Health building design method, system and terminal based on whole-process data analysis and storage medium

Abstract

The application relates to the technical field of data analysis, and discloses a health building design method, system, terminal and storage medium based on whole-process data analysis. The method comprises the following steps: obtaining vector data and remote sensing images of a target area, preprocessing the vector data and the remote sensing images, obtaining spatial information, performing space registration on the spatial information, and obtaining initial planning information; if the initial planning information is in line with rules, performing environment analysis; constructing constraint conditions and a target function according to environment information, optimizing the target function under the constraint conditions, and obtaining an initial layout scheme; performing health building environment performance simulation according to the initial layout scheme, obtaining simulation data, optimizing the initial layout scheme according to the simulation data, and obtaining a health building design scheme. Through a whole-process data interaction process, the application automatically identifies multi-modal data which have an influence on building design, realizes whole-cycle health design optimization from early planning to operation and maintenance, forms whole-process closed-loop management, and improves the health, compliance and economy of health building design.

Description

Technical Field

[0001] This invention relates to the field of data analysis technology, and in particular to a healthy building design method, system, terminal, and computer-readable storage medium based on full-process data analysis. Background Technology

[0002] In the current healthy building design process, data from different stages are fragmented, and the healthiness of the designed building layout relies on human experience for judgment. This makes it difficult to efficiently integrate model data from some mainstream design software, and there is a lack of automated identification and verification of factors such as ecological control lines, noise, and lighting during the design process.

[0003] Existing healthy building design technologies can only assist in designing a single stage or only support the import of model data in a specific format. They cannot meet the needs of healthy building design throughout the entire process and in multiple dimensions. Furthermore, they lack intelligent information integration and decision support mechanisms in material selection, construction coordination, and operation and maintenance monitoring, resulting in inaccurate healthy building design schemes.

[0004] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0005] The main objective of this invention is to provide a healthy building design method, system, terminal, and computer-readable storage medium based on full-process data analysis, aiming to solve the problem that the existing technology cannot integrate all process data of healthy building design, resulting in inaccurate final healthy building design schemes.

[0006] To achieve the above objectives, the present invention provides a healthy building design method based on full-process data analysis, the healthy building design method based on full-process data analysis comprising the following steps:

[0007] The vector data and remote sensing images of the target area are acquired, the remote sensing images are preprocessed to obtain spatial information, and the vector data and the spatial information are spatially registered to obtain the initial planning information of the target area.

[0008] A knowledge graph for healthy building design is constructed. The compliance of the initial planning information is judged based on the knowledge graph. If it is compliant, the environment of the target area is analyzed to obtain various environmental information.

[0009] Based on all the environmental information, constraints and objective functions are constructed. Under the constraints, a genetic algorithm is used to optimize the objective function to obtain an initial layout scheme.

[0010] Based on the initial layout scheme, a healthy building environment performance simulation is performed to obtain various simulation data. Corresponding weights are added to all the simulation data, and all the simulation data are integrated according to all the weights to obtain a healthy building design scheme.

[0011] In this invention, vector data and remote sensing images of a target area are acquired. The remote sensing images are preprocessed to obtain spatial information, and the vector data and spatial information are spatially registered to obtain initial planning information for the target area. A healthy building design knowledge graph is constructed, and the compliance of the initial planning information is judged based on the knowledge graph. If compliance is achieved, the environment of the target area is analyzed to obtain various environmental information. Constraints and objective functions are constructed based on all the environmental information. Under the constraints, a genetic algorithm is used to optimize the objective function to obtain an initial layout scheme. Healthy building environmental performance simulation is performed based on the initial layout scheme to obtain various simulation data. Corresponding weights are added to all the simulation data, and all the simulation data are integrated based on the weights to obtain a healthy building design scheme. This invention, through a full-process data interaction process, automatically identifies multimodal data that affects building design, realizing full-cycle healthy design optimization from early planning to operation and maintenance, forming a closed-loop management system, and improving the health, compliance, and economy of healthy building design. Attached Figure Description

[0012] Figure 1 This is a flowchart of a preferred embodiment of the healthy building design method based on full-process data analysis of the present invention;

[0013] Figure 2 This is a structural diagram of a preferred embodiment of the healthy building design system based on full-process data analysis of the present invention;

[0014] Figure 3 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, 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 of the invention and are not intended to limit the invention.

[0016] The preferred embodiment of the healthy building design method based on full-process data analysis described in this invention, such as... Figure 1 As shown, the healthy building design method based on full-process data analysis includes the following steps:

[0017] Step S10: Obtain vector data and remote sensing images of the target area, preprocess the remote sensing images to obtain spatial information, and spatially register the vector data and the spatial information to obtain the initial planning information of the target area.

[0018] The vector data includes spatial coordinates and attribute information; the spatial information includes terrain slope, building volume, and road grade. This data can be acquired using API (Application Programming Interface) secondary development technology, which develops dedicated model parsing interfaces for various design software. This provides fundamental data support for automated design and optimization at all stages of the platform's entire process, enabling seamless data flow across multiple software programs and forming a closed-loop data chain from design to operation and maintenance. Existing technologies are often limited to data processing in a single software or stage, failing to achieve full-process integration.

[0019] In one embodiment of the present invention, for custom components of a building, accurate identification can be achieved by parsing its extended data and dictionary data, thereby solving the problem of incomplete extraction of information on custom components in the prior art; for Revit models (Autodesk Revit, software specifically designed for building information modeling), data encapsulation can be performed in conjunction with BIM (Building Information Modeling) standards to ensure the complete integration of geometric, material, and attribute information, overcoming the limitations of importing from a single format.

[0020] Furthermore, the geometric parameters (coordinates, dimensions, shapes, etc.) and attribute information (component types, etc.) extracted from mainstream design software such as CAD (Computer-Aided Design) serve as the basic data support for the platform to achieve automated design and optimization at all stages of the entire process. This data can be used in the early planning stage (including compliance testing, scale testing, functional area layout design, and environmental performance simulation), construction drawing design, and the final construction stage.

[0021] Custom components (such as custom walls, doors, windows, and unit layouts) store a large amount of non-standard formatted extended information, requiring specialized algorithms to parse their extended and dictionary data. First, custom object identification and location are performed, traversing all entity objects in the CAD drawings to filter out the custom components. These objects typically have specific class names as identifiers, which are then stored in the database. A unique ID is defined for each object based on its class name to ensure data consistency. Extended data contains additional information attached to the custom objects, stored in a "application identifier + group code + value" structure (e.g., for wall thickness and material type). Parsing this data uses group code rules; for different unit layouts (healthy building functional areas), the extended data includes different identification numbers, usage functions, and area information, among other architectural design information. Furthermore, the complex attributes of custom components are stored in dictionary data. The dictionary is nested in the object data in the form of key-value pairs. After parsing, the independent parameters of each room (functional area of ​​a healthy building) can be extracted. The parsed data is verified against preset rules (such as whether the wall thickness is the standard value and whether the door and window types are in the specification list), outliers (such as negative dimensions) are removed, and unstructured data is converted into a platform-unified format (such as mapping the "Tch_Door" type to "wooden swing door" and unifying the unit of measurement to meters), which facilitates subsequent integration with BIM standards.

[0022] In another embodiment of the present invention, if a custom component is obtained through a Revit model, the parsing process requires filtering and parameter extraction. Parameters include boundaries and dimensions, walls, doors and windows, spatial relationships, etc., while attribute relationships refer to the material properties used. Finally, the extracted geometric parameters (coordinates, dimensions) are converted to a preset coordinate system (aligned with GIS (Geographic Information System) data), and attribute parameters are mapped to IFC (Industrial Foundation Class) standard fields (the output format of the BIM model). The data is encapsulated in the same format, containing the "component ID-type-geometry-attribute" association, facilitating cloud platform storage and cross-stage retrieval (such as material matching during the construction phase).

[0023] Specifically, the spatial coordinates and attribute information of the target area are obtained through a spatial data engine, and remote sensing images are acquired. The remote sensing images are analyzed to obtain the terrain slope, building volume, and road grade. The spatial coordinates, attribute information, terrain slope, building volume, and road grade are projected into the same projection coordinate system to obtain data to be registered. Multiple control points are determined in the data to be registered, and all the data to be registered are transformed according to the multiple control points to obtain unified initial planning information.

[0024] Based on the above data processing method, the spatial data engine then obtains vector data such as ecological control lines, basic farmland lines, and urban five lines (blue lines, green lines, etc.) from the city CIS database (where there are independent databases for different building types, such as healthy campuses, hospital databases, or office databases, each with its own independent database, which stores vector data corresponding to each building type; in the embodiments disclosed in this invention, healthy campuses are mainly used as an example for explanation). These data are stored in shp format (Shape file) and contain spatial coordinates and attribute information (such as control line type, management requirements, etc.).

[0025] In another embodiment of the present invention, during the GIS data integration stage of the healthy building design platform, the shapefile (shp) format is selected to store vector data such as ecological control lines and urban five-line data. Its core advantages stem from its spatial data compatibility, structural flexibility, processing efficiency, and industry adaptability, perfectly matching the platform's requirements for "multi-source data interaction, spatial analysis, and standardized applications." Furthermore, the shapefile format has strong industry universality, adapting to the full-process data integration needs. Adopting a "multi-file combination" structure (core components include .shp spatial files, .dbf attribute files, and .shx index files), it can simultaneously store the spatial coordinates and structured attribute information of vector data, perfectly matching the platform's refined requirements for control line data. Finally, the platform frequently needs to perform spatial overlay and distance calculations during the early planning (compliance verification) and scheme design (functional zoning layout) stages. The vector characteristics and indexing mechanism of the shapefile format can significantly improve analysis efficiency. The platform's GIS data (such as urban five lines that may be updated with planning adjustments) needs to be maintained in the long term. The editing characteristics of the shapefile format can reduce data management costs. When urban planning is adjusted (such as the expansion of the ecological control line), only the changed elements in the shapefile (such as newly added control areas) can be modified, without the need to re-import the full amount of data, reducing redundant operations in platform data synchronization.

[0026] Further, a remote sensing image is acquired, and noise, distortion, and bias in the remote sensing image are eliminated to obtain a target remote sensing image. Multiple target features in the target remote sensing image are determined, and each target feature is enhanced. The target remote sensing image after feature enhancement is input into a pre-constructed semantic segmentation model, which performs pixel-level classification on the processed target remote sensing image and outputs terrain slope features, building volume features, and road grade features. Error elimination processing is then performed on the terrain slope features, building volume features, and road grade features, and the processed terrain slope features, building volume features, and road grade features are vector-converted to obtain the terrain slope, building volume, and road grade.

[0027] In another embodiment of the present invention, the analysis of remote sensing satellite images can be performed by using image segmentation algorithms in computer vision to process the remote sensing images and identify features such as terrain slope, building outlines, and road networks. The terrain slope is calculated using a digital elevation model, and target detection algorithms are used to identify building volumes and road grades. These spatial information are then spatially registered with GIS vector data (alignment based on coordinate systems) to establish a database of the surrounding environment of the land parcel containing information such as geographical location, topography, and surrounding environment.

[0028] The process begins by converting remote sensing imagery into structured spatial data (such as building outline vectors and road network attributes) that can be directly used for planning and design. This process must ensure accuracy to meet the requirements of healthy building design, thus providing foundational data for subsequent processes such as illumination analysis and noise analysis. The converted data is then filtered to remove noise, distortion, and biases, improving data quality. Target features in the data are then enhanced (for example, in architectural design, features such as building edges and road textures are enhanced), improving subsequent segmentation accuracy. After identifying and segmenting prominent target features, segmentation errors (such as isolated noise points and discontinuous boundaries) are eliminated to improve feature integrity. Finally, the raster segmentation results are converted into vector data, and attribute information is extracted for use by the healthy building platform. Through the data acquisition and processing processes described in these two embodiments, the spatial coordinates, attribute information, terrain slope, building volume, and road grade required for registration are obtained.

[0029] Furthermore, during the spatial registration of data, coordinate bias is first eliminated to achieve seamless overlay and collaborative analysis of multi-source data. High-precision vector data with accurate coordinates, or accurately registered high-resolution remote sensing imagery (including geographic coordinate information) is selected as the benchmark. In the data to be registered and the reference data, obvious, easily identifiable, and spatially stable corresponding ground features are selected as control points, such as road intersections (cross or T-junctions), corners of large buildings (e.g., the four corners of a healthy building library), and river confluences. For large areas, control points should be evenly distributed throughout the study area, covering both the edges and the center. In complex terrain or deformed areas, the density of control points should be appropriately increased (e.g., at points where contour lines are distorted in mountainous areas).

[0030] During the transformation of the data to be registered based on control points, different transformation methods can be used depending on the scenario. If there are translation, rotation, and uniform scaling issues, such as directional deviations in vector data collected by UAVs due to flight attitude, a similarity transformation method can be used. This method requires at least two control points. By calculating the rotation angle and scaling ratio, data alignment is achieved. It is simple to operate, has low computational load, and is suitable for small-scale areas with uncomplicated deformations, achieving sub-meter accuracy (error < 1 meter).

[0031] Furthermore, to address issues related to data translation, rotation, non-uniform scaling, and skewness, such as trapezoidal distortion in vector data after digitizing scanned paper maps, at least three control points are required. By solving a system of linear equations, transformation parameters are determined, which can effectively handle linear deformations and is widely used in urban area data registration, achieving an accuracy within 0.5 meters.

[0032] Furthermore, for complex nonlinear deformations of data (such as distortions caused by stretching and folding of paper maps), by densely setting control points in the deformed area, the software performs elastic stretching and distortion correction on the data based on the local deformation relationship between the control points. This is suitable for the registration of severely deformed data such as historical maps, and the error can be controlled within 1-2 meters, but the computational load is relatively large.

[0033] Furthermore, for data registration in large areas with complex terrain, a high-order polynomial model (such as a quadratic or cubic polynomial) is constructed to consider the nonlinear changes of the data in the X and Y axes. This requires a large number of control points (≥6), which can effectively correct the deformation caused by the curvature of the earth and the topographic relief. It is often used for cross-regional vector data integration, and the accuracy can reach the sub-pixel level (<0.5 pixels).

[0034] After completing the selection and calibration method settings for control points, GIS software is used to calculate the residual indices of the control points, such as X residual (the difference between the actual and theoretical coordinates of the control point in the X-axis direction), Y residual (the difference between the actual and theoretical coordinates of the control point in the Y-axis direction), and root mean square error (RMS Error, an index that comprehensively reflects the errors of all control points). Generally, the RMS error is required to be less than one-third to one-half of the map scale accuracy. For example, for a 1:1000 scale map, the RMS error should be <0.33 meters to 0.5 meters. For high-resolution image registration, the RMS error needs to be controlled within 1 to 2 pixels.

[0035] Furthermore, after the transformation is completed, control points with excessive errors are deleted, and more accurate corresponding points are selected, or control points are added in the error concentration area to improve local correction accuracy. If linear transformation cannot meet the accuracy requirements, nonlinear transformation methods are tried; or the results of multiple correction methods are compared, and the scheme with the smallest error is selected. Finally, based on the selected control points and correction method, the vector data to be registered is spatially transformed, its coordinate information is updated, and the registration process is completed. After registration is completed, the correction results are saved as a new vector layer (e.g., "Registered Building Outline.shp"), and a backup of the original data is retained for subsequent comparison and traceability.

[0036] Furthermore, in GIS software, the registered data is overlaid with the reference data. By zooming in and out of the map view, it is possible to check whether key features (such as buildings and roads) are accurately aligned and whether the boundaries are seamlessly connected. Several feature points with the same name that were not registered are selected, and their coordinate differences before and after registration are compared with the reference data. The average error is calculated, and the overall accuracy of registration is evaluated to ensure that the application requirements such as healthy building design are met (e.g., building positioning error < 1 meter).

[0037] Furthermore, massive amounts of data, such as design models, material data, and supplier information, are stored using a distributed storage architecture (such as Hadoop Distributed File System, HDFS, a distributed file system). Parallel processing and analysis of this data are achieved through a distributed computing framework. Message queues are used to facilitate data interaction between multiple terminals (design terminals, construction terminals, and operation and maintenance terminals). When data is updated on one terminal, the message queue pushes the update event to other relevant terminals, triggering data synchronization. Simultaneously, a version control mechanism is introduced to record every modification to critical data such as design models, supporting version backtracking and comparative analysis to ensure data consistency and traceability.

[0038] Specifically, dedicated parsing algorithms were developed for the technical characteristics of different software. For example, for custom components, accurate identification is achieved by parsing their extended data (XData) and dictionary data, solving the problem of incomplete information extraction for custom components in existing technologies. For Revit models, data encapsulation is performed using BIM standards (IFC format) to ensure the complete integration of geometric, material, and attribute information, overcoming the limitations of importing from a single format. The process of achieving seamless data flow across multiple software programs is achieved through API secondary development and a distributed storage architecture, breaking down the barriers between design software such as CAD and Revit and GIS and remote sensing data, forming a closed-loop data chain from design to operation and maintenance. Existing technologies are mostly limited to data processing in a single software or stage, failing to achieve end-to-end integration.

[0039] Step S20: Construct a knowledge graph for healthy building design, determine the compliance of the initial planning information based on the knowledge graph, and if it complies with the rules, analyze the environment of the target area to obtain various environmental information.

[0040] The compliance verification process can be performed using a compliance verification engine. This engine constructs a pre-defined statutory planning database containing provisions from urban and rural planning, land management, and healthy building design codes, as well as planning outcomes such as local city master plans and detailed control plans. The data is stored in a relational database, and an ontology-based knowledge graph is built to clarify the relationships between planning indicators such as land use, building height, and floor area ratio, and the planning provisions. When a user uploads land parcel information, the engine uses spatial overlay analysis to overlay the parcel boundaries with control lines in the GIS database to determine if the parcel is located within ecological control lines, urban purple lines, or other controlled areas. Simultaneously, it extracts the land use, floor area ratio, and other attribute information of the parcel and compares it with the planning conditions in the statutory planning database. If discrepancies are found (e.g., the parcel is within the urban purple line but does not meet historical protection requirements), a compliance report is automatically generated, pointing out the violations, citing relevant planning provisions, and proposing specific suggestions for restricting building height and style.

[0041] Specifically, multiple planning provisions are acquired, and all the planning provisions are analyzed to obtain entity information and entity relationships. The entity information includes multiple provision types, and the entity relationships include the relationships between all provision types. Based on all provision types and entity relationships, entity attributes of all planning provisions are extracted. A healthy building design knowledge graph is constructed based on the entity information, entity relationships, and entity attributes. Spatial overlay analysis is performed on the initial planning information using the healthy building design knowledge graph to obtain a compliance report. If the compliance report indicates compliance, multiple healthy building functional zones are determined based on the initial planning information. The noise contour lines, sunshine distribution map, and traffic density values ​​for each healthy building functional zone are calculated.

[0042] Among these, constructing a planning knowledge graph is the core step in achieving automatic compliance verification. Essentially, it uses ontology technology to transform scattered clauses, indicator constraints, land use types, and other elements into a structured "entity-relationship-attribute" network, enabling computers to understand and reason about planning rules. First, the planning knowledge system is established, clarifying the core scope and objectives of the knowledge graph to ensure coverage of the key elements required for compliance verification. Then, an ontology modeling tool is used to define the "class-property-relation" framework of the knowledge graph, forming semantic specifications for the planning domain (including class hierarchy, attribute hierarchy, and relation hierarchy). The class hierarchy uses a tree structure to divide entity types, reflecting hierarchical relationships; the attribute hierarchy defines descriptive attributes for each class, recording entity characteristics; and the relation hierarchy defines the semantic relationships between classes, reflecting the planning logic. Then, knowledge is extracted from relational databases and unstructured text, transformed into an ontology-recognizable format, and natural language processing techniques are used to extract entities, relations, and attributes from the preprocessed data to form knowledge triples (subject-verb-object). A named entity recognition model is used to identify planning entities (including clause entities, indicator entities, etc.) in the text. Then, keyword matching is used to identify relationships between entities, and a relation classification model is used to process complex sentences. Entity attribute values ​​are extracted from text or structured data, and finally, conflicting data is eliminated, multi-source knowledge is integrated, and stored as a graph structure; for example, merging... Synonymous entities (such as "building height restriction" and "building height limit" are unified as "building height"), which is achieved through string similarity (such as edit distance) + domain rules (such as indicator code mapping table); to resolve relationship conflicts (such as different constraints on the same indicator by different regulations), valid relationships are preserved according to the principle of "superior law takes precedence" and "new law prevails over old law" (such as when there is a conflict between superior and subordinate regulations, the superior law shall prevail); a graph database (such as Neo4j) is used to store the knowledge graph, with entities as nodes, relationships as edges, and attributes as the attribute values ​​of nodes / edges.

[0043] Furthermore, by integrating this data, a knowledge graph for healthy building design can be generated. The completed planning knowledge graph can transform scattered legal provisions and indicator constraints into a semantic network that computers can understand. This supports the compliance verification engine to automate the entire process of "land data input → automatic matching of applicable laws and regulations → inference of indicator constraints → judgment of compliance," providing accurate legal basis for healthy building design.

[0044] Further, the planning control layer of the target area is obtained, and the planning control layer is superimposed layer by layer with the spatial coordinates in the initial planning information to obtain the first superimposed conflict information; the healthy building design knowledge graph is superimposed layer by layer with the spatial coordinates in the initial planning information to obtain the second superimposed conflict information corresponding to each planning clause; the priority of all planning clauses in the healthy building design knowledge graph is determined, and all the second superimposed conflict information is filtered according to the priority to delete duplicate superimposed conflict information; a structured compliance report is generated based on the first superimposed conflict information and all the filtered second superimposed conflict information.

[0045] Among them, spatial overlay analysis is the core process by which the compliance verification engine calculates the spatial relationship between the land parcel information uploaded by users and planning control data (such as control planning parcels, ecological control lines, urban five lines, etc.). The purpose is to accurately locate all planning constraints involved in the land parcel (such as land use compatibility requirements, building height restrictions, etc.).

[0046] Specifically, the land parcel information uploaded by users typically includes spatial boundaries (such as latitude and longitude coordinate strings, SHP files, or CAD drawings) and declared attributes (such as proposed land use, declared plot ratio, building height, etc.). This information must first be converted to a format recognizable by the engine. Then, the converted land parcel data is projected onto the unified coordinate system of the planning database to avoid overlay deviations caused by different coordinate benchmarks. Further, planning control layers related to the spatial scope of the land parcel are selected from the GIS database as "benchmark data" for overlay analysis. Then, GIS spatial analysis algorithms are used to overlay the user's land parcel with the planning control layers layer by layer, determining spatial relationships (such as "containment," "intersection," and "disjunction") and extracting associated planning constraints. For example, overlaying with the control plan land parcel layer determines basic planning indicators; overlaying with the urban five-line / ecological control line identifies spatial constraints; and overlaying with other special planning layers supplements special constraints. When multi-level overlay results show constraint conflicts (such as the control plan allowing a plot ratio of 1.5, but the ecological control line requiring a plot ratio of 0), they must be handled according to planning priority rules.

[0047] The planning priorities include legal priority (higher-level laws take precedence over lower-level laws), spatial priority (constraints that completely encompass the site take precedence over partial overlap, such as if the site is entirely within the green line, in which case all constraints of the green line apply), core area constraints take precedence over peripheral areas (e.g., constraints of the historical and cultural core area are greater than those of the construction control zone), and timeliness priority (the most recently effective planning constraints take precedence over older plans). Finally, the overlay results are organized into structured data of "spatial relationships - constraint indicators - legal basis," which serves as input for compliance verification and generates a structured overlay analysis report.

[0048] The core of the compliance verification engine is a planning knowledge graph based on ontology technology. By analyzing various provisions, it establishes semantic associations between planning indicators such as "land use" and "floor area ratio" and specific provisions (e.g., "No new non-ecological facilities are allowed within the ecological control line"), ensuring that the mapping relationship between indicators and provisions is traceable. The verification logic adopts a dual mechanism of spatial overlay analysis and attribute comparison. First, the ArcSDE engine (Spatial Database Engine) overlays the land boundary with the GIS control line to identify the spatial control area. Then, it extracts the land attributes and compares them with the statutory planning database, automatically generating a violation report containing provisions referenced. Compared with existing technologies such as manual verification or single indicator checks, the accuracy is improved by more than 40%, and the traceability is stronger.

[0049] Furthermore, the environmental information includes: noise contour lines, sunshine distribution maps, and traffic density values. In another embodiment of the invention, when performing noise simulation, the grade information of roads surrounding the plot (such as highways, main roads, secondary roads, etc.) is obtained from a GIS database, and combined with traffic flow monitoring data (obtained through API access to traffic management departments), the traffic flow density of different roads is calculated. Based on the acoustic model, considering factors such as distance attenuation, obstacle shading, and ground absorption, noise contour lines around the plot are calculated. Spatial analysis is performed on the noise contour lines and functional zones within the plot (such as noise-sensitive areas such as classrooms and dormitories) to identify the range where the noise value is lower than the allowable limit for sensitive areas (such as the noise limit of 55dB for classroom areas), which is taken as the optimal layout range and displayed through visualization technology (such as heat maps). The calculation of noise values ​​around the plot requires a three-stage process of "source strength determination → attenuation calculation → multi-source superposition," and then the compliance of noise-sensitive areas is clarified by spatial analysis and association with functional zones.

[0050] The process involves overlaying noise isosurfaces (containing different noise value ranges) with functional zoning surfaces of the land parcels (containing different sensitive areas) to generate new "overlay units." Each unit contains three core attributes: "functional type, allowable limit, and actual noise value." The overlaid unit data is then filtered according to the rule of "actual noise value ≤ allowable limit," removing units whose actual noise values ​​exceed the limit (e.g., "dormitory - 50dB limit - 53dB actual") and marking them as "areas requiring optimization." The spaces of the optimal layout range are then merged. For example, all compliant units under the same "classroom" function are merged into a continuous planar area, which is the "optimal layout range for classrooms." Similarly, the "optimal layout range for dormitories" is generated to ensure the spatial continuity of the output range, facilitating subsequent planning implementation.

[0051] Furthermore, for the sunshine simulation process, based on GIS topographic data and surrounding building height data, sunshine analysis algorithms (such as multi-point along-line sunshine analysis) are used to simulate the trajectory of sunlight at different times of the year (such as the winter solstice and the coldest day of winter). For each point within the site, the sunshine duration within the specified time period is calculated, and a sunshine duration distribution heat map is generated. Combining the sunshine requirements for functional areas such as classrooms and sports fields in the Healthy Building Design Code (such as classrooms receiving no less than 2 hours of full-window sunshine on the winter solstice), areas that meet the sunshine requirements are selected on the heat map and recommended as the optimal sites for the corresponding functional areas.

[0052] Furthermore, for traffic simulation around healthy buildings, a pedestrian and vehicle flow prediction model is constructed. This model is based on data such as population density, school size, and road capacity around the site, and uses the core algorithm of microscopic traffic simulation software to simulate the distribution of pedestrian and vehicle flow at different times. By analyzing the connection relationship between each entrance / exit of the site and the surrounding roads, the conflict probability of traffic conflict points (such as vehicle-pedestrian conflicts at intersections) is calculated. The calculation of conflict probability takes into account factors such as traffic flow, signal timing, and road geometry. Based on the magnitude of the conflict probability, the location of entrances / exits is automatically recommended (such as locations far from main road intersections with a conflict probability of less than 0.1 times / hour), and traffic organization optimization suggestions are proposed (such as setting up pedestrian crossings, speed bumps, etc.).

[0053] The automatic recommendation criteria can adopt core quantitative standards, spatial adaptation standards, traffic matching standards, and scenario-specific standards. The core quantitative standard takes "conflict probability" as the primary criterion. It uses conflict data output from micro-simulation as the core indicator and defines rigid thresholds. The thresholds include thresholds for different time periods, i.e., basic safety thresholds, which correspond to low-risk levels and can avoid traffic collisions and congestion during normal hours. Sensitive time period thresholds are for scenarios such as schools and residential areas, and require additional verification of the conflict probability during peak hours (such as school commuting times 7:30-8:30 and 16:30-17:30), requiring ≤0.05 times / hour to reduce the safety risks during periods of high pedestrian traffic.

[0054] Furthermore, the spatial adaptation standard refers to the distance requirements to surrounding roads / facilities. From the perspective of road geometry and functional area layout, it avoids direct association between entrances / exits and high-risk nodes, keeps them away from high-traffic nodes, and avoids road bottlenecks. The traffic flow matching standard is the process of matching the capacity of entrances / exits with the predicted traffic flow, ensuring that the recommended entrances / exits can accommodate the flow of people and vehicles generated by the site, avoiding congestion caused by "supply and demand mismatch", and indirectly reducing the risk of conflict. The scenario-specific standard adapts to sensitive groups and functional needs. For special scenarios such as schools, hospitals, and residential areas, it supplements differentiated standards to prioritize the safety of vulnerable traffic participants.

[0055] Step S30: Construct constraints and objective functions based on all the environmental information, and optimize the objective functions using a genetic algorithm under the constraints to obtain an initial layout scheme.

[0056] The constraints include spatial constraints, attribute constraints, and safety constraints. The total buildable floor area is calculated based on the land area and plot ratio. Combined with indicators such as per-student land area and per-student building area specified in the Healthy Building Design Knowledge Graph, a reasonable number of classes and students is calculated. For example, if the land area is 50,000 square meters and the per-student land area indicator is 15 square meters, the maximum number of students that can be accommodated is approximately 3,333. Assuming 50 students per class, 67 classes can be set up. Simultaneously, based on the number of classes and students, the required area for supporting facilities (such as canteens, restrooms, and laboratories) is calculated (e.g., a canteen area of ​​1.2 square meters per student). The required area for each supporting facility is calculated, and an investment estimate report is generated based on local construction cost indicators (e.g., 3,000 yuan per square meter). The report uses a dynamic calculation method, automatically updating the estimate results when input parameters (such as land area and plot ratio) change.

[0057] Specifically, spatial constraints and safety constraints are constructed based on all the healthy building functional zones; attribute constraints are constructed based on the noise contour lines, sunshine distribution maps, and traffic density values ​​of each healthy building functional zone; a weighted sum is performed on the noise contour lines, sunshine distribution maps, and traffic density values ​​of each healthy building functional zone, and the noise contour lines, sunshine distribution maps, and traffic density values ​​are maximized respectively, while ensuring the spatial constraints, attribute constraints, and safety constraints, to obtain the objective function; all the healthy building functional zones are encoded as initial populations, and selection, crossover, and mutation operations are performed on all the initial populations according to the spatial constraints, attribute constraints, and safety constraints to obtain optimized new populations, until the interference between all healthy building functional zones is minimized, to obtain the initial layout scheme.

[0058] Based on the aforementioned environmental data, objective functions and constraints are constructed. Objective functions include maximizing the distance between functional areas and noise sources, maximizing the sunshine duration of functional areas, and minimizing interference between functional areas. Constraints include site boundaries, building control lines, and fire safety distances. The objective function is the "indicator that needs to be maximized or minimized." Noise distance, sunshine duration, and interference levels need to be transformed into mathematical formulas. Multiple objectives are integrated through weighted summation (weights are set according to functional importance, with a total sum of 1) to ultimately obtain the maximized objective function value. Constraints are the red lines of planning regulations and actual needs. They need to be transformed into mathematical inequalities to ensure the legality and compliance of the optimization scheme, and are divided into three categories: spatial constraints, attribute constraints, and safety constraints.

[0059] Under various constraints, a genetic algorithm is used for optimization. The layout scheme of functional zones is encoded into an initial population (e.g., each representing the location coordinates of a functional zone). A new generation of population is generated through operations such as selection (roulette wheel selection based on fitness function), crossover (single-point or multi-point crossover), and mutation (randomly changing gene values). This process iteratively optimizes the layout until the optimal scheme is found. For example, the sports field is located on the west side of the plot (away from the noise of the main road on the east side), and the classroom clusters are located on the south side (to obtain optimal sunlight). To minimize interference between functional zones, a simulated annealing algorithm is used. Using the distance between functional zones and the interference coefficient (e.g., the interference coefficient of the laboratory to the classroom is 0.8) as parameters, the interference cost is calculated. By gradually decreasing the temperature (a control parameter), the layout scheme with the lowest interference cost is found, such as maintaining a distance of more than 30 meters between the laboratory and the classroom.

[0060] The core of multi-objective optimization based on genetic algorithms is the collaborative design of dynamic objective functions and constraints. Objective functions have clearly defined priorities, such as "maximizing the distance between functional areas and noise sources" (weight 30%), "maximizing sunshine duration" (weight 40%), and "minimizing functional interference" (weight 30%), with weights determined through expert surveys and verification using over 100 healthy building case studies. Algorithm parameters are used for targeted optimization. In the genetic algorithm, the initial population encoding uses a combination of "functional area coordinates + dimensions," with a crossover probability of 0.7 and a mutation probability of 0.05 to ensure population diversity. The simulated annealing algorithm has an initial temperature of 1000K and a cooling coefficient of 0.95 to avoid local optima, ultimately increasing the compliance rate of health indicators for functional area layout to 92% (compared to an average of 65% in existing technologies).

[0061] Step S40: Perform healthy building performance simulation based on the initial layout scheme to obtain various simulation data. Input all the simulation data into the corresponding responsible entity model. Each responsible entity model outputs final data based on the corresponding simulation data. Construct a healthy building design scheme based on all the final data.

[0062] Specifically, a Revit model is constructed based on the initial layout scheme. This Revit model is then converted into a scene file containing the design data, construction data, construction progress data, project management data, and supply data. The scene file is input into the constructed design entity model, construction entity model, construction progress model, project management entity model, and supply entity model, respectively. The design entity model optimizes the design data and outputs the final design data. The construction entity model optimizes the construction data and outputs the final construction data. The construction progress model optimizes the construction data and outputs the final construction data. The project management entity model optimizes the project management data and outputs the final project management data. The supply entity model optimizes the supply data and outputs the final supply data. The final design data, final construction data, final construction progress data, final project management data, and final supply data are then weighted and fused to obtain a healthy building design scheme.

[0063] The simulation data includes design data, construction data, project management data, and supply data. By constructing a hierarchical control system, these various data types are input into different models for processing. Each model is controlled by a different responsible entity, and the data within each model only includes the data corresponding to that responsible entity.

[0064] For example, in one embodiment of the present invention, the core responsibility of the design institute as the responsible entity is to design, optimize parameters, and trigger simulations for healthy building schemes. All of these data belong to design data. The design institute's data viewing permissions are limited to accessing the full set of design data, simulation results, case library, and resource library (but cannot modify them). Once the design institute has completed the design of the current building, it can upload the design model (i.e., the Revit model), adjust design parameters (such as window-to-wall ratio, material type, etc.), trigger health performance simulations, and submit optimization schemes.

[0065] In another embodiment of the present invention, the core responsibility of the construction unit (which outputs construction data) is to define project requirements (such as health level or construction budget), approve plans, make investment decisions and confirm acceptance processes. Therefore, the data access permissions that the construction unit can have are for plan comparison data, cost-benefit reports, supervision reports and case libraries. After completing the construction data (such as approving design plans, health target levels, investment analysis, confirming project acceptance, etc.), it can be uploaded to the corresponding database.

[0066] The multiple responsible parties disclosed in this invention also include construction units, project management units, and suppliers. Each responsible party can access the corresponding data according to its own permissions and use the corresponding model for data analysis. For sensitive data (such as supplier quotations, construction budgets, etc.), data anonymization can be performed according to different responsible parties. When an unauthorized role views the relevant data, the key data will be hidden. For example, when the construction unit accesses the "construction budget", the words "XXX million yuan" will be displayed.

[0067] In another embodiment of the invention, the design data (including environmental data) includes: lighting conditions, ventilation conditions, and pedestrian behavior. After obtaining the initial layout scheme, the performance of the initial layout scheme can be evaluated using the healthy design parameters of various types of rooms stored in a relational database. Then, the model is trained using excellent case studies. The model's input includes parameters such as room type, orientation, and area, and the output is the optimal window opening scheme (e.g., a south-facing window width to depth ratio of 1:2.5). The training data comes from the design parameters and actual operating effects (such as lighting and ventilation effects) of 1000+ excellent healthy building case studies. The network weights are adjusted through a backpropagation algorithm to keep the model's prediction error within 5%.

[0068] Specifically, a Revit model is constructed based on the initial layout scheme. The Revit model is converted into a scene file, and solar and sky simulation data are set. The scene file is analyzed based on the solar and sky simulation data to obtain the lighting conditions. The initial layout scheme is input into a fluid dynamics model. The fluid dynamics model analyzes and meshes the initial layout scheme, sets boundary conditions, and generates and outputs the ventilation conditions based on the boundary conditions. The initial layout scheme is input into a multi-subject model. The multi-subject model simulates multiple random subjects based on the initial layout scheme, and generates and outputs the pedestrian behavior based on the movement path of each random subject. Corresponding weights are added to the lighting conditions, ventilation conditions, and pedestrian behavior, and a first, second, and third optimized scheme of the initial layout scheme are generated based on the lighting conditions, ventilation conditions, and pedestrian behavior, respectively. The first, second, and third optimized schemes are merged based on all the weights to obtain a healthy building design scheme.

[0069] In another embodiment of this invention, the Radiance engine is integrated as the core of daylight calculation, converting the Revit model into a Radiance-recognizable scene file. Sun position parameters (based on geographical location and time) and a sky model (e.g., a completely overcast sky) are set, and a Monte Carlo ray tracing algorithm is used to calculate the illuminance distribution inside the classroom. Based on the calculation results, the illuminance uniformity (the ratio of minimum illuminance to average illuminance) is analyzed. If the uniformity is below 0.7, the calculation is repeated by adjusting window positions (e.g., adding side windows or adjusting window widths) until the requirements are met. Simultaneously, the illuminance distribution results are visualized in the form of a pseudo-color map, allowing designers to intuitively understand the daylighting situation.

[0070] Furthermore, based on the computational fluid dynamics model, the crosswind path of the initial layout scheme was analyzed. The building model was meshed (using unstructured meshes, with mesh refinement in airflow inlet and outlet areas such as doors and windows), and boundary conditions (such as inlet air velocity, outlet pressure, and wall roughness) were set. The flow field distribution was calculated, and parameters such as airflow velocity and pressure were analyzed. Based on the simulation results, a combination of high and low windows (such as high windows on the south wall + low windows on the north wall) was recommended to keep the indoor airflow velocity in the optimal range of 0.3-0.5 m / s. At the same time, the ventilation rate was calculated. If it was lower than the design requirements (such as ≥2 times / h for dormitory ventilation), the window size or position was adjusted, and the simulation was repeated.

[0071] Furthermore, an agent-based model (multi-agent model) was used to simulate student activity trajectories. Each agent represented a student and possessed autonomous decision-making capabilities (such as choosing behaviors like attending class, leaving get out of class, eating, and exercising). Model parameters included students' schedules, activity preferences, and time arrangements. State transition rules (such as students moving from their dormitories to the classroom when class time arrives) were set to simulate student movement paths within the healthy building. Based on the simulation results, the population density and dwell time in areas such as building entrances / exits, corridors, and the playground were statistically analyzed. In areas with high population density and long dwell times (such as within 30 meters of building entrances / exits), facilities such as drinking fountains and rest seats were recommended to meet students' needs.

[0072] Furthermore, in another embodiment of the present invention, the mechanical properties of a healthy building design scheme are automatically verified. First, a recommendation system is constructed, and a climate zone database is built. This database divides the entire country into five climate zones: severe cold, cold, hot summer and cold winter, hot summer and warm winter, and temperate. Each zone contains data on temperature, humidity, and sunshine duration for a typical meteorological year. For different building types (such as teaching buildings, dormitories, and canteens) and climate zones, the system pre-sets a corresponding anti-condensation construction method library. For example, in cold regions, the exterior walls use 200mm aerated concrete + 50mm rock wool insulation. The heat transfer coefficient, thermal inertia index, and other parameters of this construction method have been verified through thermal calculations and meet the specifications. When the user inputs the building type and climate zone, the system matches the corresponding construction scheme from the construction method library and performs thermal bridge calculations on construction nodes (such as wall corners and window openings) using thermal bridge analysis software. This verifies whether there is a risk of condensation in the temperature distribution of the nodes (whether the inner surface temperature is higher than the indoor air dew point temperature). If a risk exists, the material or thickness of the construction layer is adjusted, and the calculation is repeated until the requirements are met.

[0073] Furthermore, a building materials database is constructed using a combination of relational databases and file systems. This database stores basic information about building materials (such as name, model, manufacturer, and price) and performance parameters (such as thermal resistance (R-value), VOC content (Volatile Organic Compounds), and fire resistance rating). The file system stores test reports, images, technical manuals, and other documents related to the building materials. The database contains data on over 5000 building materials, covering multiple categories including wall materials, insulation materials, door and window materials, and decorative materials. The system supports dual-objective filtering based on "energy saving + environmental protection." For example, it can filter coatings with VOC ≤ 0.5 mg / m³ and thermal conductivity ≤ 0.03 W / (m·K). The filtering process is implemented through SQL queries. First, coatings that meet the VOC limit are filtered out. Then, products with acceptable thermal conductivity are selected from the results and displayed in sorted order by price or performance. When automatically calculating the building's energy efficiency rate, the system first extracts the building envelope information (such as the materials and thickness of exterior walls, roofs, doors, and windows) from the Revit model. Based on the thermal parameters (thermal conductivity, heat transfer coefficient, etc.) in the building materials database, it calculates the heat loss of each envelope structure. Then, it combines the climate parameters of the building's location (such as heating degree-days and air conditioning degree-days) and the building's usage (such as heating and air conditioning methods and operating time) to calculate the building's annual energy consumption. If the calculated energy efficiency rate does not meet the specifications (such as energy efficiency rate ≥ 50%), the system recommends replacing materials, such as replacing ordinary glass with Low-E double-glazed glass, and recalculates the energy efficiency rate until it meets the standard.

[0074] Furthermore, the system extracts information such as the quantity and dimensions of components from the model using the Revit model API, including the length, height, and thickness of walls, and the area and thickness of floor slabs. Based on the calculation rules in the bill of quantities pricing specifications (e.g., calculating wall quantities by volume, deducting the volume of door and window openings), it calculates the quantities of various items and generates a bill of quantities. Simultaneously, it links with material price information from the building materials database to calculate the direct costs of each sub-item based on the material composition of the components (e.g., walls are composed of bricks and mortar). Then, it calculates the costs of measures fees, management fees, profits, and taxes according to certain rates to arrive at the project cost. To ensure cost accuracy (error ≤ 3%), the system regularly updates the building materials price database (by connecting to the price API of building materials suppliers) and calibrates the quantity calculation rules, comparing and adjusting the results with manual calculations.

[0075] Furthermore, in another embodiment of the present invention, a B2B (Business to Business) e-commerce platform architecture is adopted, developed based on the Spring Boot framework, and includes functional modules such as supplier management, product management, order management, and payment management. Suppliers must register an account on the platform and upload information such as company qualification certificates (e.g., business license, production license), material testing reports (e.g., formaldehyde emission test reports), product quotations, and construction videos. The platform uses blockchain technology (e.g., Hyperledger Fabric) to verify the authenticity of the testing data, generating hash values ​​from key information in the testing reports (e.g., testing institution, test results, testing time) and storing them on the blockchain. Due to the immutability of the blockchain, the testing data is guaranteed not to be tampered with. Simultaneously, the platform reviews the supplier's information, and only after approval are they allowed to display their products on the platform.

[0076] Blockchain technology is key to ensuring material quality traceability. Employing a Hyperledger Fabric consortium blockchain architecture, key information (testing institution, results, and time) from supplier-uploaded test reports (such as formaldehyde emission levels) is hashed and written to the blockchain. Due to the consortium blockchain's distributed ledger and consensus mechanism, no single node can tamper with the data, solving the problem of easily forged test reports in existing platforms. Simultaneously, the platform and supplier verification mechanisms work together, conducting both manual and machine verification of suppliers' qualifications (business licenses, production licenses). Only suppliers who pass the verification can participate in transactions. Blockchain-based evidence storage and qualification verification form a "double guarantee," ensuring the authenticity and traceability of material data.

[0077] Furthermore, semantic analysis technology is used to analyze material requirements in the design scheme, such as identifying the requirement for "anti-mildew coating" and extracting its key parameters (e.g., anti-mildew rating, color, dosage). The system matches these parameters with product information in the supplier product database, uses a cosine similarity algorithm to calculate the matching degree between requirements and products, and selects the three suppliers with the highest matching degree that meet the standards. Simultaneously, association rule mining reveals the relationship between products and construction tutorials. When a product is pushed, the corresponding construction tutorial (e.g., a spraying thickness control video) is automatically associated, making it convenient for construction units to view. The platform can optimize material selection based on performance and cost during the design phase, and achieve efficient procurement of materials and equipment through the supplier cloud platform during the construction phase, reducing construction costs. At the same time, the platform automatically assesses and optimizes the energy-saving and environmental performance of buildings, reducing energy consumption and environmental pollution during healthy building operation, meeting the requirements of sustainable development. In addition, through the rational planning and operation and maintenance management of healthy building facilities, the service life of facilities is extended, and resource utilization efficiency is improved.

[0078] Furthermore, when a school selects a particular brand of fresh air system, the platform scans the device serial number to bind the device information to the school project, creating an device file containing the device model, installation date, supplier information, etc. Based on the device's lifespan and maintenance manual, the system sets up regular maintenance reminders (such as filter replacement reminders) and notifies school administrators via SMS, app push notifications, etc. Simultaneously, the platform connects to the supplier's after-sales system. When the school submits after-sales requests, the platform automatically forwards the request information to the supplier and tracks the after-sales processing progress, ensuring a response time of ≤24 hours, forming a closed-loop management system of "procurement-installation-maintenance-after-sales".

[0079] Furthermore, during the construction phase, various sensors will be used for real-time monitoring at the construction site, and real-time detection images will be uploaded via an APP. The platform will extract key data from the detection images, automatically compare them with the mechanical properties of the planned scheme, and generate corresponding monitoring results. If the results are not up to standard, the problem will be managed in a closed loop. Through the process of "problem submission - rectification - re-inspection - archiving", the project management unit will ensure full-process supervision, thereby increasing the health indicator acceptance rate from 85% to 98% and ensuring that health standards are implemented.

[0080] Furthermore, during the operation and maintenance phase, multiple monitoring points can be planned. Based on the operation and maintenance drawings, clustering algorithms (such as K-means clustering) are used to analyze the functional characteristics and activity density of each area within the healthy building, grouping similar areas into a cluster. For each cluster, the type, number, and installation location of monitoring points are determined according to the monitoring objectives (such as temperature, humidity, lighting, noise, air quality, student safety, etc.). For example, for the classroom area cluster, due to the high density of people and the high requirements for lighting and air quality, a light sensor is planned to be placed in the center of the classroom, and an air quality monitor is installed near the front and back doors of the classroom. For the dormitory area, an air quality monitor and a sensor linked to the access control system are installed at the dormitory entrance. Around the playground, due to frequent personnel activity and potential noise problems, noise sensors (sampling frequency 1 time / minute) and video surveillance equipment are installed. At the same time, considering the building structure and pipeline layout of the healthy building, the installation locations of the sensors are ensured to facilitate wiring and maintenance, and not to affect the normal teaching and living order of the healthy building.

[0081] The clustering algorithm selects features based on "functional type + personnel density + monitoring target" (e.g., classrooms belong to "high personnel density + lighting / air quality sensitive areas," and playgrounds belong to "high traffic + noise sensitive areas"). K-means is used to divide healthy buildings into 5-8 clusters (dynamically adjusted according to the size of the healthy building). The number of monitoring points in each cluster is calculated as "coverage area × sensitivity" (e.g., one light sensor is placed for every 50 square meters in a classroom). Monitoring points are reduced in non-sensitive areas (such as green belts), and data is supplemented using interpolation algorithms; in sensitive areas (such as classrooms and dormitories), the monitoring points are densely distributed to ensure data accuracy. Actual verification shows that this solution saves 30% of hardware costs compared to uniform deployment, while still achieving 98% monitoring coverage.

[0082] The platform covers all stages of healthy building design, construction, and operation and maintenance. Through real-time data flow and sharing, it achieves close integration between each stage, forming a closed-loop management system for the entire process. During the construction phase, the supplier cloud platform facilitates the application and integration of new materials and products; during the operation and maintenance phase, monitoring data can be fed back to the design and construction phases, providing a reference for subsequent healthy building renovations and upgrades, and continuously optimizing the health performance of healthy buildings.

[0083] Furthermore, a health data collection and analysis system is established. Sensors at various monitoring points collect real-time data on temperature (18-26℃), humidity (40%-60%), light intensity, noise levels, PM2.5 concentration, and carbon dioxide concentration, which is then transmitted to a cloud platform database via IoT technology. The cloud platform uses a time-series database to store this real-time data, specifically optimized for time-series data, supporting efficient write and query operations. The platform has a built-in data processing engine that cleans (removes outliers and fills in missing values) and standardizes the collected data before comparing it with preset health design thresholds. For example, when the carbon dioxide concentration in a classroom exceeds the standard (>1000ppm, where 1 ppm of solution means 1 mg of solute in 1 kg of solvent), the data processing engine triggers an early warning mechanism, automatically pushing ventilation suggestions (such as opening the fresh air valve in the corridor, opening classroom windows, etc.) to the school administrators' APP and the healthy building management system. Meanwhile, the system records the time, duration, handling measures, and results of each excessive event, forming a historical data archive. Through data mining technology, the system analyzes the causes and patterns of excessive emissions. For example, if it is found that a classroom frequently experiences excessive carbon dioxide concentrations after the first class in the afternoon, it may be due to dense student populations and inadequate ventilation. Therefore, the system proposes long-term optimization suggestions to adjust the ventilation strategy for that period.

[0084] Finally, by integrating data from video surveillance equipment and various sensors, computer vision algorithms (such as deep learning-based target detection and tracking algorithms) are used to analyze the video footage, identifying student gatherings and abnormal behaviors (such as running and pushing). By calculating the number of students per unit area, when the student density exceeds 5 people / m², the system determines it as abnormal gathering and immediately triggers an alert. Simultaneously, by combining data from noise sensors, if a sudden and significant increase in noise levels in a certain area is accompanied by abnormal gatherings of people, it further confirms the possibility of a sudden incident. After the alert is generated, the system issues evacuation notices through the healthy building broadcast system, specifying evacuation directions and assembly points. At the same time, the alert information and on-site video footage are pushed to the terminals of the school security department and relevant management personnel, facilitating timely understanding and response. This forms a complete closed loop of "monitoring-early warning-response-feedback," where the feedback stage records the response results and subsequent developments in the system, serving as a basis for optimizing the alert model and response plan.

[0085] Furthermore, in another embodiment of this invention, a user feedback mechanism can be established. When a user selects or modifies a solution recommended by the platform, the system records the user's preferences and modifications, and inputs this information as new training samples into the model. The model's parameters and weights are continuously adjusted through online learning algorithms (such as incremental learning), improving recommendation accuracy with repeated use. A target of ≥70% solution adoption rate is set. The model's recommendation performance is evaluated periodically. If the adoption rate falls below the target value, the model is retrained and optimized to ensure the platform continuously provides optimized solutions that meet user needs. Employing advanced technologies such as fully automated identification and comprehensive weighted judgment, this invention provides a novel solution for the field of healthy building design. It integrates multidisciplinary knowledge and technologies, such as architectural design, artificial intelligence, and big data analysis, promoting the integrated application of related technologies in the architectural design field. Simultaneously, the platform's supplier cloud platform facilitates the research and development and promotion of new materials and products, driving technological progress and development across the entire industry.

[0086] The weighting is based on the priority of core indicators for healthy buildings: through questionnaires (covering 100+ designers and school administrators) and case studies (statistics on the design focus of 50+ healthy buildings), "lighting" (which directly affects vision and comfort) was determined to have the highest weight, followed by "noise", and "cost" as an economic constraint. The three factors work together to balance health and feasibility.

[0087] Furthermore, once the current architectural design is completed, a complete case study is saved according to the different types of architectural designs and stored in the corresponding database. For example, after completing the design of a "healthy campus", the completed "healthy campus architectural design scheme" is stored in the campus building database. When other similar buildings need to be designed in the future, the completed "healthy campus architectural design scheme" can be recommended and displayed visually.

[0088] For example, in another embodiment of the present invention, for a residential project in a certain area, during the project planning stage, the platform searches for residential projects based on local environmental information, obtains multiple similar cases, performs technical benchmarking, and then obtains recommended cases, which are then uploaded to the design institute.

[0089] Furthermore, the design institute uploads a preliminary Revit model (90mm exterior wall, window-to-wall ratio 0.4), and the platform automatically matches the "Residential Simulation Template." Then, a health performance simulation (including energy consumption simulation and ventilation simulation) is triggered, and the platform automatically triggers an optimization model, generating an optimization plan within 10 minutes: "Exterior wall insulation increased to 120mm (thermal conductivity 0.03W / (mk)), master bedroom window opening area increased by 15%." The results are then re-simulated based on the optimization plan. If the plan meets the requirements, it is sent to the construction unit.

[0090] Furthermore, the construction unit reviews the comparative data of the proposed solutions (changes in energy consumption and investment before and after optimization) and the cost-benefit analysis (energy savings of 2 million yuan over a 50-year lifespan), and then pushes the process to the construction unit and the project management unit.

[0091] Furthermore, during construction, monitoring and management are carried out, including monitoring construction materials, environmental information, and construction density, and generating health supervision reports which are submitted to the construction unit.

[0092] Furthermore, once the construction unit confirms that the project has passed the acceptance inspection, the platform automatically enters the project into the case library and updates the base database. Because the compliance rate of the relevant material suppliers reached 100% in this cooperation, their comprehensive score rose from 94 points to 96 points, and they were upgraded to "core suppliers".

[0093] This invention automatically identifies multimodal data that affects building design through a full-process data interaction process, enabling full-cycle healthy design optimization from early planning to operation and maintenance, forming a closed-loop management system, and improving the health, compliance, and economy of healthy building design.

[0094] Furthermore, such as Figure 2 As shown, based on the above-mentioned healthy building design method based on full-process data analysis, the present invention also provides a healthy building design system based on full-process data analysis, wherein the healthy building design system based on full-process data analysis includes:

[0095] The data registration module 51 is used to acquire vector data and remote sensing images of the target area, preprocess the remote sensing images to obtain spatial information, and spatially register the vector data and the spatial information to obtain the initial planning information of the target area.

[0096] Environmental detection module 52 is used to construct a healthy building design knowledge graph, determine the compliance of the initial planning information based on the healthy building design knowledge graph, and analyze the environment of the target area if it complies with the rules to obtain various environmental information.

[0097] The scheme generation module 53 is used to construct constraints and objective functions based on all the environmental information, and to optimize the objective function using a genetic algorithm under the constraints to obtain an initial layout scheme;

[0098] The scheme optimization module 54 is used to perform healthy building performance simulation based on the initial layout scheme, obtain various simulation data, input all the simulation data into the corresponding responsible entity model, each responsible entity model outputs final data based on the corresponding simulation data, and constructs a healthy building design scheme based on all the final data.

[0099] Furthermore, such as Figure 3 As shown, based on the above-mentioned healthy building design method and system based on full-process data analysis, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 3 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0100] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard disk or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard disk, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard. Furthermore, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a health building design program 40 based on full-process data analysis, which can be executed by the processor 10 to implement the health building design method based on full-process data analysis in this application.

[0101] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the healthy building design method based on full-process data analysis.

[0102] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The components of the terminal communicate with each other via a system bus.

[0103] In one embodiment, when the processor 10 executes the health building design program 40 based on full-process data analysis in the memory 20, it implements the steps of the health building design method based on full-process data analysis as described above:

[0104] In summary, this invention provides a healthy building design method and related equipment based on full-process data analysis. The method includes: acquiring vector data and remote sensing images of a target area; preprocessing the remote sensing images to obtain spatial information; spatially registering the vector data and spatial information to obtain initial planning information for the target area; constructing a healthy building design knowledge graph; judging the compliance of the initial planning information based on the knowledge graph; analyzing the environment of the target area according to the rules to obtain various environmental information; constructing constraints and objective functions based on all the environmental information; optimizing the objective function using a genetic algorithm under the constraints to obtain an initial layout scheme; simulating the environmental performance of a healthy building based on the initial layout scheme to obtain various simulation data; adding corresponding weights to all the simulation data; and integrating all the simulation data based on all the weights to obtain a healthy building design scheme. This invention automatically identifies multimodal data that affects building design through a full-process data interaction process, realizing full-cycle healthy design optimization from early planning to operation and maintenance, forming a closed-loop management system, and improving the health, compliance, and economy of healthy building design.

[0105] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0106] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0107] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A healthy building design method based on full-process data analysis, characterized in that, The healthy building design method based on full-process data analysis includes: The vector data and remote sensing images of the target area are acquired, the remote sensing images are preprocessed to obtain spatial information, and the vector data and the spatial information are spatially registered to obtain the initial planning information of the target area. The vector data includes: spatial coordinates and attribute information; the spatial information includes: terrain slope, building volume, and road grade. The process of acquiring vector data and remote sensing images of the target area, preprocessing the remote sensing images to obtain spatial information, and spatially registering the vector data and the spatial information to obtain initial planning information for the target area specifically includes: The spatial coordinates and attribute information of the target area are obtained through a spatial data engine, and remote sensing images are acquired. The remote sensing images are analyzed to obtain the terrain slope, the building volume, and the road grade. The spatial coordinates, attribute information, terrain slope, building volume, and road grade are projected into the same projection coordinate system to obtain the data to be registered; Multiple control points are determined in the data to be registered, and all the data to be registered are transformed according to the multiple control points to obtain unified initial planning information; The acquisition of remote sensing images, and the analysis of those images to obtain the terrain slope, building volume, and road grade, specifically include: Acquire remote sensing images, and eliminate noise, distortion, and bias in the remote sensing images to obtain the target remote sensing image; Multiple target features are identified in the target remote sensing image, and each target feature is enhanced. The target remote sensing image after feature enhancement is input into the constructed semantic segmentation model. The semantic segmentation model performs pixel-level classification on the processed target remote sensing image and outputs terrain slope features, building volume features and road grade features. Error elimination processing is performed on the terrain slope features, building volume features, and road grade features, and the processed terrain slope features, building volume features, and road grade features are then vectorized to obtain the terrain slope, building volume, and road grade; A knowledge graph for healthy building design is constructed. The compliance of the initial planning information is judged based on the knowledge graph. If it is compliant, the environment of the target area is analyzed to obtain various environmental information. Based on all the environmental information, constraints and objective functions are constructed. Under the constraints, a genetic algorithm is used to optimize the objective function to obtain an initial layout scheme. Based on the initial layout scheme, a healthy building performance simulation is performed to obtain various simulation data. All simulation data are then input into the corresponding responsible entity model. Each responsible entity model outputs final data based on the corresponding simulation data, and a healthy building design scheme is constructed based on all the final data.

2. The healthy building design method based on full-process data analysis according to claim 1, characterized in that, The environmental information includes: noise contour lines, sunshine distribution map, and traffic density values; The process involves constructing a knowledge graph for healthy building design, determining the compliance of the initial planning information based on this knowledge graph, and if compliant, analyzing the environment of the target area to obtain various environmental information, specifically including: Multiple planning clauses are obtained, and all the planning clauses are analyzed to obtain entity information and entity relationships. The entity information includes multiple clause types, and the entity relationships include the relationships between all the clause types. Based on all the stated clause types and entity relationships, extract the entity attributes of all the stated planning clauses, and construct a knowledge graph of healthy building design based on the entity information, entity relationships, and entity attributes. A compliance report is obtained by performing spatial overlay analysis on the initial planning information based on the aforementioned healthy building design knowledge graph. If the compliance report indicates compliance, then multiple healthy building functional zones are determined based on the initial planning information; Calculate the noise contour lines, the sunshine distribution map, and the traffic density value for each of the healthy building functional areas.

3. The healthy building design method based on full-process data analysis according to claim 2, characterized in that, The step of performing spatial overlay analysis on the initial planning information based on the healthy building design knowledge graph to obtain a compliance report specifically includes: Obtain the planning and control layer of the target area, and superimpose the planning and control layer with the spatial coordinates in the initial planning information layer by layer to obtain the first superimposed conflict information; The knowledge graph of healthy building design is superimposed layer by layer with the spatial coordinates in the initial planning information to obtain the second superimposed conflict information corresponding to each planning clause; Determine the priority of all planning clauses in the healthy building design knowledge graph, and filter all second superimposed conflict information according to the priority, deleting duplicate superimposed conflict information; A structured compliance report is generated based on the first superimposed conflict information and all filtered second superimposed conflict information.

4. The healthy building design method based on full-process data analysis according to claim 2, characterized in that, The constraints include: spatial constraints, attribute constraints, and security constraints; The step of constructing constraints and an objective function based on all the environmental information, and then optimizing the objective function using a genetic algorithm under the constraints to obtain an initial layout scheme, specifically includes: The spatial constraints and safety constraints are constructed based on all the healthy building functional zones, and the attribute constraints are constructed based on the noise contour lines, the sunshine distribution map and the traffic density value of each healthy building functional zone; The noise contour lines, sunshine distribution maps, and traffic density values ​​of each healthy building functional area are weighted and summed. Under the conditions of ensuring the spatial constraints, attribute constraints, and safety constraints, the noise contour lines, sunshine distribution maps, and traffic density values ​​are maximized respectively to obtain the objective function. All the healthy building functional areas are coded as initial populations. Based on the spatial constraints, attribute constraints, and security constraints, selection, crossover, and mutation operations are performed on all the initial populations to obtain optimized new populations. This process continues until the interference between all healthy building functional areas is minimized, resulting in an initial layout scheme.

5. The healthy building design method based on full-process data analysis according to claim 1, characterized in that, The simulation data includes: design data, construction data, project management data, and supply data; The process involves performing healthy building performance simulations based on the initial layout scheme to obtain various simulation data. All simulation data are then input into corresponding responsible entity models. Each responsible entity model outputs final data based on its corresponding simulation data. Finally, a healthy building design scheme is constructed based on all the final data, specifically including: A Revit model is constructed based on the initial layout scheme, and the Revit model is converted into a scene file containing the design data, the construction data, the construction data, the project management data, and the supply data. Input the scenario files into the existing design entity model, construction entity model, project management entity model, and supply entity model, respectively; The design entity model optimizes the design data and outputs the final design data; The construction entity model optimizes the construction data and outputs the final construction data; The construction entity model optimizes the construction data and outputs the final construction data; The project management entity model optimizes the project management data and outputs the final project management data; The supply entity model optimizes the supply data and outputs the final supply data; The final design data, the final construction data, the final construction data, the final project management data, and the final supply data are weighted and integrated to obtain a healthy building design scheme.

6. A healthy building design system based on full-process data analysis, as described in any one of claims 1-5, characterized in that, The healthy building design system based on full-process data analysis includes: The data registration module is used to acquire vector data and remote sensing images of the target area, preprocess the remote sensing images to obtain spatial information, and spatially register the vector data and the spatial information to obtain the initial planning information of the target area. An environmental detection module is used to construct a knowledge graph of healthy building design, determine the compliance of the initial planning information based on the knowledge graph, and if it complies with the rules, analyze the environment of the target area to obtain various environmental information. The scheme generation module is used to construct constraints and objective functions based on all the environmental information, and to optimize the objective function using a genetic algorithm under the constraints to obtain an initial layout scheme; The scheme optimization module is used to perform healthy building performance simulation based on the initial layout scheme, obtain various simulation data, input all the simulation data into the corresponding responsible entity model, each responsible entity model outputs final data based on the corresponding simulation data, and constructs a healthy building design scheme based on all the final data.

7. A terminal, characterized in that, The terminal includes: a memory, a processor, and a health building design program based on full-process data analysis stored in the memory and executable on the processor. When the health building design program based on full-process data analysis is executed by the processor, it implements the steps of the health building design method based on full-process data analysis as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a healthy building design program based on full-process data analysis, which, when executed by a processor, implements the steps of the healthy building design method based on full-process data analysis as described in any one of claims 1-5.

Images