Campus functional group carbon emission prediction and optimization method

Abstract

The invention discloses a carbon emission prediction and optimization method for a campus functional group. The technical problem that in the prior art, when carbon emission of grouped buildings is predicted, specific analysis and processing are not carried out on functions, forms and regional characteristics of campus groups, and consequently the prior art cannot be completely suitable for predicting carbon emission of the campus groups is solved. The building carbon emission intensity rapid prediction based on the spatial form parameters of the functional group is realized by utilizing the spatial form and the building power consumption characteristic of the functional group of the cold region campus and the regional characteristic and the climate characteristic of the cold region city and combining a performance simulation method and a machine learning method. And the spatial morphological parameters of the functional group are automatically optimized under the carbon reduction target, so that an architect is effectively assisted to carry out low-carbon design of the functional group of the campus in the cold region in a limited design period. The method is mainly used for predicting carbon emission of campus functional groups.

Description

Technical Field

[0001] This invention relates to the field of low-carbon technology, specifically to a method for predicting and optimizing carbon emissions from campus functional clusters. Background Technology

[0002] Campus functional clusters refer to groups of buildings in the overall campus plan that are centrally located according to functional type, provide unified service facilities, share site resources, and are constructed simultaneously or in phases. Their boundaries are naturally enclosed by campus roads, green spaces, or underground pipe networks, and they share a unified function, such as teaching, research and experimentation, student dormitories, administrative offices, and living facilities. Carbon emissions from campus functional clusters refer to all direct and indirect greenhouse gas emissions generated by the operation of buildings within a complete operating year.

[0003] Carbon emission prediction and optimization of campus functional clusters has become a key link in reducing the carbon footprint of university operations, thus achieving campus-scale carbon reduction goals at the planning stage. Existing technologies, such as CN117648872B, disclose a "multi-objective optimization method for energy conservation and carbon reduction in urban blocks based on machine learning." This method takes urban blocks as the object, establishes simulations of building energy consumption and photovoltaic power generation, and then couples an ensemble learning surrogate model with the NSGA-II genetic algorithm to perform multi-objective optimization of building energy consumption, photovoltaic power generation, and carbon emissions in the urban block, ultimately outputting a Pareto front solution.

[0004] However, when this technology is directly applied to campus functional cluster scenarios: its proxy model is constructed using a block-scale one-time sampling-batch training method, which leads to the "averaging" of the mapping relationship between spatial morphology parameters and carbon emissions. When there are differences in the functional types of clusters, the model prediction error is significantly amplified. At the same time, multi-objective optimization needs to simultaneously weigh the three objectives of energy consumption, production capacity, and carbon emissions. The optimization results are diluted in response to the single requirement of "minimizing carbon emissions," and cannot provide a spatial morphology solution with the lowest carbon emissions under the condition of a long heating season in cold regions. Summary of the Invention

[0005] To address the issue that existing technologies for predicting carbon emissions from campus functional clusters do not specifically analyze and process the functional, morphological, and regional characteristics of campus clusters, thus making them unsuitable for predicting carbon emissions from campus clusters, this invention proposes a method for predicting and optimizing carbon emissions from campus functional clusters.

[0006] This invention is achieved through the following scheme:

[0007] A method for predicting and optimizing carbon emissions from campus functional clusters includes:

[0008] S100: Within the visual programming platform, construct a 3D model based on campus building geometry data and divide it into functional clusters; construct an energy consumption simulation model and a solar power generation potential simulation model;

[0009] S200: Input the energy consumption driving parameters of each cluster of buildings into the energy consumption simulation model, and input the solar power generation boundary parameters into the solar power generation potential simulation model. Based on the output data of the two models and urban data, calculate the carbon emission intensity of the cluster of buildings.

[0010] S300: Select spatial morphological parameters that have a significant impact on carbon emission intensity from the three-dimensional model as sensitive parameters;

[0011] S400: Using sensitive parameters, urban data, and cluster functional types as inputs, and cluster building carbon emission intensity as output, a cluster carbon emission intensity prediction model is constructed.

[0012] S500: The cluster carbon emission intensity prediction model is iteratively optimized by using a swarm intelligence evolutionary algorithm within the initial values ​​and constraints of sensitive parameters to determine the optimal cluster carbon emission model.

[0013] S600: Adopting a cluster carbon emission optimization model, with the single objective of minimizing building carbon emission intensity, it optimizes the target campus functional clusters and outputs their final spatial morphology parameters and the corresponding minimum building carbon emission intensity.

[0014] Furthermore, the functional grouping includes: teaching building group, library group, laboratory building group, dormitory building group, canteen group, and activity center group.

[0015] Furthermore, the building energy consumption driving parameters include operating condition parameters, time constraint parameters, building envelope parameters, and external condition parameters;

[0016] The operating parameters include: personnel utilization rate, lighting power, electrical equipment power, ventilation intensity, and cooling and heating temperature setpoints;

[0017] The time constraint parameters include: two operating modes, working state and rest state; the start and end times of the heating season and cooling season; and the winter and summer vacation periods.

[0018] The external condition parameters include: the three-dimensional spatial morphology of the cluster and the meteorological data of the city where it is located;

[0019] The total energy consumption includes: annual cooling energy consumption, annual heating energy consumption, annual lighting energy consumption, and annual electrical equipment energy consumption of the cluster.

[0020] Furthermore, the solar power generation boundary parameters include solar system parameters, solar installation condition parameters, and external condition parameters;

[0021] The system parameters of the solar energy system include: photovoltaic module conversion efficiency and system loss efficiency;

[0022] The solar energy installation condition parameters include: roof installation coefficient and facade installation coefficient;

[0023] The external condition parameters include: the three-dimensional spatial morphology of the cluster and the meteorological data of the city where it is located.

[0024] Furthermore, the city data includes: city number, city meteorological data, and city carbon emission factor;

[0025] The urban carbon emission factors include: thermal carbon emission factors and electricity carbon emission factors.

[0026] Furthermore, the spatial morphology parameters in the three-dimensional spatial morphology model include total building area, site area, site length-to-width ratio, plot ratio, building density, average building height, maximum building height, building height standard deviation, shape coefficient, perimeter-to-area ratio, facade-to-roof area ratio, building orientation, number of buildings, average building length, average building depth, building height-to-depth ratio, sky exposure coefficient, and sky viewing angle coefficient.

[0027] Furthermore, the method for filtering the sensitive parameters includes:

[0028] Correlation analysis was conducted between carbon emission intensity and cluster spatial morphology parameters. Candidate parameters were selected based on the criteria that the correlation coefficient was greater than a preset threshold A and the significance coefficient was less than a preset threshold B.

[0029] Then, multiple linear regression analysis was performed on carbon emission intensity and candidate parameters, and sensitive parameters were screened out based on the variance inflation factor being greater than the preset threshold C.

[0030] Furthermore, the construction of the cluster carbon emission intensity prediction model also includes training the cluster carbon emission intensity prediction model using at least three supervised machine learning algorithms, the training methods including:

[0031] 1) Divide the sample dataset into training set, validation set and test set according to a preset ratio;

[0032] 2) Within the hyperparameter search range, perform hyperparameter tuning for each machine learning algorithm;

[0033] 3) Use the training set to complete model training, use the validation set to optimize hyperparameters, use the test set to evaluate prediction accuracy and computational cost, and select the campus functional cluster carbon emission prediction model with the best overall performance.

[0034] Furthermore, the method for determining the cluster carbon emission optimization model using swarm intelligence evolutionary algorithm includes:

[0035] 1) Couple the carbon emission intensity prediction model with each intelligent optimization algorithm to construct different candidate carbon emission optimization models;

[0036] 2) Input the initial values ​​and limiting ranges of the spatial morphology parameters of the clusters into each candidate carbon emission optimization model;

[0037] 3) Iterate through each candidate carbon emission optimization model and output the final spatial morphology parameters, building carbon emission intensity, and the numerical changes of both during the iteration process;

[0038] 4) Compare the optimization performance of each candidate optimization model and determine the one with the best performance as the optimization model for carbon emissions from university campus buildings in cold regions.

[0039] Furthermore, during the iterative operation of each candidate carbon emission optimization model, the hyperparameters of each intelligent optimization algorithm are tuned, and an elite retention mechanism is used to maintain the optimal solution across generations. The optimal model is selected by using the optimization effect of carbon emission intensity, running time, and resource consumption as evaluation indicators.

[0040] The beneficial effects of this invention are:

[0041] This invention takes into account the spatial morphology and building power consumption characteristics of functional clusters in cold-region campuses, as well as the regional and climatic characteristics of cold-region cities. It combines the advantages of performance simulation methods and machine learning methods to achieve rapid prediction of building carbon emission intensity based on the spatial morphology parameters of functional clusters, and automatic optimization of the spatial morphology parameters of functional clusters under carbon reduction targets. This effectively assists architects in carrying out low-carbon design of functional clusters in cold-region campuses within a limited design cycle. Attached Figure Description

[0042] 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.

[0043] Figure 1 This is a schematic diagram illustrating the implementation process of one embodiment of the present invention;

[0044] Figure 2 This is a flowchart of a method according to one embodiment of the present invention;

[0045] Figure 3 This is a schematic diagram of the functional grouping of a university campus in a cold region according to one embodiment of the present invention;

[0046] Figure 4 This is a schematic diagram of the spatial morphological parameters of functional clusters in a university campus in a cold region, according to one embodiment of the present invention.

[0047] Figure 5This is a schematic diagram showing the changes in personnel utilization and lighting power of each functional group in the working state in one embodiment of the present invention.

[0048] Figure 6 This is a schematic diagram showing the changes in the power of personnel and electrical equipment and the ventilation intensity of each functional group in the working state in one embodiment of the present invention;

[0049] Figure 7 This is a schematic diagram of the heating and cooling temperature settings of each functional group in the working state in one embodiment of the present invention.

[0050] Figure 8 This is a schematic diagram showing the changes in personnel utilization and lighting power of each functional group in a resting state in one embodiment of the present invention.

[0051] Figure 9 This is a schematic diagram showing the changes in the power of electrical equipment and ventilation intensity of personnel in each functional group during a resting state in one embodiment of the present invention;

[0052] Figure 10 This is a schematic diagram of the heating and cooling temperature settings for each functional group in a resting state according to one embodiment of the present invention.

[0053] Figure 11 This is a schematic diagram illustrating the iterative optimization process of spatial morphological parameters of functional groups in the ANN-DE model in one embodiment of the present invention;

[0054] Figure 12 This is a schematic diagram illustrating the iterative optimization process of spatial morphological parameters of functional groups in the ANN-GA model in one embodiment of the present invention. Detailed Implementation

[0055] 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. Specific Implementation Method 1

[0057] This is a flowchart illustrating the method of this embodiment. This embodiment provides an implementation process for a method for predicting and optimizing carbon emissions from campus functional clusters, as described in this invention:

[0058] S100: Within the visual programming platform, construct a 3D model based on campus building geometry data and divide it into clusters according to function; construct an energy consumption simulation model and a solar power generation potential simulation model.

[0059] Specifically, vector data of the base contours and elevation data of buildings on university campuses in cold regions are extracted, and the data is input into a visualization programming platform to establish a three-dimensional spatial morphological model of the campus buildings. Based on the functional type of the buildings, the buildings are divided into different clusters, and the functional type of each cluster is marked in the three-dimensional spatial morphological model.

[0060] refer to Figure 3 In this implementation, campus buildings are divided according to their functional types, and the resulting clusters include: Teaching Building Cluster (TBC), Library Cluster (LC), Laboratory Building Cluster (EBC), Dormitory Building Cluster (DBC), Dining Hall Cluster (CBC), and Activity Center Cluster (ACC).

[0061] Use a visual programming platform to build energy consumption simulation models and solar power generation potential simulation models.

[0062] S200: Input the energy consumption driving parameters of each cluster of buildings into the energy consumption simulation model, and input the solar power generation boundary parameters into the solar power generation potential simulation model. Based on the output data of the two models and urban data, calculate the carbon emission intensity of the cluster of buildings.

[0063] Specifically, the operating condition parameters, time constraint parameters, maintenance structure parameters, and external condition parameters of different functional groups are input into the energy consumption simulation model, and the model outputs the total annual energy consumption and energy intensity of each building functional group.

[0064] The solar system parameters, solar installation condition parameters, and external condition parameters are input into the solar power generation potential simulation model, and the model outputs the annual solar power generation and solar power intensity of the functional cluster.

[0065] The simulated solar power generation potential model helps planners maximize solar energy self-sufficiency and reduce grid dependence while meeting energy consumption and carbon emission constraints. The simulated annual power generation can be directly used to calculate energy substitution emission reductions, thereby offsetting building operation carbon emissions and achieving carbon emission reduction.

[0066] In practical applications, if the target campus does not have solar power generation facilities, the relevant values ​​for this part can be set to 0.

[0067] Then, based on the total energy consumption of each campus building functional group output by the energy consumption simulation model, the annual solar power generation output by the solar power generation potential simulation model, the thermal carbon emission factor of the city, and the electricity carbon emission factor of the city, the building carbon emission intensity of the campus building functional group is calculated.

[0068] The building energy consumption driving parameters include operating condition parameters, time constraint parameters, building envelope parameters, and external condition parameters. The solar power generation boundary parameters include solar system parameters, solar installation condition parameters, and external condition parameters.

[0069] The operating parameters include: personnel utilization rate, lighting power, electrical equipment power, ventilation intensity, and cooling and heating temperature setpoints. The baseline values ​​are shown in Table 1.

[0070] The time constraint parameters include: two operating modes: working and resting; the start and end times of the heating and cooling seasons; and the winter and summer vacation periods. The changes in operating parameters between working and resting modes are as follows: Figure 5-9 As shown.

[0071] The parameters of the enclosure structure are shown in Table 2.

[0072] The external condition parameters include: the three-dimensional spatial morphology of the cluster and the meteorological data of the city where it is located.

[0073] The total energy consumption includes: annual cooling energy consumption, annual heating energy consumption, annual lighting energy consumption, and annual electrical equipment energy consumption of the cluster.

[0074] The formula for calculating the energy intensity is:

[0075] ;

[0076] in, It is the building's total energy consumption, which includes building cooling energy consumption, heating energy consumption, lighting energy consumption, and electrical equipment energy consumption. It is the total building area.

[0077] The system parameters of the solar energy system include: photovoltaic module conversion efficiency and system loss efficiency.

[0078] The solar energy installation condition parameters include: roof installation coefficient and facade installation coefficient.

[0079] The formula for calculating the building's solar power generation intensity is as follows:

[0080] ;

[0081] ;

[0082] Where SEGI is the building-integrated solar power intensity, and I is the solar irradiance (kWh / m²) on the surface of the photovoltaic module. 2 / y), obtained through solar radiation simulation. It refers to the conversion efficiency of the photovoltaic system. It refers to the loss efficiency of the photovoltaic system. It refers to the installable area of ​​photovoltaic panels. It is the roof area. It refers to the area of ​​the exterior walls. It is the roof installation coefficient. It is the external wall installation coefficient.

[0083] In this embodiment .

[0084] The city data includes: city number, city meteorological data, and city carbon emission factor.

[0085] The urban carbon emission factors include: thermal carbon emission factors and electricity carbon emission factors.

[0086] The formula for calculating the building carbon emission intensity is as follows:

[0087] ;

[0088] in, It is the energy intensity of building cooling, lighting and equipment. ), Building heating energy intensity ( ). For the carbon emission factor of electricity ( ), Thermal carbon emission factor ( In this embodiment, .

[0089] S300: Select spatial morphological parameters that have a significant impact on carbon emission intensity from the three-dimensional model as sensitive parameters.

[0090] Specifically, the method for filtering the sensitive parameters includes:

[0091] Correlation analysis was conducted between carbon emission intensity and cluster spatial morphology parameters. Candidate parameters were selected based on the criteria that the correlation coefficient was greater than a preset threshold A and the significance coefficient was less than a preset threshold B.

[0092] Then, multiple linear regression analysis was performed on carbon emission intensity and candidate parameters, and sensitive parameters were screened out based on the variance inflation factor being greater than the preset threshold C.

[0093] In this embodiment, threshold A is 0.4, threshold B is 0.1, and threshold C is 5.

[0094] The spatial morphological parameters in the three-dimensional spatial morphological model include total building area, site area, site length-to-width ratio, plot ratio, building density, average building height, maximum building height, standard deviation of building height, shape coefficient, perimeter-to-area ratio, facade-to-roof area ratio, building orientation, number of buildings, average building length, average building depth, building height-to-depth ratio, sky exposure coefficient, and sky viewing angle coefficient. The spatial morphological parameters of functional clusters in the campus of a university in a cold region are as follows: Figure 4 As shown.

[0095] The formula for calculating the total building area FA is as follows:

[0096] ;

[0097] in, is the area of ​​each floor of the building, and n is the number of floors.

[0098] The formula for calculating the site aspect ratio SAR is as follows:

[0099] ;

[0100] Where L is the length of the field and W is the width of the field.

[0101] The formula for calculating the Floor Area Ratio (FAR) is as follows:

[0102] ;

[0103] in, This refers to the area of ​​the site.

[0104] The formula for calculating the building density (BCR) is as follows:

[0105] ;

[0106] in, N represents the building's footprint, and N represents the number of buildings within the cluster.

[0107] The formula for calculating the average building height ABH is as follows:

[0108] ;

[0109] in, The height of the building.

[0110] The formula for calculating the maximum building height (MBH) is as follows:

[0111] ;

[0112] The formula for calculating the standard deviation of building height (SDBH) is as follows:

[0113] ;

[0114] in, It is the average height of the building.

[0115] The formula for calculating the body shape coefficient SF is as follows:

[0116] ;

[0117] Where S is the surface area of ​​the building in contact with the outdoor atmosphere, and V is the volume of the building.

[0118] The formula for calculating the perimeter-area ratio PAR is:

[0119] ;

[0120] Where P is the length of the building's exterior wall.

[0121] The formula for calculating the facade roof area ratio (FRAR) is as follows:

[0122] ;

[0123] in, For building facade area, This refers to the roof area of ​​the building.

[0124] The formula for calculating the average building length ABL is as follows:

[0125] ;

[0126] in, The length of the building.

[0127] The formula for calculating the average building depth (ABD) is as follows:

[0128] ;

[0129] in, The depth of the building;

[0130] The formula for calculating the building height-to-depth ratio (BHDR) is as follows:

[0131] ;

[0132] The formula for calculating the Sky Exposure Factor (SEF) is as follows:

[0133] ;

[0134] The formula for calculating the Sky View Factor (SVF) is as follows:

[0135] ;

[0136] S400: Using sensitive parameters, city data, and cluster functional types as inputs, and cluster building carbon emission intensity as output, a cluster carbon emission intensity prediction model is constructed.

[0137] The construction of the cluster carbon emission intensity prediction model also includes training the cluster carbon emission intensity prediction model using at least three supervised machine learning algorithms. The training methods include:

[0138] 1) Divide the sample dataset into training set, validation set and test set according to a preset ratio;

[0139] 2) Within the hyperparameter search range, perform hyperparameter tuning for each machine learning algorithm;

[0140] 3) Use the training set to complete model training, use the validation set to optimize hyperparameters, use the test set to evaluate prediction accuracy and computational cost, and select the campus functional cluster carbon emission prediction model with the best overall performance.

[0141] In this embodiment, three supervised machine learning algorithms were used: artificial neural network, support vector machine, and random forest. These three algorithms were used to train the carbon emission intensity prediction model. The search range of hyperparameters is shown in Table 3, where ANN, SVM, and SF represent artificial neural network, support vector machine, and random forest machine learning algorithms, respectively. The prediction accuracy is determined by... Joint decision: The coefficient of determination ranges from 0 to 1, with values ​​closer to 1 indicating higher prediction accuracy. RMSE is the root mean square error, and MAE is the mean absolute error; smaller RMSE and MAE values ​​indicate higher prediction accuracy. Computational cost is determined by training time, prediction time, and memory usage.

[0142] S500: Utilizes a swarm intelligence evolutionary algorithm to iteratively optimize within the initial values ​​and constraints of sensitive parameters, thereby determining the optimal model for cluster carbon emissions.

[0143] The method for determining the optimal model for cluster carbon emissions using swarm intelligence evolutionary algorithms includes:

[0144] 1) Couple the carbon emission intensity prediction model with each intelligent optimization algorithm to construct different candidate carbon emission optimization models;

[0145] 2) Input the initial values ​​and limiting ranges of the spatial morphology parameters of the clusters into each candidate carbon emission optimization model;

[0146] 3) Iterate through each candidate carbon emission optimization model and output the final spatial morphology parameters, building carbon emission intensity, and the numerical changes of both during the iteration process;

[0147] 4) Compare the optimization performance of each candidate optimization model and determine the one with the best performance as the optimization model for carbon emissions from university campus buildings in cold regions.

[0148] During the iterative operation of each candidate carbon emission optimization model, the hyperparameters of each intelligent optimization algorithm are tuned, and an elite retention mechanism is used to maintain the optimal solution for each generation. The optimal model is selected by using the optimization effect of carbon emission intensity, running time and resource consumption as evaluation indicators.

[0149] Specifically, the optimization objective is to minimize carbon emission intensity. Candidate carbon emission optimization models are iteratively run, with hyperparameters adjusted in each iteration, generating many different combinations of spatial morphological parameters. Each combination is considered a "child generation." Among all children, only the "optimal child" with the lowest carbon emission intensity is retained and directly enters the next generation. Other children are eliminated. The vacated positions from eliminated children are used to generate new children through random mutation, continuing the trial-and-error process. After repeating this process multiple times, if the carbon emission intensity can no longer be reduced, the minimum carbon emission condition is considered to have been reached. The percentage of carbon emission optimization, the runtime of this run, and resource consumption are recorded at this point.

[0150] After each candidate carbon emission optimization model has completed the above operations, the carbon emission optimization percentage, running time, and resource consumption of each candidate model are used as evaluation indicators. These three indicators are converted into dimensionless scores in the range of 0-1, and then weighted and summed according to predetermined weights to obtain the score. The model with the highest score among all candidate models is selected as the optimal carbon emission optimization model.

[0151] The spatial morphological parameters at this point can be used as the optimized design standard.

[0152] In this embodiment, the swarm intelligence evolutionary algorithm includes the genetic algorithm GA, the particle swarm optimization algorithm PSO, and the differential evolution algorithm DE. The search range of hyperparameters is shown in Table 4.

[0153] S600: Adopting a cluster carbon emission optimization model, with the single objective of minimizing building carbon emission intensity, it optimizes the target campus functional clusters and outputs their final spatial morphology parameters and the corresponding minimum building carbon emission intensity.

[0154] To verify the beneficial effects of the present invention, the following experiments were conducted:

[0155] This embodiment uses a cluster of teaching buildings on a university campus in Beijing as an example, and conducts an experiment using the campus functional cluster carbon emission prediction and optimization method described in this invention, with reference to... Figure 2 This is a flowchart of this embodiment. The experimental steps are as follows:

[0156] Step 1: Extract the building base contour vector data and elevation data of 1711 functional clusters of university campuses in cold regions, establish a three-dimensional spatial morphology model on the visualization programming platform, and label the functional type of the cluster. At the same time, extract the meteorological data of the city where the functional cluster is located.

[0157] Step 2: In the visual programming platform, pick up the spatial morphological parameters of the functional groups of the university campus in the cold region, including total building area, site area, site length-to-width ratio, plot ratio, building density, average building height, maximum building height, building height standard deviation, shape coefficient, perimeter-to-area ratio, facade-to-roof area ratio, building orientation, number of buildings, average building length, average building depth, building height-to-depth ratio, sky exposure coefficient, and sky viewing angle coefficient.

[0158] Step 3: Use a visual programming platform to build an energy consumption simulation model for the functional clusters of a university campus in a cold region. Based on the functional type of the building clusters, input baseline values ​​for parameters such as occupancy rate, lighting power, electrical equipment power, ventilation intensity, and cooling and heating temperature setpoints, along with their variations during working and rest periods. Set the heating season, cooling season, and university holidays for the year. Set the building envelope parameters. Input the three-dimensional spatial form of the functional clusters and the meteorological parameters of the city into the energy consumption simulation model, and output the annual cooling energy consumption, heating energy consumption, lighting energy consumption, electrical equipment energy consumption, and energy intensity of the functional clusters.

[0159] Step 4: Use a visual programming platform to build a simulation model of the solar power generation potential of functional clusters on a university campus based on spatial morphology. Set the photovoltaic module conversion efficiency, system loss efficiency, roof installation coefficient, and facade installation coefficient. Input the three-dimensional spatial morphology of the functional clusters and the meteorological parameters of the city into the solar power generation potential model, and output the annual solar power generation and solar power intensity of the functional clusters.

[0160] Step 5: Based on the annual cooling energy consumption, heating energy consumption, lighting energy consumption, electrical equipment energy consumption and solar power generation of the functional cluster, and combined with the thermal carbon emission factor and electricity carbon emission factor of the city, calculate the building carbon emission intensity of the functional cluster.

[0161] Step 6: Through correlation analysis and multiple linear regression analysis, we can identify the spatial morphological parameters that affect the carbon emission intensity of buildings in the campus functional clusters, i.e., sensitive parameters, including: total building area, average building height, standard deviation of building height, shape coefficient, facade-to-roof area ratio, and building height-to-depth ratio.

[0162] Step 7: Using the aforementioned sensitive parameters, city data, and functional type as inputs, and building carbon emission intensity as output, establish a prediction model for the building carbon emission intensity of campus functional clusters. Train the prediction model using artificial neural networks, support vector machines, and random forest machine learning algorithms. Comparisons show that the prediction model with the best prediction accuracy for campus functional cluster building carbon emission intensity—an ANN model with 100, 70, and 70 neurons in the hidden layers—achieved a prediction accuracy of 95.37%. The model's alpha = 0.0001, learning_rate = 0.01, max_iter = 1000, and patience = 15.

[0163] Step 8: Combine the carbon emission prediction model for functional cluster buildings with the optimal prediction accuracy with optimization algorithms such as genetic algorithm, particle swarm optimization, and differential evolution algorithm to establish an optimization model for carbon emissions of functional cluster buildings in university campuses in cold regions. Input the initial values ​​and constraints of the spatial morphology parameters of the campus functional clusters, and output the final spatial morphology parameters, building carbon emission intensity, and the changes in their values ​​during iterative optimization.

[0164] The total construction area of ​​this complex is 86,840 square meters. The values ​​remain unchanged during the optimization process. The initial value for the average building height is 80m, with a limiting range of 4-80m. The initial value for the standard deviation of building height is 1.43m, with a limiting range of 0-28m. The initial value for the shape coefficient is 0.16, with a limiting range of 0.08-0.62. The initial value for the facade-to-roof area ratio is 11.68, with a limiting range of 0.25-11. The initial value for the building height-to-depth ratio is 0.24, with a limiting range of 0.08-2.5.

[0165] Ultimately, the best-performing optimization models for carbon emissions from university campuses in cold regions were selected, including the ANN-GA model with a popsize of 50 and the ANN-DE model with a popsize of 4m. The ANN-GA model has a max_iter of 100, a mutation of 0.2, and a crossover rate of 0.8; the ANN-DE model has a mutation of 0.7, a recombination of 0.8, and a max_iter of 100.

[0166] Step 9: Optimize the target campus functional clusters using the best-performing optimization model, reducing the building carbon emission intensity from 51.07. It dropped to -86.54 Furthermore, the optimization process of the spatial morphological parameters of functional clusters, such as... Figure 11 , Figure 12As shown, the average building height was optimized to 4m, the standard deviation of building height was optimized to 0m, the shape coefficient was optimized to 0.62, the ratio of facade to roof area was optimized to 0.25, and the building height-to-depth ratio was optimized to 1.56.

[0167] Table 1 Power Consumption Parameter Settings for Functional Clusters in Universities in Cold Regions

[0168]

[0169] Table 2. Parameter Setting Table for Envelope Structure of Functional Clusters in Universities in Cold Regions

[0170]

[0171] Table 3. Hyperparameter search range settings for different machine learning algorithms.

[0172]

[0173] Table 4. Hyperparameter search range settings for different optimization algorithms

[0174]

[0175] The above description of the disclosed embodiments enables those skilled in the art to make or use the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A method for predicting and optimizing carbon emissions from functional clusters within a campus, characterized in that, include: S100: Within the visual programming platform, a 3D model is constructed based on the geometric data of campus buildings and groups are divided according to function; Construct energy consumption simulation models and solar power generation potential simulation models; S2 00: Input the energy consumption driving parameters of each cluster of buildings into the energy consumption simulation model, and input the solar power generation boundary parameters into the solar power generation potential simulation model. Based on the output data of the two models and urban data, calculate the carbon emission intensity of the cluster of buildings. S300: Select spatial morphological parameters that have a significant impact on carbon emission intensity from the three-dimensional model as sensitive parameters; S400: Using sensitive parameters, urban data, and cluster functional types as inputs, and cluster building carbon emission intensity as output, a cluster carbon emission intensity prediction model is constructed. S500: The cluster carbon emission intensity prediction model is iteratively optimized by using a swarm intelligence evolutionary algorithm within the initial values ​​and constraints of sensitive parameters to determine the optimal cluster carbon emission model. S600: Adopting a cluster carbon emission optimization model, with the single objective of minimizing building carbon emission intensity, it optimizes the target campus functional clusters and outputs their final spatial morphology parameters and the corresponding minimum building carbon emission intensity.

2. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The functional grouping includes: teaching building group, library group, laboratory building group, dormitory building group, canteen group, and activity center group.

3. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The building energy consumption driving parameters include operating condition parameters, time constraint parameters, building envelope parameters, and external condition parameters; The operating parameters include: personnel utilization rate, lighting power, electrical equipment power, ventilation intensity, and cooling and heating temperature setpoints; The time constraint parameters include: two operating modes, working state and rest state; the start and end times of the heating season and cooling season; and the winter and summer vacation periods. The external condition parameters include: the three-dimensional spatial morphology of the cluster and the meteorological data of the city where it is located; The total energy consumption includes: annual cooling energy consumption, annual heating energy consumption, annual lighting energy consumption, and annual electrical equipment energy consumption of the cluster.

4. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The solar power generation boundary parameters include solar system parameters, solar installation condition parameters, and external condition parameters. The system parameters of the solar energy system include: photovoltaic module conversion efficiency and system loss efficiency; The solar energy installation condition parameters include: roof installation coefficient and facade installation coefficient; The external condition parameters include: the three-dimensional spatial morphology of the cluster and the meteorological data of the city where it is located.

5. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The city data includes: city number, city meteorological data, and city carbon emission factor; The urban carbon emission factors include: thermal carbon emission factors and electricity carbon emission factors.

6. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The spatial morphology parameters in the three-dimensional spatial morphology model include total building area, site area, site length-to-width ratio, plot ratio, building density, average building height, maximum building height, building height standard deviation, shape coefficient, perimeter-to-area ratio, facade-to-roof area ratio, building orientation, number of buildings, average building length, average building depth, building height-to-depth ratio, sky exposure coefficient, and sky viewing angle coefficient.

7. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The methods for filtering the sensitive parameters include: Correlation analysis was conducted between carbon emission intensity and cluster spatial morphology parameters. Candidate parameters were selected based on the criteria that the correlation coefficient was greater than a preset threshold A and the significance coefficient was less than a preset threshold B. Then, multiple linear regression analysis was performed on carbon emission intensity and candidate parameters, and sensitive parameters were screened out based on the variance inflation factor being greater than the preset threshold C.

8. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The construction of the cluster carbon emission intensity prediction model involves training the model using at least three supervised machine learning algorithms after its construction. The training methods include: 1) Divide the sample dataset into training set, validation set and test set according to a preset ratio; 2) Within the hyperparameter search range, perform hyperparameter tuning for each machine learning algorithm; 3) Use the training set to complete model training, use the validation set to optimize hyperparameters, use the test set to evaluate prediction accuracy and computational cost, and select the campus functional cluster carbon emission prediction model with the best overall performance.

9. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 1, characterized in that, The method for determining the optimal model for cluster carbon emissions using swarm intelligence evolutionary algorithms includes: 1) Couple the carbon emission intensity prediction model with each intelligent optimization algorithm to construct different candidate carbon emission optimization models; 2) Input the initial values ​​and limiting ranges of the spatial morphology parameters of the clusters into each candidate carbon emission optimization model; 3) Iterate through each candidate carbon emission optimization model and output the final spatial morphology parameters, building carbon emission intensity, and the numerical changes of both during the iteration process; 4) Compare the optimization performance of each candidate optimization model and determine the one with the best performance as the optimization model for carbon emissions from university campus buildings in cold regions.

10. The method for predicting and optimizing carbon emissions from campus functional clusters according to claim 9, characterized in that, During the iterative operation of each candidate carbon emission optimization model, the hyperparameters of each intelligent optimization algorithm are tuned, and an elite retention mechanism is used to maintain the optimal solution for each generation. The optimal model is selected by using the optimization effect of carbon emission intensity, running time and resource consumption as evaluation indicators.

Images