating and cooling system energy consumption are calculated based on the real-time system load and system energy efficiency ratio. Finally, a detailed energy consumption sequence of 8760 data points arranged chronologically is generated. This detailed energy consumption sequence forms the basis for subsequent identification and analysis of high-energy-consumption characteristic periods.

[0092] In practical implementation, extracting high-energy-consumption characteristic periods from detailed energy consumption sequences involves sliding window analysis and energy consumption composition analysis. The scheme generation module performs sliding window analysis on the hourly total energy consumption sequence for the whole year. The length of the sliding window can be set to 6 hours, 12 hours, or 24 hours. The scheme generation module moves the sliding window step by step along the time sequence, calculates the average total energy consumption within each window, and identifies continuous time windows where the average total energy consumption within the window exceeds a preset high-energy-consumption threshold as high-energy-consumption time windows. The preset high-energy-consumption threshold can be set to 1.5 times or 2 times the average hourly energy consumption of the building, depending on the building type and scale. In some embodiments, the scheme generation module further analyzes the energy consumption composition within each high-energy-consumption time window. The scheme generation module parses the detailed energy consumption sequence and calculates the proportion of cooling energy consumption, heating energy consumption, lighting energy consumption, and equipment energy consumption in the total energy consumption within each identified high-energy-consumption time window. The scheme generation module categorizes high-energy-consumption time windows into three types based on the dominance of various energy consumption categories: those dominated by cooling energy consumption, those dominated by heating energy consumption, and those dominated by overall energy consumption. Dominance by overall energy consumption means that no single energy consumption component exceeds 50%, but rather multiple energy consumption items are collectively at a high level. The scheme generation module records the start and end times, outdoor weather conditions, building indoor load status, and thermal state of the building envelope for each high-energy-consumption characteristic period. This information collectively constitutes the simulation boundary conditions used in subsequent analyses. Optionally, see Table 2, which shows a segment of a high-energy-consumption characteristic period identified.

[0093] Table 2: Results of High Energy Consumption Period Identification

[0094] Time period number Start and end times Average total energy consumption per window (kWh) Dominant energy consumption type Average outdoor dry-bulb temperature (°C) Indoor population density (people / m²) H1 July 15th 13:00-18:00 185.6 Cooling energy consumption is the main factor. 35.2 0.08 H2 January 20, 08:00-12:00 162.3 Heating energy consumption is the main factor -2.1 0.10 H3 May 10, 09:00-15:00 120.7 Comprehensive energy consumption is the dominant factor 22.5 0.09

[0095] It can be understood that an exemplary formula for calculating the average total energy consumption within a window in sliding window analysis is as follows:

[0096]

[0097] in: Represents a sliding window Average total energy consumption within, This represents the number of hours contained in the sliding window. Represents the starting hour number of the window. Represents the first in the detailed energy consumption sequence The total energy consumption value per hour. In some embodiments, when high-energy-consumption time windows are adjacent or overlap, the scheme generation module will merge them into a longer high-energy-consumption characteristic period. The scheme generation module will output all extracted high-energy-consumption characteristic periods and their corresponding simulated boundary conditions as a structured dataset for subsequent energy-saving strategy analysis and prediction. Optionally, the simulated boundary condition dataset includes not only meteorological and load data, but also key thermal state parameters of the building envelope during the period, such as the average temperature of the inner surface of the envelope in each orientation and the solar heat gain of the windows. These parameters are directly read from the simulation results of the building energy consumption analysis model.

[0098] In one embodiment of the present invention, the training and prediction process of the ensemble learning model used in the scheme generation module is as follows: A prediction model pool is constructed, consisting of multiple building energy consumption simulation engines with different principles or settings as base learners. These base learners collectively form the ensemble learning model. Detailed simulation data or actual monitored operational data of historical building projects are used as training samples. Each sample contains specific simulation boundary conditions, the equipment control strategies attempted or applied under these boundary conditions, and the energy consumption change resulting from applying the strategy compared to the baseline strategy. The ensemble learning model is trained using these samples, enabling the model to learn the complex nonlinear mapping relationship between "simulated boundary conditions" and "candidate equipment control strategies" to "energy-saving potential". When applied to new projects, the simulation boundary conditions corresponding to the extracted high-energy-consumption characteristic periods are input into the trained ensemble learning model, and a series of candidate equipment operation control strategies are generated by enumerating or searching the strategy space. The ensemble learning model outputs a rapid prediction value of the energy-saving potential that each candidate strategy can bring under the specific boundary conditions. After obtaining the predicted energy-saving potential values ​​of each candidate strategy, a multi-objective decision-making method is used to screen the optimized equipment operation strategy scheme. Multiple decision objectives are defined, primarily including maximizing energy efficiency, minimizing investment or retrofit costs required for strategy implementation, and minimizing the impact of the strategy on indoor thermal comfort indicators. The energy-saving potential predicted by the ensemble learning model is transformed into energy-saving effect metrics. These metrics are then combined with a pre-defined database of equipment strategy implementation costs and simulated impact data on thermal environment indicators such as indoor temperature and PMV to comprehensively evaluate the specific performance of each candidate strategy across the three decision objectives of energy efficiency, cost, and comfort. A Pareto optimal solution selection algorithm is applied to identify a set of non-dominated solutions from all candidate strategies, forming the Pareto front solution set. Any solution in this set is superior to other solutions on at least one objective and not inferior to other solutions on other objectives. Based on the preset preference weights during the design phase, such as a greater emphasis on energy saving or cost control, the final solution that best matches the design intent is selected from the Pareto optimal solution set as the recommended equipment operation strategy optimization scheme.

[0099] In practice, the training and prediction of the ensemble learning model follows a complete process from data preparation to model application. The ensemble learning model consists of a prediction model pool containing multiple different building energy consumption simulation engines as base learners. These base learners can be based on simplified models with different physical principles or data-driven models with varying levels of detail. The ensemble learning model uses detailed simulation data or actual operational monitoring data from historical building projects as training samples. Each training sample is a data record containing the simulated boundary conditions under specific operating conditions, the equipment control strategies adopted under those conditions, and the energy consumption change after applying the equipment control strategies compared to the baseline strategy. In practice, the simulated boundary conditions include meteorological parameters during high-energy-consumption periods, indoor occupant and equipment loads, and the thermal state of the building envelope; the equipment control strategies can include adjusting the air conditioning set temperature, regulating the fresh air volume, and controlling shading devices; the energy consumption change is the difference in energy consumption before and after implementing the strategy. The training process of an ensemble learning model involves enabling the model to learn the complex mapping relationship between "simulated boundary conditions and candidate device control strategies" and "energy-saving potential." The training process uses optimization algorithms to adjust the weights or ensemble methods of each base learner, minimizing the error between the overall predicted output of the ensemble learning model and the actual change in energy consumption.

[0100] In practical implementation, applying the trained ensemble learning model to predict energy-saving potential involves input and enumeration calculations. The scheme generation module inputs the simulated boundary conditions corresponding to the extracted high-energy-consumption characteristic periods into the trained ensemble learning model. Simultaneously, the scheme generation module enumerates or searches a series of candidate equipment control strategies from the policy library. It can be understood that the enumerated candidate equipment control strategies cover a feasible operating range; for example, the air conditioner set temperature changes from 24°C to 28°C in 0.5°C steps, and the fresh air exchange rate changes from 0.5 times / hour to 2 times / hour in 0.2 times / hour steps. For each input combination of "simulated boundary conditions + candidate equipment control strategies," the ensemble learning model outputs a predicted value. The predicted value represents the energy-saving potential that implementing this candidate equipment control strategy under the specific boundary conditions can bring. The energy-saving potential is usually expressed as energy saving or energy saving percentage within that period. In some embodiments, the ensemble learning model integrates the prediction results of each base learner through weighted averaging or stacked generalization. Optionally, the scheme generation module compiles all candidate equipment control strategies and their corresponding predicted energy-saving potentials into a list, which serves as input data for multi-objective decision-making.

[0101] In practical implementation, using a multi-objective decision-making method to screen equipment operation strategy optimization schemes requires clearly defined objectives and evaluation criteria. The scheme generation module defines multiple decision objectives, including maximizing energy saving, minimizing investment costs, and minimizing the impact on indoor thermal comfort. The module converts the energy-saving potential predicted by the ensemble learning model into quantifiable energy-saving values, such as annual electricity cost savings. It also incorporates a pre-set equipment strategy implementation cost database, containing hardware modification costs or software debugging costs required to implement various control strategies. Furthermore, it combines simulation data obtained through rapid simulation of the strategy's impact on the indoor predicted average voting index to evaluate the specific performance of each candidate equipment control strategy across the three decision objectives: energy saving, investment cost, and thermal comfort impact. To facilitate comparisons between objectives of different dimensions, the scheme generation module needs to normalize the original values ​​of each objective, converting them to the range of 0 to 1. In some embodiments, the scheme generation module applies a Pareto optimal solution screening algorithm to find a set of non-dominated solutions from all candidate strategies. The core of the Pareto optimal solution screening algorithm is to compare the merits of any two strategies across various objectives. Given a strategy A and another strategy B, if strategy A is not inferior to strategy B on all objectives, and strategy A is strictly superior to strategy B on at least one objective, then strategy A is said to dominate strategy B. Strategies not dominated by any other strategy constitute the Pareto optimal solution set. The solution generation module iterates through and compares all candidate strategies, selecting all non-dominated strategies to form the Pareto optimal solution set. The solution generation module selects the final solution from the Pareto optimal solution set according to the preference weights preset during the design phase. These preference weights represent the decision-maker's emphasis on each objective in vector form. For example, a weight vector (0.6, 0.3, 0.1) indicates a higher emphasis on energy saving, followed by cost, with relatively lower attention paid to thermal comfort.

[0102] In practical implementation, the final optimized equipment operation strategy is determined by calculating a weighted comprehensive value within the Pareto optimal solution set. The solution generation module calculates a weighted comprehensive evaluation value for each solution in the Pareto optimal solution set. The formula for calculating the weighted comprehensive evaluation value is as follows:

[0103]

[0104] in: Represents the Pareto optimal solution set of the th The solution (i.e. the first solution) The weighted comprehensive evaluation value of (each candidate strategy); This represents the total number of decision-making objectives. ; Representing the Preset preference weights for each decision objective; It is the first The solution is at the th solution. The normalized performance value for each decision objective. The solution generation module compares the weighted comprehensive evaluation value of all solutions in the Pareto optimal solution set. Select the weighted comprehensive evaluation value The maximum solution is taken as the final optimized device operation strategy. Optional, normalized performance value. The calculation uses a linear scaling transformation method. For objectives that need to be maximized, such as energy saving, For objectives that need to be minimized, such as investment costs and thermal comfort impact, ,in It is the first The solution is at the th solution. The original values ​​on each target and These are all solutions at the th... The maximum and minimum values ​​for each objective. It can be understood that the final selected equipment operation strategy optimization scheme is the result of comprehensively considering absolute performance and the decision-maker's subjective preferences. The scheme generation module transmits the final selected equipment operation strategy optimization scheme, along with its corresponding energy-saving effect prediction, cost estimation, and thermal comfort impact assessment, as a structured output to the subsequent integration module.

[0105] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A building energy saving optimization design system based on BIM technology, characterized in that, include: The data integration module is used to import the initial BIM design model of the building and identify and extract the building envelope information, spatial division information and equipment system information in the initial BIM design model; The environment simulation module is used to parse the geographical location, geometric attributes and material attributes of building components from the initial BIM design model, and combine them with a preset external environment parameter database to generate building external micro-environment field data. The building envelope optimization module is used to input the building external microenvironment field data, building envelope structure information and spatial division information into an improved particle swarm optimization algorithm. The improved particle swarm optimization algorithm dynamically adjusts the interaction network of the particle swarm according to the spatial topological relationship of the building components, and iteratively searches to obtain an optimized scheme of building envelope construction parameters that meets the preset thermal performance constraints. The model update module is used to update the initial BIM design model by applying the optimization scheme of the building envelope construction parameters, and generate a thermal performance optimization model. The energy consumption analysis module is used to acquire preset hourly data on personnel activities and equipment operating loads inside the building, and integrate it with the thermal performance optimization model to form a building energy consumption analysis model that includes dynamic internal disturbances. The scheme generation module generates a comprehensive energy-saving design scheme for buildings based on the building energy consumption analysis model. 2.The BIM-based building energy saving optimization design system according to claim 1, characterized in that, The environmental simulation module generates micro-environmental field data of the building's external environment, including: Based on the geographical location of the building components, typical annual meteorological data of the area where the building is located are obtained from a preset external environmental parameter database, including hourly dry-bulb temperature, wet-bulb temperature, direct solar radiation, diffuse radiation, wind speed and wind direction. Using computational fluid dynamics principles and combining the geometric properties of the building components, the wind field distribution and solar radiation shading around the building are simulated. The simulated wind field, temperature field, and radiation field data are coupled with the material properties of the building's outer surface to obtain the hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution of each outer surface of the building, thus forming the building's external microenvironment field data. 3.The BIM-based building energy saving optimization design system according to claim 2, characterized in that, The environmental simulation module couples the simulated wind field, temperature field, and radiation field data with the material properties of the building's exterior surface to obtain the hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution of each exterior surface, forming the building's external microenvironmental field data, including: The hourly wind speed, air temperature, and total solar radiation irradiance at the location of the building's outer surface are obtained from the simulation results of the external microenvironment field data. Based on the material properties of the building's exterior surface, determine its solar radiation absorptivity, long-wave emissivity, roughness, and the azimuth and tilt angles determined by the geometric properties of the building components. Based on the total solar radiation illuminance, the azimuth and tilt angle of the building's outer surface, calculate the corrected direct solar radiation illuminance and diffuse solar radiation illuminance received by the building's outer surface. The solar radiation gain is calculated by combining the solar radiation absorptivity of the building's exterior surface. The hourly solar heat gain coefficient is calculated by using the ratio of the solar radiation gain to the corrected total solar irradiance. The hourly convective heat transfer coefficient is calculated by combining the formula for calculating the convective heat transfer coefficient of the building's exterior surface with the hourly wind speed and the roughness of the building's exterior surface as input variables. Using the surface heat balance equation, with the solar radiation heat gain, convective heat transfer, radiative heat transfer with the surrounding environment, and the heat conduction process of the building's outer surface as input, the hourly surface temperature distribution of each outer surface of the building is obtained by iterative solution. The hourly solar heat gain coefficient, convective heat transfer coefficient, and surface temperature distribution are associated with the corresponding building exterior surface components and organized according to time series to generate the building exterior microenvironment field data. 4.The BIM-based building energy saving optimization design system according to claim 1, characterized in that, The improved particle swarm optimization algorithm in the enclosure optimization module dynamically adjusts the interaction network of the particle swarm based on the spatial topological relationship of building components, including: Define each particle in the particle swarm as a candidate combination of enclosure structure construction parameters; Based on the initial BIM design model, the spatial adjacency and thermal bridge relationships between wall, roof, and window components in the building envelope are analyzed, and a building component adjacency diagram is constructed. With reference to the structure of the adjacency graph of the building components, the interaction network of the particle swarm is initialized so that the particle parameters corresponding to the components that are adjacent in space or have thermal correlation have a higher connection probability in the interaction network. During the particle swarm optimization process, the overall thermal performance of the parameter combination represented by each particle is evaluated, and its fitness is calculated. The updating of particle velocity and position depends not only on the individual historical best solution and the group historical best solution, but also on the range of neighborhood best solutions for each particle, which is limited to the solutions of other particles directly connected to the current particle in the dynamic interaction network. As the iteration proceeds, based on the convergence of particle fitness, the connection weights between different particles in the interaction network are dynamically sparsified or strengthened, guiding the search to achieve a balance between parameter space with superior local performance and global exploration, and finally outputting the optimized scheme for the construction parameters of the enclosure structure. 5.The BIM-based building energy saving optimization design system according to claim 1, characterized in that, The scheme generation module generates a comprehensive building energy-saving design scheme based on the building energy consumption analysis model, including: The building energy consumption analysis model is subjected to hourly dynamic energy consumption simulation throughout the year to calculate a detailed energy consumption sequence including cooling energy consumption, heating energy consumption and lighting energy consumption. High-energy-consumption characteristic periods are extracted from the detailed energy consumption sequence, and the simulated boundary conditions corresponding to the high-energy-consumption characteristic periods are input into an ensemble learning model. The ensemble learning model is trained to predict the energy-saving potential of different equipment control strategies. Based on the prediction results of the ensemble learning model, a set of optimized equipment operation strategies are selected using a multi-objective decision-making method. The optimization scheme for the building envelope construction parameters is combined with the optimization scheme for the equipment operation strategy to generate the final comprehensive building energy-saving design scheme. 6.The BIM-based building energy saving optimization design system according to claim 5, characterized in that, The building energy consumption analysis model is used to perform hourly dynamic energy consumption simulation throughout the year, including: Set the simulation time step to hours and the simulation period to years; Within each simulation time step, the external disturbance data, hourly personnel activity and equipment operating load data, and thermal parameters of the building envelope are taken as inputs from the external microenvironment field data of the building. Solve the heat balance equations for each zone of the building to calculate the indoor air temperature, the inner surface temperature of each building envelope, and the real-time load of the heating and cooling system for each zone of the building at the current simulation time step. The energy consumption of the lighting system, equipment system, heating system, and cooling system are accumulated and calculated to generate the detailed hourly energy consumption sequence. 7.The BIM-based building energy saving optimization design system according to claim 5, characterized in that, Extracting high-energy-consumption characteristic periods from the detailed energy consumption sequence includes: A sliding window analysis is performed on the detailed energy consumption sequence to identify continuous time windows in which the total energy consumption exceeds a preset high energy consumption threshold; Within each high-energy-consumption time window, further analyze the energy consumption composition to distinguish the time period types dominated by cooling energy consumption, heating energy consumption, or overall energy consumption. Record the start and end times, outdoor weather conditions, building indoor load conditions, and thermal state of the building envelope for each high energy consumption characteristic period as the simulation boundary conditions.

8. The building energy-saving optimization design system based on BIM technology according to claim 5, characterized in that, The training and prediction of the ensemble learning model includes: Construct a pool of prediction models that use multiple energy consumption simulation engines as base learners; Using simulation data or actual operation data of historical building projects as training samples, each sample includes the simulated boundary conditions under specific working conditions, the equipment control strategy adopted under the corresponding working conditions, and the energy consumption change corresponding to the equipment control strategy. The ensemble learning model is trained to learn the complex mapping relationship from simulated boundary conditions and candidate policies to energy-saving potential; In application, the simulated boundary conditions of the high energy consumption period are input into the trained ensemble learning model, and a series of candidate equipment control strategies are enumerated or searched. The ensemble learning model outputs a predicted value of the energy-saving potential of each candidate strategy.

9. The building energy-saving optimization design system based on BIM technology according to claim 8, characterized in that, The process of using a multi-objective decision-making method to select a set of optimized equipment operation strategies includes: Define multiple decision objectives, including maximizing energy efficiency, minimizing investment costs, and minimizing the impact on indoor thermal comfort; The energy-saving potential predicted by the integrated learning model is transformed into energy-saving effect, and the performance of each candidate strategy is evaluated on various decision objectives by combining the preset strategy implementation cost data and the simulation data of the strategy's impact on the indoor thermal environment. By applying the Pareto optimal solution selection algorithm, a set of non-dominated solutions is found from all candidate strategies, forming the Pareto optimal solution set; Based on the preference weights preset during the design phase, a final solution is selected from the Pareto optimal solution set as the optimization scheme for the equipment operation strategy.

10. The BIM-based building energy-saving optimization design system according to claim 9, characterized in that, The optimization scheme for the building envelope construction parameters is combined with the optimization scheme for the equipment operation strategy to generate the final comprehensive building energy-saving design scheme, including: The technical feasibility of the optimized structural parameters of the enclosure structure is verified to ensure that the selected structural parameters comply with local specifications, the materials are available, and the construction is feasible. The system compatibility verification of the equipment operation strategy optimization scheme is performed to ensure that the strategy does not exceed the physical operating limits of the equipment system. The two verified solutions are integrated in terms of time and measure type to generate an implementation list that includes passive technical measures and active control strategies. In the implementation list, each measure is marked with its associated BIM components, systems, expected energy-saving contribution, estimated cost, and implementation stage, forming a structured document of the comprehensive building energy-saving design scheme.

Images