CN-121980928-A - Balcony arbor layout optimization method based on conditional diffusion generation model

CN121980928ACN 121980928 ACN121980928 ACN 121980928ACN-121980928-A

Abstract

The invention provides a balcony arbor layout optimization method based on a conditional diffusion generation model, which comprises the following steps of S1, acquiring summer climate parameters of an area where a fourth-generation residential project is located and the dimensional and position relation of an air garden type balcony and an indoor space, generating project scene description data according to the dominant wind direction in summer, the opening direction of the air garden type balcony and the communication condition of the air garden type balcony and all main rooms, and establishing an air garden type balcony arbor layout automatic generation and optimization method capable of considering the dominant wind direction ventilation channel connectivity and the thermal comfort requirements of multiple rooms under the condition of giving the summer climate and the building geometry of the fourth-generation residential building. The invention can optimize the configuration of the garden-type balcony arbor while ensuring the smooth ventilation channel, improve the indoor thermal comfort in summer and reduce manual trial calculation.

Inventors

LIU WEIGUO
LI ZEHAO
ZHOU SHAXU
XIE WENHAI

Assignees

吉首大学

Dates

Publication Date: 20260505
Application Date: 20260109

Claims (8)

1. A balcony arbor layout optimization method based on a conditional diffusion generation model is used for a fourth-generation residence air garden balcony and is characterized by comprising the following steps: s1, acquiring summer climate parameters of an area where a fourth-generation residential project is located and the dimensional and position relation of an air garden type balcony and an indoor space, and generating project scene description data according to the dominant wind direction in summer, the opening direction of the air garden type balcony and the communication condition of the air garden type balcony and each main room; S2, constructing a conditional diffusion arbor layout generation model according to the air garden balcony dividing units, the dominant wind direction and the room communication condition in project scene description data, taking whether tree species, tree species types and tree specifications of each air garden balcony dividing unit are used as generation variables in the conditional diffusion arbor layout generation model, taking the connectivity of a ventilation channel from outdoor through the air garden balcony to an indoor opening of the dominant wind direction as constraint, determining layout preference parameters according to a thermal comfort evaluation result of the air garden arbor layout in historical engineering, and generating a project arbor layout generation model; S3, generating a plurality of groups of initial layouts of the air garden type balcony trees according to project scene description data and a project tree layout generation model, judging each group of initial layouts of the air garden type balcony trees according to connectivity of the ventilation channels, screening the initial layouts of the air garden type balcony trees by using layout preference parameters, and removing layouts which do not meet connectivity constraints of the ventilation channels of the dominant wind direction, so as to obtain an air garden type balcony tree candidate layout set; S4, generating a corresponding building and air garden type balcony arbor integrated three-dimensional model for each air garden type balcony arbor candidate layout according to the air garden type balcony arbor candidate layout set and the size and position relation between the air garden type balcony and the indoor space; S5, carrying out computational fluid dynamics simulation according to the building and air garden type balcony arbor integrated three-dimensional model and summer climate parameters, and calculating the wind speed, the temperature and the humidity of each main room under each air garden type balcony arbor candidate layout to generate indoor environment response data; S6, calculating a thermal comfort evaluation index of each aerial garden type balcony arbor candidate layout according to the indoor environment response data, and comprehensively evaluating by adopting an entropy weight-rubbing posis method based on the thermal comfort evaluation index and preset index weight to obtain an aerial garden type balcony arbor layout sequencing result; S7, determining the air garden type balcony arbor layout meeting the thermal comfort requirement according to the air garden type balcony arbor layout sorting result and a preset thermal comfort threshold value, forming a target air garden type balcony arbor layout scheme, and outputting the target air garden type balcony arbor layout scheme.
2. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the step S1 is specifically as follows: Acquiring summer climate parameters of a region where a fourth-generation residential project is located, wherein the summer climate parameters comprise a summer dominant wind direction, a dominant wind speed level, a typical summer outdoor temperature and a typical summer outdoor relative humidity, and finishing the summer climate parameters into summer climate parameter data; Acquiring the dimensional position relation between an air garden type balcony and an indoor space, wherein the dimensional position relation comprises a building plane position, an air garden type balcony boundary line, an air garden type balcony opening position, plane positions of main rooms and door and window positions connected with the air garden type balcony, dividing the air garden type balcony into a plurality of air garden type balcony dividing units according to the air garden type balcony boundary line, and determining the plane size, the height and the center coordinates of each air garden type balcony dividing unit to form air garden type balcony dividing unit data; Calculating the opening direction of each air garden type balcony according to the summer dominant wind direction, the air garden type balcony opening position and the door and window position in summer climate parameter data, calculating the geometric communication path between each air garden type balcony dividing unit and each main room according to the air garden type balcony dividing unit data, and determining the communication condition of each air garden type balcony dividing unit and each main room to form air garden type balcony communication relation data; And uniformly coding the main wind direction in summer, the opening direction of the air garden type balcony and the communication condition of the air garden type balcony and each main room according to the summer climate parameter data, the air garden type balcony dividing unit data and the air garden type balcony communication relation data to generate project scene description data.
3. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the step S2 is specifically: Extracting the plane position, the height and the connection relation between each air garden type balcony dividing unit and each main room according to the air garden type balcony dividing unit, the dominant wind direction and the room connection condition in the project scene description data, generating air garden type balcony dividing unit layout state basic data, generating scene condition basic data according to the dominant wind direction in summer and the dominant wind speed level in summer in the project scene description data, and simultaneously acquiring historical engineering air garden type balcony arbor layout thermal comfort evaluation data; Generating a scene condition vector according to scene condition basic data, generating a layout state vector sequence according to layout state basic data of an air garden type balcony dividing unit, arranging the layout state vector sequence according to the projection sequence of a dominant wind direction on an air garden type balcony plane, and splicing the scene condition vector and the layout state vector in a numerical dimension to form an input vector sequence of a conditional diffusion arbor layout generation model; Setting a one-dimensional convolution feature extraction layer in a conditional diffusion arbor layout generation model, taking an input vector sequence as input, carrying out convolution feature calculation on a plurality of air garden type balcony dividing units adjacent along a dominant wind direction, and generating an intermediate feature vector sequence, wherein the intermediate feature vector sequence is used for representing the influence of air flow path caused by the combination of the air garden type balcony dividing units on the dominant wind direction; Setting a ventilation channel grading layer in a conditional diffusion arbor layout generation model, taking an intermediate feature vector sequence as input, calculating a ventilation channel grading value of each air garden type balcony dividing unit in the main wind direction, and synthesizing the ventilation channel grading value and the intermediate feature vector in the numerical dimension to form a channel feature vector sequence, wherein the channel feature vector sequence is used for reflecting the connectivity of the ventilation channel from the main wind direction to an indoor opening through the air garden type balcony; Generating layout preference parameters according to thermal comfort evaluation results corresponding to different arbor positions, tree types and tree specifications in historical engineering air garden type arbor layout thermal comfort evaluation data, inputting the layout preference parameters into a layout preference adjusting layer in a conditional diffusion arbor layout generation model, and carrying out weighted adjustment on channel feature vector sequences to obtain channel feature vector sequences with the layout preference parameters, wherein the channel feature vector sequences with the layout preference parameters are used for guiding an air garden type balcony dividing unit to be close to a layout mode with a good historical thermal comfort evaluation result; Setting an output control layer in a conditional diffusion arbor layout generation model, taking a channel characteristic vector sequence with layout preference parameters as input, and calculating whether tree species control values, tree species type control values and tree specification control values of each air garden type balcony dividing unit through a plurality of layers of fully connected neurons to form a layout control value set, wherein the layout control value set is used for updating whether tree species, tree species types and tree specifications of the air garden type balcony dividing units in each time step of a diffusion process; And determining a layout updating mode of the conditional diffusion arbor layout generation model according to the layout control value set and the basic data of the layout state of the air garden balcony dividing unit, solidifying network parameters of the conditional diffusion arbor layout generation model after the layout updating mode is stable, and generating the project arbor layout generation model.
4. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the method is characterized in that scene condition vectors for reasoning are generated according to air garden balcony dividing units, dominant wind directions and room communication conditions in project scene description data, and whether the initial tree species, the initial tree species type and the initial tree specification of all the air garden balcony dividing units are planted or not is set to form a corresponding initial layout state vector sequence; inputting a scene condition vector and an initial layout state vector sequence into a project arbor layout generation model, updating the layout state vector sequence on a plurality of diffusion time steps according to a layout updating mode in the project arbor layout generation model, adjusting whether tree species, tree species types and tree specifications of an air garden type balcony dividing unit are planted or not according to a layout control value set in each diffusion time step to obtain a plurality of groups of layout updating results, and converting each group of layout updating results into a plurality of groups of air garden type arbor initial layouts; Inputting a plurality of groups of initial layouts of the air garden type balcony trees and scene condition vectors into ventilation channel evaluation layers in a project tree layout generation model, calculating the ventilation channel evaluation value of each air garden type balcony dividing unit under the corresponding initial layouts of the air garden type balcony trees, and carrying out accumulated calculation on the ventilation channel evaluation values of each air garden type balcony dividing unit according to the dominant wind direction to form ventilation channel connectivity evaluation data of each group of initial layouts of the air garden type balcony trees; Comparing the ventilation channel connectivity evaluation data with a preset ventilation channel connectivity threshold, and determining the corresponding air garden type balcony arbor initial layout as a ventilation channel qualified layout when the ventilation channel connectivity evaluation data meets the preset ventilation channel connectivity threshold to form a ventilation channel qualified air garden type balcony arbor initial layout set; Calculating layout preference scoring values of the initial layout of the air garden type balcony arbor qualified by each group of ventilation channels according to the initial layout set and layout preference parameters of the air garden type balcony arbor qualified by the ventilation channels, and determining the layout preference scoring values according to whether tree species, tree species types and tree specifications in an air garden type balcony dividing unit are matched with the degree that thermal comfort evaluation in historical engineering reaches a preset thermal comfort standard layout; And comparing the layout preference scoring value with a layout preference scoring threshold, reserving the initial layout of the air garden type balcony arbor corresponding to the qualified ventilation channel when the layout preference scoring value is greater than or equal to the layout preference scoring threshold, and merging the reserved initial layouts of the air garden type balcony arbor qualified by the ventilation channel to generate an air garden type balcony arbor candidate layout set.
5. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the step S4 is specifically: Analyzing the dimensional and positional relationship between the air garden balcony and the indoor space into plane boundary lines and height information of the outer wall, the floor slab, the air garden balcony ground and the inner surfaces of all main rooms, and generating a building three-dimensional basic model comprising a building main body and the indoor space according to the plane boundary lines and the height information; Determining the positions of the corresponding air garden type balcony dividing units in the building three-dimensional basic model as tree planting points according to whether the tree is planted, the tree type and the tree specification of each air garden type balcony dividing unit in each air garden type balcony arbor candidate layout set, and determining trunk height and crown volume parameters according to the tree specification to generate a corresponding arbor three-dimensional object set; And superposing the arbor three-dimensional object set corresponding to each aerial garden type balcony arbor candidate layout into a building three-dimensional basic model, and processing geometric connection between the arbor three-dimensional object and the aerial garden type balcony ground, the outer wall and the indoor opening to form a building and aerial garden type balcony arbor integrated three-dimensional model for computational fluid mechanics simulation.
6. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein each building and the air garden balcony arbor integrated three-dimensional model are in butt joint with summer climate parameters, outdoor incoming flow boundaries are determined according to a summer dominant wind direction, incoming flow boundaries and far field boundaries are assigned according to a summer dominant wind speed level, a typical summer outdoor temperature and a typical summer outdoor relative humidity, indoor air areas of main rooms are calibrated in the building and air garden balcony arbor integrated three-dimensional model, and calculation domains and boundary condition data for calculating hydrodynamic simulation are formed; Generating computational fluid mechanics computational grids according to the computational domain and boundary condition data, inputting the computational fluid mechanics computational grids and summer climate parameters into a computational fluid mechanics solver together, and calculating a steady-state flow field and a temperature-humidity field of each building and an air garden type balcony arbor integrated three-dimensional model to obtain simulation result data comprising a wind speed field, a temperature field and a humidity field outside the building and inside each main room; And according to simulation result data and indoor air areas of main rooms calibrated in the building and air garden type balcony arbor integrated three-dimensional model, carrying out statistical calculation on a wind speed field, a temperature field and a humidity field of each main room in the corresponding indoor air area to form a wind speed, temperature and humidity numerical value set of each main room under each air garden type balcony arbor candidate layout, and taking the wind speed, temperature and humidity numerical value set as indoor environment response data.
7. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the step S6 is specifically: Taking the values of the wind speed, the temperature and the humidity of each air garden type balcony arbor candidate layout in the indoor environment response data as input, calculating a wind speed comfort level index, a temperature comfort level index and a humidity comfort level index for each main room according to the thermal comfort intervals respectively set for the wind speed, the temperature and the humidity, and summarizing according to the room and the index types to form a thermal comfort evaluation index set of each air garden type balcony arbor candidate layout; The method comprises the steps of forming an evaluation matrix from a thermal comfort evaluation index set of all aerial garden type arbor candidate layouts according to index types, calculating entropy weight of each thermal comfort evaluation index according to numerical distribution conditions of each thermal comfort evaluation index in the evaluation matrix on different aerial garden type arbor candidate layouts, and synthesizing the entropy weight with preset index weights to obtain comprehensive weight of each thermal comfort evaluation index; The entropy weight rubbing posis method is adopted, an evaluation matrix and comprehensive weights are used as input, the closeness value of each air garden type balcony arbor candidate layout relative to an ideal thermal comfortable layout and a non-ideal thermal comfortable layout is calculated, and all the air garden type balcony arbor candidate layouts are ordered according to the closeness value, so that an air garden type balcony arbor layout ordering result is obtained.
8. The balcony arbor layout optimization method based on the conditional diffusion generation model of claim 1, wherein the method is characterized in that according to the closeness value of each aerial garden type balcony arbor candidate layout and the corresponding thermal comfort evaluation index in the aerial garden type balcony arbor layout sequencing result, combining the thermal comfort requirements of all main rooms, and refining the preset thermal comfort threshold value into a preset thermal comfort threshold value set aiming at all main rooms and all thermal comfort evaluation indexes; Comparing the sorting result of the air garden type balcony arbor layout with a preset thermal comfort threshold value set, judging whether each thermal comfort evaluation index of the air garden type balcony arbor candidate layout in each main room is larger than or equal to a corresponding preset thermal comfort threshold value or not for each air garden type balcony arbor candidate layout, and incorporating the air garden type balcony arbor candidate layout into the air garden type balcony arbor layout set meeting the thermal comfort requirement when all the thermal comfort evaluation indexes are larger than or equal to the corresponding preset thermal comfort threshold values; and selecting the first aerial garden type balcony arbor candidate layout ordered in the aerial garden type balcony arbor layout ordering result from the aerial garden type balcony arbor layout set meeting the thermal comfort requirement according to the aerial garden type balcony arbor layout set meeting the thermal comfort requirement and the aerial garden type balcony arbor layout ordering result, forming a target aerial garden type balcony arbor layout scheme, and outputting the target aerial garden type balcony arbor layout scheme.

Description

Balcony arbor layout optimization method based on conditional diffusion generation model Technical Field The invention relates to the technical field of building environment, in particular to a balcony arbor layout optimization method based on a conditional diffusion generation model. Background With the continuous acceleration of the urban process in China, the urban population density is continuously improved, and residential buildings not only need to meet basic living demands, but also evolve towards ecology, comfort and diversification. The traditional high-rise residence is difficult to match pursuit of residents on natural living experience due to the limitations of strong sealing property and insufficient greening, and the aerial garden type high-rise residence becomes an important direction for breaking the contradiction by virtue of the innovative form of vertical greening and three-dimensional courtyard, namely, the novel high-rise residence expands the outdoor activity field of residents and improves the ecological environment of urban cement forest by vegetation coverage by integrating an open garden space in a high-rise structure, thereby realizing dual promotion of living functions and ecological value, being rapidly popularized in China in recent years, and continuously climbing market acceptance and policy support strength. In the practice of domestic high-rise garden type house, the 'fourth generation house' is a very representative realization form, takes a 'double-layer through high-rise garden and staggered layer overhanging structure' as a core design, and configures independent courtyard space for each house through staggered layout of a parity layer balcony, thereby enhancing the living experience of the 'house has a house', further playing the roles of sun shading, dust stagnation and cooling of vertical greening by means of large-area earthing and green planting, and having remarkable value for beautifying the space line of a city and improving living quality. However, fourth-generation residences generally face the problems of strong solar radiation, remarkable urban heat island effect and the like in summer, and the air garden type balcony space is widely used for planting arbor to improve sun shading, reduce heat gain of an enclosure structure and improve living comfort. Meanwhile, the garden balcony serves as an important buffer space between the outside of the building and the indoor space, and the opening direction, the communication relation with doors and windows of a main room and the greening arrangement form of the garden balcony can obviously influence the natural ventilation path under the dominant wind direction and the indoor wind speed, temperature and humidity distribution. Therefore, on the premise of ensuring landscapes and greening quantity, how to comprehensively arrange the balcony arbor layout of the air garden and ventilation and thermal environment in summer becomes a key problem in the design of building environment. In the prior art, the fourth-generation residential air garden balcony greening is mainly based on standard treatises and empirical greening guidelines, and the number, variety and general planting positions of arbor are given according to the area of the air garden balcony or sunlight conditions, and the method of configuring the greening by simple sunshade coefficients or greening rate indexes is also available. Part of researches utilize wind environment simulation or building energy consumption simulation to carry out computational fluid mechanics or thermal engineering simulation on a small number of given air garden type balcony planting schemes, and then a designer carries out multi-round manual adjustment according to the results. Some parameterized design and optimization methods can adjust vegetation positions and heights on the dimension of an air garden balcony, but usually only take sun shading, landscapes or single indoor indexes as evaluation basis, and lack integrated coupling with indoor multi-room thermal comfort indexes. The prior art generally lacks a mechanism for systematically associating and restricting the main wind direction from the outdoor ventilation channel to the indoor opening through the garden balcony, and the specific position, tree species and specification of the garden balcony arbor, so that the greening arrangement and ventilation path continuity are difficult to be simultaneously considered. Meanwhile, the existing method is mostly based on single or few environmental indexes, lacks a multi-room and multi-index thermal comfort comprehensive evaluation and sequencing means based on computational fluid mechanics results, and is difficult to automatically screen a scheme which meets ventilation and thermal comfort requirements in a plurality of alternative air garden type balcony arbor layouts. Therefore, a balcony arbor layout optimization method based on a conditional diffusion generation model is urgentl