A method and system for estimating the order of magnitude of the parameters of a building resistance-capacitance model

Abstract

The application belongs to the technical field of energy management, and discloses a building resistance-capacitance model parameter order-of-magnitude estimation method and system, which comprises the following steps: clustering and constructing a building geometric size dataset for the obtained urban building contour information dataset; simulating each type of building according to the clustering results and the geometric size dataset to obtain training data of the corresponding electrical equivalent resistance-capacitance (RC) model; setting a parameter order-of-magnitude range according to the training data, performing three rounds of iterative screening on the parameter order-of-magnitude in the parameter order-of-magnitude range based on a genetic algorithm, and obtaining a parameter order-of-magnitude combination corresponding to each type of building; constructing an RC model parameter order-of-magnitude prediction model based on the parameter order-of-magnitude combination; inputting the geometric information of a target building into the parameter order-of-magnitude prediction model that has completed optimization verification, and outputting a target parameter order-of-magnitude of the RC model corresponding to the target building; and the application can quickly obtain the parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

Description

Technical Field

[0001] This invention relates to the field of energy management technology, and in particular to a method and system for predicting the order of magnitude of parameters in a building resistance-capacity model. Background Technology

[0002] Renewable energy sources such as wind power and photovoltaics have a certain degree of randomness and volatility, which affects the stability and reliability of power supply systems.

[0003] Virtual power plants, as a new type of energy management system, not only enhance the flexibility and stability of the power grid but also address the supply-demand imbalance in the power system, promoting efficient operation and improved reliability. Building air conditioning systems are a crucial component of virtual power plants, fully utilizing the building's thermal storage capacity or dedicated energy storage equipment to flexibly adjust the timing and power consumption of air conditioning loads. Typically, the thermal dynamics of a building are analyzed and simulated through the development and utilization of electrical equivalent resistance-capacitance (RC) models.

[0004] In existing technologies, using the electrical equivalent resistance-capacitance (ERC) model to reflect the impact of building structures on temperature changes requires the identification of ERC model parameters. However, this identification process is typically time-consuming. This is mainly because when using genetic algorithms to optimize model parameters, finding the global optimum requires setting a reasonable and effective order of magnitude for the model parameters. However, the differences in the physical properties of different buildings necessitate setting specific parameter orders of magnitude for each building during the model parameter identification process. Currently, the setting of the parameter order of the ERC model largely relies on exhaustive or empirical methods, consuming significant computational resources and time, resulting in low efficiency in model parameter identification and limiting the rapid application of the ERC model in engineering practice.

[0005] The existing technology still suffers from the problem of low efficiency in setting the order of magnitude of parameters in the electrical equivalent resistance-capacitance model. Therefore, the existing technology needs to be improved. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to provide a method and system for predicting the order of magnitude of parameters in a building resistance-capacitance model, in order to address the shortcomings of existing technologies and solve the problem of low efficiency in setting the order of magnitude of parameters in existing electrical equivalent resistance-capacitance models.

[0007] The technical solution adopted by this invention to solve the technical problem is as follows:

[0008] In a first aspect, the present invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, including:

[0009] Obtain a dataset of urban building outline information, cluster the dataset, and construct a dataset of building geometric dimensions based on the clustering results;

[0010] Based on the clustering results and the geometric size dataset, simulations are performed on each type of building to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0011] Based on the training data, the parameter order of magnitude range of the electrical equivalent resistance-capacitance model is set, and the parameter order of magnitude in the parameter order of magnitude range is iteratively screened in three rounds based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0012] Based on the aforementioned combination of parameters, a parameter order-of-magnitude prediction model is constructed for the electrical equivalent resistance-capacitance model corresponding to each type of building, and the optimization and verification of the parameter order-of-magnitude prediction model are completed.

[0013] Obtain the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has been optimized and verified, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

[0014] In one implementation, the step of clustering the urban building outline information dataset and constructing a geometric dimension dataset of buildings based on the clustering results includes:

[0015] Determine the initial number of clusters;

[0016] Based on the initial cluster number, the urban building outline information data is clustered to obtain a clustering result containing the geometric dimension data of each type of building.

[0017] A dataset of building geometry dimensions is constructed based on the clustering results.

[0018] In one implementation, the step of simulating each type of building based on the clustering results and the geometric size dataset to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building includes:

[0019] A Python script is created based on the clustering results and the geometric dimension dataset. The Python script connects to the model information files of the building environment analysis software and the building energy consumption analysis software, respectively, and is used to input the geometric dimension dataset into the building environment analysis software and the building energy consumption analysis software to realize automatic simulation of each type of building.

[0020] Execute the Python script to obtain the training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0021] In one implementation, the three-round iterative screening includes a first round of screening, which includes:

[0022] By arranging and combining the order of magnitude of parameters within the range of parameter order of magnitude, we can obtain the combination of parameter order of magnitude for each type of building.

[0023] The genetic algorithm is used to perform a first preset number of operations on the combination of the order of magnitude of the parameters, and the result of each operation is used as the first operation result.

[0024] Calculate the root mean square error of the first operation result, discard the first operation result whose root mean square error does not meet the preset root mean square error threshold, and use the combination of the order of magnitude of the parameters corresponding to the retained first operation result as the first round of screening result.

[0025] In one implementation, the three-round iterative screening includes a second round of screening, which includes:

[0026] The genetic algorithm performs a second preset number of operations on the combination of parameters of the order of magnitude in the first round of screening results, and the result of each operation is used as the second operation result.

[0027] Calculate the root mean square error of the second operation result, and remove the second operation results whose root mean square error does not meet the preset root mean square error threshold to obtain all the second operation results of the parameter order of magnitude combination;

[0028] The parameter combinations whose number of second operation results does not meet the preset threshold for the number of second operation results are removed, and the remaining parameter combinations are used as the second round of screening results.

[0029] In one implementation, the three-round iterative screening includes a third round of screening, which includes:

[0030] The parameter order-of-magnitude combinations in the second round of screening results are screened in a preset stage; wherein, the same screening steps are performed in each stage, the parameter order-of-magnitude combinations to be screened in the first stage are defined as the parameter order-of-magnitude combinations in the second operation result, and the screening results of each remaining stage are used as the parameter order-of-magnitude combinations to be screened in the next stage.

[0031] The third round of screening ends when the screening result of any stage meets either the first preset condition or the second preset condition; the first preset condition is that the number of parameter order-of-magnitude combinations meets a preset number threshold; the second preset condition is that the extreme values ​​of parameter order-of-magnitude combinations meet a preset extreme value threshold.

[0032] In one implementation, the screening steps at any stage of the third round of screening include:

[0033] Obtain the order of magnitude combination of parameters to be screened, and perform a third preset number of operations on the order of magnitude combination of parameters to be screened using a genetic algorithm, and take the result of each operation as the third operation result;

[0034] Calculate the root mean square error of the third operation result, sort the root mean square errors of all the third operation results in ascending order, and take the root mean square errors of the first preset percentage as the first sample set.

[0035] The first sample set is then subjected to extreme value screening, extreme value ratio screening, and mean absolute error screening in sequence to obtain the screening results.

[0036] Secondly, the present invention provides a system for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising:

[0037] The building data acquisition module is used to acquire a dataset of urban building outline information, cluster the dataset of urban building outline information, and construct a dataset of building geometric dimensions based on the clustering results.

[0038] The simulation module is used to simulate each type of building based on the clustering results and the geometric size dataset, and to obtain the training data of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0039] The iterative screening module is used to set the parameter order of magnitude range of the electrical equivalent resistance-capacitance model according to the training data, and to perform three rounds of iterative screening of the parameter order of magnitude within the parameter order of magnitude range based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0040] The model building module is used to construct a parameter order-of-magnitude prediction model of the electrical equivalent resistance-capacitance model corresponding to each type of building based on the parameter order-of-magnitude combination, and to complete the optimization and verification of the parameter order-of-magnitude prediction model.

[0041] The result output module is used to acquire the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has completed optimization and verification, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

[0042] Thirdly, the present invention provides a terminal, comprising: a processor and a memory, wherein the memory stores a prediction program for the order of magnitude of building resistance-capacitance model parameters, and the prediction program for the order of magnitude of building resistance-capacitance model parameters, when executed by the processor, is used to implement the operation of the prediction method for the order of magnitude of building resistance-capacitance model parameters as described in the first aspect.

[0043] Fourthly, the present invention also provides a computer-readable storage medium storing a program for estimating the order of magnitude of building resistance-capacitance model parameters. When executed by a processor, the program for estimating the order of magnitude of building resistance-capacitance model parameters is used to implement the operation of the method for estimating the order of magnitude of building resistance-capacitance model parameters as described in the first aspect.

[0044] The present invention, by employing the above technical solution, has the following effects:

[0045] This invention provides a method and system for predicting the order of magnitude of parameters in a building's resistive-capacitive (RC) model, comprising: acquiring a dataset of urban building outline information; clustering the urban building outline information dataset; constructing a geometric dimension dataset of buildings based on the clustering results; performing simulation on each type of building based on the clustering results and the geometric dimension dataset to obtain training data for the corresponding electrical equivalent resistive-capacitive (ERC) model; setting a range of parameter order of magnitude for the ERC model based on the training data; performing three rounds of iterative filtering on the parameter order of magnitude within the range of parameter order of magnitude using a genetic algorithm to obtain a combination of parameter order of magnitude for the ERC model corresponding to each type of building; constructing a parameter order of magnitude prediction model for the ERC model corresponding to each type of building based on the parameter order of magnitude combination, and completing the optimization and verification of the parameter order of magnitude prediction model; acquiring the geometric information of the target building; inputting the geometric information into the optimized and verified parameter order of magnitude prediction model; and outputting the target parameter order of magnitude of the ERC model corresponding to the target building. This invention can quickly obtain the parameter order of magnitude of the ERC model of a target building. Attached Figure Description

[0046] 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 some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0047] Figure 1 This is a flowchart of the method for predicting the order of magnitude of the building resistance-capacitance model parameters in this invention.

[0048] Figure 2 This is a schematic diagram of the 2R2C model structure used in one implementation of the present invention.

[0049] Figure 3 This is a schematic diagram of the 3R2C model structure used in one implementation of the present invention.

[0050] Figure 4 This is a schematic diagram of the 3R3C model structure used in one implementation of the present invention.

[0051] Figure 5 This is a histogram of the prediction accuracy of the 2R2C model parameters in one implementation of the present invention.

[0052] Figure 6 This is a histogram of the prediction accuracy of the 3R2C model parameters in one implementation of the present invention.

[0053] Figure 7 This is a histogram of the prediction accuracy of the 3R3C model parameters in one implementation of the present invention.

[0054] Figure 8 This is a functional schematic diagram of the terminal in one implementation of the present invention.

[0055] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0057] Exemplary methods

[0058] In existing technologies, by developing and utilizing the electrical equivalent resistance-capacitance (ERC) model to simulate and analyze the thermal dynamics of buildings, the ERC model can accurately reflect the response of building structures to temperature changes, thereby optimizing building energy efficiency and power system supply and demand scheduling. However, when constructing the ERC model, if thermal simulation data from multiple buildings is required as a basis, existing research typically necessitates modeling and processing each building individually. This process is repetitive and tedious, consuming significant manpower and time, resulting in substantial efficiency losses.

[0059] Furthermore, using the electrical equivalent RC model to reflect the impact of building structures on temperature changes requires the identification of RC model parameters. However, this identification process is usually time-consuming, mainly because when using a genetic algorithm (GA) to optimize model parameters, finding the global optimum requires setting a reasonable and effective order of magnitude of model parameters. However, due to the differences in the physical properties of different buildings, their corresponding RC model parameters also exhibit individual characteristics. This necessitates setting the parameter order of magnitude specifically for different buildings during the model parameter identification process. Currently, the parameter order of magnitude setting for the electrical equivalent RC model largely relies on exhaustive or empirical methods. These methods consume significant computational resources and time, ultimately leading to low efficiency in model parameter identification and limiting the rapid application of the electrical equivalent RC model in engineering practice.

[0060] To address the above technical problems, this invention provides a method and system for predicting the order of magnitude of parameters in a building's resistive-capacitive model. The method includes: acquiring a dataset of urban building outline information; clustering the urban building outline information dataset; constructing a geometric dimension dataset of buildings based on the clustering results; performing simulation on each type of building based on the clustering results and the geometric dimension dataset to obtain training data for the electrical equivalent resistive-capacitive model corresponding to each type of building; setting a range of parameter order of magnitude for the electrical equivalent resistive-capacitive model based on the training data; performing three rounds of iterative filtering on the parameter order of magnitude within the range of parameter order of magnitude using a genetic algorithm to obtain a combination of parameter order of magnitude for the electrical equivalent resistive-capacitive model corresponding to each type of building; constructing a parameter order of magnitude prediction model for the electrical equivalent resistive-capacitive model corresponding to each type of building based on the parameter order of magnitude combination, and completing the optimization and verification of the parameter order of magnitude prediction model; acquiring the geometric information of the target building; inputting the geometric information into the optimized and verified parameter order of magnitude prediction model; and outputting the target parameter order of magnitude of the electrical equivalent resistive-capacitive model corresponding to the target building. This invention can quickly obtain the parameter order of magnitude of the electrical equivalent resistive-capacitive model of a target building.

[0061] like Figure 1 As shown, this embodiment of the invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising the following steps:

[0062] Step S100: Obtain a dataset of urban building outline information, cluster the dataset of urban building outline information, and construct a dataset of building geometric dimensions based on the clustering results.

[0063] Specifically, in one implementation of this embodiment, step S100 includes the following steps:

[0064] Step S101: Obtain the urban building outline information dataset.

[0065] In this embodiment, urban building outline information is obtained, and the geometric dimension data of each representative building in the city is used as a subset to construct an urban building outline information dataset.

[0066] In this embodiment, the obtained urban building outline information dataset is the urban building outline information dataset of City A. The geometric dimension data of each representative building in City A is obtained by using the Geographic Information System (GIS) application ArcGIS Pro to construct the urban building outline information dataset of City A.

[0067] In this embodiment, the urban building outline information dataset is the measured data of urban buildings. By using the measured data of urban buildings, the prediction accuracy of the prediction model of the parameter order of magnitude in subsequent steps can be improved.

[0068] In this embodiment, for ease of explanation, the electrical equivalent resistance-capacitance model, i.e., the electrical equivalent RC model, is abbreviated as RC model.

[0069] Step S102: Determine the initial number of clusters.

[0070] In this embodiment, based on the decreasing trend of the total sum of squared errors (SSE) of the dataset corresponding to the contour information as the number of clusters changes, the "elbow" position where the SSE decrease rate slows down significantly is identified, and the number of clusters corresponding to this position is determined as a reasonable initial number of clusters, as shown in the following formula:

[0071] ;

[0072] in, For the sum of squared errors, This represents the initial number of clusters.

[0073] Step S103: Based on the initial cluster number, cluster the urban building outline information data to obtain clustering results containing the geometric dimension data of each type of building.

[0074] In this embodiment, the clustering algorithm used is the K-means clustering algorithm. Based on the initial number of clusters, the urban building outline information is clustered to obtain clustering results containing building geometric dimension data, i.e., clusters. Each cluster corresponds to the geometric dimension data of a representative type of building.

[0075] Step S104: Construct a dataset of building geometry dimensions based on the clustering results.

[0076] In this embodiment, a building geometry dataset containing all clustering results is constructed, which is a representative building geometry dataset.

[0077] like Figure 1 As shown, this embodiment of the invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising the following steps:

[0078] Step S200: Simulate each type of building based on the clustering results and the geometric size dataset to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0079] In this embodiment, a script is created to automatically simulate each type of building based on a geometric size dataset, thereby automatically acquiring training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0080] In this embodiment, the training data is thermal data, which includes outdoor dry-bulb temperature, indoor temperature, cooling capacity, indoor heat gain, etc. This data is used to lay a solid foundation for identifying the order of magnitude of parameters in the RC model, reduce repetitive data acquisition and processing work, and improve data preparation efficiency.

[0081] Specifically, in one implementation of this embodiment, step S200 includes the following steps:

[0082] Step S201: Create a Python script based on the clustering results and the geometric dimension dataset; the Python script connects to the model information files of the building environment analysis software and the building energy consumption analysis software respectively, and is used to input the geometric dimension dataset into the building environment analysis software and the building energy consumption analysis software to realize automatic simulation of each type of building.

[0083] In this embodiment, the clustering result is the geometric dimension data of a representative type of building. The purpose of creating the Python script is to realize the automatic simulation of each type of building.

[0084] In this embodiment, the created Python script connects to the model information files of the building environment analysis software and the building energy consumption analysis software respectively. By inputting the geometric dimension dataset into the building environment analysis software and the building energy consumption analysis software, automatic simulation of each type of building is achieved.

[0085] In this embodiment, the selected building environment analysis software is Designbuilder, and the selected building energy consumption analysis software is EnergyPlus. A Python script is created using the Python programming language to connect the model information files of Designbuilder and EnergyPlus. Specifically, the model information file is a model information IDF file.

[0086] Step S202: Execute the Python script to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0087] In this embodiment, by executing a Python script, the simulation of multiple buildings in the geometric size dataset is automatically performed, and the training data of the corresponding RC model for each type of building is obtained.

[0088] Specifically, a Python script is executed to automatically modify the model information IDF file of the building energy consumption simulation software Designbuilder. At the same time, EnergyPlus is called to simulate the energy consumption of each simulation model. Then, the training data required for the RC model of different building simulation models is automatically organized and saved, including running time, outdoor dry-bulb temperature, indoor temperature, cooling capacity, indoor heat gain, and solar radiation transmitted through windows.

[0089] like Figure 1 As shown, this embodiment of the invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising the following steps:

[0090] Step S300: Set the parameter order of magnitude range of the electrical equivalent resistance-capacitance model according to the training data, and perform three rounds of iterative screening of the parameter order of magnitude in the parameter order of magnitude range based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0091] In this embodiment, the selected RC models include multiple types. A genetic algorithm is used to automatically identify the order of magnitude of parameters of multiple RC models. Based on the training data, the corresponding range of the order of magnitude of parameters of the RC models is set, and the genetic algorithm is iterated. According to the iteration results and the identification process, the order of magnitude of parameters of multiple buildings and multiple RC models is determined.

[0092] In this embodiment, the RC model includes the 2R2C model, the 3R2C model, and the 3R3C model.

[0093] Specifically, in one implementation of this embodiment, step S300 includes the following steps:

[0094] Step S301: Set the order of magnitude range of the parameters of the electrical equivalent resistance-capacitance model according to the training data;

[0095] In this embodiment, the selected RC model includes various types, such as Figure 2 The diagram shown is a schematic of the 2R2C model structure used in this embodiment. The 2R2C model mainly considers the heat transfer and heat storage / release effects between the outdoor environment, the external envelope (including windows), and the indoor air. The corresponding formula for this model is as follows:

[0096] ;

[0097] ;

[0098] ;

[0099] ;

[0100] in, Heat transfer between the external envelope and indoor air, measured in W; Heat transfer between outdoor air and the building envelope, expressed in W; Indoor air temperature unit, in °C; For the equivalent temperature node of the external envelope, the unit is °C; Outdoor air temperature is the unit of temperature (°C). The equivalent resistance of indoor air is expressed in Ω. The equivalent resistance of the outer casing structure is expressed in Ω. Cooling capacity of a building's indoor air conditioning system per unit time, expressed in W; The indoor heat gain per unit time in a building, measured in W; Time, in seconds; The equivalent capacitance of indoor air is expressed in F (F). This is the equivalent capacitance of the outdoor building envelope, expressed in F.

[0101] like Figure 3 The diagram shown is a schematic of the 3R2C model used in this embodiment. The 3R2C model mainly considers the heat transfer and heat storage / release effects between the outdoor environment, the exterior wall envelope, the exterior glass windows, and the indoor air. The corresponding formula for this model is as follows:

[0102] ;

[0103] ;

[0104] ;

[0105] ;

[0106] ;

[0107] in, Heat transfer between outdoor air and windows, measured in W; The equivalent resistance of the exterior wall glass window is expressed in Ω.

[0108] like Figure 4 The diagram shown is a schematic of the 3R3C model structure used in this embodiment. The 3R3C model mainly considers the heat transfer and heat storage / release effects between the outdoor environment, the external envelope (including windows), indoor air, and indoor thermal mass. The corresponding formula for this model is as follows:

[0109] ;

[0110] ;

[0111] ;

[0112] ;

[0113] ;

[0114] ;

[0115] in, Heat transfer between indoor air and indoor thermal mass, measured in W; Indoor thermogravimetric temperature node, unit: °C; The equivalent resistance of indoor thermal mass is expressed in Ω. The total amount of solar radiation transmitted through the window, expressed in W. The equivalent capacitance of the indoor thermal mass is expressed in F.

[0116] In this embodiment, the parameter range of the RC model is set according to the training data. Specifically, a suitable RC model is selected for each type of building, the training data of the building is input into the selected RC model, and the corresponding parameter range is set.

[0117] In this embodiment, a genetic algorithm is used to perform three rounds of iterative screening on the order of magnitude of the parameters within the range of parameter magnitudes, including a first round of screening, a second round of screening, and a third round of screening; wherein, the first round of screening is an initial screening, the second round of screening is a robust screening, and the third round of screening is a phased progressive screening.

[0118] Step S302: Perform the first round of screening of the parameter magnitude range based on the genetic algorithm.

[0119] In this embodiment, the first round of screening is the initial screening, which obtains the set parameter range for all buildings, obtains all parameter ranges, and performs initial screening on these parameter ranges.

[0120] Specifically, in one implementation of this embodiment, step S302 includes the following steps:

[0121] Step S3021: Arrange and combine the order of magnitude of parameters within the range of parameter order of magnitude to obtain the combination of parameter order of magnitude for each type of building.

[0122] In this embodiment, all parameter orders of magnitude within the obtained parameter order of magnitude range are arranged and combined to generate all possible combinations of parameter order of magnitude. Each combination of parameter order of magnitude corresponds to a combination of parameter order of magnitude for a type of building, but a type of building can correspond to multiple combinations of parameter order of magnitude, thereby obtaining the combination of parameter order of magnitude for each type of building.

[0123] Step S3022: Perform a first preset number of operations on the combination of parameters using a genetic algorithm, and use the result of each operation as the first operation result.

[0124] In this embodiment, the first preset number of times is 1, and the genetic algorithm is performed once on all combinations of parameters of the same order of magnitude, and the result of this operation is taken as the first operation result.

[0125] Step S3023: Calculate the root mean square error of the first operation result, discard the first operation result whose root mean square error does not meet the preset root mean square error threshold, and use the combination of the order of magnitude of the parameters corresponding to the retained first operation result as the first round of screening result.

[0126] In this embodiment, the root mean square error of the first calculation result is calculated using the following formula:

[0127] ;

[0128] in, This is the root mean square error.

[0129] In this embodiment, the preset root mean square error threshold is 0.3. The first calculation result whose root mean square error does not meet the preset root mean square error threshold is discarded, i.e., discarded. The first operation result is used to obtain the retained first operation result. The parameter order of magnitude combinations corresponding to all retained first operation results are obtained to obtain the first round of screening results. The first round of screening results is a feasible parameter order of magnitude candidate set.

[0130] In this embodiment, the first round of screening can quickly eliminate obviously unreasonable assumptions about the order of magnitude of parameters, narrow the search space, and improve the efficiency of subsequent optimization.

[0131] Step S303: Perform a second round of screening based on the combination of parameters of the order of magnitude in the first round of screening results using a genetic algorithm.

[0132] In this embodiment, the second round of screening is a robust screening, which obtains all combinations of parameter magnitudes from the results of the first round of screening, that is, obtains a candidate set of parameter magnitudes, and performs robust screening on the candidate set of parameter magnitudes.

[0133] Specifically, in one implementation of this embodiment, step S303 includes the following steps:

[0134] Step S3031: Perform a second preset number of operations on the parameter order-of-magnitude combination in the first round of screening results using a genetic algorithm, and use the result of each operation as the second operation result.

[0135] In this embodiment, the second preset number of times is 20 times. The genetic algorithm is used to perform 20 operations on all combinations of parameters in the obtained parameter order of magnitude candidate set. The result of each operation is used as the second operation result. That is, for a single parameter order of magnitude combination, there are 20 second operation results.

[0136] Step S3032: Calculate the root mean square error of the second operation result, and remove the second operation results whose root mean square error does not meet the preset root mean square error threshold to obtain all the second operation results of the parameter order of magnitude combination.

[0137] In this embodiment, the root mean square error of the second calculation result is calculated. The preset root mean square error threshold is 0.3. Second calculation results whose root mean square error does not meet the preset root mean square error threshold are discarded, i.e., discarded. The second operation result is retained. The second result of the operation.

[0138] In this embodiment, the second calculation result is further filtered using the R and C values ​​of the RC model, retaining only the values ​​in the second calculation result. The second result of the operation.

[0139] Step S3033: Remove parameter combinations whose number of second operation results does not meet the preset threshold for the number of second operation results, and use the remaining parameter combinations as the second round of screening results.

[0140] In this embodiment, the number of second operation results for each parameter order-of-magnitude combination is 20. The number of second operation results retained for each parameter order-of-magnitude combination after elimination in step S3032 is the number of qualified times for that parameter order-of-magnitude combination. The preset threshold for the number of second operation results is 10. Parameter order-of-magnitude combinations whose number of second operation results does not meet the preset threshold are eliminated. That is, parameter order-of-magnitude combinations whose number of qualified times in 20 genetic algorithm operations is greater than or equal to 10 are retained. All retained parameter order-of-magnitude combinations are used as the second round of screening results, i.e., the robust set.

[0141] In this embodiment, a robust set can be obtained through the second round of screening. By statistically analyzing the distribution of repeated experiments, the stability of the model under random initial values ​​and population perturbations can be verified, which helps to eliminate solutions with strong randomness.

[0142] Step S304: Perform a third round of screening based on the combination of parameters of the second round of screening results using a genetic algorithm.

[0143] In this embodiment, the third round of screening is a phased and progressive dynamic screening that obtains all parameter order-of-magnitude combinations in the robust set and performs phased and progressive screening on these parameter order-of-magnitude combinations.

[0144] Specifically, in one implementation of this embodiment, step S304 includes the following steps:

[0145] Step S3041: Perform preset stage screening on the parameter order-of-magnitude combinations in the second round screening results, i.e., the robust set; wherein, the same screening steps are performed in each stage, the parameter order-of-magnitude combinations to be screened in the first stage are defined as the parameter order-of-magnitude combinations in the second operation results, and the screening results of each remaining stage are used as the parameter order-of-magnitude combinations to be screened in the next stage.

[0146] The screening steps at each stage include:

[0147] Step a: Obtain the order of magnitude combination of parameters to be screened, and perform a third preset number of operations on the order of magnitude combination of parameters to be screened using a genetic algorithm, and take the result of each operation as the third operation result.

[0148] In this embodiment, the order of magnitude combination of the parameters to be screened in the first stage comes from the robust set obtained from the second screening, while the order of magnitude combination of the parameters to be screened in the second stage and any subsequent stage comes from the screening result of the previous stage.

[0149] In this embodiment, the third preset number of times is 200 times. The genetic algorithm is used to perform 200 operations on the obtained combinations of parameters to be screened. The result of each operation is used as the third operation result. That is, for each combination of parameters to be screened, there are 200 third operation results.

[0150] Step b: Calculate the root mean square error of the third operation result, sort the root mean square errors of all the third operation results in ascending order, and take the root mean square errors of the first preset percentage as the first sample set.

[0151] In this embodiment, the preset percentage is 25%, or 1 / 4, for calculating the root mean square error of the third calculation result. All third operation results are calculated based on their root mean square error. The values ​​are sorted in ascending order, and the R and C values ​​of the corresponding architectural RC models for the third calculation result are recorded. Based on the principles of small error and appropriate sample size, the values ​​are then sorted from the results. Take the first 1 / 4 As the first sample set, denoted as Each first sample in the first sample set corresponds to a combination of orders of magnitude of the parameters to be screened for a type of building, and a first sample set is generated at each stage for the three-level screening in subsequent steps.

[0152] Step c: Perform extreme value screening, extreme value ratio screening, and mean absolute error screening on the first sample set in sequence to obtain the screening results.

[0153] In this embodiment, the first sample set is processed sequentially. The screening process involves three layers: extreme value screening, extreme value ratio screening, and mean absolute error screening, to obtain the screening results.

[0154] Specifically, in one implementation of this embodiment, step c includes the following steps:

[0155] Step c1: Perform extreme value filtering on the first sample in the first sample set to obtain the second sample set.

[0156] In this embodiment, extreme value filtering is performed on the first sample in the first sample set, i.e., elimination. The first sample contains a lot of extreme values. Samples with a lot of extreme values ​​have poor numerical stability and need to be eliminated.

[0157] In this embodiment, if there exists any parameter column in the combination of parameter orders of magnitude corresponding to the first sample... If a parameter column is deemed unqualified, then that parameter column is considered unqualified; for any combination of parameter numbers, if the number of samples corresponding to the unqualified parameter column is greater than [a certain value], then [the parameter column is considered unqualified]. When that happens, remove the parameter order of magnitude combination corresponding to that parameter column.

[0158] In this embodiment, the first sample corresponding to the retained combination of parameter orders of magnitude is denoted as the second sample, thus obtaining the second sample set. .

[0159] Step c2: Perform extreme value ratio screening on the second samples in the second sample set to obtain the third sample set.

[0160] In this embodiment, extreme value ratio screening is performed on the second samples in the second sample set, i.e., elimination. The second sample whose extreme value ratio does not meet the preset extreme value ratio threshold, or whose extreme value ratio fluctuates too much, needs to be eliminated if its corresponding parameter order of magnitude combination is not met.

[0161] In this embodiment, if there exists an extreme value ratio of any parameter series in the combination of parameter orders of magnitude corresponding to the second sample... If the ratio exceeds the preset extreme value threshold, the parameter column is considered unqualified, and the corresponding parameter order of magnitude combination is removed; the second extreme value threshold is 10.

[0162] In this embodiment, the second sample corresponding to the retained parameter order-of-magnitude combination is denoted as the third sample, thus obtaining the third sample set. .

[0163] Step c3: Perform mean absolute error screening on the third samples in the third sample set to obtain the screening results.

[0164] In this embodiment, the mean absolute error is used to filter the third samples in the third sample set. First, the mean absolute error of each column in the combination of parameter orders of magnitude corresponding to the third sample needs to be calculated. The calculation formula is as follows:

[0165] ;

[0166] in, This represents the mean absolute error.

[0167] In this embodiment, for any combination of parameters of any magnitude, if any parameter column exists in the combination of parameters of any magnitude, its If the value is greater than the mean absolute error threshold, it means that most of the values ​​in the parameter series have a large average deviation from the mean. The combination of the corresponding parameter series is considered an unstable combination and is removed. The mean absolute error threshold is 150.

[0168] In this embodiment, the order of magnitude combination of parameters corresponding to the retained third sample is obtained to obtain the screening result containing all retained order of magnitude combinations of parameters.

[0169] Step S3042: When the screening result of any stage meets either the first preset condition or the second preset condition, the third round of screening ends; the first preset condition is that the number of parameter order-of-magnitude combinations meets a preset number threshold; the second preset condition is that the extreme values ​​of parameter order-of-magnitude combinations meet a preset extreme value threshold.

[0170] In this embodiment, when the screening result of any stage meets either the first preset condition or the second preset condition, the third round of screening ends; otherwise, the screening result is used as the order of magnitude combination of the parameters to be screened in the next stage, and the three-layer screening of the next stage is started.

[0171] In this embodiment, for the screening results of any stage, the total number of extreme value samples of the combined order of magnitude of the retained parameters is calculated. And define two robustness indicators.

[0172] The first indicator is the maximum mean absolute error of the combination of parameter orders of magnitude. This is used to measure the overall dispersion of combinations of orders of magnitude of this parameter, and the corresponding calculation formula is as follows:

[0173] ;

[0174] The second indicator is the minimum safe distance between the mean of each parameter in the combination of parameter orders of magnitude and the boundary. This is used to assess physical plausibility, and the corresponding calculation formula is as follows:

[0175] ;

[0176] Based on the two robustness indicators mentioned above and the total number of extreme value samples of each parameter order of magnitude retained. The optimal combination of parameter orders of magnitude is further determined by the first and second preset conditions.

[0177] In this embodiment, the first preset condition is that the number of parameter order-of-magnitude combinations meets a preset quantity threshold.

[0178] Specifically, the preset quantity threshold is 1, meaning that if there is only one combination of parameter order of magnitude in the screening results of any stage, then that combination of parameter order of magnitude is selected as the optimal combination of parameter order of magnitude.

[0179] In this embodiment, the second preset condition is that the extreme values ​​of the parameter order of magnitude combination meet the preset extreme value threshold.

[0180] Specifically, if the screening result at any stage does not meet the first preset condition, or if the screening result at any stage contains a combination of more than one parameter order of magnitude, the second preset condition is used to judge the screening result. The second preset condition includes a preset extreme value threshold. .

[0181] In this embodiment, if all parameter order-of-magnitude combinations in the screening results satisfy the condition that the total number of extreme values ​​is greater than the extreme value threshold, then the screening results will be used as the parameter order-of-magnitude combinations to be screened in the next stage, and the three-layer screening in the next stage will be started.

[0182] If any one or more parameter combinations in the filtering results satisfy the condition that the total number of extreme values ​​is less than the extreme value threshold, then the corresponding one or more parameter combinations are retained.

[0183] Furthermore, when the number of parameters retained is a single combination, it is taken as the optimal number of parameters.

[0184] When there are multiple combinations of the order of magnitude of the parameters to be retained, the combinations are further filtered based on the first and second indicators, and the filtering rules are as follows:

[0185] when When, select The combination of parameter orders of magnitude corresponding to the maximum value is the optimal combination of parameter orders of magnitude.

[0186] when At that time, if Select The minimum value corresponds to the optimal combination of parameter orders of magnitude.

[0187] when At that time, if The optimal parameter order of magnitude combination is selected by minimizing the extreme values.

[0188] Step S305 yields the order of magnitude combination of parameters for the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0189] In this embodiment, the optimal combination of parameter magnitudes is used as the combination of parameter magnitudes for the RC model of the corresponding building category, thereby obtaining the combination of parameter magnitudes for the RC model corresponding to each building category.

[0190] In this embodiment, by identifying the order of magnitude of various RC model parameters, 69 sets of representative building geometry corresponding to the order of magnitude of three RC electrical equivalent model parameters are finally obtained, which serve as the data basis for the prediction model.

[0191] like Figure 1 As shown, this embodiment of the invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising the following steps:

[0192] Step S400: Based on the combination of the order of magnitude of the parameters, construct the order of magnitude prediction model of the electrical equivalent resistance-capacitance model corresponding to each type of building, and complete the optimization and verification of the order of magnitude prediction model of the parameter.

[0193] In this embodiment, based on the optimal parameter order-of-magnitude combination for each type of building, order-of-magnitude prediction models for various electrical equivalent RC model parameters are constructed. A unified feature construction, logarithmic transformation, and order-of-magnitude determination process is adopted for all three electrical equivalent RC models: 2R2C, 3R2C, and 3R3C. First, based on the building's geometric information (length, width, height), two derived features, volume and surface area, are constructed as input features. Then, a log10 spatial transformation is performed on all RC parameters, mapping the original physical quantities to the form of "significant logarithmic value + order of magnitude," transforming the order of magnitude of each parameter into a learnable natural number. Subsequently, a nested cross-validation framework consisting of an outer 5-fold repeated layer and an inner 3-fold layer is used to independently perform hyperparameter grid search for each target parameter and select the optimal model, ensuring the objectivity and generalization ability of the model selection process. After determining the optimal model, the predicted value is inversely transformed from the logarithmic space to the original space and written in scientific notation. By simultaneously checking whether the significant digits of the predicted value fall within the interval [1, 1000) and whether its exponent part is consistent with the true order of magnitude, it is determined whether the predicted order of magnitude hits the true order of magnitude. Finally, the order of magnitude prediction accuracy of each RC parameter on the training set and the test set is obtained.

[0194] In this embodiment, the order of magnitude of parameters from three RC electrical equivalent models corresponding to 69 representative building geometries is used as the data basis for the prediction model. The prediction results are as follows: Figure 5 ,Figure 6 , Figure 7 As shown, where, Figure 5 Histogram of prediction accuracy for the 2R2C model with parameters of varying magnitudes; Figure 6 Histogram of prediction accuracy for the 3R2C model parameters by order of magnitude; Figure 7 The histogram shows the prediction accuracy of the 3R3C model parameters by order of magnitude.

[0195] It should be noted that, Figure 5 , Figure 6 , Figure 7 In the figure, the horizontal axis represents the electrical equivalent RC parameters to be predicted by different RC models, and the overall prediction result obtained by averaging the sample sizes of all parameters; the vertical axis represents the order-of-magnitude prediction accuracy of the corresponding parameter on the test set, in percentage; each bar represents the order-of-magnitude hit rate of the order-of-magnitude prediction model proposed in this invention on the corresponding parameter, and the vertical line above represents the 95% confidence interval calculated based on the binomial distribution, which is used to reflect the statistical stability of the prediction performance of the parameter.

[0196] from Figure 5 , Figure 6 , Figure 7 It can be seen that the order-of-magnitude prediction performance of each RC parameter under the prediction method of the order of magnitude of the building resistance-capacitance model parameters is good.

[0197] Among them, the 2R2C order-of-magnitude prediction model has electrical equivalent RC parameters , , , The overall prediction accuracy reached 100%, with corresponding 95% confidence intervals of [100.00%, 100.00%], indicating that the prediction results are highly stable.

[0198] 3R2C order-of-magnitude prediction model in , , , , The overall prediction accuracy also reached 100%, and the 95% confidence interval remained at [100.00%, 100.00%], further verifying the consistency and stability of the model.

[0199] For the higher-dimensional 3R3C model, its electrical equivalent RC parameters , , , , , The overall prediction accuracy was 92.1%, 88.2%, 89.1%, 99.1%, 98.5%, 91.5%, and 93.1%, respectively, with corresponding 95% confidence intervals of [89.18%, 94.93%], [84.81%, 91.66%], [85.81%, 92.43%], [98.12%, 100.00%], [97.25%, 99.81%], [88.50%, 94.44%], and [90.57%, 95.35%]. These confidence intervals all exhibited relatively narrow ranges, indicating that the model's prediction results across different parameter dimensions have strong statistical credibility and stability.

[0200] In summary, from the perspective of model prediction performance, the parameter order-of-magnitude prediction models constructed for the three electrical equivalent RC models of 2R2C, 3R2C and 3R3C all show high and stable prediction capabilities.

[0201] like Figure 1 As shown, this embodiment of the invention provides a method for predicting the order of magnitude of parameters in a building resistance-capacitance model, comprising the following steps:

[0202] Step S500: Obtain the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has completed optimization and verification, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

[0203] In this embodiment, the geometric information of the target building is obtained, the geometric information is input into the optimized and verified RC model, and the target parameter order of magnitude of the predicted target building RC model is output. The output result is the target parameter order of magnitude of multiple RC models. The most suitable target parameter order of magnitude of the RC model can be selected according to the geometric information of the target building.

[0204] This embodiment achieves the following technical effects through the above technical solution:

[0205] This invention also provides a method for estimating the order of magnitude of parameters in a building resistance-capacitance (RC) model, enabling efficient acquisition of thermal data from building simulation models. By using Python code to connect to the building energy consumption simulation software Designbuilder and EnergyPlus, efficient and automated building simulation and acquisition of corresponding RC model training data are achieved. This allows for the rapid acquisition of RC model training data, including outdoor dry-bulb temperature, indoor temperature, cooling capacity, and indoor heat gain, providing a solid data foundation for identifying the order of magnitude of RC model parameters. This reduces repetitive data acquisition and processing work, improving data preparation efficiency.

[0206] It can also obtain the order of magnitude of parameters for multiple electrical equivalent RC models with a single click. Through a designed automated identification process for RC model parameter order of magnitude, after setting the range of parameter order of magnitude for the RC model, a genetic algorithm and automated identification process can quickly determine the order of magnitude of the RC model parameters. Based on this, by training multiple buildings and various RC models, a dataset of order of magnitude parameters for multiple electrical equivalent RC models is obtained. A machine learning model is then constructed, its hyperparameters are optimized, and model cross-validation is performed, ultimately generating a predictive model for the order of magnitude of parameters for multiple electrical equivalent RC models. This model allows for the one-click acquisition of the order of magnitude of parameters for multiple electrical equivalent RC models, significantly reducing computational resource and time consumption and improving the efficiency of RC model parameter identification.

[0207] Exemplary device

[0208] Based on the above embodiments, the present invention also provides a system for predicting the order of magnitude of building resistance-capacitance model parameters, comprising:

[0209] The building data acquisition module is used to acquire a dataset of urban building outline information, cluster the dataset of urban building outline information, and construct a dataset of building geometric dimensions based on the clustering results.

[0210] The simulation module is used to simulate each type of building based on the clustering results and the geometric size dataset, and to obtain the training data of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0211] The iterative screening module is used to set the parameter order of magnitude range of the electrical equivalent resistance-capacitance model according to the training data, and to perform three rounds of iterative screening of the parameter order of magnitude within the parameter order of magnitude range based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building.

[0212] The model building module is used to construct a parameter order-of-magnitude prediction model of the electrical equivalent resistance-capacitance model corresponding to each type of building based on the parameter order-of-magnitude combination, and to complete the optimization and verification of the parameter order-of-magnitude prediction model.

[0213] The result output module is used to acquire the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has completed optimization and verification, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

[0214] Based on the above embodiments, the present invention also provides a terminal, the principle block diagram of which can be as follows: Figure 8 As shown.

[0215] The terminal includes: a processor, a memory, an interface, a display screen, and a communication module connected via a system bus; wherein, the processor of the terminal provides computing and control capabilities; the memory of the terminal includes a computer-readable storage medium and internal memory; the computer-readable storage medium stores an operating system and computer programs; the internal memory provides an environment for the operation of the operating system and computer programs in the computer-readable storage medium; the interface is used to connect to external devices; the display screen is used to display relevant information; and the communication module is used to communicate with a cloud server or other devices.

[0216] When executed by a processor, this computer program is used to implement a method for estimating the order of magnitude of parameters in a building resistance-capacitance model.

[0217] It will be understood by those skilled in the art that Figure 8 The schematic diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0218] In one embodiment, a terminal is provided, comprising: a processor and a memory, the memory storing a prediction program for building resistance-capacitance model parameters on the order of magnitude, the prediction program for building resistance-capacitance model parameters being executed by the processor to implement the above-described prediction method for building resistance-capacitance model parameters on the order of magnitude.

[0219] In one embodiment, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a prediction program for building resistance-capacitance model parameters on the order of magnitude, the prediction program for building resistance-capacitance model parameters being executed by a processor for implementing the above-described method for predicting building resistance-capacitance model parameters on the order of magnitude.

[0220] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, database, or other media used in the embodiments provided by this invention can include both non-volatile and volatile memory.

[0221] In summary, this invention provides a method and system for predicting the order of magnitude of parameters in a building's resistive-capacitive model, comprising: acquiring a dataset of urban building outline information; clustering the urban building outline information dataset; constructing a geometric dimension dataset of buildings based on the clustering results; performing simulation on each type of building based on the clustering results and the geometric dimension dataset to obtain training data for the electrical equivalent resistive-capacitive model corresponding to each type of building; setting a range of parameter order of magnitude for the electrical equivalent resistive-capacitive model based on the training data; performing three rounds of iterative filtering on the parameter order of magnitude within the range of parameter order of magnitude based on a genetic algorithm to obtain a combination of parameter order of magnitude for the electrical equivalent resistive-capacitive model corresponding to each type of building; constructing a parameter order of magnitude prediction model for the electrical equivalent resistive-capacitive model corresponding to each type of building based on the parameter order of magnitude combination, and completing the optimization and verification of the parameter order of magnitude prediction model; acquiring the geometric information of the target building; inputting the geometric information into the optimized and verified parameter order of magnitude prediction model; and outputting the target parameter order of magnitude of the electrical equivalent resistive-capacitive model corresponding to the target building. This invention can quickly obtain the parameter order of magnitude of the electrical equivalent resistive-capacitive model of the target building.

[0222] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A method for predicting the order of magnitude of parameters in a building resistance-capacitance model, characterized in that, include: Obtain a dataset of urban building outline information, cluster the dataset, and construct a dataset of building geometric dimensions based on the clustering results; Based on the clustering results and the geometric size dataset, simulations are performed on each type of building to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building. Based on the training data, the parameter order of magnitude range of the electrical equivalent resistance-capacitance model is set, and the parameter order of magnitude in the parameter order of magnitude range is iteratively screened in three rounds based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building. Based on the aforementioned combination of parameters, a parameter order-of-magnitude prediction model is constructed for the electrical equivalent resistance-capacitance model corresponding to each type of building, and the optimization and verification of the parameter order-of-magnitude prediction model are completed. Obtain the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has been optimized and verified, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building; The three-round iterative screening includes a first round of screening, a second round of screening, and a third round of screening; wherein, the first round of screening is an initial screening, the second round of screening is a robust screening, and the third round of screening is a phased progressive screening; In the third round of screening, the screening steps at any stage include: Obtain the order of magnitude combination of parameters to be screened, and perform a third preset number of operations on the order of magnitude combination of parameters to be screened using a genetic algorithm, and take the result of each operation as the third operation result; Calculate the root mean square error of the third operation result, sort the root mean square errors of all the third operation results in ascending order, and take the root mean square errors of the first preset percentage as the first sample set. The first sample set is then subjected to extreme value screening, extreme value ratio screening, and mean absolute error screening in sequence to obtain the screening results.

2. The method for predicting the order of magnitude of building resistance-capacitance model parameters according to claim 1, characterized in that, The process of clustering the urban building outline information dataset and constructing a geometric dimension dataset of buildings based on the clustering results includes: Determine the initial number of clusters; Based on the initial cluster number, the urban building outline information data is clustered to obtain a clustering result containing the geometric dimension data of each type of building. A dataset of building geometry dimensions is constructed based on the clustering results.

3. The method for predicting the order of magnitude of building resistance-capacitance model parameters according to claim 1, characterized in that, The step of simulating each type of building based on the clustering results and the geometric size dataset to obtain training data for the electrical equivalent resistance-capacitance model corresponding to each type of building includes: A Python script is created based on the clustering results and the geometric dimension dataset. The Python script connects to the model information files of the building environment analysis software and the building energy consumption analysis software, respectively, and is used to input the geometric dimension dataset into the building environment analysis software and the building energy consumption analysis software to realize automatic simulation of each type of building. Execute the Python script to obtain the training data for the electrical equivalent resistance-capacitance model corresponding to each type of building.

4. The method for predicting the order of magnitude of building resistance-capacitance model parameters according to claim 1, characterized in that, The three-round iterative screening includes a first round of screening, which includes: By arranging and combining the order of magnitude of parameters within the range of parameter order of magnitude, we can obtain the combination of parameter order of magnitude for each type of building. The genetic algorithm is used to perform a first preset number of operations on the combination of the order of magnitude of the parameters, and the result of each operation is used as the first operation result. Calculate the root mean square error of the first operation result, discard the first operation result whose root mean square error does not meet the preset root mean square error threshold, and use the combination of the order of magnitude of the parameters corresponding to the retained first operation result as the first round of screening result.

5. The method for predicting the order of magnitude of building resistance-capacitance model parameters according to claim 4, characterized in that, The three-round iterative screening includes a second round of screening, which includes: The genetic algorithm performs a second preset number of operations on the combination of parameters of the order of magnitude in the first round of screening results, and the result of each operation is used as the second operation result. Calculate the root mean square error of the second operation result, and remove the second operation results whose root mean square error does not meet the preset root mean square error threshold to obtain all the second operation results of the parameter order of magnitude combination; The parameter combinations whose number of second operation results does not meet the preset threshold for the number of second operation results are removed, and the remaining parameter combinations are used as the second round of screening results.

6. The method for predicting the order of magnitude of building resistance-capacitance model parameters according to claim 5, characterized in that, The three-round iterative screening includes a third round of screening, which includes: The parameter order-of-magnitude combinations in the second round of screening results are screened in a preset stage; wherein, the same screening steps are performed in each stage, the parameter order-of-magnitude combinations to be screened in the first stage are defined as the parameter order-of-magnitude combinations in the second operation result, and the screening results of each remaining stage are used as the parameter order-of-magnitude combinations to be screened in the next stage. The third round of screening ends when the screening result of any stage meets either the first preset condition or the second preset condition; the first preset condition is that the number of parameter order-of-magnitude combinations meets a preset number threshold; the second preset condition is that the extreme values ​​of parameter order-of-magnitude combinations meet a preset extreme value threshold.

7. A system for predicting the order of magnitude of parameters in a building resistance-capacitance model, used to implement the method for predicting the order of magnitude of parameters in a building resistance-capacitance model as described in any one of claims 1-6, characterized in that, include: The building data acquisition module is used to acquire a dataset of urban building outline information, cluster the dataset of urban building outline information, and construct a dataset of building geometric dimensions based on the clustering results. The simulation module is used to simulate each type of building based on the clustering results and the geometric size dataset, and to obtain the training data of the electrical equivalent resistance-capacitance model corresponding to each type of building. The iterative screening module is used to set the parameter order of magnitude range of the electrical equivalent resistance-capacitance model according to the training data, and to perform three rounds of iterative screening of the parameter order of magnitude within the parameter order of magnitude range based on the genetic algorithm to obtain the parameter order of magnitude combination of the electrical equivalent resistance-capacitance model corresponding to each type of building. The model building module is used to construct a parameter order-of-magnitude prediction model of the electrical equivalent resistance-capacitance model corresponding to each type of building based on the parameter order-of-magnitude combination, and to complete the optimization and verification of the parameter order-of-magnitude prediction model. The result output module is used to acquire the geometric information of the target building, input the geometric information into the parameter order-of-magnitude prediction model that has completed optimization and verification, and output the target parameter order-of-magnitude of the electrical equivalent resistance-capacitance model corresponding to the target building.

8. A terminal, characterized in that, include: The processor and memory, wherein the memory stores a program for estimating the order of magnitude of building resistance-capacitance model parameters, the program for estimating the order of magnitude of building resistance-capacitance model parameters being executed by the processor to implement the operation of the method for estimating the order of magnitude of building resistance-capacitance model parameters as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program for estimating the order of magnitude of building resistance-capacitance model parameters. When executed by a processor, the program for estimating the order of magnitude of building resistance-capacitance model parameters is used to implement the operation of the method for estimating the order of magnitude of building resistance-capacitance model parameters as described in any one of claims 1-6.

Images