Method and device for predicting heating carbon emission of urban residences in cold region

Through Python language and multivariate linear regression algorithm, combined with Sobol global sensitivity analysis, a carbon emission prediction model for urban residential heating in cold areas was constructed, which solved the problem of poor applicability and accuracy of prediction models in the existing technology, and achieved more efficient design guidance and process simplification.

CN118917476APending Publication Date: 2024-11-08ZHEJIANG CONSTR INVESTMENT INNOVATION TECH CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410990323.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-08

AI Technical Summary

Technical Problem

In the prior art, the carbon emission forecast model for urban residential heating in cold areas is poor in applicability and accuracy, the dynamic calculation method is time-consuming and complex in operation, and it cannot provide architects with timely and effective design guidance.

Method used

Using Python to write steady-state calculation methods, combined with Sobol global sensitivity analysis and multivariate linear regression algorithm, build a heating carbon emission prediction model that transforms from steady-state to dynamics, simplify the basic residential model, generate case databases, analyze design parameters and meteorological parameters On the influence of calorific consumption, a conversion coefficient prediction model is established.

Benefits of technology

It improves the accuracy and applicability of heating carbon emission forecasts, provides timely and effective architectural design guidance, simplifies the residential design process, and reduces the computational complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118917476A_ABST
    Figure CN118917476A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of carbon emission prediction, solves the problems that in the prior art, a prediction model is poor in applicability and accuracy, a dynamic calculation method is long in time consumption, modeling calculation needs to be conducted on a specific platform, and operation is complex, and discloses a prediction method and device for urban residence heating carbon emission in a cold region. The method comprises the following steps: building basic residence models of different regions, calculating to obtain annual accumulated steady-state heat consumption per unit area, quantitatively analyzing and simplifying the basic residence models, calculating the simplified annual accumulated steady-state heat consumption per unit area and dynamic heat consumption per unit area, carrying out correlation analysis, and calculating the dynamic heat consumption per unit area. An optimal conversion coefficient prediction model is constructed, tested and screened, and a heating carbon emission prediction model based on residential space design parameters, enclosure thermal performance parameters and meteorological parameters is established according to the optimal conversion coefficient prediction model, so that an influence mechanism of different residential design parameters on heating carbon emission can be better disclosed; and the prediction accuracy and applicability are improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of carbon emission prediction, and in particular to a method and device for predicting carbon emissions from urban residential heating in cold regions. Background Art

[0002] Urban residences in cold regions mainly use centralized heating, which accounts for the majority of residential operation energy consumption and is also the main source of heating carbon emissions. Therefore, it is important to predict, evaluate and analyze heating carbon emissions of urban residences in cold regions during the design phase to guide the optimization of residential design.

[0003] The process of building a prediction model based on field measurements involves multi-source heterogeneous data and multi-modal data, and its collection and integration process is complex and cannot cover everything. In addition, most of the current prediction model construction processes directly establish a mapping between measured data and residential space design parameters. The heating carbon emission prediction model based on this mapping relationship ignores the basic principles of building thermal balance, which affects the applicability and accuracy of the prediction model. Secondly, according to relevant standard requirements, the dynamic energy consumption calculation method based on energy consumption simulation software replaces the steady-state calculation method based on four arithmetic operations and building design parameters. The dynamic calculation method takes a long time, requires modeling and calculation on a specific platform, is complicated to operate, and cannot provide architects with timely and effective architectural design guidance, which is not conducive to architectural creation. Summary of the invention

[0004] The purpose of this application is to overcome the problems in the prior art that the applicability and accuracy of the prediction models are poor, the dynamic calculation methods are time-consuming, and the modeling calculations and operations on a specific platform are complex, and to provide a method and device for predicting carbon emissions from urban residential heating in cold regions.

[0005] In the first aspect, a method for predicting carbon emissions from urban residential heating in cold regions is provided, comprising:

[0006] Constructing basic residential models in different regions, evolving the basic models to batch generate urban residential cases in cold regions, and constructing a case database based on the urban residential cases in cold regions;

[0007] Based on the case database, a steady-state calculation method is written in Python to calculate the annual cumulative steady-state heat consumption per unit area of ​​urban residential cases in cold regions;

[0008] The influence of different residential space design parameters, envelope structure thermal performance parameters and meteorological parameters on steady-state heat consumption is quantitatively analyzed through Sobol global sensitivity analysis, and the design parameters with low sensitivity are screened and eliminated to simplify the basic residential model;

[0009] The simplified basic model is evolved to batch generate residential cases in different preset cities in cold zones A and B, and the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​the residential cases in cold areas are calculated, where zones A and B are determined according to the zones where the cities are located in the case database;

[0010] The ratio of the steady-state heat consumption to the dynamic heat consumption is used as a conversion coefficient, and the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified residential basic model is analyzed, wherein the design parameters are a general term for the space design parameters and the thermal performance parameters of the enclosure structure;

[0011] The urban residential cases in cold regions are divided into a test set and a validation set, and a conversion coefficient prediction model based on residential space design parameters and meteorological parameters is constructed by using a multivariate linear regression algorithm in combination with the test set, and the prediction performance of the conversion coefficient prediction model is tested and evaluated through the validation set to screen out the best conversion coefficient prediction model;

[0012] A heating carbon emission prediction model based on residential design parameters and meteorological parameters is established according to the optimal conversion coefficient prediction model, and the heating carbon emission prediction model is used to predict heating carbon emissions of urban residential buildings in cold regions.

[0013] In some possible implementations, basic residential models for different regions are constructed, including: selecting a preset region, sorting out and establishing a basic residential model, integrating spatial design parameters in the basic residential model according to the basic residential model, splitting and reorganizing the spatial design parameters to form a plurality of design schemes, combining the design schemes, quantifying the sensitivity of different spatial design parameters, thermal performance parameters of the envelope structure, and meteorological parameters in the steady-state heat consumption calculation process, and simplifying the basic residential model according to the sensitivity of the residential design parameters to obtain the basic residential model.

[0014] In some possible implementations, the space design parameters include at least one of the following: number of building floors, building floor height, external wall area, number of functional rooms, building width, building depth, ventilation volume, floor area, external window information, roof information, external door information and total building area. Among them, the number of functional rooms includes the number of different functional rooms on the standard residential floor and the total number of different functional rooms in the residential building, the external window information includes the area and heat transfer coefficient of the external window, the external door information includes the area of ​​the external door and whether the external door has one door or two doors, etc., the ventilation volume is taken as 0.6 times the value submitted by the building, and the roof information includes the area and heat transfer coefficient of the roof.

[0015] In some possible implementations, writing a steady-state calculation method based on the case database includes:

[0016] Define the value ranges of space design parameters and thermal performance parameters of enclosure structures;

[0017] Generate independent unit housing space design parameters and townhouse space design parameters;

[0018] Combine the independent unit residential space design parameters and the townhouse space design parameters for inspection and screening, and add other combination parameters required for heat consumption calculation to form a cold region residential design data set;

[0019] The meteorological parameters of the target city are defined, and the residential design data sets in cold regions are respectively integrated with the corresponding thermal performance parameters of the enclosure structure according to the differences in the energy-saving requirements for construction in different climate zones, so as to form residential design data sets for different preset cities in cold regions;

[0020] Calculate various heat consumption items according to the residential design data set of the preset city respectively, and integrate the calculated various heat consumption items into the residential design steady-state heat consumption data set of the target city;

[0021] Based on the residential design steady-state heat consumption dataset of the target city, the heat consumption calculation results can be output by changing the space design parameters and the thermal performance parameters of the envelope structure.

[0022] In some possible implementations, the calculation formula for the accumulated steady-state heat consumption and dynamic heat consumption per unit area throughout the year is:

[0023] SqH=0.001·24·q H t (1)

[0024] DqH=0.001·24·q H ·t / φ (2)

[0025] Among them, SqH is the cumulative steady-state heat consumption per unit area throughout the year, DqH is the cumulative dynamic heat consumption per unit area throughout the year, t is the number of heating days, and qH is the heat consumption index.

[0026] In some possible implementations, the heating carbon emission prediction model is expressed as:

[0027] y heating =1000·EF i ·SqH / (η i ·q·φ)=24·EF i ·q H ·t / (η i ·q·φ) (3)

[0029] Among them, y heating is the annual heating carbon emission per unit building area of ​​the house; ηi Indicates the comprehensive efficiency of the heating system; q is the calorific value conversion coefficient; EF i is the energy carbon emission factor.

[0030] In a second aspect, a device for predicting carbon emissions from urban residential heating in cold regions is provided, comprising:

[0031] A construction module is used to construct a basic residential model in different regions, and evolve the basic model through Python language to generate urban residential cases in cold regions, and construct a case database based on the urban residential cases in cold regions;

[0032] The first calculation module is used to write a steady-state calculation method using Python based on the case database and calculate the annual cumulative steady-state heat consumption per unit area of ​​each urban residential case in cold regions;

[0033] A simplification module is used to quantify the influence of spatial design parameters, thermal performance parameters of envelope structures and meteorological parameters on steady-state heat consumption of different houses through Sobol global sensitivity analysis, and to screen and eliminate parameters with low sensitivity to simplify the basic model of the house;

[0034] The second calculation module is used to evolve the simplified basic model to batch generate residential cases in different preset cities in cold areas A and B, and calculate the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​the urban residential cases in the cold areas;

[0035] An analysis module, used to use the ratio of the steady-state heat consumption to the dynamic heat consumption as a conversion coefficient, and to analyze the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified residential basic model, wherein the design parameters include space design parameters and thermal performance parameters of the enclosure structure;

[0036] A screening module is used to divide the urban residential cases in cold regions into a test set and a validation set, and to construct a conversion coefficient prediction model based on residential design parameters and meteorological parameters by using a multivariate linear regression algorithm in combination with the test set, and to test and evaluate the prediction performance of the conversion coefficient prediction model by using the validation set, so as to screen out the best conversion coefficient prediction model;

[0037] The prediction module is used to establish a heating carbon emission prediction model based on residential design parameters and meteorological parameters according to the optimal conversion coefficient prediction model, and use the heating carbon emission prediction model to predict the heating carbon emission of urban residential buildings in cold areas.

[0038] In a third aspect, a computer program product comprising instructions is provided, which, when executed on a computer, enables the computer to execute the method described in the first aspect.

[0039] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable medium stores a program code for execution by a device, wherein the program code includes steps for executing the method in any one of the implementations of the first aspect described above.

[0040] In a fifth aspect, an electronic device is provided, comprising a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction, when executed by the processor, implements a method as in any one of the implementations in the first aspect above.

[0041] The present application has the following beneficial effects: the present application combines Python language programming with steady-state calculation methods, compares dynamic calculation results, analyzes the sensitivity of different parameters such as urban meteorology, residential space design and thermal performance to heat consumption, establishes a method path for transformation from steady state to dynamic, and relies on a multivariate linear regression algorithm to construct a prediction model for heating carbon emissions, providing design guidance and data support for low-carbon residences in cold areas, which can better reveal the influence mechanism of different design parameters on heating carbon emissions, and improve the accuracy and applicability of the prediction method, and the present application can be consistent with the residential design process and the architect's design model, so as to provide architects with timely and effective architectural design guidance, which is conducive to architectural design creation. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] The drawings constituting a part of the present application are used to provide a further understanding of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application.

[0043] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0044] Figure 1 It is a flow chart of the method for predicting carbon emissions from urban residential heating in cold regions of the present application;

[0045] Figure 2 This is a simplified flow chart of the method for predicting carbon emissions from urban residential heating in cold regions according to Example 1 of the present application;

[0046] Figure 3 It is a flow chart of the steady-state calculation method in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0047] Figure 4It is an abstract simplified flow chart of the basic residential model in the method for predicting carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0048] Figure 5 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 1 ;

[0049] Figure 6 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 2 ;

[0050] Figure 7 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 3 ;

[0051] Figure 8 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 4 ;

[0052] Fig. 9 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 5 ;

[0053] Fig.10 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 6 ;

[0054] Fig.11 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 7 ;

[0055] Fig.12 The column chart of the sensitivity results of the calculation parameters in the heat consumption calculation in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application is Figure 8 ;

[0056] Fig.13 It is a histogram of the sensitivity of the enclosure structure area in the heat consumption calculation in the prediction method for carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0057] Fig.14The calculation model M for establishing the heat consumption index of residential buildings in cold regions in the prediction method for carbon emissions from urban residential heating in cold regions of Example 1 of the present application is hs Flowchart of

[0058] Fig.15 It is a statistical chart of the hourly usage rate of various types of equipment and the room occupancy rate of personnel in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0059] Fig.16 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 1 ;

[0060] Fig.17 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 2 ;

[0061] Fig.18 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 3 ;

[0062] Fig.19 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 4 ;

[0063] Fig. 20 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 5 ;

[0064] Fig.21 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 6 ;

[0065] Fig. 22 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 7 ;

[0066] Fig.23 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 8 ;

[0067] Fig.24 is the numerical distribution of the conversion coefficient φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application Figure 9 ;

[0068] Fig.25 It is a statistical graph of the correlation between the spatial design parameters and the meteorological parameters and φ in the prediction method for carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0069] Fig.26 is a graph of iwine, te and φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0070] Fig. 27 is a graph of S, V and φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0071] Fig.28 is the curve of Se, Ve and φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0072] Fig.29 is a curve diagram of Sw_e, Sw_s and φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0073] Fig.30 is a curve diagram of SqH, qH and φ in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0074] Fig.31 It is a statistical graph of the standardized coefficients of the independent variables of the conversion coefficient prediction model in the prediction method for carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0075] Fig.32 It is a statistical graph of a training data set of a machine learning algorithm model for predicting conversion coefficients in the method for predicting carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0076] Fig.33 is the conversion coefficient prediction R of different machine learning algorithm models in the prediction method of carbon emissions from urban residential heating in cold regions in Example 1 of the present application 2 Test result graph;

[0077] Fig.34 This is a MAE test result diagram of the conversion coefficient prediction of different machine learning algorithm models in the prediction method of carbon emissions from urban residential heating in cold areas in Example 1 of the present application;

[0078] Fig.35 This is a graph showing the MSE test results of the conversion coefficient prediction of different machine learning algorithm models in the prediction method for carbon emissions from urban residential heating in cold regions in Example 1 of the present application;

[0079] Fig.36 This is a structural block diagram of a device for predicting carbon emissions from urban residential heating in cold regions according to Example 2 of the present application;

[0080] Fig.37 It is a schematic diagram of the internal structure of the electronic device of Example 5 of the present application.

[0081] Reference numerals:

[0082] 100, construction module; 200, first calculation module; 300, simplification module; 400, second calculation module; 500, analysis module; 600, screening module; 700, prediction module. DETAILED DESCRIPTION

[0083] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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 creative work are within the scope of protection of the present invention.

[0084] Example 1

[0085] A method for predicting carbon emissions from urban residential heating in cold regions, as described in Example 1 of the present application, Figure 1As shown, it includes: constructing basic residential models in different regions, and evolving the basic model through Python language to batch generate urban residential cases in cold regions, and constructing a case database based on the urban residential cases in cold regions; using Python to write a steady-state calculation method based on the case database and calculate the annual cumulative steady-state heat consumption per unit area of ​​each urban residential case in cold regions; quantitatively analyzing the influence of different spatial design parameters and meteorological parameters on steady-state heat consumption through Sobol global sensitivity analysis, screening and eliminating design parameters with low sensitivity to simplify the basic residential model; evolving the simplified basic model to batch generate residential cases in different preset cities in cold regions, and recalculating the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​urban residential cases in cold regions; using the ratio of the steady-state heat consumption to the dynamic heat consumption as a conversion coefficient, and analyzing the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified basic residential model; The urban residential cases in cold regions are divided into a test set and a validation set. A conversion coefficient prediction model based on residential design parameters and meteorological parameters is constructed by using a multiple linear regression algorithm in combination with the test set. The prediction performance of the conversion coefficient prediction model is tested and evaluated through the validation set to screen out the best conversion coefficient prediction model. A heating carbon emission prediction model based on design parameters and meteorological parameters is established according to the best conversion coefficient prediction model. The heating carbon emission prediction model is used to predict heating carbon emissions of urban residential buildings in cold regions. This method combines Python language programming with a steady-state calculation method, compares dynamic calculation results, analyzes the sensitivity of different parameters such as urban meteorology, residential space design and thermal performance to heat consumption, establishes a method path for transformation from steady state to dynamic, and builds a heating carbon emission prediction model based on a multiple linear regression algorithm. This provides design guidance and data support for low-carbon residential buildings in cold regions, can better reveal the influence mechanism of different design parameters on heating carbon emissions, and improve the accuracy and applicability of the prediction method.

[0086] Specifically, Figure 2 A flow chart showing a method for predicting carbon emissions from urban residential heating in cold regions in Application Example 1 is shown, comprising:

[0087] Residents' energy consumption behavior has a great impact on residential energy consumption. However, in the process of residential design, architects cannot predict the future living style of residents in advance, and can only make assumptions and analyses according to ideal working conditions. Therefore, in the performance-based design of buildings, steady-state or dynamic calculation methods are still needed to provide a basis for scheme optimization. Although the results of steady-state and dynamic calculations are different, the trends of building heat consumption reflected by the two are similar. Except for high-altitude strong radiation areas, the results of steady-state and dynamic calculations of urban residences in other medium and low altitude cold areas are not much different. It can be seen that the steady-state calculation method can provide certain support for the prediction of heating energy consumption and carbon emissions of urban residences in cold areas. In addition, according to the current building energy conservation standards, heating energy consumption in cold areas needs to be calculated based on the annual cumulative heat consumption, comprehensive efficiency of heating system, standard coal calorific value and comprehensive power generation coal consumption calculated by dynamic simulation software. This standard does not take into account the impact of hourly changes in the comprehensive efficiency of the heating system. On the basis of the principle of self-comparison, it makes it more operational to implement building heating energy consumption prediction and low-carbon design evaluation from the design level. Therefore, this embodiment analyzes the relationship between urban meteorological parameters, space design parameters, thermal performance parameters of envelope structures, building heat consumption indicators, dynamic cumulative heat consumption and heating carbon emissions in cold areas in a manner of transforming from steady-state calculation to dynamic calculation.

[0088] like Figure 2 As shown in the figure, in this process, since the steady-state calculation method only involves four arithmetic operations, the Python language can be used to establish the calculation model M of the residential heat consumption index in cold areas. hs , the annual steady-state total heat consumption per unit building area of ​​different residential cases SqH is calculated in batches by "adjusting parameters". Secondly, the corresponding residential cases are modeled through DesignBuilder software simulation to calculate the annual dynamic cumulative heat consumption per unit building area DqH. The ratio φ of SqH to DqH is the coefficient for converting steady-state calculation to dynamic calculation (recorded as conversion coefficient φ). Subsequently, the influence mechanism of basic parameters such as urban meteorology, space design, thermal performance and their different combination parameters on the conversion coefficient φ is analyzed, and the conversion coefficient prediction model M is established based on this. φ Finally, M hs and M φ Combined with the above, the indirect prediction of carbon emissions from urban residential heating in cold regions can be achieved.

[0089] S100. Construction of basic model: The climate conditions and economic culture of cities in cold regions are different. In this embodiment, Taiyuan, Chengde, Lanzhou, Weifang in cold region A, and Beijing, Jinan, Xi'an, Zhengzhou in cold region B, a total of 8 cities are selected as target cities for heating carbon emission prediction analysis. The steady-state calculation method of the total heat consumption of building envelope structure (QHT) and the total heat consumption of air infiltration (QINF) per unit time is as follows: Figure 3As shown. In which, the heat consumption per unit building area per unit time is represented by q instead of Q to distinguish them.

[0090] The steady-state calculation method involves a large number of urban meteorological parameters, space design parameters and thermal performance parameters of the enclosure structure. In order to quantify the influence mechanism of various parameters on various heat consumption indicators of the residence, such as Figure 4 As shown, in this embodiment, the residential scheme is abstracted into a cubic residential basic model, and the following settings are made: 1) Residents in cold northern regions have a high demand for north-south oriented residential buildings, so the residential basic model is set to be north-south oriented; 2) Super high-rise residential buildings with a building height of more than 100m and low-rise residential buildings with 1 to 3 floors are not the main body of current urban residential construction, so the number of building floors is set to 4 to 30 floors; 3) The building floor height is set to 2.8 to 3.3m; 4) An enclosed balcony is used, and the balcony space is often connected to the main functional rooms such as the living room. The bedroom, living room, kitchen and bathroom are heated rooms, and other auxiliary rooms are treated as non-heated rooms. Therefore, the total heat consumption of air infiltration per unit building area per unit time q INF When the ventilation volume V h The value is taken as 0.60 times of the building volume V; 5) The location of the residence is set as the inland area of ​​the city, and the influence of wind force addition rate is not considered; 6) The height of the roof stairwell is set to be the same as the building floor height, the residential stairwell is placed on the north side to reduce the impact on the south-facing lighting, and the proportion of the north-facing external windows of the residential stairwell is set to be the same as the north-facing window-to-wall ratio of the residential building; 7) All external doors of the building are single-door; 8) Special spatial forms such as bay windows and projecting floor slabs are not considered.

[0091] according to Figure 4 The residential design scheme parameterization process shown in the figure reversely constructs the basic residential models of different regions, including: selecting a preset region, abstracting and summarizing and establishing a basic residential model, sorting out the spatial design parameters of the basic residential model according to the basic residential model, and splitting and reorganizing the spatial design parameters to form multiple residential design schemes.

[0092] S200, parameter sensitivity analysis:

[0093] In steady-state calculations, the ranges of values ​​for basic parameters and combined parameters (collectively referred to as calculation parameters) vary greatly. w Average heat transfer coefficient K of the external window win The value can be taken within the range of "20% improvement in thermal performance" in the "Green Building Evaluation Standard" to the "Trade-off Judgment Limit" in the "General Specification for Building Energy Conservation and Renewable Energy Utilization" (Range 1), and can also be taken between "Prescriptive Indicators" and "Trade-off Judgment Limit" (Range 2). Average roof heat transfer coefficient K r , ground average heat transfer coefficient K fThe SHGC of the external window cannot be weighed and judged, and can only be taken in the range of "thermal performance improvement of 20%" to "prescriptive index". In order to analyze the influence of different calculation parameters on the steady-state heat consumption, this embodiment uses Sobol global sensitivity analysis to quantify parameter sensitivity, and the quantitative indicators include the overall sensitivity index (ST index) and the first-order sensitivity index (S1 index).

[0094] The analysis results are as follows Figure 5-Figure 12 As shown, the average heat transfer coefficient of the building envelope and the width of the building exterior door d w and building exterior door height d h The total steady-state heat consumption per unit time of the residence is Q H Or residential heat consumption per unit building area per unit time (i.e. residential steady-state heat consumption index) q H Compared with range 1, when the average heat transfer coefficient of the exterior wall K w Average heat transfer coefficient K of the external window win When the value is within range 2, the impact on the corresponding heat consumption is smaller. For cost control considerations, simply reducing the heat transfer coefficient of the enclosure structure is not the primary strategy to improve the green performance of the building. In view of this, this embodiment uses the average heat transfer coefficient K of the exterior wall of the basic residential model as w Set to 0.45, the average heat transfer coefficient of the exterior window K win is set to 2.00, the external window comprehensive solar heat gain coefficient SHGC is set to 0.55, and the ground average heat transfer coefficient K f Set to 0.50, the average heat transfer coefficient of the roof K r Set to 0.25 (cold A zone, i.e. K ra =0.25) and 0.30 (cold B zone, i.e. K rb =0.25), the average heat transfer coefficient of the outer door K d Set to 2.00. The building exterior door is simplified to an opaque insulation door with a width of 1.10m and a height of 2.40m. Furthermore, the exterior wall area and corresponding heat consumption of the east-west heating room and stairwell, and the exterior window area and corresponding heat consumption of the east-west heating room and stairwell are all calculated for q H or Q H The impact is lower, such as Fig.13 To further simplify the basic model, this embodiment sets the window-to-wall area ratio of the east-facing residence as e Window-to-wall ratio to the west w Same as the area of ​​the east-facing window S wine Replace the west-facing window area S winw , using the east exterior wall area S we , East-facing heating room exterior wall area S wec And the exterior wall area of ​​the east staircase S wes Replacing the west exterior wall area S ww, West-facing heating room exterior wall area S wwc And the exterior wall area of ​​the west staircase S wws .

[0095] S300, steady-state calculation compilation:

[0096] In this embodiment, according to Fig.14 The process shown in the figure establishes the calculation model M of the heat consumption index of residential buildings in cold areas hs . First, define the value range of basic parameters for residential space design and thermal performance of envelope structures. Secondly, generate independent unit residential design data Ds and townhouse design data Dm. Among them, Ds traverses the basic parameters of the basic model's space design through the nested "for loop" command, records them in the array parameters, and then reconstructs them with "DataFrame". The basic model of Dm can be regarded as a number of independent unit residential buildings connected by east-west facades. The total length of the building width, the area of ​​the first floor external door of the building, and the area of ​​the external door, external wall and external window of the roof staircase are multiplied according to the number of units. For the change of east-west wall properties, such as the transformation of the external wall into the internal wall or the integration of the east-west external windows, it can be achieved by adjusting the design parameters of the east-west facade of the basic model. Subsequently, the independent unit residential design data Ds and the townhouse design data Dm are merged, tested and screened, and other combination parameters required for heat consumption calculation are supplemented to form the residential design data set Dr in cold areas. On this basis, the meteorological parameters of the target cities are defined. According to the differences in building energy-saving requirements in different climate zones, the residential design dataset Dr in cold regions is integrated with the corresponding basic parameters of thermal performance of the envelope structure to form the residential design dataset Dra in cold zone A and the residential design dataset Drb in cold zone B. Finally, the heat consumption of each item is calculated based on the data information of the residential design dataset Dra in cold zone A and the residential design dataset Drb in cold zone B, and then integrated into the residential design steady-state heat consumption dataset Dhs of the eight target cities. The residential design steady-state heat consumption dataset Dhs of the eight target cities reflects the changes in residential steady-state heat consumption under different design and climate conditions. By changing the basic parameters of residential space design and thermal performance of the envelope structure, the calculation results of each item of heat consumption can be output quickly and in batches.

[0097] S400, steady state to dynamic transformation:

[0098] According to the basic residential model, a total of 3,550 residential simulation samples were established, including 1,820 residential samples in Beijing, 380 residential samples in Jinan, 216 residential samples in Xi'an and Zhengzhou, 232 residential samples in Taiyuan and Chengde, 229 residential samples in Lanzhou, and 225 residential samples in Weifang. Since the elevator rooms and corridors in the building are not heated, the connected traffic spaces are merged in the simulation samples. The indoor space is merged and divided into bedrooms, living rooms, bathrooms, kitchens, etc., and the main functional rooms such as bedrooms and living rooms are arranged on the south side as much as possible, and each bathroom has an external window.

[0099] The basic parameters of space design are determined according to Appendix 4. Among them, the proportion of staircase protruding wall t u Q H and Q H The sensitivity of is very low, and the correlation between the two is weak. The area of ​​the exterior wall of the east-facing staircase S w_es , the exterior wall area of ​​the west staircase S w_ws , Heat consumption of the exterior wall of the east-facing staircase Q Hwalles Heat consumption of the west staircase exterior wall Q Hwallws Q H The sensitivity is very low and is related to q H There is no high correlation between them. u , S w_es With S w_ws The impact on residential heat consumption is limited and the correlation is weak. Therefore, in order to reduce the interference of insensitive parameters, this embodiment simulates the t u Set to 0.

[0100] The annual dynamic cumulative heat load DqH of each simulation sample is simulated and calculated using DesignBuilder. During the simulation process, the use conditions of each simulation sample must be set according to the General Specification for Building Energy Conservation and Renewable Energy Utilization GB55015. Among them, the per capita building area is 25m 2 / person, ventilation rate is 0.5 times / h. The power density of electrical equipment in each room is 3.8W / m 2 , lighting power density is 5W / m 2 The illumination of each room is set to 75lx for the bedroom, 50lx for the corridor, and 100lx for other rooms. The hourly utilization rate of various equipment and the occupancy rate of people in the room are as follows: Fig.15 During the heating period, bedrooms, living rooms, kitchens, bathrooms and other indoor rooms are heated 24 hours a day, and the indoor temperature in the heating area is 18-26℃.

[0101] Furthermore, the days for calculating the heating period are the days with an annual average daily temperature of no more than 5°C calculated using the sliding average method, which is only used in energy-saving design stages such as steady-state heat consumption calculations, rather than the statutory heating days in various places. According to relevant regulations, the actual heating days in the eight cities are different, and most of them are more than the calculated heating period days (as shown in Table 1 below). Among them, the actual heating period in the four cities in the cold B zone is the same. The actual heating period in Taiyuan, Chengde and Lanzhou in the cold A zone is the same. The actual heating period in Weifang is relatively similar to that of the four cities in the cold B zone. In addition, the people's governments of each city will also adjust the heating period time according to actual conditions such as meteorology.

[0102] Table 1: Comparison of heating periods in various target cities

[0103]

[0104]

[0105] It should be noted that the actual heating period of the eight cities in Table 1 above is calculated based on the "Beijing Heating Management Measures", "Jinan City Urban Central Heating Management Regulations", "Xi'an Central Heating Regulations", "Zhengzhou City Urban Heating and Heat Use Management Measures", "Taiyuan City Urban Heating Management Regulations", "Chengde City Urban Heating Regulations", "Lanzhou City Heating and Heat Use Regulations" and "Weifang City Heating Regulations".

[0106] Therefore, this embodiment calculates the dynamic cumulative heat load of the simulated samples throughout the year according to the actual heating days of the eight cities. In addition, the annual steady-state total heat consumption per unit building area of ​​the residential building SqH and the annual steady-state total heat consumption of the residential building SQH can be calculated according to formulas (1) to (3). Finally, this embodiment uses the ratio of SqH to DqH of each simulated sample as the conversion coefficient φ from the steady-state total heat consumption to the dynamic cumulative heat load. By analyzing the distribution characteristics of the conversion coefficient and the influence mechanism of meteorological parameters and residential spatial design parameters on the conversion coefficient, a conversion coefficient prediction model can be established. Conversely, for the DqH of a residential design scheme to be solved, it can be predicted according to formula (3) without having to use simulation software modeling and calculation.

[0107] SqH=0.001·24·q H t (1)

[0108] SQH=0.001·24·q H ·S·t (2)

[0109]

[0110] Where, t is the actual number of heating days, which is taken from Table 1, unit: day; SQH is the total steady-state heat consumption of the house throughout the year, kWh; SqH is the total steady-state heat consumption of the house per unit building area throughout the year, kWh / m 2 ; DqH is the cumulative heat load per unit building area, kWh / m 2 ;q H It is the heat consumption index, W / m 2 ; S is the total building area, m 2 .

[0111] According to statistics, the overall distribution range of the conversion coefficient φ is 0.56 to 1.70, with an average of 0.84, a median of 0.81, and a standard deviation of 0.157. Figure 16-Figure 24 As shown. In addition, the conversion coefficients of Beijing's residential simulation samples are concentrated around 0.800. The conversion coefficients of Taiyuan's residential simulation samples have the largest distribution range, and their average value and standard deviation are relatively high, reaching 1.213 and 0.144 respectively. The average value of the conversion coefficients of Weifang's residential simulation samples is relatively the lowest, only 0.686, and the standard deviation of the conversion coefficients is low. The distribution range of the conversion coefficients of Xi'an's residential simulation samples is relatively the smallest, and its standard deviation is only 0.067, indicating that the degree of dispersion of the conversion coefficient distribution of Xi'an's residential buildings is relatively the lowest among the eight target cities.

[0112] S500, conversion coefficient prediction:

[0113] In this embodiment, SPSS was used to select 1640 Beijing simulation samples, 340 Jinan simulation samples, 195 Xi'an and Zhengzhou simulation samples, 210 Taiyuan and Chengde simulation samples, and 205 Lanzhou and Weifang simulation samples from the above 3550 residential simulation samples, a total of 3200 residential simulation samples from different cities as prediction samples for building a heating carbon emission prediction model, and other simulation samples were used as test samples. Through "Spearman correlation analysis", the correlation coefficient and significance level of the conversion coefficient φ of the prediction sample and different spatial design parameters or meteorological parameters were calculated in this embodiment, such as Fig.25 As shown. The total building area S, building volume V, and shape coefficient V were analyzed by SPSS "curve evaluation" e , East facing exterior wall area S w_e , average solar radiation heat of the outer surface of the east-facing window during the heating period iwin e , average outdoor temperature during heating period t e , annual steady-state total heat consumption per unit building area SqH, heat consumption index q H The relationship between parameters and φ. Among them, the quadratic or cubic curve model is in iwin e The performance of the cubic curve model is relatively good in describing the relationship between t eIt has a better performance in describing the relationship with φ, such as Fig.26 In addition, the power model has e , V, V e , S w_e , S w_s The S model and logarithmic model have relatively good performance in describing the relationship between S and φ, V and φ, S w_e The inverse model also performs well in describing the relationship with φ. e The exponential model performs well in describing the relationship with φ; e The relationship description with φ is better, such as Figure 27-Figure 29 shown.

[0114] like Fig.30 As shown in the figure, the cubic curve model, quadratic curve model and power model have better performance in describing the relationship between SqH and φ, while the exponential model, cubic curve model and power function can describe the relationship between qH and φ relatively better.

[0115] According to the ranking of the fitting performance of the calculation parameter "curve evaluation" from high to low, a better calculation parameter conversion model can be screened out as the independent variable for building a heating carbon emission prediction model. e , SqH, t e and q H Can be converted into corresponding square and cube values ​​(iwin e 2 With iwin e 3 , SqH 2 With SqH 3 ,t e 2 With t e 3 ,q H 2 With q H 3 ); S, V, S w_e Can be converted to base 10 logarithm (S lg 、V lg and S welg );S e and S w_s Can be converted into reciprocal (S er and S wsr ). In addition, some parameters can also be converted into power function values, S-curve function values ​​or exponential function values, as shown in Table 2 below. pf , S epf 、V pf 、V epf , Swepf , S wspf , SqH pf ,q Hpf are the power function values ​​of the corresponding parameters respectively; N asc , B hsc , S sc , S esc 、V sc 、V esc , S wesc , S wssc are the S-curve function values ​​of the corresponding parameters; V eef and q Hef V e and q H The exponential function value of .

[0116] Table 2: Conversion of some calculation parameters

[0117] Calculation parameters Power Model S Model Exponential Model <![CDATA[N a ]]> — <![CDATA[N asc =and 1.734 / Na ]]> — <![CDATA[B h ]]> — <![CDATA[B hsc =and 5.143 / bh ]]> — S <![CDATA[S pf =S -0.154 ]]> <![CDATA[S sc =and 545.732 / S ]]> — <![CDATA[S e ]]> <![CDATA[S epf =S e -0.180 ]]> <![CDATA[S esc =and 541.157 / Se ]]> — V <![CDATA[V pf =V -0.150 ]]> <![CDATA[V sc =and 1540.523 / V ]]> — <![CDATA[V e ]]> <![CDATA[V epf =V e 0.439 ]]> <![CDATA[V esc =and -0.107 / Ve ]]> <![CDATA[V eef =and 1.735·Ve ]]> <![CDATA[S w_e ]]> <![CDATA[S wepf =S w_e -0.162 ]]> <![CDATA[S wesc =and 63.536 / Sw_e ]]> — <![CDATA[S w_s ]]> <![CDATA[S wspf =S w_s -0.152 ]]> <![CDATA[S wssc =and 72.774 / Sw_s ]]> — SqF <![CDATA[SqH pf =SqH 0.473 ]]> — — <![CDATA[q H ]]> <![CDATA[q Hpf =q H 0.482 ]]> — <![CDATA[q Hef =and 0.065·qH ]]>

[0118] In order to make the conversion coefficient prediction method more interpretable, a multivariate linear regression model is used to fit the conversion coefficient. The fitting process screens the independent variables in a "step-by-step" manner to ensure their significance, and removes or integrates the independent variables in combination with collinearity diagnosis to ensure their independence. After regression fitting and comparative analysis, a total of 10 conversion coefficient prediction models were established in this embodiment. Among the independent variables of the prediction model, the meteorological parameters of each city have been determined in the project site selection stage, and the shape scale parameters and part of the heat consumption calculation results in the space design parameters can also be extracted from the steady-state heat consumption calculation. The number of various functional rooms can be increased and determined according to the deepening of the design. According to the type and number of independent variables, these 10 conversion coefficient prediction models can be sorted into M z1 ~M z10 , expressed by equations (4) to (13):

[0119]

[0120]

[0121]

[0122]

[0123]

[0124]

[0125]

[0126]

[0127]

[0128]

[0129]

[0130] Where: y z1 ~y z10 is the conversion coefficient.

[0131] From the perspective of model composition, the independent variables of each prediction model include t, ε we , ε r ,t e 2 , L h , B h 、ww s and SqH pf And these 8 parameters can form the conversion coefficient prediction model M z1 . M z2 ~M z10 In M z1 Different shape scale parameters or functional configuration parameters are added on the basis of the above. The influence of different independent variables on the conversion coefficient is as follows: Fig.31 shown.

[0132] Among them, t, SqH pf and ε r It has a higher positive impact on the prediction results, and the influence intensity (the absolute value of the influence intensity can be used to represent the parameter sensitivity) decreases successively; ε we It has a high negative impact on the prediction results. In addition, ww s The positive impact on the prediction results is slightly higher than that of B h ; L h The reverse impact on the prediction results is slightly greater than t e 2 Furthermore, w and N sth The positive impact on the prediction results, and t e 、ww e 、ww n 、N std 、N stw and N str The reverse impact on the prediction results is relatively low. pf , ε we and ε r It is the main parameter to ensure that the conversion coefficient prediction model achieves a good fitting effect. e 2 The impact on the conversion coefficient prediction results is slightly higher than ww s, the influence of function configuration parameters on prediction results is slightly higher than t e 、ww n and ww e .

[0133] S600, fitting performance test:

[0134] In terms of fitting performance, the R 2 All reached above 0.960, as shown in Table 3. More independent variables do not necessarily reduce the fitting error of the prediction model. z7 With M z10 R 2 The highest, and the fitting error is the lowest. z1 R 2 The lowest, and the fitting error is the highest. w(compared with M z1 With M z2 )、ww n (Compared to M z2 With M z3 , or compared with M z5 With M z9 ), N sth (Compared to M z2 With M z4 ) and other three parameters can slightly improve the fitting performance of the prediction model.

[0135] Table 3: Fitting performance of the conversion coefficient prediction model

[0136] Model Number <![CDATA[R 2 ]]> MAE MAPE MSE <![CDATA[M z1 ]]> 0.963 0.022 2.65% 0.001 <![CDATA[M z2 ]]> 0.967 0.021 2.51% 0.001 <![CDATA[M z3 ]]> 0.968 0.021 2.47% 0.001 <![CDATA[M z4 ]]> 0.969 0.021 2.44% 0.001 <![CDATA[M z5 ]]> 0.970 0.020 2.39% 0.001 <![CDATA[M z6 ]]> 0.971 0.020 2.32% 0.001 <![CDATA[M z7 ]]> 0.972 0.019 2.27% 0.001 <![CDATA[M z8 ]]> 0.968 0.021 2.47% 0.001 <![CDATA[M z9 ]]> 0.971 0.020 2.36% 0.001 <![CDATA[M z10 ]]> 0.972 0.019 2.27% 0.001

[0137] In addition, the M z4 With M z5 、M z6 、M z7 By comparison, it can be found that N std 、N str and N stw For the prediction model R 2 The lifting effect of M is reduced successively. z3 With M z8 , and M z7 With M z10 By comparison, the independent variable t e and ww e It has almost no effect on the fitting performance of the prediction model. Among all the prediction models, M z5 、M z6 、M z9 The design depth requirement is higher, while other models have lower design depth requirements. z3 The fitting performance is the highest among the four models that do not use functional configuration parameters as independent variables; M z4 In Mz3 The number of households in the standard layer was added to the independent variable of M, which slightly improved the fitting performance; z7 Then, adding the number of households and bedrooms on the standard floor to the independent variables can achieve the best fitting effect, although the design depth requirement is not high. z5 、M z6 、M z9 The fitting performance of M z7 .

[0138] S700, Prediction Performance Test:

[0139] Using M z1 ~M z10 Calculate the predicted conversion coefficient of 350 test samples. The predicted R of each model 2 , MAE, MAPE and MSE are shown in Table 4. The prediction performance and fitting performance of each model are relatively close. Since the conversion coefficient prediction models are all based on the calculation of steady-state heat consumption, the difference in the requirements for design depth is mainly reflected in the functional configuration parameters. z6 、M z7 and M z10 Prediction R 2 The highest, and M z7 The prediction error is the lowest, and it can achieve more effective conversion coefficient prediction in each stage of residential design.

[0140] Table 4: Prediction performance of the conversion coefficient prediction model

[0141] Model Number <![CDATA[R 2 ]]> MAE MAPE MSE <![CDATA[M z1 ]]> 0.964 0.021 2.52% 0.001 <![CDATA[M z2 ]]> 0.967 0.021 2.40% 0.001 <![CDATA[M z3 ]]> 0.967 0.021 2.40% 0.001 <![CDATA[M z4 ]]> 0.968 0.020 2.35% 0.001 <![CDATA[M z5 ]]> 0.970 0.020 2.31% 0.001 <![CDATA[M z6 ]]> 0.971 0.019 2.24% 0.001 <![CDATA[M z7 ]]> 0.971 0.019 2.20% 0.001 <![CDATA[M z8 ]]> 0.967 0.021 2.45% 0.001 <![CDATA[M z9 ]]> 0.970 0.020 2.32% 0.001 <![CDATA[M z10 ]]> 0.971 0.019 2.21% 0.001

[0142] S800, the best heating carbon emission prediction model:

[0143] Based on the conversion coefficient prediction model, a carbon emission prediction model M for urban residential heating in cold regions in winter is established. heating , which is expressed as formula (14). In addition, since M z7 The conversion coefficient prediction performance of is the best and has the highest fit with the architectural design process. Therefore, substituting formula (10) into formula (14), we can get M 28 The model function is expressed as formula (15).

[0144]

[0145] y heating =24·EF i ·q H ·t / {η i ·q·[0.011·t-20.604·ε we +13.507·ε r-

[0146]

[0147] In the formula, y heating is the annual heating carbon emission per unit building area of ​​the house, kgCO2 / m 2 ; η i It represents the comprehensive efficiency of the heating system. When a coal-fired boiler is used, the value is 0.81, and when a gas-fired boiler is used, the value is 0.85. q is the calorific value conversion coefficient, which is 277777.78 kWh / TJ; EF i is the energy carbon emission factor. When the heat source is standard coal, the value is 94.44tCO2 / TJ. When the heat source is natural gas, the value is 55.54tCO2 / TJ.

[0148] Formula (15) is a complex nonlinear calculation model involving many independent variables. The sensitivity of each independent variable is shown in Table 5. In general, the total length of the building width w is the most sensitive to the prediction results of heating carbon emissions, followed by the total length of the building depth d and the energy carbon emission factor EF i Actual heating days t, Proportion of staircase protruding wall t u 、Unit number U n , heating system comprehensive efficiency η i , average outdoor temperature during heating period t e The sensitivity of the staircase width swd and staircase depth sdp is relatively low, and the sensitivity of other calculation parameters can be ignored. It can be seen that the building width and depth are the main contents of residential low-carbon design.

[0149] Table 5: Sensitivity of independent variables in the heating carbon emission prediction model

[0150]

[0151]

[0152] In this embodiment, the prediction performance of the prediction model established in this study is demonstrated by comparing it with other different machine learning algorithm models. Since the prediction of the conversion coefficient is the core of the technical path of this study, different machine learning algorithms and independent variable data sets are used to establish a machine learning-based conversion coefficient prediction model. On the basis of Mz1~Mz10, according to the type of independent variables and their influence on the prediction results, a total of 34 groups of independent variable data sets Dz1~Dz34 are sorted out, such as Fig.32 shown.

[0153] The independent variables in each data set include t, ε we , ε r ,t e2 , L h , B h 、ww s and SqH pf The differences in 8 parameters are reflected in w, ww n 、N sth 、N std 、N stw and N str Among them, N str Contains N sth 、N std and N stw Therefore, the case where these four parameters exist in the independent variable at the same time is not considered. In addition, in order to facilitate the linear regression model M z1 ~M z10 In the process of research and development, the independent variables of these 10 models were extracted. According to the order of the variables in the independent variable data set, these 10 groups of independent variables are recorded as D z1 (From M z1 )、D z9 (From M z2 )、D z10 (From M z4 )、D z11 (From M z7 )、D z12 (From M z10 )、D z13 (From M z5 )、D z14 (From M z6 )、D z26 (From M z3 )、D z27 (From M z8 ) and D z30 (From M z9 ).

[0154] GridSearchCV and cross-validation were used to optimize hyperparameters and data sets. Each data set was divided into 4 training sets and 1 test set. Each data set was combined with machine learning algorithms such as ridge regression, K nearest neighbor, linear support vector machine, decision tree, random forest, Adaboost, gradient boosting tree and multi-layer perceptron, and R 2 As the basis for optimization, the hyperparameter optimization range of each machine learning algorithm is set as shown in Table 6 below.

[0155] Table 6: Hyperparameter values ​​for various machine learning algorithms

[0156]

[0157]

[0158] Finally, this study tested the prediction performance of different machine learning algorithm models such as R2, MAE, MSE, etc. through 350 test samples, and compared them with Mz1~Mz10. After grid search and cross-validation, the best prediction performance that can be achieved by different machine learning algorithm models and data sets is as follows: Figure 33 to Figure 35 shown.

[0159] In contrast, the prediction performance of decision tree and linear support vector machine models is poor and cannot be used for effective prediction of conversion coefficients. In the other six machine learning algorithm models, different machine learning algorithms are applicable to different independent variable data sets. Conversely, the prediction performance of different machine learning algorithms on different independent variable data sets is also different.

[0160] First, the average R of the random forest model when using different data sets for prediction 2 The highest, and higher than the linear regression model. z25 , D z31 , D z34 , D z5 , D z7 , D z16 , D z17 and D z24 Random Forest Model in R 2 Both reached 0.991. For the independent variables of the linear regression model, D z14 , D z11 and D z12 A relatively high R can be achieved in random forest models 2 In addition, the average R of the gradient boosting tree model when using different data sets for prediction is 2 It is close to the linear regression model. z14 The linear regression model M z6 Compared with D z14 Prediction of Gradient Boosted Tree Model in R 2 In addition to the higher, other gradient boosting tree models can achieve higher prediction R after fitting with the linear regression model independent variables. 2 Furthermore, D z13 Average prediction R across different machine learning algorithm models 2 The highest, with the highest matching degree with different machine learning algorithms. Finally, in the prediction R 2 In terms of volatility, the prediction R of 34 sets of independent variable data sets in the random forest model 2The volatility is the smallest, followed by the Adaboost model. On the contrary, the multi-layer perceptron model has the largest volatility. Secondly, the random forest model has the smallest average prediction error. z16 , D z17 , D z24 , D z25 , D z31 and D z34 The MAE of the random forest model is only 0.009. After the random forest model is fitted with the independent variables of the linear regression model, the prediction error can be reduced. z14 The prediction error of the gradient boosting tree model is slightly higher than that of the linear regression model M z6 In addition, all other linear regression model independent variables can achieve prediction error reduction in the gradient boosting tree model. z11 and D z13 The average prediction error among different machine learning algorithm models is the lowest. Compared with other machine learning algorithms, the prediction error fluctuation of 34 sets of independent variable data sets in Adaboost, Random Forest and K nearest neighbor models is smaller.

[0161] In summary, in addition to the linear regression model, the random forest and gradient boosting tree models can also effectively predict the conversion coefficient, and the prediction performance of the random forest model is better than that of the linear regression model. However, the difference in prediction performance between the random forest model and the linear regression model is small. Therefore, the linear regression model M z1 ~M z10 It can provide architects with more intuitive guidance in predicting carbon emissions from residential heating.

[0162] Example 2

[0163] like Fig.36 As shown, a prediction device for carbon emissions from urban residential heating in cold regions according to Example 2 of the present application includes:

[0164] A construction module 100 is used to construct basic residential models in different regions, evolve the basic models to generate urban residential cases in cold regions, and construct a case database based on the urban residential cases in cold regions;

[0165] The first calculation module 200 is used to compile a steady-state calculation method based on the case database and calculate the annual cumulative steady-state heat consumption per unit area of ​​each urban residential case in cold regions;

[0166] The simplification module 300 is used to quantitatively analyze the influence of different residential design parameters and meteorological parameters on steady-state heat consumption through Sobol global sensitivity analysis, and to screen and eliminate design parameters with low sensitivity to simplify the basic residential model;

[0167] The second calculation module 400 is used to evolve the simplified basic model to batch generate residential cases of different preset cities in cold regions, and calculate the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​the residential cases in cold regions;

[0168] An analysis module 500 is used to use the ratio of the steady-state heat consumption to the dynamic heat consumption as a conversion coefficient, and to analyze the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified basic residential model;

[0169] A screening module 600 is used to divide the urban residential cases in cold regions into a test set and a validation set, and to construct a conversion coefficient prediction model based on residential design parameters and meteorological parameters by using a multivariate linear regression algorithm in combination with the test set, and to test and evaluate the prediction performance of the conversion coefficient prediction model by using the validation set, so as to screen out the best conversion coefficient prediction model;

[0170] The prediction module 700 is used to establish a heating carbon emission prediction model based on residential space design parameters, envelope structure thermal performance parameters and meteorological parameters according to the optimal conversion coefficient prediction model, and use the heating carbon emission prediction model to predict urban residential heating carbon emissions in cold areas.

[0171] It should be noted that other specific implementations of the device for predicting carbon emissions from urban residential heating in cold regions in the embodiment of the present invention can refer to the specific implementations of the method for predicting carbon emissions from urban residential heating in cold regions mentioned above, and will not be repeated here to avoid redundancy.

[0172] Example 3

[0173] A computer program product involved in Example 3 of the present application is used to store a computer program. When the computer program runs on a computer, the method in any one of the implementation modes in Example 1 of the present application is implemented.

[0174] Example 4

[0175] A computer-readable storage medium according to Embodiment 4 of the present application, wherein the computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing a method in any one of the implementations in Embodiment 1 of the present application;

[0176] Among them, the computer-readable storage medium can be a read-only memory (ROM), a static storage device, a dynamic storage device or a random access memory (RAM); the computer-readable storage medium can store program code, and when the program stored in the computer-readable storage medium is executed by the processor, the processor is used to execute the steps of the method in any one of the implementation methods in Example 1 of the present application.

[0177] Example 5

[0178] like Fig.37 As shown, an electronic device involved in Embodiment 5 of the present application includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, and when the program or instruction is executed by the processor, a method in any one of the implementations in Embodiment 1 of the present application is implemented;

[0179] Among them, the processor can adopt a general-purpose central processing unit (CPU), a microprocessor, an application specific integrated circuit (ASIC), a graphics processing unit (GPU) or one or more integrated circuits to execute relevant programs to implement the method in any one of the implementation methods in Example 1 of the present application.

[0180] The processor may also be an integrated circuit electronic device with signal processing capability. In the implementation process, each step of the method in any implementation of Embodiment 1 of the present application may be completed by an integrated logic circuit of hardware in the processor or by instructions in software form.

[0181] The above-mentioned processor may also be a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The methods, steps and logic block diagrams disclosed in the embodiments of the present application may be implemented or executed. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in the embodiments of the present application may be directly embodied as being executed by a hardware decoding processor, or may be executed by a combination of hardware and software modules in a decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. The storage medium is located in a memory, and the processor reads the information in the memory, and completes the functions required to be performed by the unit included in the data processing device of the embodiment of the present application in combination with its hardware, or executes the method in any one of the implementation modes in the embodiment 1 of the present application.

[0182] The above are only preferred specific implementations of the present application; however, the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical solution and its improved ideas of the present application within the technical scope disclosed in the present application, which should be included in the protection scope of the present application.

Claims

1. A method for predicting carbon emissions from urban residential heating in cold regions, characterized in that: include: Constructing basic residential models in different regions, and evolving the basic models through Python language to batch generate urban residential cases in cold regions, and constructing a case database based on the urban residential cases in cold regions; Based on the case database, a steady-state calculation method is written in Python to calculate the annual cumulative steady-state heat consumption per unit area of ​​urban residential cases in cold regions; The influence of spatial design parameters, thermal performance parameters of envelope structures and meteorological parameters on steady-state heat consumption of different houses is quantitatively analyzed through Sobol global sensitivity analysis, and parameters with low sensitivity are screened and eliminated to simplify the basic model of the house. The simplified basic model is evolved to batch generate residential cases in different preset cities in cold zones A and B, and the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​urban residential cases in cold regions are calculated; The ratio of the steady-state heat consumption to the dynamic heat consumption is used as a conversion coefficient, and the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified basic residential model is analyzed, wherein the design parameters include space design parameters and thermal performance parameters of the enclosure structure; The urban residential cases in cold regions are divided into a test set and a validation set, a conversion coefficient prediction model based on residential design parameters and meteorological parameters is constructed by using a multivariate linear regression algorithm in combination with the test set, and the prediction performance of the conversion coefficient prediction model is tested and evaluated through the validation set to screen out the best conversion coefficient prediction model; A heating carbon emission prediction model based on residential design parameters and meteorological parameters is established according to the optimal conversion coefficient prediction model, and the heating carbon emission prediction model is used to predict heating carbon emissions of urban residential buildings in cold regions.

2. The method for predicting carbon emissions from urban residential heating in cold regions according to claim 1 is characterized in that: Constructing basic residential models in different regions includes: selecting a preset region, sorting out and establishing a basic residential model, integrating the spatial design parameters in the basic residential model according to the basic residential model, splitting and reorganizing the spatial design parameters to form a plurality of design schemes, combining the design schemes, quantifying the sensitivity of different spatial design parameters, thermal performance parameters of the enclosure structure and meteorological parameters in the steady-state heat consumption calculation process, and simplifying the basic residential model according to the sensitivity of the residential design parameters to obtain the basic residential model.

3. The method for predicting carbon emissions from urban residential heating in cold regions according to claim 2 is characterized in that: The space design parameters include at least one of the following: number of building floors, building height, exterior wall areas in each direction, building width, building depth, number of functional rooms, ventilation volume, floor area, roof information, exterior window information, exterior door information and total building area.

4. The method for predicting carbon emissions from urban residential heating in cold regions according to claim 1 is characterized in that: Writing steady-state calculation methods based on the case database includes: Define the value ranges of space design parameters and thermal performance parameters of enclosure structures; Generate independent unit housing space design parameters and townhouse space design parameters; Combine the independent unit residential space design parameters and the townhouse space design parameters for inspection and screening, and add other combination parameters required for heat consumption calculation to form a cold region residential design data set; The meteorological parameters of different target cities are defined, and the residential design data sets in cold regions are respectively integrated with the corresponding thermal performance parameters of the enclosure structure according to the differences in the energy-saving requirements for construction in different climate zones, so as to form residential design data sets for different preset cities in cold regions; Calculate various heat consumption items according to the residential design data set of the preset city respectively, and integrate the calculated various heat consumption items into the residential design steady-state heat consumption data set of the target city; Based on the residential design steady-state heat consumption dataset of the target city, the calculation results of various heat consumption can be output by changing the space design parameters and the thermal performance parameters of the envelope structure.

5. The method for predicting carbon emissions from urban residential heating in cold regions according to claim 1 is characterized in that: The calculation formula for the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area is: SqH=0.001·24·q H ·t (1) DqH=0.001·24·q H ·t / φ (2) Among them, SqH is the cumulative steady-state heat consumption per unit area throughout the year, DqH is the cumulative dynamic heat consumption per unit area throughout the year, t is the number of heating days, and qH is the heat consumption index.

6. The method for predicting carbon emissions from urban residential heating in cold regions according to claim 5 is characterized in that: The expression of the heating carbon emission prediction model is: y heating =1000·EF i ·SqH / (η i ·q·φ)=24·EF i ·q H ·t / (η i ·q·φ) (3) Among them, y heating is the annual heating carbon emission per unit building area of ​​the house; η i Indicates the comprehensive efficiency of the heating system; q is the calorific value conversion coefficient; EF i is the energy carbon emission factor.

7. A prediction device for carbon emissions from urban residential heating in cold regions, characterized in that: include: A construction module is used to construct a basic residential model in different regions, and evolve the basic model through Python language to generate urban residential cases in cold regions, and construct a case database based on the urban residential cases in cold regions; The first calculation module is used to write a steady-state calculation method using Python based on the case database and calculate the annual cumulative steady-state heat consumption per unit area of ​​each urban residential case in cold regions; A simplification module is used to quantify the influence of spatial design parameters, thermal performance parameters of envelope structures and meteorological parameters on steady-state heat consumption of different houses through Sobol global sensitivity analysis, and to screen and eliminate parameters with low sensitivity to simplify the basic model of the house; The second calculation module is used to evolve the simplified basic model to batch generate residential cases in different preset cities in cold areas A and B, and calculate the annual cumulative steady-state heat consumption and dynamic heat consumption per unit area of ​​the urban residential cases in the cold areas; An analysis module, used to use the ratio of the steady-state heat consumption to the dynamic heat consumption as a conversion coefficient, and to analyze the correlation between the numerical distribution range of the conversion coefficient and the design parameters and meteorological parameters of the simplified residential basic model, wherein the design parameters include space design parameters and thermal performance parameters of the enclosure structure; A screening module is used to divide the urban residential cases in cold regions into a test set and a validation set, and to construct a conversion coefficient prediction model based on residential design parameters and meteorological parameters by using a multivariate linear regression algorithm in combination with the test set, and to test and evaluate the prediction performance of the conversion coefficient prediction model by using the validation set, so as to screen out the best conversion coefficient prediction model; The prediction module is used to establish a heating carbon emission prediction model based on residential design parameters and meteorological parameters according to the optimal conversion coefficient prediction model, and use the heating carbon emission prediction model to predict the heating carbon emission of urban residential buildings in cold areas.

8. A computer program product, characterized in that The computer program product stores computer instructions, and when the computer instructions are executed by a processor, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a program code for execution by a device, wherein the program code includes steps for executing the method according to any one of claims 1 to 6.

10. An electronic device, characterized in that: The electronic device includes a processor, a memory, and a program or instruction stored in the memory and executable on the processor, wherein the program or instruction implements the method according to any one of claims 1 to 6 when executed by the processor.