Parametric Design Method for Existing Buildings Based on BIM

Abstract

The present invention discloses a parametric design method for existing buildings based on BIM, which specifically relates to the field of building optimization design, and includes the following steps: first, data of the existing buildings is collected, a BIM model is established, and its daily usage is simulated to obtain a noise data set of each area; whether noise pollution exists is determined by the noise data, and if so, the area is divided into pollution areas and the data is summarized to generate a pollution information set; cluster analysis is performed based on the information set to generate a pollution level form, and a pre-trained machine learning model is used to generate a comprehensive pollution coefficient; according to the coefficient, the average or maximum noise is selected as the optimization target, and the optimal combination of optimization parameters is found using a genetic algorithm, and finally an optimized design scheme for the entire building is formed; the present invention ensures differentiated design for areas with different noise levels, which can not only meet the mandatory sound insulation requirements of high-noise areas, but also moderately optimize areas with lower noise.

Description

Technical Field

[0001] The present invention relates to the field of building optimization design, and more specifically, to a BIM-based parametric design method for existing buildings. Background Art

[0002] In modern urban construction and development, noise pollution has become a critical issue affecting building comfort and safety. Existing buildings, in particular, face the challenge of noise pollution during daily use due to a lack of adequate consideration of noise control during the initial design phase. Traditional noise control methods often rely on the simple installation of soundproofing materials or partial wall modifications, lacking systematic assessment and dynamic optimization of the entire building.

[0003] With the widespread adoption of BIM (Building Information Modeling) technology, building design and management are gradually becoming digital and intelligent. BIM models enable data-driven management and analysis of building structures, material usage, and daily operations. In the field of noise pollution control, BIM technology provides strong support for parametric optimization design. However, existing technologies for noise pollution assessment and control still lack a systematic approach that leverages big data and intelligent algorithms for multi-dimensional analysis and optimization. Therefore, a BIM-based parametric design method for noise control in existing buildings is proposed. This approach leverages data acquisition, pollution assessment, machine learning, and genetic algorithms to achieve dynamic optimization and cost control of building noise. Summary of the Invention

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] The BIM-based parametric design method for existing buildings includes the following steps:

[0006] Collect data from existing buildings and build BIM models. Simulate the daily usage of existing buildings and divide the BIM models into different areas to obtain noise data sets for the target areas.

[0007] Determine whether the target area is noise-polluted based on the noise dataset of the target area. If noise pollution exists, the area is divided into a polluted area, and the noise datasets of all polluted areas are aggregated to obtain a pollution information set.

[0008] A cluster analysis operation is performed based on the pollution information set to obtain a pollution level table for the target area. The pollution level table for the target area is then fed into a pre-trained machine learning model to generate a comprehensive pollution coefficient. Based on the comprehensive pollution coefficient, the average noise or maximum noise is used as the optimization target.

[0009] The optimization target of the target area is simulated in the established BIM model, and the genetic algorithm is used to find the optimization parameter combination that minimizes the cost of achieving the optimization target. The optimization parameter combinations of all polluted areas are summarized to obtain the optimized design plan for the entire building.

[0010] In a preferred embodiment, the noise data set of the target area consists of multiple noise data groups, each noise data group corresponds to a time period of data collection, and all time periods have no time overlap and cover the entire preset sampling window.

[0011] In a preferred embodiment, judging whether a target area is noise-contaminated based on a noise dataset of the target area refers to:

[0012] The evaluation is performed with reference to the preset noise standard, and the noise data measured in the target area is compared with the preset noise standard. If the noise data measured in the target area is higher than the preset noise standard, it indicates that there is noise pollution. If the noise data measured in the target area is not higher than the preset noise standard, it indicates that there is no noise pollution.

[0013] In a preferred embodiment, performing a cluster analysis operation based on the pollution information set refers to:

[0014] Step 1: Determine the number of cluster centers, i.e., the K value;

[0015] Step 2: Initialize cluster centers: randomly select K initial cluster centers, each cluster center represents a noise level interval;

[0016] Step 3: Calculate the distance between each region's data point and the cluster center: Use the Euclidean distance formula to calculate the distance between each region's noise measurement value and each cluster center;

[0017] Step 4: Classify data points: Classify the noise data points in each area to the nearest cluster center to form multiple noise level area clusters, each area cluster represents a noise interval;

[0018] Step 5: Update the cluster center: Calculate the average value of all data points in each noise region cluster and update the cluster center of each cluster;

[0019] Step 6. Iterate until convergence: Repeat steps 2 to 5 until the cluster center no longer changes or the preset number of iterations is reached.

[0020] In a preferred embodiment, determining the number of cluster centers, i.e., the K value, refers to:

[0021] There are multiple preset interval levels for representing noise levels, and the total number of interval levels is the number of cluster centers.

[0022] In a preferred embodiment, the pollution level table of the target area includes:

[0023] The interval level corresponding to each noise level region cluster and the total number of data in each noise level region cluster.

[0024] In a preferred embodiment, the pollution level table of the target area is fed into a pre-trained machine learning model to generate a comprehensive pollution coefficient:

[0025] The pre-trained machine learning models are:

[0026] ; n represents the total number of noise level region clusters, i represents the index number of the noise level region cluster, represents the total number of data in the i-th noise level region cluster, represents the interval level corresponding to the i-th noise level region cluster, represents the preset influence coefficient corresponding to the i-th noise level area cluster. All influence coefficients are positive and their sum is one. Represents the comprehensive pollution coefficient.

[0027] In a preferred embodiment, using average noise or maximum noise as an optimization target based on the comprehensive pollution coefficient means:

[0028] The comprehensive pollution coefficient is compared with the preset pollution threshold. If the comprehensive pollution coefficient is greater than the preset pollution threshold, the maximum noise of the target area is used as the optimization target. If the comprehensive pollution coefficient is less than or equal to the preset pollution threshold, the average noise of the target area is used as the optimization target.

[0029] In a preferred embodiment, using a genetic algorithm to find an optimal parameter combination that minimizes the cost of achieving the optimization goal refers to:

[0030] Encoding and initial population: Encode the optimization parameter combination into chromosome form and randomly generate m chromosomes as the initial population;

[0031] Fitness evaluation: Taking the optimization goal as the constraint condition in fitness evaluation, the fitness evaluation formula is established:

[0032] ; Indicates optimized material type The amount of money spent, m represents the type of optimized material The total number of species, is the fitness value;

[0033] Selection operation: Use the roulette wheel selection method to select children as new parents;

[0034] Crossover operation: randomly exchange data in different parent chromosomes;

[0035] Mutation operation: randomly select data from different offspring chromosomes for adjustment;

[0036] Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal optimization parameter combination.

[0037] The technical effects and advantages of the present invention are as follows:

[0038] The present invention comprehensively collects and analyzes noise data of existing buildings based on the BIM model. By using cluster analysis and machine learning models, it can accurately identify noise pollution areas and their specific noise levels, avoiding the defects of inaccurate judgment of noise problems in traditional methods.

[0039] By incorporating a genetic algorithm, the present invention automatically generates an optimal noise control solution, finding the optimal balance between cost and noise control effectiveness. Compared to traditional sound insulation stacking, the present invention achieves optimal noise control with minimal modification costs, offering greater intelligence and cost-effectiveness.

[0040] The present invention flexibly selects average noise or maximum noise as the optimization target based on the noise data and pollution level of each area, ensuring differentiated design for areas with different noise levels. It can not only meet the mandatory sound insulation requirements in high-noise areas, but also moderately optimize areas with lower noise levels.

[0041] Through the BIM model, this invention integrates noise data, pollution information sets, and optimization solutions to generate a comprehensive building noise optimization design solution. This method not only simplifies the design process but also improves the overall efficiency of noise optimization design.

[0042] The present invention uses the comprehensive pollution coefficient as the judgment basis for optimal design, ensuring that excessive sound insulation design is not performed in areas with relatively light noise pollution, thereby avoiding unnecessary waste of resources and improving the economic benefits of architectural design. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] In order to facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings;

[0044] Figure 1 This is a schematic diagram of the BIM-based parametric design method for existing buildings in the present invention. DETAILED DESCRIPTION

[0045] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0046] Reference Figure 1 The following examples were obtained:

[0047] Example 1: A BIM-based parametric design method for existing buildings, comprising the following steps:

[0048] Data is collected from existing buildings and a BIM model is created. Simulations are conducted on the building's daily usage, and the BIM model is divided into multiple building zones to generate noise datasets for the target zones. Data collection (such as noise data) is used to obtain actual usage data for the existing building. This data is integrated into the BIM model to create a digital model that reflects the building's actual conditions. The BIM model visualizes the building's structure and environment, providing foundational data for subsequent design optimization. Simulations of daily usage are then used to analyze the noise generated by these conditions and the impact of the noise in the building's surroundings. This helps identify operational issues, namely noise pollution, and ensures that optimization plans are tailored to actual needs. Buildings are divided into zones based on function or space to facilitate management and targeted optimization. For example, separate zones can be used for offices and public areas. Sensors are deployed in different zones to capture noise data and generate noise datasets for each zone. This step forms the basis for determining whether noise pollution exists in the area. Data collection accurately reflects the noise conditions in different zones.

[0049] Based on the target area's noise dataset, the target area is determined to be noisy. If so, the area is classified as a polluted area. The noise datasets for all polluted areas are aggregated to form a pollution information set. By analyzing the noise data, it is determined whether the noise level in each area exceeds the standard threshold. If the noise level in a particular area is high, it can be determined to be noisy. The results of this step determine whether noise optimization is required in that area. The data for all noise-polluted areas is aggregated to form a pollution information set containing the noise characteristics of each polluted area. This provides the data foundation for subsequent noise level classification and optimization design.

[0050] A cluster analysis is performed based on the pollution information set to generate a pollution level table for the target area. This table is then fed into a pre-trained machine learning model to generate a comprehensive pollution coefficient. Based on this comprehensive pollution coefficient, the optimization objective is determined as either average noise or maximum noise. Cluster analysis is then used to categorize the polluted area into different noise levels, generating a pollution level table. This allows for differentiation between areas with varying pollution levels, facilitating the design of targeted optimization measures. For example, low- and high-pollution areas may require different treatment approaches. By feeding the pollution level table into the pre-trained machine learning model, a comprehensive pollution coefficient is generated for each area. This coefficient reflects the overall pollution situation in the area and can be used to determine the optimization approach. The magnitude of the comprehensive pollution coefficient determines whether to use average noise or maximum noise as the optimization objective. For areas with low pollution coefficients, average noise may be sufficient, while for areas with high pollution coefficients, maximum noise is used as the optimization objective to ensure effective sound insulation even in high-noise conditions.

[0051] The optimization objectives for the target area are simulated within the established BIM model. A genetic algorithm is used to find the optimal parameter combination that minimizes the cost of achieving the optimization objective. The optimized parameter combinations for all contaminated areas are then aggregated to produce an optimized design for the entire building. Noise optimization simulations are performed within the BIM model, and the effects of different optimization measures are estimated using the digital model. This allows for the evaluation of noise control effectiveness before actual renovations are carried out, saving time and costs. A genetic algorithm is used to find the optimal combination of noise and sound insulation measures, ensuring that renovation costs are minimized while achieving noise control objectives. The genetic algorithm can intelligently select among multiple options to identify a design that is both noise-optimized and cost-effective. The optimization results for all contaminated areas are integrated to generate a noise control design for the entire building. This solution incorporates specific optimization measures for different areas to ensure a comprehensive improvement in the noise environment of the entire building, and the final design results can be visualized through the visualization capabilities of the BIM model.

[0052] The noise data set of the target area consists of multiple noise data groups, each of which corresponds to a time period for data collection. All time periods have no time overlap and together cover the entire preset sampling window. Each noise data group corresponds to an independent time period, ensuring that data collection in different time periods does not overlap. This method can ensure that the noise data covers the entire preset sampling window and provides specific noise information for each time period, thereby conducting a comprehensive assessment of the noise situation in the area. The non-overlap of all time periods means that duplicate recording or omission of data can be avoided, ensuring the integrity and reliability of the time series. The segmented collection of noise data sets makes the data more orderly and provides a detailed data foundation for subsequent noise analysis, pollution identification and optimization decisions. By collecting noise data in multiple independent time periods, the noise changes in the area in different time periods can be captured, because noise usually exhibits different levels at different times of the day, and the standards for judging the existence of noise pollution in different areas and time periods are different. For details, please see the following description:

[0053] Judging whether a target area has noise pollution based on a noise data set of the target area means: performing an assessment with reference to a preset noise standard, comparing the noise data measured in the target area with the preset noise standard; if the noise data measured in the target area is higher than the preset noise standard, it indicates that noise pollution exists; if the noise data measured in the target area is not higher than the preset noise standard, it indicates that no noise pollution exists.

[0054] This step compares the noise data of the target area with the preset noise standards to clearly determine whether the area is noise polluted. This standard-based comparison provides an objective and quantitative basis for judgment, avoiding subjectivity or ambiguity, and ensuring an accurate assessment of the regional noise situation.

[0055] By determining whether noise pollution exists, we can identify areas requiring further noise control or sound insulation design, allowing us to propose targeted optimization solutions. For polluted areas, we can further analyze noise levels and ultimately design appropriate noise reduction measures. For unpolluted areas, we can avoid unnecessary renovation investments.

[0056] It should be noted that after the target area is divided into a polluted area, since the noise data set of the target area is composed of multiple noise data groups, during the subsequent aggregation, the noise data groups without noise pollution will be eliminated, and only the noise data groups with noise pollution will be retained to reduce the burden of subsequent analysis tasks.

[0057] Noise standards are typically set based on regulations, environmental requirements, and building usage. Standards set noise limits based on the needs of different environments, functions, or areas. Here are some common noise standard categories:

[0058] Outdoor noise standards: Generally, daytime and nighttime noise limits are set based on different environmental functional zones. For example, the daytime noise limit in residential areas is ≤55dB, and the nighttime limit is ≤45dB. The daytime noise limit in industrial areas is ≤65dB, and the nighttime limit is ≤55dB.

[0059] Indoor noise standards: Indoor noise standards for residential, office, hospital, school and other buildings:

[0060] Office area: ≤45dB (daytime). Classroom or conference room: ≤40dB (daytime). Medical area: ≤35dB (daytime).

[0061] Therefore, time division is performed at the beginning of data collection. Each noise data group corresponds to a data collection time period. All time periods have no time overlap and together cover the entire preset sampling window.

[0062] Cluster analysis based on contaminated information sets refers to:

[0063] Step 1: Determine the number of cluster centers, or K value. This means presetting multiple noise level intervals, with the total number of intervals representing the number of cluster centers. Determining the K value is the first step in cluster analysis and indicates how many intervals the noise level will be divided into. By presetting multiple noise level intervals, the noise pollution profile of a region can be further refined. A larger K value results in a more detailed division of noise intervals, allowing for a more accurate reflection of regional noise pollution differences.

[0064] Step 2: Initialize cluster centers: Randomly select K initial cluster centers, each representing a noise level interval. Initialize the K cluster centers to represent the noise level intervals in the initial state. By randomly selecting cluster centers, each cluster center represents a preliminary estimate of a different noise level. This randomness provides diverse initial conditions for the subsequent clustering process, helping the algorithm find a more reasonable cluster structure.

[0065] Step 3: Calculate the distance from each region's data point to the cluster center: Use the Euclidean distance formula to calculate the distance between each region's noise measurement and each cluster center. By calculating the distance between each region's noise measurement and each cluster center, we can determine which range the noise level in each region is closer to. Euclidean distance is a commonly used metric that accurately reflects the similarities or differences between different noise data points in geometric space.

[0066] Step 4: Classify Data Points: Classify noise data points in each region to the nearest cluster center, forming multiple noise level clusters. Each cluster represents a noise interval. Furthermore, classify noise data points in each region into the category of the nearest cluster center, forming multiple noise level clusters. Each cluster represents an area with a different noise level. This step can group areas with similar noise characteristics together, helping to understand pollution distribution.

[0067] Step 5: Update the cluster centers: Calculate the average value of all data points in each noise region cluster and update the cluster center of each cluster. By calculating the average value of the data points in each noise region cluster, the cluster center value can be updated to more accurately reflect the noise characteristics of the region. Updating the cluster center is to adjust the representativeness of each noise cluster so that the cluster center gradually approaches the actual distribution of the data.

[0068] Step 6: Iterate until convergence: Repeat steps 2 through 5 until the cluster centers no longer change, or until the preset number of iterations is reached. Repeat the initialization, distance calculation, classification, and update steps until the cluster centers no longer change significantly, indicating that the algorithm has found a relatively stable noise distribution pattern. This step ensures the reliability and consistency of the clustering results, and the resulting noise level classification accurately reflects the noise pollution in the target area.

[0069] The pollution level table of the target area includes: the interval level corresponding to each noise level area cluster and the total number of data in each noise level area cluster.

[0070] Feeding the pollution level form of the target area into the pre-trained machine learning model to generate the comprehensive pollution coefficient means: the pre-trained machine learning model is:

[0071] ; n represents the total number of noise level region clusters, i represents the index number of the noise level region cluster, represents the total number of data in the i-th noise level region cluster, represents the interval level corresponding to the i-th noise level region cluster, represents the preset influence coefficient corresponding to the i-th noise level area cluster. All influence coefficients are positive and their sum is one. Represents the comprehensive pollution coefficient. It is used to amplify the impact of the number of noise data on the comprehensive pollution coefficient, reflecting that the area with more data contributes more to the pollution coefficient. Logarithmic transformation of the noise level is used to smooth out extreme noise data while still maintaining a large interval difference. Further enhance the impact of high noise areas to ensure that high noise areas have a significant impact on pollution. The larger the comprehensive pollution coefficient, the more serious the noise pollution.

[0072] Using average noise or maximum noise as the optimization target based on the comprehensive pollution coefficient means:

[0073] The comprehensive pollution coefficient is compared with the preset pollution threshold. If the comprehensive pollution coefficient is greater than the preset pollution threshold, the maximum noise of the target area is used as the optimization target. If the comprehensive pollution coefficient is less than or equal to the preset pollution threshold, the average noise of the target area is used as the optimization target.

[0074] It should be noted that both the average noise and the maximum noise are obtained by eliminating the noise data groups without noise pollution and retaining only the noise data groups with noise pollution. In the noise optimization process, whether to use the average noise or the maximum noise as the optimization target depends on the overall assessment of the noise environment and the comprehensive consideration of the building usage.

[0075] Average noise represents the overall noise level of an area over a longer period of time and can better reflect the area's typical noise exposure. If an area has relatively low noise levels most of the time, with only brief peaks, using average noise can help avoid over-designing sound insulation measures due to a few outliers.

[0076] Sporadic noise is a short-lived peak that doesn't represent the long-term noise pollution in an area. Calculating the average noise level can smooth out the effects of these sporadic noise events without requiring excessive structural optimization to mitigate short bursts of high noise levels.

[0077] Maximum noise levels are often short-term peaks, while the goal of building optimization is to address long-term, sustained noise conditions. Using maximum noise as the sole optimization target can lead to overdesign and unnecessary increased sound insulation costs. This is especially true when noise peaks are only occasional, short-lived events, so adopting the maximum noise level isn't necessarily the most economical solution.

[0078] When the comprehensive pollution coefficient is below a certain threshold, it means that the pollution situation in the area is relatively controllable. In this case, the noise pollution may not be severe, and using average noise as the optimization target is sufficient to meet the noise requirements of the area. Optimizing building structures using maximum noise is often done to address high-pollution or severe noise areas, but if the comprehensive pollution coefficient is below the judgment threshold, this level of sound insulation optimization is generally not necessary.

[0079] The total number of data points and the interval level within each noise level cluster have a direct impact on the overall contamination coefficient. In some areas, the noise data is relatively evenly distributed, with few extreme high-noise intervals. If the data volume within these clusters is large and the noise level is relatively concentrated, using average noise can better reflect the actual noise distribution in that area. This optimized design can both control noise contamination and avoid over-design due to a few extreme values.

[0080] In highly polluted areas, where noise issues are severe and persistent, it's recommended to use maximum noise as the optimization target to ensure that the building structure provides adequate noise isolation even under the most severe conditions. In areas with lower comprehensive pollution coefficients, where noise issues are less severe and more sporadic, using average noise optimization can ensure results while avoiding unnecessary costs.

[0081] Using genetic algorithms to find the optimal parameter combination that minimizes the cost of achieving the optimization goal refers to:

[0082] Coding and initial population: The optimization parameter combination is encoded into chromosome form and m chromosomes are randomly generated as the initial population; each chromosome represents a possible solution. Through encoding, different optimization parameter combinations can be expressed in mathematical form, allowing the genetic algorithm to operate and optimize these combinations.

[0083] Fitness evaluation: Taking the optimization goal as the constraint condition in fitness evaluation, the fitness evaluation formula is established:

[0084] ; Indicates optimized material type The amount of money spent, m represents the type of optimized material The total number of species, is the fitness value; the formula calculates the total cost of the optimization plan. The lower the fitness value, the more the optimization combination meets the cost-saving requirements.

[0085] Selection operation: Use the roulette wheel selection method to select offspring as new parents; chromosomes with high fitness have a greater chance of being selected, which ensures that better parameter combinations can be inherited to the next generation, gradually optimizing the overall population.

[0086] Crossover operation: Randomly swap the data in the chromosomes of different parents; by crossing the chromosome data of two different parents, a new offspring is generated. This simulates genetic recombination in biology, aiming to find a better combination by recombining existing solutions.

[0087] Mutation: Randomly selects data from different daughter chromosomes for adjustment; randomly selects certain genes within a chromosome for fine-tuning to increase population diversity. Mutation prevents the algorithm from becoming trapped in local optima, ensuring that more of the solution space is explored to find the global optimal solution. The mutation probability is typically set between 1% and 5% to maintain population diversity.

[0088] Iteration and termination conditions: When the pre-set termination condition is reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal optimization parameter combination. This ensures that the algorithm ends within a reasonable time and outputs the optimal parameter combination solution.

[0089] The above formulas are all dimensionless and numerical calculations. The formulas are obtained by collecting a large amount of data and performing software simulation to obtain the most recent real situation. The preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0090] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0091] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0092] Those skilled in the art will clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0093] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. The BIM-based parametric design method for existing buildings is characterized by: The following steps are involved: Collect data from existing buildings and build BIM models. Simulate the daily usage of existing buildings and divide the BIM models into different areas to obtain noise data sets for the target areas. Determine whether the target area is noise-polluted based on the noise dataset of the target area. If noise pollution exists, the area is divided into a polluted area, and the noise datasets of all polluted areas are aggregated to obtain a pollution information set. A cluster analysis operation is performed based on the pollution information set to obtain a pollution level table for the target area. The pollution level table for the target area is then fed into a pre-trained machine learning model to generate a comprehensive pollution coefficient. Based on the comprehensive pollution coefficient, the average noise or maximum noise is used as the optimization target. The optimization objectives of the target area are simulated in the established BIM model, and the genetic algorithm is used to find the optimal parameter combination that minimizes the cost of achieving the optimization objective. The optimized parameter combinations of all polluted areas are summarized to obtain the optimal design scheme for the entire building. Feeding the pollution level table of the target area into the pre-trained machine learning model to generate a comprehensive pollution coefficient means: The pre-trained machine learning models are: ; n represents the total number of noise level region clusters, i represents the index number of the noise level region cluster, represents the total number of data in the i-th noise level region cluster, represents the interval level corresponding to the i-th noise level region cluster, represents the preset influence coefficient corresponding to the i-th noise level area cluster. All influence coefficients are positive and their sum is one. represents the comprehensive pollution coefficient; Using average noise or maximum noise as the optimization target based on the comprehensive pollution coefficient means: The comprehensive pollution coefficient is compared with the preset pollution threshold. If the comprehensive pollution coefficient is greater than the preset pollution threshold, the maximum noise of the target area is used as the optimization target. If the comprehensive pollution coefficient is less than or equal to the preset pollution threshold, the average noise of the target area is used as the optimization target. Using genetic algorithms to find the optimal parameter combination that minimizes the cost of achieving the optimization goal refers to: Encoding and initial population: Encode the optimization parameter combination into chromosome form and randomly generate m chromosomes as the initial population; Fitness evaluation: Taking the optimization goal as the constraint condition in fitness evaluation, the fitness evaluation formula is established: ; Indicates optimized material type The amount of money spent, m represents the type of optimized material The total number of species, is the fitness value; Selection operation: Use the roulette wheel selection method to select children as new parents; Crossover operation: randomly exchange data in different parent chromosomes; Mutation operation: randomly select data from different offspring chromosomes for adjustment; Iteration and termination conditions: When the pre-set termination conditions are reached, the chromosome with the highest fitness is selected from the final population for decoding to obtain the optimal optimization parameter combination.

2. The BIM-based parametric design method for existing buildings according to claim 1, characterized in that: The noise data set of the target area consists of multiple noise data groups, each noise data group corresponds to a time period of data collection, and all time periods have no time overlap and cover the entire preset sampling window.

3. The BIM-based parametric design method for existing buildings according to claim 2, characterized in that: Judging whether a target area is noise-polluted based on its noise dataset refers to: The evaluation is performed with reference to the preset noise standard, and the noise data measured in the target area is compared with the preset noise standard. If the noise data measured in the target area is higher than the preset noise standard, it indicates that there is noise pollution. If the noise data measured in the target area is not higher than the preset noise standard, it indicates that there is no noise pollution.

4. The BIM-based parametric design method for existing buildings according to claim 3 is characterized in that: Cluster analysis based on contaminated information sets refers to: Step 1: Determine the number of cluster centers, i.e., the K value; Step 2: Initialize cluster centers: randomly select K initial cluster centers, each cluster center represents a noise level interval; Step 3: Calculate the distance between each region's data point and the cluster center: Use the Euclidean distance formula to calculate the distance between each region's noise measurement value and each cluster center; Step 4: Classify data points: Classify the noise data points in each area to the nearest cluster center to form multiple noise level area clusters, each area cluster represents a noise interval; Step 5: Update the cluster center: Calculate the average value of all data points in each noise region cluster and update the cluster center of each cluster; Step 6. Iterate until convergence: Repeat steps 2 to 5 until the cluster center no longer changes or the preset number of iterations is reached.

5. The BIM-based parametric design method for existing buildings according to claim 4 is characterized in that: Determining the number of cluster centers, that is, the K value, refers to: There are multiple preset interval levels for representing noise levels, and the total number of interval levels is the number of cluster centers.

6. The BIM-based parametric design method for existing buildings according to claim 5, characterized in that: The pollution level table for the target area includes: The interval level corresponding to each noise level region cluster and the total number of data in each noise level region cluster.

Images