A building health performance optimization design method based on spatial thermal autonomy

Abstract

The application provides a building health performance optimization design method based on space thermal autonomy. The method realizes comprehensive and overall evaluation of user thermal comfort by integrating a thermal comfort comprehensive evaluation model in the calculation process. Meanwhile, a space-time graph model is applied to train a prediction model of space thermal autonomy, and the dynamic characteristics of the indoor environment in the time dimension and the space layout are accurately captured and analyzed in combination with the user activity law, so that the problems of large prediction error and large computing power demand of the current space thermal autonomy are solved. The implementation of the application can greatly improve the efficiency and accuracy of the prediction of the space thermal autonomy, provide technical support for optimizing the space thermal autonomy in the building design process, improve the passive design strategy of the building, enhance the thermal regulation capacity of the building itself, improve the living experience of the occupants, save energy and reduce carbon, and promote the development of the building design in a more healthy, green and efficient direction.

Description

Technical Field

[0001] This invention relates to the field of building technology, and in particular to a building health performance optimization design method based on spatial thermal autonomy. Background Technology

[0002] With societal development, the public's demands for quality of life are constantly increasing, and creating a healthy and comfortable indoor environment has become a core pursuit in the field of architectural design. The thermal comfort of the indoor environment is closely related to the comfort and overall satisfaction of residents, as well as the overall energy consumption of the building. Therefore, accurately predicting and optimizing the thermal comfort of buildings has become an indispensable part of current architectural design.

[0003] In traditional building design workflows, thermal comfort assessments are typically based on thermal comfort models with fixed formulas (such as predicted average vote PMV), combined with hourly calculations using simulated thermal environment data, and finally integrated into general annual indicators, such as the percentage of time spent outside the PMV range or the proportion of time spent outside the set operating temperature range. These traditional thermal comfort indicators often limit thermal comfort conditions to a narrow temperature range. However, this constant environmental model not only fails to effectively promote the occupants' own thermal balance regulation capabilities but also leads to frequent reliance on active control systems such as air conditioning, significantly exacerbating building energy consumption. Research shows that under naturally ventilated conditions, users, through their body's thermal regulation capabilities and thermal adaptation behaviors, can tolerate a wider comfort temperature range and greater temperature fluctuations compared to non-naturally ventilated rooms. Furthermore, natural ventilation conditions are beneficial for improving indoor environmental quality and occupants' health.

[0004] Against this backdrop, the concept and application of Spatial Thermal Autonomy (sTA) are particularly important. Spatial Thermal Autonomy represents the percentage of a building's area that passively meets or exceeds a given thermal comfort standard. During its calculation, thermal comfort evaluation methods can be selected and adjusted according to design requirements. Appropriate methods can fully consider residents' thermal regulation and adaptation capabilities, breaking through the narrow thermal comfort temperature range defined by traditional thermal comfort indices and expanding the acceptable temperature fluctuation range for residents. Furthermore, its calculation process no longer divides the building into single thermal zones, but rather selects appropriately sized grids to divide the space according to building usage requirements, assesses the thermal comfort of each grid, and thus captures the temporal and spatial changes of the indoor environment, addressing the current limitation in building design that insufficiently considers the uneven distribution of the thermal environment within indoor spaces.

[0005] However, in the current prediction of spatial thermal autonomy, traditional thermal comfort methods cannot comprehensively evaluate actual thermal comfort perception, resulting in a significant discrepancy with users' actual experiences. Furthermore, current calculations do not consider differences in human activity time, space, and thermal demands, leading to accumulated errors that result in substantial discrepancies between the final spatial thermal autonomy calculation and reality. In addition, spatial thermal autonomy calculations require year-round, hourly, gridded simulations of the building's spatial thermal environment, resulting in a massive computational burden. Performing a single spatial thermal autonomy assessment of a building consumes considerable computational resources and time, making its application in the optimization of building design schemes extremely difficult.

[0006] Therefore, in order to address the problems of large prediction errors, high computational requirements, and long time consumption in building space thermal autonomy prediction, it is urgent to develop a building space thermal autonomy prediction model that is oriented towards actual use scenarios and can comprehensively evaluate thermal comfort. This model can not only comprehensively consider the actual situation and fully take into account the thermal comfort needs of residents, broaden the range of thermal comfort for residents, and improve their overall health level, but also achieve accurate and efficient prediction of building space thermal autonomy. Summary of the Invention

[0007] This invention aims to propose a building health performance optimization design method based on spatial thermal autonomy. By incorporating a comprehensive thermal comfort evaluation model into the calculation process, a comprehensive assessment of user thermal comfort is achieved. Simultaneously, a spatiotemporal graph model is applied to train a spatial thermal autonomy (sTA) prediction model. Combined with user activity patterns, this method accurately captures and analyzes the dynamic characteristics of the indoor environment in terms of time and spatial layout, solving the problems of large prediction errors and high computational requirements in current spatial thermal autonomy prediction methods. The implementation of this invention can significantly improve the efficiency and accuracy of building spatial thermal autonomy prediction, providing technical support for optimizing spatial thermal autonomy during building design. This improves passive design strategies, enhances the building's own thermal regulation capabilities, improves the resident experience, and simultaneously saves energy and reduces carbon emissions, promoting the development of building design towards a healthier, greener, and more efficient direction.

[0008] This invention is achieved through the following technical solution: This invention proposes a building health performance optimization design method based on spatial thermal autonomy, the method comprising the following steps:

[0009] S1. Construction of a thermal comfort model combining multimodal deep learning and fuzzy neural networks;

[0010] Step S1 specifically involves:

[0011] S1.1: Collection and database construction of regional human physiological indicators based on high-precision physiological sensing technology;

[0012] S1.2: Construct a database integrating subjective and objective perceptions of regional people's emotions, cognitive states, and thermal sensations;

[0013] S1.3: Construction of a regional climate time series prediction model based on multi-sensor fusion data;

[0014] S1.4: Collection of indoor thermal environment parameters of typical buildings in the region;

[0015] S1.5: Construction of a regional thermal comfort model based on multimodal data and comprehensive regional population indicators;

[0016] S2. Construction of a building indoor thermal environment and heat demand prediction model based on spatiotemporal graph convolutional network;

[0017] Step S2 specifically involves:

[0018] S2.1: Construction of a dataset of spatiotemporal thermal environments of typical buildings in the region;

[0019] S2.2: Regional Population Behavior - Hot Demand Collection and Dataset Construction;

[0020] S2.3: Deep learning-driven multidimensional data feature extraction and preprocessing;

[0021] S2.4: Training of a building indoor spatiotemporal thermal environment-heat demand prediction model based on spatiotemporal graph convolutional network;

[0022] S3. Construction of a prediction model for the thermal autonomy of building space based on spatiotemporal Transformer;

[0023] Step S3 specifically involves:

[0024] S3.1: Construction of a dataset on the thermal autonomy of architectural spaces;

[0025] S3.2: Temporal and spatial codes are concatenated with input features;

[0026] S3.3: Training of a spatial thermal autonomy prediction model based on spatiotemporal Transformer;

[0027] S3.4: Performance verification of the space thermal autonomy prediction model.

[0028] Furthermore, in step S1.2, based on the data collected by the recording equipment and near-infrared brain functional imaging equipment, language feature recognition tools, time series analysis, frequency domain analysis, event correlation analysis and optical path correction algorithms are used to analyze the language expression and brain activity status of people in different seasons, usage scenarios, activity states and thermal environment conditions. Combined with the subjective thermal sensation, thermal comfort, related cognitive ability tests and emotional state of people under different thermal environment conditions collected by subjective questionnaire survey, a database of thermal sensation and related emotional cognitive state of people in the region is established.

[0029] Furthermore, in step S1.5, a multimodal dataset is constructed based on the indoor and outdoor thermal environment conditions, physiological indicators, subjective thermal sensations, emotions, and cognitive states collected in the previous steps. By applying a fuzzy logic system and defining fuzzy rules and membership functions, all thermal comfort-related indicators in the multimodal data are comprehensively analyzed to establish a comprehensive thermal comfort model for local residents, so as to achieve a comprehensive evaluation based on objective physiological and psychological health data, cognitive abilities, and subjective evaluation data.

[0030] Further, step S2.1 specifically includes:

[0031] S2.1.1: Acquisition of typical regional building design parameters based on lidar;

[0032] S2.1.2: Construction of building form and energy consumption model;

[0033] S2.1.3: Simulation of spatial thermal environment parameters based on high-performance computing (HPC) clusters.

[0034] Further, step S2.2 specifically includes:

[0035] S2.2.1: On-site collection of regional personnel behavior and physiological indicators;

[0036] S2.2.2: Regional personnel behavior-heat demand cluster analysis; specifically: cluster analysis of different behavior types of regional personnel, and determine the heat demand coefficient for different groups, different activity intensities and their physiological states and clothing thermal resistance by combining physiological indicators;

[0037] S2.2.3: Establishment of the spatiotemporal mapping relationship between regional personnel behavior and heat demand; specifically, mapping the time and space information of regional personnel behavior to heat demand coefficients to construct a spatiotemporal matrix of regional personnel behavior and heat demand.

[0038] Further, step S2.4 specifically involves: based on the dataset, applying a spatiotemporal graph convolutional network to construct a building indoor spatiotemporal thermal environment-thermal demand prediction model, transforming the building floor plan into a graph structure, with the grid as nodes to describe the topological structure of the space, and using the thermal demand, activities, and number of people at different times of each node as features for convolutional processing in the time dimension.

[0039] Further, step S3.1 specifically includes:

[0040] S3.1.1: Acquisition of building indoor spatiotemporal thermal environment - heat demand dataset;

[0041] S3.1.2: Prediction of building indoor spatial thermal comfort index;

[0042] S3.1.3: Calculation of thermal autonomy of building interior space.

[0043] Further, step S3.3 specifically involves: constructing a building interior space thermal autonomy prediction model based on the temporal-spatial distribution data of building interior space thermal environment parameters and personnel thermal demand, combined with a spatiotemporal Transformer model; using a multi-head attention mechanism, spatial attention is used to capture spatial heat conduction, temporal attention is used to model the impact of daily and seasonal changes on the thermal environment and personnel behavior, and for each time step, a spatial-temporal attention matrix is ​​used to capture the correlation of each spatiotemporal location.

[0044] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the building health performance optimization design method based on spatial thermal autonomy.

[0045] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the building health performance optimization design method based on spatial thermal autonomy.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0047] This invention provides a building health performance optimization design method based on spatial thermal autonomy. It aims to improve the prediction efficiency and accuracy of building spatial thermal autonomy, enabling iterative optimization of building spatial thermal autonomy during the design process. This expands the scope of passive control methods used in buildings, improving both thermal comfort and energy efficiency. Through improvements to the thermal comfort evaluation model, this invention can more comprehensively and accurately reflect the thermal comfort needs of residents in different regions, climates, and age groups, expanding the temperature range for building thermal comfort. By applying a spatiotemporal diagram model, it can more efficiently predict the building's thermal environment, heat demand distribution, and calculate building spatial thermal autonomy, allowing building designs to better identify, utilize, or improve thermal non-uniformity, thereby effectively reducing localized thermal discomfort and improving overall thermal comfort performance.

[0048] By applying the design method of this invention, architectural schemes can anticipate and optimize building performance under different thermal environments during the design phase. By promoting passive design strategies (such as improving building envelope, shading design, and rationalizing spatial layout), reliance on traditional active temperature control systems like air conditioning is reduced, effectively lowering energy consumption for building thermal environment regulation, optimizing building energy performance, and enhancing long-term adaptability and climate resilience. Furthermore, with the rapid development and widespread adoption of the Internet of Things and smart homes, this invention can further optimize architectural design schemes by combining the local thermal environment regulation advantages of personal comfort systems, driving buildings towards a more sustainable, intelligent, and adaptable direction. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0050] Figure 1 This is a flowchart of a building health performance optimization design method based on spatial thermal autonomy, as described in this invention. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Combination Figure 1This invention proposes a building health performance optimization design method based on spatial thermal autonomy, the method comprising the following steps:

[0053] S1. Construction of a thermal comfort model combining multimodal deep learning and fuzzy neural networks;

[0054] Step S1 specifically involves:

[0055] S1.1: Collection and database construction of regional human physiological indicators based on high-precision physiological sensing technology;

[0056] S1.2: Constructing a database integrating subjective and objective perceptions of regional personnel's emotions, cognitive states, and thermal sensations; In step S1.2, based on data collected by recording equipment and near-infrared brain functional imaging (fNIRS) equipment, language feature recognition tools, time series analysis, frequency domain analysis, event correlation analysis, and optical path correction algorithms are used to analyze the language expression and brain activity states of regional personnel under different seasons, usage scenarios, activity states, and thermal environment conditions. Combined with subjective questionnaire surveys that collect subjective thermal sensations, thermal comfort, related cognitive ability tests, and emotional states of personnel under different thermal environment conditions, a database of regional personnel's thermal sensations and related emotional and cognitive states is established.

[0057] S1.3: Construction of a regional climate time series prediction model based on multi-sensor fusion data;

[0058] S1.4: Collection of indoor thermal environment parameters of typical buildings in the region;

[0059] S1.5: Construction of a comprehensive thermal comfort model for regional personnel based on multimodal data; In step S1.5, a multimodal dataset is constructed based on the indoor and outdoor thermal environment conditions, physiological indicators, subjective thermal sensations, emotions, and cognitive states collected in the previous steps; A fuzzy logic system is applied to comprehensively analyze all thermal comfort-related indicators in the multimodal data by defining fuzzy rules and membership functions, and a comprehensive thermal comfort model for regional personnel is established to achieve a comprehensive evaluation based on objective physiological and psychological health data, cognitive abilities, and subjective evaluation data;

[0060] S2. Construction of a building indoor thermal environment and heat demand prediction model based on spatiotemporal graph convolutional network;

[0061] Step S2 specifically involves:

[0062] S2.1: Construction of a spatiotemporal thermal environment dataset of typical buildings in the region; Step S2.1 specifically includes:

[0063] S2.1.1: Acquisition of typical regional building design parameters based on lidar;

[0064] S2.1.2: Construction of building form and energy consumption model;

[0065] S2.1.3: Simulation of spatial thermal environment parameters based on high-performance computing (HPC) clusters;

[0066] S2.2: Regional Population Behavior - Hot Demand Collection and Dataset Construction; Step S2.2 specifically includes:

[0067] S2.2.1: On-site collection of regional personnel behavior and physiological indicators;

[0068] S2.2.2: Regional personnel behavior-heat demand cluster analysis; specifically: cluster analysis of different behavior types of regional personnel, and determine the heat demand coefficient for different groups, different activity intensities and their physiological states and clothing thermal resistance by combining physiological indicators;

[0069] S2.2.3: Establishment of the spatiotemporal mapping relationship between regional personnel behavior and heat demand; specifically, mapping the time and space information of regional personnel behavior to heat demand coefficients to construct a spatiotemporal matrix of regional personnel behavior and heat demand.

[0070] S2.3: Deep learning-driven multidimensional data feature extraction and preprocessing;

[0071] S2.4: Training a building indoor spatiotemporal thermal environment-heat demand prediction model based on spatiotemporal graph convolutional networks; Specifically, step S2.4 involves: on the dataset, applying spatiotemporal graph convolutional networks (ST-GCN) to construct a building indoor spatiotemporal thermal environment-heat demand prediction model, transforming the building floor plan into a graph structure, with grids as nodes to describe the topological structure of the space, and using the heat demand, activities, and number of people at different times of each node as features, and performing convolution processing in the time dimension;

[0072] S3. Construction of a prediction model for the thermal autonomy of building space based on the spatiotemporal Transformer (ST-Transformer);

[0073] Step S3 specifically involves:

[0074] S3.1: Construction of the thermal autonomy dataset for building spaces; Step S3.1 specifically includes:

[0075] S3.1.1: Acquisition of building indoor spatiotemporal thermal environment - heat demand dataset;

[0076] S3.1.2: Prediction of building indoor spatial thermal comfort index;

[0077] S3.1.3: Calculation of thermal autonomy of building interior spaces;

[0078] S3.2: Temporal and spatial codes are concatenated with input features;

[0079] S3.3: Training a spatial thermal autonomy prediction model based on a spatiotemporal Transformer (ST-Transformer); Step S3.3 specifically involves: constructing a spatial thermal autonomy prediction model based on the temporal-spatial distribution data of building indoor space thermal environment parameters and personnel thermal demand, combined with a spatiotemporal Transformer model; using a multi-head attention mechanism, spatial attention is used to capture spatial heat conduction, temporal attention is used to model the impact of daily and seasonal changes on the thermal environment and personnel behavior, and for each time step, a spatial-temporal attention matrix is ​​used to capture the correlation of each spatiotemporal location;

[0080] S3.4: Performance verification of the space thermal autonomy prediction model.

[0081] Example

[0082] The present invention will be described in detail below with reference to specific embodiments:

[0083] according to Figure 1 As shown, this invention provides a building health performance optimization design method based on spatial thermal autonomy, comprising the following steps:

[0084] Step 1: Construct a thermal comfort model combining multimodal deep learning and fuzzy neural networks;

[0085] Step 1 specifically involves:

[0086] Step 1.1: Collection and database construction of regional population physiological indicators based on high-precision physiological sensing technology;

[0087] Step 1.2: Construct a database integrating subjective and objective perceptions of regional people's emotions, cognitive states, and thermal sensations;

[0088] Step 1.3: Construction of a regional climate time-series prediction model based on multi-sensor fusion data;

[0089] Step 1.4: Collection of indoor thermal environment parameters for typical buildings in the region;

[0090] Step 1.5: Construction of a regional thermal comfort model based on multimodal data and comprehensive regional population indicators;

[0091] Step 1.1 specifically includes:

[0092] First, the climate zone of the building is determined. Climate zone division is based on the climate characteristics of the region, ensuring effective architectural design optimization for different climatic conditions. After determining the climate zone, representative subjects from each age group are selected for physiological data collection based on the age distribution of local residents. To ensure the representativeness and scientific rigor of the sampling, this invention employs random sampling, randomly selecting eligible subjects from each age group and inviting them to participate in data collection. This step ensures that the subject group is broadly representative in terms of age structure, accurately reflecting the differences in physiological responses to indoor thermal environments among different age groups.

[0093] During physiological data collection, portable physiological data acquisition devices are used for real-time monitoring. These devices continuously record multiple physiological parameters of the subjects, including but not limited to skin temperature, heart rate, skin conductivity, and respiratory rate. The collection of this physiological data covers all four seasons to comprehensively assess the impact of seasonal climate change on the human body, ensuring the comprehensiveness and representativeness of the data. Physiological data is monitored in real-time via sensors and subjected to time-series analysis to assess the immediate impact of the thermal environment on physiological responses. The collected physiological data will be used to establish a regional physiological database containing physiological response data for people of different ages and climate zones. Through long-term monitoring and data accumulation, a scientific basis can be provided for subsequent building design, helping to better optimize thermal comfort and energy efficiency.

[0094] Step 1.2 specifically includes:

[0095] In step S1.2, based on data collected by the recording device and near-infrared brain functional imaging (fNIRS) device, the facial expressions, language expressions, and brain activity of the subjects in different seasons, usage scenarios, and thermal environments are analyzed using language emotion recognition tools such as OpenSMILE, time series analysis, frequency domain analysis, event correlation analysis, and optical path correction algorithms to assess their emotional state and cognitive abilities. Specifically, the recording device is used to record the subjects' language expressions in thermal environments, and the OpenSMILE tool is used to perform emotion analysis on the speech signals to assess the subjects' language emotion changes; the near-infrared brain functional imaging device is used to monitor the subjects' brain activity in thermal environments, and combined with EEG analysis and cerebral blood flow change data, time series analysis and frequency domain analysis methods are used to assess the subjects' cognitive state and psychological load; at the same time, event correlation analysis technology is applied to analyze the subjects' response patterns to changes in the thermal environment to further confirm the relationship between their emotional fluctuations and cognitive changes.

[0096] In addition, data was collected through a combination of on-site surveys and online questionnaires to analyze participants' subjective thermal sensations, thermal comfort, related cognitive ability tests, and emotional surveys under different thermal environmental conditions. Specifically, a multi-dimensional questionnaire was designed and distributed, including participants' self-ratings of thermal sensations and thermal comfort under different thermal environments, their work efficiency evaluations, and emotional states, to assess their adaptability to the current environment. Combined with cognitive ability tests (such as reaction time, memory, and concentration), the impact of the thermal environment on participants' thinking and attention was further analyzed; simultaneously, participants' fatigue was investigated to assess the physiological and psychological fatigue that may result from prolonged exposure to specific thermal environments. Through the integration of these multiple data sources, a database of regional participants' thermal sensations and related emotional and cognitive states was ultimately constructed, providing scientific data support for subsequent personalized thermal comfort prediction models.

[0097] Step 1.3 specifically includes:

[0098] In step S1.3, data collected based on multi-sensor fusion technology is obtained using devices such as temperature sensors, humidity sensors, wind speed sensors, and radiation temperature sensors to collect outdoor thermal environment data for the region. Specifically, temperature sensors are used to collect real-time air temperature data, humidity sensors are used to monitor changes in air humidity, wind speed sensors are used to detect wind speed fluctuations, and radiation temperature sensors are used to obtain the distribution of radiant heat energy. These data collectively reflect the current thermal environment status of the region. By fusing data from these different types of sensors, and using multi-sensor data fusion algorithms for data preprocessing, noise filtering, and signal integration, high-precision environmental parameter data is ensured.

[0099] Next, a time-series data prediction algorithm is constructed using a combination of Autoregressive Integral Moving Average (ARIMA) model, Long Short-Term Memory (LSTM) network, and Convolutional Neural Network (CNN). Specifically, the ARIMA model is used to model the long-term trends of environmental data, identifying and fitting the long-term variation patterns of variables such as regional temperature, humidity, wind speed, and radiation; the LSTM network is used to capture long-term dependencies in time-series data, making it particularly suitable for predicting dynamic changes in climate data; and the CNN is used to extract local features from multi-dimensional sensor data, further optimizing the model's prediction accuracy. By combining these three methods, the constructed regional climate model can efficiently perform deep learning and analysis of multi-dimensional environmental variables, thereby predicting thermal environment changes under different seasons and climatic conditions.

[0100] This model analyzes the changing trends of environmental variables such as regional temperature, humidity, wind speed, and radiation. By combining ARIMA for modeling long-term trends, CNN for extracting local features, and LSTM for capturing long-term dependencies, it can accurately predict thermal environment changes over a future period. Ultimately, the model's predictions will be used to provide accurate regional climate forecasts, aiding in thermal environment control and optimization decisions.

[0101] Step 1.5 specifically involves:

[0102] In step S1.5, subjective thermal sensations, thermal comfort, cognitive ability tests, and emotional responses of individuals under different seasons and usage scenarios are first collected and combined with corresponding indoor and outdoor thermal environment data (temperature, humidity, wind speed, etc.). An integrated multi-input system is then used to standardize the data, ensuring that data from different dimensions (such as body temperature, emotion, and cognitive ability) are on the same scale. Next, fuzzy sets are defined based on the characteristics of the input variables, and a membership function is defined for each fuzzy set to represent the membership degree of each input variable. Then, fuzzy rules are constructed based on empirical data, and different fuzzy logics are applied to distinguish the differences in thermal demand under different activity conditions. Finally, the fuzzy results of the output variables are derived using the Mamdani inference method, resulting in the Comprehensive Comfort Index (CCI).

[0103] Step 2: Constructing a building indoor thermal environment-heat demand prediction model based on spatiotemporal graph convolutional networks;

[0104] Step 2 specifically involves:

[0105] Step 2.1: Construction of a spatiotemporal thermal environment dataset of typical buildings in the region;

[0106] Step 2.2: Regional Population Behavior - Hot Demand Collection and Dataset Construction;

[0107] Step 2.3: Deep learning-driven multidimensional data feature extraction and preprocessing;

[0108] Step 2.4: Training of the building indoor spatiotemporal thermal environment-heat demand prediction model based on spatiotemporal graph convolutional network.

[0109] Step 2.1 specifically involves:

[0110] Step 2.1.1: Obtaining typical regional building design parameters based on lidar;

[0111] Step 2.1.2: Building Form and Energy Consumption Model Construction;

[0112] Step 2.1.3: Simulation of spatial thermal environment parameters based on high-performance computing (HPC) clusters;

[0113] Step 2.1.1 specifically includes:

[0114] LiDAR technology is applied to obtain typical architectural design parameters in a region. The LiDAR equipment is used to scan and collect data on the building facade, spatial structure and surrounding environment. Design parameters such as the building's geometry, size, floor height and window position are extracted. Based on this, a dataset of typical architectural design parameters in the region is constructed. Data cleaning and standardization are then performed to ensure the accuracy and consistency of the architectural design parameters.

[0115] Step 2.2 specifically involves:

[0116] Step 2.2.1: On-site collection of behavioral and physiological indicators of local personnel;

[0117] Step 2.2.2: Regional personnel behavior-heat demand cluster analysis;

[0118] Step 2.2.3: Establishing the spatiotemporal mapping relationship between regional personnel behavior and heat demand;

[0119] Step 2.2.1 specifically involves:

[0120] Throughout the year, data on human behavior, clothing, and physiological indicators (respiratory rate and heart rate) of typical buildings in the region were collected on-site, and the data were labeled according to time and space.

[0121] Step 2.2.2 specifically involves:

[0122] Based on respiratory rate and heart rate, the metabolic rate under different behaviors and clothing conditions is calculated and clustered to correspond different behaviors and clothing conditions to a heat demand index.

[0123] Step 2.2.3 specifically involves:

[0124] By mapping the temporal and spatial information of the behavior of local people to the heat demand coefficient, a spatiotemporal matrix of local people's behavior and heat demand is constructed.

[0125] Step 2.3 specifically involves: based on the type of input data, applying techniques such as standardization, time series analysis, and convolutional neural networks to construct preprocessing schemes for different types of data, obtaining optimized architectural design parameters and architectural floor plan data. Specifically, for architectural design parameters, standardization is applied combined with feature selection methods to remove low-relevance or redundant features, resulting in effective design parameter data. For architectural floor plans, the floor plans are converted into image data, and feature extraction is performed using a convolutional neural network (CNN). The CNN model automatically extracts spatial features from the floor plans, such as the shape, size, and location of different areas like rooms, corridors, and windows. If the architectural floor plans are complex or the image data is limited, an image segmentation model can be used to segment different areas, clearly defining the boundaries of each spatial area of ​​the building.

[0126] Step 2.4 specifically involves: based on the dataset, applying Spatio-temporal Graph Convolutional Networks (ST-GCN) to construct a building indoor spatio-temporal thermal environment-thermal demand prediction model. The building floor plan is transformed into a graph structure, with the grid as nodes to describe the topological structure of the space. At the same time, the thermal demand, activities, and number of people at each node at different times are used as features and convolutional processing is performed in the time dimension.

[0127] Step 3: Construct a prediction model for the thermal autonomy of building space based on spatiotemporal Transformer;

[0128] Step 3 specifically involves:

[0129] Step 3.1: Construction of the thermal autonomy dataset for architectural spaces;

[0130] Step 3.2: Concatenate temporal and spatial codes with input features;

[0131] Step 3.3: Training a spatial thermal autonomy prediction model based on the spatiotemporal Transformer (ST-Transformer);

[0132] Step 3.4: Performance verification of the space thermal autonomy prediction model.

[0133] Step 3.1 specifically involves:

[0134] Step 3.1.1: Obtaining the building's indoor spatiotemporal thermal environment - heat demand dataset;

[0135] Step 3.1.2: Prediction of the spatial-temporal thermal comfort index of building interiors;

[0136] Step 3.1.3: Calculation of thermal autonomy of building interior space.

[0137] Step 3.1.1 specifically includes:

[0138] Substituting typical building design parameters of different functional types in the region into the building indoor spatiotemporal thermal environment-heat demand prediction model obtained in step 2.4, the spatiotemporal distribution matrix of building indoor thermal environment-heat demand is obtained.

[0139] Step 3.1.2 specifically includes:

[0140] Substituting the characteristics of thermal environment parameters and thermal demand coefficients from the building indoor thermal environment-thermal demand spatiotemporal distribution matrix into the comprehensive thermal comfort model in step 1.5, we obtain the building indoor thermal comfort spatiotemporal distribution matrix.

[0141] Step 3.1.3 specifically involves:

[0142] Spatial thermal autonomy is calculated based on hourly thermal comfort conditions at grid points. Spatial thermal autonomy is defined as "the percentage of building area where a thermal zone achieves or exceeds a given thermal comfort standard solely through passive means," and the hourly spatial thermal autonomy index (sTA) is also calculated. t The calculation formula is as follows:

[0143]

[0144] Among them, A i It is the local area of ​​the space represented by grid point i, and the unit is m. 2 ;1 {i∈comfort} It is an indicator function; it equals 1 if grid point i meets the comfort criterion, and 0 otherwise. total It is the total area of ​​the space, and the unit is m. 2 .

[0145] Next, the annual space thermal autonomy (sTA) is calculated. annual The calculation formula is:

[0146]

[0147] Where, τ comfort It is the hourly spatial comfort threshold, representing the minimum percentage of the area (or grid point i) that a given area must meet the defined comfort criteria to be considered thermally autonomous at hour t. It is 1, if in hour t, the hourly spatial thermal autonomy sTA t Greater than or equal to τ comfort Otherwise, it is 0; T year It is the total number of hours in a year (generally 8760 hours in non-leap years).

[0148] Step 3.2 specifically involves:

[0149] Spatiotemporal information is introduced by location encoding the model input dataset (thermal environment parameters, heat demand coefficient, etc.), and sine and cosine encoding is used to represent time to ensure that the model can recognize the order of time. Spatial coordinates are used to assign a location code to each spatial point.

[0150] Step 3.3 specifically involves:

[0151] Based on the temporal-spatial distribution data of building interior space thermal environment parameters and occupants' thermal demands, and combined with a spatiotemporal Transformer model, a predictive model for building interior space thermal autonomy is constructed. First, a Transformer architecture is designed, employing a multi-head attention mechanism. Spatial attention is used to capture spatial heat conduction, while temporal attention is used to model the impact of daily and seasonal variations on the thermal environment and occupants' behavior. For each time step, a spatial-temporal attention matrix is ​​used to capture the correlation between each spatiotemporal location. Finally, a fully connected layer is applied to map the model's output to the prediction target—spatial thermal autonomy.

[0152] Step 3.4 specifically involves:

[0153] During training, mean squared error (MSE) loss is used to measure the difference between the model's predicted annual thermal comfort ratio and the actual value. The Adam optimizer is used to train the model, and hyperparameters such as the learning rate are continuously adjusted to improve the model's prediction accuracy.

[0154] The present invention also proposes an electronic device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the building health performance optimization design method based on spatial thermal autonomy.

[0155] The present invention also proposes a computer-readable storage medium for storing computer instructions, which, when executed by a processor, implement the steps of the building health performance optimization design method based on spatial thermal autonomy.

[0156] The memory in this application embodiment can be volatile memory or non-volatile memory, or it can include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous linked dynamic random access memory (SLDRAM), and direct rambus RAM (DRRAM). It should be noted that the memory used in the methods described in this invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0157] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions. When the computer instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device such as a server or data center that integrates one or more available media. The available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., high-density digital video discs (DVDs)), or semiconductor media (e.g., solid-state disks (SSDs)).

[0158] In implementation, each step of the above method can be completed by integrated logic circuits in the processor's hardware or by instructions in software. The steps of the method disclosed in the embodiments of this application can be directly implemented by a hardware processor, or by a combination of hardware and software modules in the processor. The software modules can reside in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads information from the memory and, in conjunction with its hardware, completes the steps of the above method. To avoid repetition, detailed descriptions are omitted here.

[0159] It should be noted that the processor in the embodiments of this application can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method embodiments can be completed by the integrated logic circuitry in the processor's hardware or by instructions in software form. The processor can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly embodied as being executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor. The software modules can be located in random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, registers, or other mature storage media in the art. This storage medium is located in memory, and the processor reads the information in the memory and, in conjunction with its hardware, completes the steps of the above methods.

[0160] The above provides a detailed description of the building health performance optimization design method based on spatial thermal autonomy proposed in this invention. Specific examples have been used to illustrate the principles and implementation methods of this invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of this invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this invention. Therefore, the content of this specification should not be construed as a limitation of this invention.

Claims

1. A building health performance optimization design method based on spatial thermal autonomy, characterized in that, The method includes the following steps: S1. Construction of a thermal comfort model combining multimodal deep learning and fuzzy neural networks; Step S1 specifically involves: S1.1: Collection and database construction of regional human physiological indicators based on high-precision physiological sensing technology; S1.2: Construct a database integrating subjective and objective perceptions of regional people's emotions, cognitive states, and thermal sensations; S1.3: Construction of a regional climate time series prediction model based on multi-sensor fusion data; S1.4: Collection of indoor thermal environment parameters of typical buildings in the region; S1.5: Construction of a regional thermal comfort model based on multimodal data and comprehensive regional population indicators; S2. Construction of a building indoor thermal environment and heat demand prediction model based on spatiotemporal graph convolutional network; Step S2 specifically involves: S2.1: Construction of a dataset of spatiotemporal thermal environments of typical buildings in the region; S2.2: Regional Population Behavior - Hot Demand Collection and Dataset Construction; S2.3: Deep learning-driven multidimensional data feature extraction and preprocessing; S2.4: Training of a building indoor spatiotemporal thermal environment-heat demand prediction model based on spatiotemporal graph convolutional network; S3. Construction of a prediction model for the thermal autonomy of building space based on spatiotemporal Transformer; Step S3 specifically involves: S3.1: Construction of a dataset on the thermal autonomy of architectural spaces; Step S3.1 specifically includes: S3.1.1: Acquisition of building indoor spatiotemporal thermal environment - heat demand dataset; S3.1.2: Prediction of building indoor spatial thermal comfort index; S3.1.3: Calculation of thermal autonomy of building interior spaces; S3.2: Temporal and spatial codes are concatenated with input features; S3.3: Training of a spatial thermal autonomy prediction model based on spatiotemporal Transformer; S3.4: Performance verification of the space thermal autonomy prediction model.

2. The method according to claim 1, characterized in that, In step S1.2, based on the data collected by the recording equipment and near-infrared brain functional imaging equipment, the language expression and brain activity of the local people under different seasons, usage scenarios, activity states and thermal environment conditions are analyzed by language feature recognition tools, time series analysis, frequency domain analysis, event correlation analysis and optical path correction algorithms. Combined with the subjective thermal sensation, thermal comfort, related cognitive ability test and emotional state of the people under different thermal environment conditions collected by the subjective questionnaire survey, a database of the thermal sensation and related emotional cognitive state of the local people is established.

3. The method according to claim 1, characterized in that, In step S1.5, a multimodal dataset is constructed based on the indoor and outdoor thermal environment conditions, physiological indicators, subjective thermal sensations, emotions, and cognitive states collected in the previous steps. By applying a fuzzy logic system and defining fuzzy rules and membership functions, all thermal comfort-related indicators in the multimodal data are comprehensively analyzed to establish a comprehensive thermal comfort model for local residents, so as to achieve a comprehensive evaluation based on objective physiological and psychological health data, cognitive abilities, and subjective evaluation data.

4. The method according to claim 1, characterized in that, Step S2.1 specifically includes: S2.1.1: Acquisition of typical regional building design parameters based on lidar; S2.1.2: Construction of building form and energy consumption model; S2.1.3: Simulation of spatial thermal environment parameters based on high-performance computing (HPC) clusters.

5. The method according to claim 1, characterized in that, Step S2.2 specifically includes: S2.2.1: On-site collection of regional personnel behavior and physiological indicators; S2.2.2: Regional personnel behavior-heat demand cluster analysis; specifically: cluster analysis of different behavior types of regional personnel, and determine the heat demand coefficient for different groups, different activity intensities and their physiological states and clothing thermal resistance by combining physiological indicators; S2.2.3: Establishment of the spatiotemporal mapping relationship between regional personnel behavior and heat demand; specifically, mapping the time and space information of regional personnel behavior to heat demand coefficients to construct a spatiotemporal matrix of regional personnel behavior and heat demand.

6. The method according to claim 1, characterized in that, Specifically, step S2.4 involves: based on the dataset, constructing a building indoor spatiotemporal thermal environment-heat demand prediction model using a spatiotemporal graph convolutional network, transforming the building floor plan into a graph structure, with the grid as nodes to describe the topological structure of the space, and using the heat demand, activities, and number of people at different times of each node as features for convolution processing in the time dimension.

7. The method according to claim 1, characterized in that, Step S3.3 specifically involves: constructing a building interior space thermal autonomy prediction model based on the temporal-spatial distribution data of building interior space thermal environment parameters and personnel thermal demand, combined with a spatiotemporal Transformer model; using a multi-head attention mechanism, spatial attention is used to capture spatial heat conduction, temporal attention is used to model the impact of daily and seasonal changes on the thermal environment and personnel behavior, and for each time step, a spatial-temporal attention matrix is ​​used to capture the correlation of each spatiotemporal location.

8. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1-7.

9. A computer-readable storage medium for storing computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the steps of the method according to any one of claims 1-7.

Images