Zero-carbon building design method based on AI optimization

Abstract

The invention belongs to the technical field of zero-carbon buildings, and particularly relates to a zero-carbon building design method based on AI optimization. According to the AI optimization-based zero-carbon building design method provided by the invention, meteorological data of a region where a target building is located is systematically collected and analyzed, a detailed building model and environment model database is constructed, real-time monitoring data of internal and external environments of the building is collected, and key energy consumption indexes of the building during operation are recorded; and a machine learning model is trained by using the comprehensive data, so that the energy efficiency is improved and the carbon emission is reduced in the building design process. According to the method, not only are the characteristics of the building considered, but also the influence of the external environment and the actual operation condition is fully considered, a more scientific reference basis is provided for the building design stage, and environment protection and energy conservation of the building are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of zero-carbon buildings, and specifically relates to a zero-carbon building design method based on AI optimization. Background Art

[0003] The traditional architectural design method mainly relies on the experience and expertise of designers, as well as the standard specifications of the construction industry. This method is largely based on qualitative evaluation rather than quantitative analysis. At present, the traditional architectural design process has limitations in the following aspects.

[0004] The first aspect is the lack of precise environmental adaptability. Traditional architectural design often fails to fully utilize the environmental characteristics of the project location, such as climate conditions and terrain characteristics, to optimize the energy performance of the building. This makes it difficult for buildings to effectively utilize natural resources (such as solar energy, natural ventilation, etc.) during actual use, thereby increasing energy consumption and carbon emissions.

[0005] The second aspect is that design decisions rely more on subjectivity. In the traditional design process, many design decisions are based on the designer's personal experience and preferences rather than data-driven analysis. This subjective design approach may result in the design failing to meet the expected performance standards in actual applications, especially in terms of energy efficiency.

[0006] The third aspect is the limited use of data. Although modern architectural design has introduced computer-aided design (CAD) tools and other digital means, it is still insufficient in data collection and analysis. Traditional design methods often ignore the collection and analysis of actual operating data of buildings, resulting in a lack of feedback mechanism in the design process, which affects the overall performance of the construction project and the realization of sustainable development goals. Summary of the invention

[0007] In view of the shortcomings of the prior art, the present invention provides a zero-carbon building design method based on AI optimization.

[0008] The zero-carbon building design method based on AI optimization provided by the present invention systematically collects and analyzes meteorological data in the area where the target building is located, constructs a detailed building model and environmental model database, collects real-time monitoring data of the internal and external environment of the building, records key energy consumption indicators of the building during operation, and uses these comprehensive data to train machine learning models, thereby achieving improved energy efficiency and reduced carbon emissions in the building design process.

[0009] The zero-carbon building design method based on AI optimization provided by the present invention comprises the following steps: Step a. Collecting meteorological data of the area where the target building is located; Step b. constructing a building model database including geometric information, material properties and facility configuration; Step c. constructing an environmental model database covering terrain, vegetation and adjacent building features; Step d collects real-time monitoring data of the internal and external environment of the target building; Step e. Record the energy consumption indicators of the building during operation; Step f. Training a machine learning model using the data obtained in the above steps (a)-(e); Step g. Predict energy consumption indicators based on the trained model to optimize the design of zero-carbon buildings.

[0010] As a preferred solution for the zero-carbon building design method based on AI optimization, in step a, the meteorological data includes data on the temperature field of the area where the building is located changing over time, data on the humidity field changing over time, data on the wind speed field changing over time, data on the solar radiation intensity changing over time, and data on the solar radiation angle changing over time.

[0011] As a preferred solution for the zero-carbon building design method based on AI optimization, in step b, the building geometry information includes at least the building's floor plan, floor height, wall thickness, roof height, and the location and size of doors and windows; the building material properties include at least the thermal conductivity, density, and specific heat capacity of the materials used in each part; the building facility configuration includes at least the location and working mode of the shading system, ventilation equipment, heating equipment, and cooling equipment in the building.

[0012] As a preferred solution of the zero-carbon building design method based on AI optimization, in step c, the terrain data includes at least altitude, slope, and slope direction; the vegetation data includes at least vegetation type and distribution density information; and the neighboring building data includes at least height, position, and shape data of neighboring buildings.

[0013] As a preferred solution of the zero-carbon building design method based on AI optimization, in step c, it also includes obtaining enhanced environmental data, using the model data of the building and adjacent buildings and the terrain data to create a virtual model of the building and its surrounding environment in a three-dimensional modeling software, and simulating the relationship data of the building's sunlight path and shadow area changing over time as the enhanced environmental data.

[0014] As a preferred solution of the zero-carbon building design method based on AI optimization, in step d, real-time monitoring data of the building is collected, and the real-time monitoring data includes at least temperature, humidity and light data; the real-time monitoring data is obtained by installing sensors in the internal and external areas of the building, and the sensors include temperature sensors, humidity sensors and light sensors.

[0015] As a preferred solution for the zero-carbon building design method based on AI optimization, in step e, building operation data is collected, and the operation data at least includes electricity consumption and gas consumption; the electricity consumption and gas consumption are recorded by meters distributed at different locations of the building, and stored according to location coordinates and timestamps; the electricity consumption and gas consumption data are recorded according to their respective location coordinates and corresponding time series; the electricity consumption and gas consumption data are used to calculate the total energy consumption at each time point.

[0016] As a preferred solution for the zero-carbon building design method based on AI optimization, in step f, the collected data is preprocessed, and the data at least includes collected meteorological, building models, environmental models, real-time monitoring data and building operation data; the preprocessing includes loading data in CSV format, removing rows containing missing values, converting time series data into the form of supervised learning problems, and standardizing all features using MinMaxScaler.

[0017] As a preferred solution for the zero-carbon building design method based on AI optimization, in step f, it includes building a machine learning model, which uses a long short-term memory network to predict energy demand in a future time period; the model construction includes creating a sequential model, adding an LSTM layer to the model to learn time dependency, and adding a fully connected layer to output total energy consumption; the model training is carried out using training data, the training rounds are set to multiple iterations, and the model is predicted and evaluated, and the evaluation uses the root mean square error to measure the performance of the model.

[0018] As a preferred scheme for the zero-carbon building design method based on AI optimization, in step g, the optimization process includes inputting different design schemes into a trained model, wherein the design schemes at least involve the selection of building materials, the improvement of building envelope structures and the configuration adjustment of energy systems; the model predicts the corresponding energy consumption according to the input design parameters; and by comparing and analyzing the prediction results of different design schemes, the design options that can maximize energy efficiency are identified.

[0019] Beneficial Effects The present invention provides a zero-carbon building design method based on AI optimization, which helps to improve the scientific and intelligent level of building design. By systematically collecting and analyzing meteorological data in the area where the target building is located, building a detailed building model and environmental model database, collecting real-time monitoring data of the internal and external environment of the building, and recording key energy consumption indicators during operation, this method greatly enhances the objectivity and scientificity of design decisions, reduces the subjectivity in traditional design, and makes the design scheme more in line with actual needs and meets the expected standards in terms of energy efficiency. In addition, the machine learning model trained with these comprehensive data can predict future energy needs, thereby allowing energy efficiency to be optimized in the early stages of design, identifying and selecting design schemes that can maximize energy efficiency, thereby reducing the overall carbon emissions of the building and supporting the design concept of sustainable development.

[0020] The method of the present invention improves the accurate grasp of environmental adaptability during the building design stage, and enhances the utilization rate of data through the feedback mechanism of real-time monitoring data and energy consumption indicators during operation, so that designers can continuously adjust and improve the design plan in actual operation. More importantly, this method can predict the energy consumption of different plans in the early stage of design, which helps to plan energy supply in advance, avoid energy waste, and discover potential energy saving opportunities, thereby reducing the carbon footprint of the building before it is put into use, and promoting the realization and development of green zero-carbon buildings. DETAILED DESCRIPTION

[0021] The present invention is further illustrated by specific examples below. These examples are exemplary and are intended to illustrate the problem and explain the present invention, but are not intended to be limiting.

[0022] The present invention provides a zero-carbon building design method based on artificial intelligence optimization, which realizes the improvement of energy efficiency and the reduction of carbon emissions in the building design process through systematic data collection and analysis combined with machine learning technology. By collecting comprehensive meteorological data of the area where the target building is located, constructing a building model database including geometric information, material properties and facility configuration, establishing an environmental model database covering terrain, vegetation and neighboring building characteristics, collecting real-time monitoring data of the internal and external environment of the building, recording the key energy consumption indicators of the building during operation, and using the above data to train the machine learning model, the purpose of accurately predicting and optimizing the building design can be achieved, providing a reference basis for the building design stage, and promoting the development of environmentally friendly and energy-saving buildings.

[0023] A zero-carbon building design method based on AI optimization includes the following steps: Step S1: Collect meteorological data Collect meteorological data of the area where the existing buildings are located, including temperature, humidity, wind speed, and solar radiation intensity. By collecting meteorological data, obtain the temperature field data T(x, y, z, t) of the area where the building is located, the humidity field data H(x,y, z, t), the wind speed field data V_x (x, y, z, t), V_y (x, y,z, t), V_z (x, y, z, t), the solar radiation intensity data R(t), and the solar radiation angle θ_1(t) and θ_2(t) data.

[0024] Step S2: Constructing a building model database Collect building geometry information: Collect geometric parameters of the building's floor plan, floor height, wall thickness, roof height, and the location and size of doors and windows. The floor plan is represented by a three-dimensional array, where the first dimension represents the floor number and the last two dimensions represent the floor plan of each floor. For each floor, a two-dimensional array is used to represent the layout of the floor, and different rooms or functional areas are assigned different identifiers. Specifically, it is defined as layout[layer][i][j], where layer represents the floor number, and layout[layer] is a two-dimensional array, where the value at the (i,j) position represents the space type or room identifier at that position. The boundary definition of a room or functional area is represented by a series of coordinate points, and each room or functional area is defined by a list of points, which are connected in sequence to form a closed polygon. Use a dictionary to store the room boundary information of each layer, such as room_boundaries[layer]= {room_id: [(x1, y1), (x2,y2), ..., (xn, yn)]}, where layer represents the floor number, room_id is the unique identifier of the room, and (xi,yi) are the point coordinates on the boundary. This representation can adapt to the design differences of different floors, and the shape of the room or functional area can be accurately described by point coordinates, making the model both flexible and accurate. Wall thickness is represented by a dictionary, where the key is the wall number or position, and the value is its thickness, represented as wall_thickness= {'wall1': t1, 'wall2': t2, ...}. The roof height is a numerical value roof_height; the position and size of doors and windows are represented by a list, each element is a tuple (position, size), where position is the position coordinate of the door and window (x, y), and size is the size of the door and window (width, height).

[0025] Collect building material properties: Enter the material properties used in the building walls, roofs, floors, etc., including thermal conductivity, density, and specific heat capacity. Specifically, material properties are represented by a dictionary, where the key is the material name and the value is a tuple containing these properties. The expression is as follows: material_properties = {'material1':(conductivity1, density1, specific_heat1),'material2': (conductivity2,density2, specific_heat2), ...}.

[0026] Collect the so-called facility configuration information of the building: including determining the location and working mode of the shading system, ventilation equipment, heating and cooling equipment in the building. Specifically, the facility configuration information is represented by a nested dictionary, where the outer key is the device type, such as shading, ventilation, etc., the inner key is the specific device location, and the value is the device's working mode or other related parameters. The expression is as follows: facility_configurations = {'shading_system': {'device1':(location1, mode1), 'device2': (location2,mode2)}, 'ventilation': {'device1':(location1, mode1), 'device2': (location2, mode2)}, ...}.

[0027] Step S3: Constructing an environmental model database of a building Collect terrain data: Get information about the terrain and geographic features of the building, including altitude, slope, and aspect. Terrain data is represented by a multidimensional array, where the first two dimensions represent the X and Y coordinates of the space, the third dimension represents the altitude Z, and additional dimensions are added to represent the slope and aspect. Defined as: terrain_data = [[[altitude(x, y), slope(x, y), aspect(x, y)]for y in range(y_max)]for x inrange(x_max)]. Here altitude(x, y) represents the altitude at the coordinate (x, y), slope(x, y) represents the slope at the coordinate (x, y) (which can be expressed in percentage or radians), aspect(x, y) represents the slope at the coordinate (x, y) (in degrees, 0° points to due north and increases clockwise), and x_max and y_max represent the maximum X and Y coordinates of the terrain grid, respectively.

[0028] Collect vegetation data: Get information about the type and density of vegetation around buildings. Vegetation data is represented by a two-dimensional array, and each element contains information about the type and density of vegetation. Define the type of vegetation as an integer code, and the density is directly represented by a floating point number. Defined as: vegetation_data = [[(species_code(x, y),density(x, y)) for y in range(y_max)]for x in range(x_max)]. Here species_code(x, y) represents the code of the vegetation type at the coordinate (x, y), and density(x, y) represents the density of vegetation at the coordinate (x, y).

[0029] Collect neighboring building data: Get the height, position, and shape data of neighboring buildings. The neighboring building data is represented by a list, and each element is a tuple containing the position, height, and shape information of the neighboring building. The position is represented by the center point coordinates, the height is a numerical value, and the shape is described by a set of coordinate points. Defined as: adjacent_buildings = [((center_x, center_y), height, [(x1, y1), (x2, y2), ..., (xn, yn)])]. Here (center_x, center_y) represents the center coordinates of the building, height represents the height of the building, and [(x1, y1), (x2,y2), ..., (xn, yn)] represents the vertex coordinate set of the building outline.

[0030] Obtain enhanced environmental data: Using the model data of the building and its neighboring buildings and terrain data, create a virtual model of the building and its surrounding environment in the 3D modeling software, and simulate the relationship data of the building's sunlight path and shadow area over time. Defined as: sun_path = [sun_radiation_angles(t) for t in time_intervals], shadow_regions = [calculate_shadow_region(t) for t in time_intervals]. Here sun_radiation_angles(t) represents the solar radiation angle at time t, time_intervals represents a series of time intervals, and calculate_shadow_region(t) represents the calculation of the shadow area at time t (returning a list containing the coordinates of the shadow area).

[0031] Step S4: Collect real-time monitoring data of buildings Collect real-time monitoring data of buildings: Install sensors in existing zero-carbon buildings, covering the entire interior and exterior areas of the building, including: temperature sensors, humidity sensors, and light sensors. Deploy an IoT platform to integrate all sensor data to realize real-time data collection and storage. Through the temperature sensor, obtain the data of the temperature field of the interior and exterior surfaces of the building over time, and obtain the data of the humidity field of the interior and exterior surfaces of the building over time. For each type of sensor data, define a four-dimensional array to represent their readings over time. Specifically, the reading of a specific sensor at a certain point in time is represented as: temperature sensor reading: T(x_s, y_s, z_s, t_s), where (x_s, y_s, z_s) is the location coordinate of the sensor and t_s is the timestamp of the reading; humidity sensor reading: H(x_s, y_s, z_s, t_s); light sensor reading: I(x_s, y_s, z_s, t_s). Stored as: sensor_readings[t] = { 'temperature': T(x, y, z, t), 'humidity': H(x, y, z, t), 'illumination': I(x,y, z, t)}.

[0032] Step S5: Collect building operation data Collect records of building operation data, including records of electricity consumption and gas consumption. Electricity consumption is represented as an array containing location coordinates and timestamps. If there are multiple electricity meters distributed in different locations of the building, define electricity consumption as: electricity_consumption = { (x1, y1, z1): [E1(t1), E1(t2), ..., E1(tn)], (x2, y2, z2): [E2(t1), E2(t2), ..., E2(tn)], ...}. Here (x1, y1, z1) are the location coordinates of the electricity meter, and E1(t1), E1(t2), ..., E1(tn) are the electricity consumption at each time point t1, t2, ..., tn at that location. Gas consumption is also expressed in a similar way: gas_consumption = { (x1, y1, z1): [G1(t1),G1(t2), ..., G1(tn)], (x2, y2, z2): [G2(t1), G2(t2), ..., G2(tn)], ...}. (x1, y1, z1) are the location coordinates of the gas meter, and G1(t1), G1(t2), ..., G1(tn) are the gas consumption at each time point t1, t2, ..., tn. According to the energy summary of the consumed electricity and gas data (uniformly converted to kilowatt-hours), the total energy consumption data at each time point is obtained: TEC(t1), TEC (t2), ..., TEC (tn).

[0033] Step S6: Machine learning model training A time series forecasting model, Long Short-Term Memory Network (LSTM), is used to predict energy demand in future time periods.

[0034] Data preprocessing Load the CSV format data collected and sorted from step S1 to step S5; remove the rows containing missing values; convert the time series data into the form of a supervised learning problem, that is, use the data of multiple consecutive time points as input features to predict the value of the next time point; use MinMaxScaler to scale all features to between 0 and 1 for data standardization to speed up the convergence of the model.

[0035] Constructing input and output data Construct the input feature matrix from the data collected in steps S1 to S4. Each row represents a sample and each column represents a feature. Construct the target vector from the total energy consumption data collected in step S5. Each sample corresponds to the total energy consumption in a certain period in the future.

[0036] Building the LSTM model Create a Sequential Model; add an LSTM layer to the model, which will receive input data and learn temporal dependencies from it; add a fully connected layer (Dense Layer) with only one neuron to output a single value, namely the total energy consumption. Define the loss function as Mean Squared Error (MSE) and choose Adam as the optimizer.

[0037] Training the model Use the training data to train the model, set the training rounds (epochs) to 100 times, and the batch size (batch size) to 1 so that the model can be trained data by data.

[0038] Prediction and evaluation Use the trained model to predict the data of the training set and the test set. Restore the model prediction results and the actual results from the standardized state back to the scale of the original data. Use the root mean square error (RMSE) to evaluate the performance of the model on the training set and the test set.

[0039] Step S7: Optimizing the design of zero-carbon buildings based on the trained model Different design options are input into the trained model. These options may involve the selection of building materials, improvement of building envelopes, configuration adjustment of energy systems, etc. The model predicts the corresponding energy consumption based on the input design parameters. By comparing and analyzing the prediction results of different design options, the design options that maximize energy efficiency are identified.

[0040] The above embodiments are exemplary, and their purpose is to illustrate the technical concept and features of the present invention so that people familiar with the technology in this field can understand the content of the present invention and implement it accordingly, and they cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A zero-carbon building design method based on AI optimization, characterized by: The following steps are involved: Step a. Collecting meteorological data of the area where the target building is located; Step b. constructing a building model database including geometric information, material properties and facility configuration; Step c. constructing an environmental model database covering terrain, vegetation and adjacent building features; Step d collects real-time monitoring data of the internal and external environment of the target building; Step e. Record the energy consumption indicators of the building during operation; Step f. Training a machine learning model using the data obtained in step ae above; Step g. Predict energy consumption indicators based on the trained model to optimize the design of zero-carbon buildings.

2. The zero-carbon building design method based on AI optimization according to claim 1 is characterized by: In step a, the meteorological data includes the data of the temperature field changing with time in the area where the building is located, the data of the humidity field changing with time, the data of the wind speed field changing with time, the data of the solar radiation intensity changing with time, and the data of the solar radiation angle changing with time.

3. The zero-carbon building design method based on AI optimization according to claim 1 is characterized in that: In step b, the building geometry information includes at least the building's floor plan, floor height, wall thickness, roof height, and the locations and sizes of doors and windows; the building material properties include at least the thermal conductivity, density, and specific heat capacity of the materials used in each part; and the building facility configuration includes at least the location and working mode of the shading system, ventilation equipment, heating equipment, and cooling equipment in the building.

4. The zero-carbon building design method based on AI optimization according to claim 1 is characterized in that: In step c, the terrain data at least includes altitude, slope, and slope direction; the vegetation data at least includes vegetation type and distribution density information; and the neighboring building data at least includes height, position, and shape data of neighboring buildings.

5. The zero-carbon building design method based on AI optimization according to claim 4 is characterized in that: In step c, it also includes obtaining enhanced environmental data, using the model data of the building and adjacent buildings and the terrain data to create a virtual model of the building and its surrounding environment in a three-dimensional modeling software, and simulating the relationship data of the building's sunlight path and shadow area changing over time as the enhanced environmental data.

6. The zero-carbon building design method based on AI optimization according to claim 1 is characterized in that: In step d, real-time monitoring data of the building is collected, and the real-time monitoring data includes at least temperature, humidity and light data; the real-time monitoring data is obtained by installing sensors in the interior and exterior areas of the building, and the sensors include temperature sensors, humidity sensors and light sensors.

7. The zero-carbon building design method based on AI optimization according to claim 1 is characterized in that: In step e, building operation data is collected, and the operation data includes at least electricity consumption and gas consumption; the electricity consumption and gas consumption are recorded by meters distributed in different locations of the building, and stored according to location coordinates and timestamps; the electricity consumption and gas consumption data are recorded according to their respective location coordinates and corresponding time series; the electricity consumption and gas consumption data are used to calculate the total energy consumption at each time point.

8. The zero-carbon building design method based on AI optimization according to claim 1 is characterized by: In step f, the collected data is preprocessed, and the data at least includes the collected meteorological data, building model, environmental model, real-time monitoring data and building operation data; the preprocessing includes loading data in CSV format, removing rows containing missing values, converting time series data into the form of supervised learning problems, and using MinMaxScaler to standardize all features.

9. The zero-carbon building design method based on AI optimization according to claim 8 is characterized in that: In step f, it includes building a machine learning model, which uses a long short-term memory network to predict energy demand in a future time period; the model building includes creating a sequential model, adding an LSTM layer to the model to learn time dependency, and adding a fully connected layer to output total energy consumption; the model training is performed using training data, the training rounds are set to multiple iterations, and the model is predicted and evaluated, and the evaluation uses the root mean square error to measure the performance of the model.

10. The zero-carbon building design method based on AI optimization according to claim 1 is characterized by: In step g, the optimization process includes inputting different design schemes into a trained model, wherein the design schemes at least involve the selection of building materials, the improvement of building envelope structures, and the configuration adjustment of energy systems; the model predicts the corresponding energy consumption according to the input design parameters; and by comparing and analyzing the prediction results of different design schemes, identifying the design options that can maximize energy efficiency.