Fabricated building construction period carbon emission quantification and prediction method based on knowledge graph

Abstract

The invention discloses a knowledge graph-based prefabricated building construction period carbon emission quantification and prediction method, which specifically comprises the steps of S1, prefabricated building construction period carbon emission knowledge graph visualization research, S2, prefabricated building construction period carbon emission quantification research, S3, prefabricated building construction period carbon emission quantification case analysis, and S4, prefabricated building construction period carbon emission quantification prediction. S4, carbon emission prediction and research in the construction period of the fabricated building. According to the method, an influence factor system and an influence mechanism of the carbon emission in the construction period of the fabricated building are defined, and the source and proportion of the carbon emission of the construction module in each stage of the construction period of the fabricated building are defined; the carbon emission in the construction period of the fabricated building is predicted and researched based on four machine learning algorithms, the optimal prediction model is determined, the model refinement degree is high, the model boundary condition, the space-time process and the quantitative statistics module have small deviation from the actual situation of the fabricated building, the model parameters have the optimization iteration process, the stability is good, and the prediction accuracy is high.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon emissions during building construction, and specifically to a method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on a knowledge graph. Background Art

[0002] As one of the main sources of carbon emissions, the construction industry is a key area for carbon emission reduction.

[0003] The construction period of prefabricated buildings has the characteristics of short carbon emission time, high emission concentration, and great difficulty in quantification. However, there are relatively few studies on the quantification of carbon emissions during the construction period of prefabricated buildings in China. Existing studies are mostly focused on the building operation stage, and the carbon emission factor system and influencing mechanism during the construction period are still unclear. The existing carbon emission quantification model has not conducted a correlation analysis of the influencing factors for the characteristics of the construction period of prefabricated buildings. There are large deviations between the model boundary conditions, spatiotemporal processes, and quantitative statistical modules and the actual situation of prefabricated buildings. The existing carbon emission prediction research model structure introduces many empirical parameters, and there is no relevant parameter optimization iterative process. The model is not very refined and stable. Summary of the invention

[0004] The purpose of the present invention is to provide a method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on a knowledge graph to solve the above defects.

[0005] In order to achieve the above object, the present invention provides the following technical solutions:

[0006] The method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph includes the following steps:

[0007] S1. Visualization of knowledge graph of carbon emissions during the construction period of prefabricated buildings:

[0008] Retrieve literature on carbon emissions during the construction period of prefabricated buildings from the CNKI database and the WoS core collection database, and select and construct a knowledge graph visualization dataset of carbon emissions during the construction period of prefabricated buildings; Based on the CiteSpace knowledge graph visualization software, deeply analyze the research hotspot keywords in the field of carbon emissions during the construction period of prefabricated buildings, obtain the research dynamics in the field of carbon emissions during the construction period of prefabricated buildings, the evolution path of keywords and their mutual relationships, and identify the research hotspots and future trends in this field;

[0009] S2. Quantification of carbon emissions during the construction period of prefabricated buildings:

[0010] Based on the comparative analysis of carbon emission quantification methods during the construction period, a carbon emission quantification model for prefabricated building construction period based on the combination of process analysis method and carbon emission factor method is established; carbon emission factors in current specifications and literature are analyzed and quantified, and a carbon emission factor library for prefabricated building construction period is constructed; a carbon emission quantification formula for prefabricated building construction period is constructed, and a matrix product quantification program is developed using Python programming language to improve the efficiency of carbon emission quantification;

[0011] S3. Case analysis of carbon emissions during the construction period of prefabricated buildings:

[0012] A case study of a prefabricated public parking building in Chongqing was conducted to analyze the established carbon emission quantification model for prefabricated buildings during construction. The carbon sources at the construction site were identified by three methods: construction site carbon source identification, engineering data carbon source identification, and drone carbon source identification. The carbon emission data during the construction period were collected, and the carbon emissions during the construction period of prefabricated buildings were calculated using the matrix product quantification method in different modules. The overall characteristics of carbon emissions during the construction period of prefabricated buildings and the characteristics of construction modules were analyzed based on visualization analysis and time series analysis methods, and emission reduction measures were formulated according to the analysis of carbon emission characteristics.

[0013] S4. Study on prediction of carbon emissions during the construction period of prefabricated buildings:

[0014] Select a machine learning algorithm to build a carbon emission prediction model for the construction period of prefabricated buildings; combine knowledge graph visualization and carbon emission feature analysis results to identify key influencing factors of carbon emissions of prefabricated buildings; expand learning samples based on the thin plate spline interpolation algorithm, use grid optimization and k-fold cross validation to find the optimal hyperparameters in the model, and use this to establish a carbon emission prediction model; use root mean square error and goodness of fit coefficient as evaluation indicators to compare the prediction effects of machine learning prediction models, and determine the carbon emission prediction model with the best prediction performance, thereby realizing the entire process of quantification and prediction of carbon emissions during the construction period of prefabricated buildings based on knowledge graphs.

[0015] Preferably, the S2 step is as follows:

[0016] S21. Carbon emission quantification basis: Establish the basic framework of the carbon emission quantification model during the construction period of prefabricated buildings, and clarify the basis for quantification and the system boundaries;

[0017] S22. Improvement of carbon emission quantification methods: Carry out modular quantification research on carbon emissions during the construction period of prefabricated buildings. Based on the sub-projects of prefabricated buildings, comprehensively use the process analysis method and the carbon emission factor method to improve the carbon emission quantification model during the construction period of traditional buildings, and establish a carbon emission quantification model for the construction period of prefabricated buildings based on the combination of the process analysis method and the carbon emission factor method.

[0018] S23. Carbon emission quantitative index system: Quantitative index analysis of carbon sources at the construction site of prefabricated buildings during construction period, clarify the specific sources of carbon sources, and then introduce Pareto analysis method to classify and analyze carbon sources at the construction site;

[0019] S24. Selection and quantification of carbon emission factors: A method based on a combination of process analysis and carbon emission factor method is used to quantify the carbon emissions during the construction period of prefabricated buildings, and the values ​​of all carbon emission factors in the carbon emission quantification model are determined to construct a carbon emission factor library for the construction period of prefabricated buildings;

[0020] S25. Construction of carbon emission quantification model: Construct a carbon emission quantification formula for the construction period of prefabricated buildings, and use Python programming language to develop a matrix multiplication quantification program to improve the efficiency of carbon emission quantification.

[0021] Preferably, in step S22, the carbon emission quantification model for the construction period of prefabricated buildings includes: earthwork and foundation pit engineering module, foundation and pile foundation engineering module, reinforced concrete engineering module, prefabricated engineering module, decoration and renovation engineering module, electromechanical and equipment installation engineering module, and construction waste treatment engineering module.

[0022] Preferably, in step S24, the carbon emission factors include electricity, fossil energy, building materials, construction machinery, transportation and construction personnel.

[0023] Preferably, in step S25, the carbon emission quantification matrix of the prefabricated building during the construction period is as shown in formula (22):

[0024]

[0025] In the formula, E j is the consumption matrix of the jth construction module, M ji is the consumption of the i-th building material in the j-th construction module, J ji Y is the construction shift of the i-th machine in the j-th construction module, ji is the i-th type of transport equipment shift in the j-th construction module, R ji The i-th construction worker team of the j-th construction module; F j is the j-th construction module matrix, F mji is the carbon emission factor of the i-th material of the j-th construction module, F Jji is the carbon emission factor of the i-th construction machinery in the j-th construction module, F Yji is the carbon emission factor of the i-th transportation vehicle in the j-th construction module, F Rji is the carbon emission factor of the i-th construction worker team in the j-th construction module; i is a, b, c…g respectively;

[0026] Convert a matrix into a matrix containing only its diagonal data using the Kronecker delta function and the given matrix Hadamard product operation. The final quantization result is shown in formula (25):

[0027]

[0028] In the formula, C z is the total carbon emissions, C1, C2, …, C i Represents the carbon emissions of each construction module.

[0029] Preferably, the step S4 is specifically as follows:

[0030] S41. Comparative analysis of carbon emission prediction models:

[0031] Through comparative analysis of building carbon emission prediction methods, a machine learning algorithm is selected to construct a carbon emission prediction model for prefabricated building construction period. The machine learning algorithm includes support vector regression, BP neural network, extreme learning machine, and random forest algorithm.

[0032] S42. Sources and correlation analysis of learning sample data for prediction model:

[0033] Combining the knowledge graph visualization and carbon emission characteristic analysis results, the key influencing factors of carbon emissions of prefabricated buildings are identified, and the Pearson correlation test is applied to test the degree of correlation between the influencing factors and carbon emissions per unit area;

[0034] S43. Expansion and normalization of prediction model learning samples:

[0035] Expand the learning samples based on the thin plate spline interpolation algorithm and normalize the learning samples;

[0036] S44. Construction and evaluation of carbon emission prediction model:

[0037] A prediction model is established based on the machine learning algorithm. First, the hyperparameters of a single model and its optimization range are determined. The optimal hyperparameters of the model are found through grid optimization and k-fold cross-validation method to establish a carbon emission prediction model for the construction period of prefabricated buildings. The root mean square error and goodness of fit coefficient are used as evaluation indicators to compare the prediction effects of four machine learning prediction models, and the carbon emission prediction model for the construction period of prefabricated buildings with the best prediction performance is determined, thereby realizing the entire process of quantification and prediction of carbon emissions during the construction period of prefabricated buildings based on knowledge graphs.

[0038] The beneficial effects of the present invention are:

[0039] (1) The method of the present invention uses the CNKI database and the WoS core collection database to search for literature in the field of carbon emissions during the construction period of prefabricated buildings, and then performs a visual analysis of the knowledge graph based on the CiteSpace software. It can deeply analyze the core keywords and their interrelationships in the field of carbon emissions during the construction period of prefabricated buildings. The research content is comprehensive and can provide guidance for subsequent research.

[0040] (2) The method of the present invention improves the traditional carbon emission quantification method, establishes a carbon emission quantification model based on process analysis method and carbon emission factor method, establishes a carbon emission factor library suitable for prefabricated buildings, and constructs a carbon emission quantification formula for prefabricated building construction modules, thereby clarifying the influencing factor system and influencing mechanism of carbon emissions during the construction period of prefabricated buildings, providing a reference for carbon emission reduction during the construction period of prefabricated buildings;

[0041] (3) The method of the present invention analyzes the process correlation characteristics of the carbon emission quantification model case during the construction period of prefabricated buildings, clarifies the sources and proportions of carbon emissions from construction modules at each stage during the construction period of prefabricated buildings, and formulates emission reduction measures in a targeted manner to improve the emission reduction effect;

[0042] (4) The method of the present invention is based on the factors affecting carbon emissions, and uses four machine learning algorithms to conduct a predictive study on carbon emissions during the construction period of prefabricated buildings. The optimal prediction model is determined through comparative analysis. The obtained prediction model has a high degree of refinement, and the model boundary conditions, spatiotemporal processes, and quantitative statistical modules have little deviation from the actual situation of prefabricated buildings. The model parameters have an iterative optimization process, good stability, and high prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 : Technical roadmap of the method of the present invention;

[0044] Figure 2 : Improved framework diagram of the carbon emission quantification model for prefabricated buildings during construction period of the present invention;

[0045] Figure 3 : Correlation analysis diagram of factors affecting carbon emissions during the construction period of prefabricated buildings of the present invention;

[0046] Figure 4 : A flow chart of the carbon emission prediction model construction of the present invention;

[0047] Figure 5 : Carbon emission prediction result diagram of four machine learning prediction models of the present invention;

[0048] Figure 6 : Comparison chart of predicted values ​​and actual values ​​of the four machine learning models of the present invention. DETAILED DESCRIPTION

[0049] Embodiment 1:

[0050] like Figure 1-6 As shown in the figure, the method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph includes the following steps:

[0051] S1. Visualization of knowledge graph of carbon emissions during the construction period of prefabricated buildings:

[0052] Literature on carbon emissions during the construction period of prefabricated buildings was retrieved from the CNKI database and the WoS core collection database, and a knowledge graph visualization dataset of carbon emissions during the construction period of prefabricated buildings was constructed. Based on the CiteSpace knowledge graph visualization software, the research hotspot keywords in the field of carbon emissions during the construction period of prefabricated buildings were analyzed to obtain the research dynamics, keyword evolution paths and their interrelationships in the field of carbon emissions during the construction period of prefabricated buildings, and to identify the research hotspots and future trends in this field.

[0053] S2. Quantification of carbon emissions during the construction period of prefabricated buildings:

[0054] The basic framework of the carbon emission quantification model for the construction period of prefabricated buildings was established, and the basis and system boundaries of quantification were clarified; based on the comparative analysis of carbon emission quantification methods during the construction period of buildings, a carbon emission quantification model for the construction period of prefabricated buildings based on the combination of process analysis method and carbon emission factor method was established; carbon emission factors in current specifications and literature were analyzed and quantified, and a carbon emission factor library for the construction period of prefabricated buildings was constructed; a formula for carbon emission quantification during the construction period of prefabricated buildings was constructed, and a matrix product quantification program was developed using the Python programming language to improve the efficiency of carbon emission quantification.

[0055] The details are as follows:

[0056] S21. Basis for quantification of carbon emissions during the construction period of prefabricated buildings:

[0057] (1) Basis for quantification of carbon emissions:

[0058] Since there are many factors affecting carbon emissions during the construction period of prefabricated buildings, it is difficult to quantify carbon emissions from all construction processes. In terms of time, compared with cast-in-place construction, prefabricated construction focuses mainly on the installation of components, which greatly shortens the construction period. Therefore, the characteristics of carbon emissions during the construction period of prefabricated buildings are shorter cycles, more concentrated intensity, and greater difficulty in quantification. For the spatial dimension, in addition to conventional construction materials, prefabricated components are the core components of prefabricated buildings, and their carbon emissions are a unique consideration for prefabricated buildings. As a new type of building, it is difficult to accurately quantify carbon emissions in different construction scenarios using a single method. By investigating relevant prefabricated construction sites, including construction site inspections, project bills of quantities, and construction plan analysis, and referring to relevant national and industry standards, a basis for carbon emission quantification is established, as shown in Table 3.

[0059] Table 3. Basis for quantification of carbon emissions during the construction period of prefabricated buildings

[0060]

[0061] (2) Carbon emission quantification system boundaries:

[0062] Prefabricated concrete buildings were selected as the research object to quantitatively analyze their carbon emissions during the construction period. Carbon emissions during the construction period of prefabricated buildings refer to the carbon emissions generated by the consumption of materials, construction machinery, energy, and construction personnel activities during the construction process. Different system boundary ranges will affect the quantitative results of carbon emissions. Therefore, when conducting quantitative analysis of carbon emissions during the construction period of prefabricated buildings, it is necessary to accurately define the system boundaries of the model, including spatial boundaries and time boundaries. The construction period of prefabricated buildings is a complex system, whose inputs include building materials, construction machinery, transportation equipment, construction personnel, and various types of energy, and whose outputs include construction waste and carbon emissions.

[0063] The time boundary of carbon emission quantification is the construction stage of prefabricated buildings, including the foundation construction stage, the main construction stage, the decoration and renovation construction stage, the electromechanical and equipment installation construction stage, and the construction waste treatment stage. The research content of carbon emission quantification covers the consumption of building materials, material transportation, construction machinery, and carbon emissions of construction personnel. The spatial scope of carbon emission quantification covers the construction site of prefabricated buildings and the transportation routes of materials and prefabricated components.

[0064] S22. Improvement of the method for quantifying carbon emissions during the construction period of prefabricated buildings:

[0065] A modular quantitative study on carbon emissions during the construction period of prefabricated buildings was conducted. Based on the sub-projects of prefabricated buildings, the process analysis method and the carbon emission factor method were comprehensively used to improve the carbon emission quantitative model during the construction period of traditional buildings. This method can more accurately analyze and quantify the carbon emissions during the construction period of prefabricated buildings by analyzing the sub-projects of prefabricated building construction in detail. The framework for improving the quantitative model of carbon emissions during the construction period of prefabricated buildings is as follows: Figure 2 shown.

[0066] According to the General Specification for Construction Quality Control of Buildings and Municipal Engineering GB55032-2022

[107] and the Specification for Pricing of Bill of Quantities for Construction Engineering GB50500-2022

[108] , with the construction of the main structure of prefabricated buildings as the core, the carbon emission quantification model during the construction period of prefabricated buildings is divided into earthwork and foundation pit engineering module, foundation and pile foundation engineering module, reinforced concrete engineering module, prefabricated engineering module, decoration and renovation engineering module, electromechanical and equipment installation engineering module, and construction waste treatment module. The process analysis method and carbon emission factor method are applied to each module to conduct carbon source analysis and carbon emission quantification, so as to achieve detailed quantification of carbon emissions during the construction period of prefabricated buildings and ensure the accuracy and reliability of the data.

[0067] S23. Quantitative indicator system for carbon emissions during the construction period of prefabricated buildings:

[0068] S231, Carbon source quantitative index analysis:

[0069] In the carbon source analysis of construction sites, carbon emissions are divided into direct carbon emissions, embodied carbon emissions and indirect carbon emissions. Direct carbon emissions refer to carbon emissions generated by direct energy consumption during the construction process. Embedded carbon emissions are also called material carbon emissions, which mainly refer to the carbon emissions inherent in the materials used in the construction process. Indirect carbon emissions refer to carbon emissions that are associated with construction activities at the construction site but are not directly generated by on-site construction activities.

[0070] Analyze the specific sources of carbon emissions during the construction period of prefabricated buildings. Material carbon emissions refer to the carbon emissions generated by material consumption, which are embedded carbon emissions. Transportation carbon emissions include the process of transporting materials and prefabricated components from the factory to the construction site. Transportation carbon emissions are direct carbon emissions. Construction machinery carbon emissions refer to the carbon emissions generated during the use of construction machinery, including direct carbon emissions and indirect carbon emissions. Construction personnel carbon emissions refer to the carbon emissions generated by construction personnel during various construction activities on site during the construction of prefabricated buildings, which are direct carbon emissions.

[0071] S232, Carbon source quantitative index code:

[0072] The Pareto analysis method (ABC classification method) is introduced to analyze the carbon sources at the construction site. This method is based on the main technical or economic characteristics of things to classify and distinguish between key and non-key factors, so as to adopt different management strategies. In general, factors are divided into three categories: Class A factors (main influencing factors), with a cumulative occurrence frequency of 0% to 80%; Class B factors (secondary influencing factors), with a cumulative occurrence frequency of 80% to 90%; Class C factors (general influencing factors), with a cumulative occurrence frequency of 90% to 100%.

[0073] In terms of energy consumption, machine hours, material consumption and construction cost, more than 90% of the carbon sources generated in the corresponding construction modules will be included in the scope of quantification of carbon emissions during the construction period of prefabricated buildings.

[0074] S24. Selection and quantification of carbon emission factors:

[0075] The method for quantifying carbon emissions during the construction period of prefabricated buildings adopts a method based on a combination of process analysis and carbon emission factor methods. Carbon emission factor refers to the amount of carbon emissions generated by a certain energy or resource consumption per unit mass or volume. These carbon emission factors include carbon emission factors of electricity, fossil energy, building materials, transportation, and construction personnel, which form the basis for quantifying carbon emissions during the construction period of prefabricated buildings. A comprehensive analysis of commonly used carbon emission factors during the construction period of prefabricated buildings is conducted to screen out basic data suitable for prefabricated buildings, and determine the relevant carbon emission factor values ​​in the model.

[0076] (1) Electricity carbon emission factor: The electricity carbon emission factor for quantifying carbon emissions during the construction period of prefabricated buildings adopts the national unified electricity carbon emission factor. The latest national average emission factor of the power grid is 0.5703kgCO2e / kWh (0.5703 kilograms of carbon emissions per kilowatt-hour).

[0077] (2) Carbon emission factor of fossil energy: The carbon emission factor of fossil energy consumed by relevant mechanical equipment during the construction period of prefabricated buildings is selected according to the "Building Carbon Emission Calculation Standard" GB / T51366-2019. The average lower calorific value refers to the "General Rules for Comprehensive Energy Consumption Calculation" GB / T2589-2020. The carbon emission factor of fossil energy is calculated by quantitative formula, see Table 4.

[0078] Table 4. Carbon emission factors of fossil energy

[0079]

[0080] (3) Carbon emission factors of building materials: The carbon emission factors of building materials refer to the "Building Carbon Emission Calculation Standard" GBT51366-2019, combined with the raw material measurement values ​​in relevant research literature, for comprehensive analysis and calculation. According to the quantitative carbon source analysis of carbon emissions of prefabricated buildings in step S2.4, the carbon emission factors of some construction raw materials are shown in Table 5, and other finished materials and semi-finished materials are calculated on this basis.

[0081] Table 5. Carbon emission factors of building raw materials

[0082]

[0083] (4) Construction machinery carbon emission factor: During the construction of prefabricated buildings, construction machinery and equipment consume a large amount of fossil energy and electricity, making the carbon emission factor of construction machinery a key parameter for assessing environmental impact. According to the "Building Carbon Emission Calculation Standard"

[0084] The carbon emission factor of prefabricated construction machinery and equipment is calculated based on the energy consumption data of construction machinery and equipment per shift in GBT51366-2019, combined with the carbon emission factors of electricity and fossil energy.

[0085] (5) Transportation carbon emission factor: Material transportation in prefabricated buildings is divided into two main links: horizontal transportation and vertical transportation. The carbon emission factors of various transportation modes are shown in Table 6.

[0086] Table 6 Carbon emission factors of various modes of transportation

[0087]

[0088] (6) Carbon emission factor for construction workers: The carbon emission of construction workers is obtained by multiplying the number of working hours per unit by the carbon emission intensity per unit of working hours. The average annual carbon emission of Chinese construction workers is 1.30tCO2 equivalent, which is converted to 3.56kgCO2e / person / work day. In the actual quantification process, the carbon emission of labor should be adjusted according to the labor intensity, and the adjustment coefficient can be between 0.8 and 1.2. This method does not consider the differences between different types of work and different intensities.

[0089] S25. Construction of a quantitative model for carbon emissions during the construction period of prefabricated buildings:

[0090] The quantitative formula of the prefabricated carbon emission model is shown in formula (2). The quantitative formula includes seven modules during the construction period of prefabricated buildings. The formulas of the carbon emission quantitative model of the seven quantitative modules are shown in formulas (3) to (17).

[0091] C z =C tf +C jc +C gh +C zp +C zs +C jd +C k (2)

[0092] Where: C z is the total carbon emission during the construction period of prefabricated buildings, kgCO2e; C tf is the total carbon emission of earthwork and foundation pit construction module; C jc is the total carbon emission of foundation and pile construction modules, kgCO2e; C gh is the total carbon emission of reinforced concrete construction module, kgCO2e; C zpis the total carbon emission of prefabricated construction modules, kgCO2e; C zs is the total carbon emission of the decoration and renovation engineering modules during the construction period of prefabricated buildings, kgCO2e; C jd is the total carbon emission of the mechanical and electrical installation construction module, kgCO2e; C lc is the total carbon emission of the construction waste treatment module, kgCO2e;

[0093] Earthwork and foundation pit engineering module:

[0094] C tf =C ty +C tm +C my +C tj +C tr (3)

[0095]

[0096] Where: C ty Carbon emissions from earthwork transportation of earthwork and foundation pit engineering modules, kgCO2e; C tm is the carbon emission of material consumption, kgCO2e; C my Carbon emissions from material transportation, kgCO2e; C tj is the carbon emission of construction machinery, kgCO2e; C tr is artificial carbon emissions, kgCO2e; V t is the earthwork excavation volume, m 3 ; V c is the load volume of the dump truck, m 3 ; E a F is the energy consumption of earthwork transport vehicles, kg; e is the carbon emission factor of vehicle energy consumption, kgCO2e / kg; M ai is the consumption of the i-th material, t; F mai is the carbon emission factor of the i-th material, kgCO2e / t; D ai is the transportation distance of the i-th material, km; F yai is the carbon emission factor of transport vehicles, kgCO2e / (t·km); J ai is the number of shifts used for the i-th type of machinery, shifts; F jai is the carbon emission per shift of the i-th type of machinery, kgCO2e / shift; Q ra is labor consumption, man-day; F ra is the artificial carbon emission factor, kgCO2e / man-day.

[0097] Foundation and Pile Engineering Module:

[0098] C jc =Cjm +C jy +C jj +C jr (5)

[0099]

[0100] Where: C jm Carbon emissions from the consumption of foundation and pile foundation engineering module materials; C jy Carbon emissions from material transportation; C jj is the carbon emission of construction machinery; C jr is artificial carbon emissions; M bi is the amount of material consumed; F mbi is the carbon emission factor of the i-th material; D bi is the distance of transportation of the i-th material; F ybi is the carbon emission factor of transportation vehicles; J bi F is the number of hours used for the i-th machine; jbi is the carbon emission per shift of the i-th machine; Q rb F is labor consumption; rb is the artificial carbon emission factor.

[0101] Reinforced Concrete Engineering Module:

[0102] C gh =C gm +C gy +C gj +C gr (7)

[0103]

[0104] Where: C gm Carbon emissions from reinforced concrete engineering module material consumption; C gy is the carbon emissions from transportation; C gj is the carbon emission of construction machinery; C gr is artificial carbon emissions; M ci is the amount of material consumed; F mci is the carbon emission factor of the i-th material; D ci is the transportation distance of the i-th material; F yci is the carbon emission factor of transportation vehicles; J ci F is the number of hours used by the i-th machine in the module; jci is the carbon emission per shift of the i-th machine in the module; Q rc F is the labor consumption of the module; rc is the artificial carbon emission factor of the module.

[0105] Prefabricated engineering modules:

[0106] C zp =C zg +C zm +C zy +C zj +C zr (9)

[0107]

[0108] Where: C zg Carbon emissions from materials consumption and transportation for prefabricated construction modules; C zm C is the carbon emission of material consumption; zy Carbon emissions for material transportation; C zj is the carbon emission of construction machinery; C zr is artificial carbon emissions; M g F is the amount of material consumed for prefabricated components; gm is the carbon emission factor of prefabricated component materials; D gi F is the transportation distance of the i-th prefabricated component; ygi M is the carbon emission factor of the i-th prefabricated component transportation vehicle; di is the amount of material consumed; F mdi is the carbon emission factor of the i-th material; D mdi is the transportation distance of the i-th material; F ydi Carbon emission factor of materials transportation vehicles for prefabricated construction modules; J di The number of shifts used by the i-th type of machinery in the prefabricated construction module; F jdi Carbon emission factor of the i-th material consumed by the prefabricated construction module; Q rd Labor consumption of prefabricated construction modules; F rd Artificial carbon emission factors of prefabricated construction modules.

[0109] Decoration and renovation project module:

[0110] C zs =C sm +C sy +C sj +C sr (11)

[0111]

[0112] Where: C sm Carbon emissions from the consumption of modular materials for decoration and renovation projects; C sy Carbon emissions for material transportation; C sj is the carbon emission of construction machinery; C sr is artificial carbon emissions; M ei is the amount of material consumed; F meiis the carbon emission factor of the i-th material; D mei is the distance of transportation of the i-th material; F yei is the carbon emission factor of transportation vehicles; J ei F is the number of hours used for the i-th machine; jei is the carbon emission per shift of the i-th machine; Q re F is labor consumption; re is the artificial carbon emission factor.

[0113] Mechanical and electrical equipment installation engineering module:

[0114] C jd =C dm +C dy +C dj +C dr (13)

[0115]

[0116] Where: C dm Carbon emissions from material consumption for mechanical and electrical installation engineering modules; C dy Carbon emissions for material transportation; C dj is the carbon emission of construction machinery; C dr is artificial carbon emissions; M fi is the amount of material consumed; F mfi is the carbon emission factor of the i-th material consumption; D mfi is the distance of transportation of the i-th material; F yfi is the carbon emission factor of transportation vehicles; J fi F is the number of hours used for the i-th machine; fi is the carbon emission per shift of the i-th machine; Q rf F is labor consumption; rf is the artificial carbon emission factor.

[0117] Construction waste treatment engineering module:

[0118] Construction waste refers to the general term for construction waste, metal, concrete and other waste materials generated during the construction process. Construction waste has been considered in the earthwork transportation construction stage and is not measured in the waste treatment process. The remaining construction solid waste generation can be estimated according to formula (15).

[0119] M gi =S×m×k (15)

[0120] C lc =C lj +C ly (16)

[0121]

[0122] Where: M gi is the amount of construction waste generated; S is the building area; m is the base of waste generated per unit building area, and the steel-concrete structure building is taken as 0.03t / m 2

[105] ; k is the correction coefficient for garbage generation, which is 1.10-1.20 for cities with relatively fast economic development; 1.00-1.10 for economically developed cities; and 0.8-1.00 for ordinary cities; C lj Carbon emissions of garbage loading machinery and equipment for construction waste treatment engineering module; C ly Carbon emissions from garbage transportation equipment; D mgi is the distance for transporting the i-th type of construction waste; F ygi is the carbon emission factor of transportation vehicles; J gi F is the number of hours used for the i-th machine; gi is the carbon emission factor per shift of the i-th machine.

[0123] The quantification method of the carbon emission model during the construction period of prefabricated buildings is based on the process analysis method and the carbon emission factor method. ji , J ji , Y ji , R ji ) is defined as a primitive. Primitives are independent of each other, and each primitive has a specific meaning. The consumption of carbon emission sources of any construction module can be accurately calculated by the primitive number. If there is a default value in the carbon source analysis, it can be supplemented in subsequent case applications.

[0124] In order to make the quantitative model more complete, a module consumption sequence is established based on primitives. The module consumption sequence includes the building materials (M ji ), Construction machinery (J ji ), transportation equipment (Y ji ), and construction teams (R ji ). The j-th construction module consumes sequence H j , as shown in formula (18).

[0125]

[0126] Where: H j is the consumption sequence of the jth construction module; M ji is the consumption of the i-th building material in the j-th construction module; J ji Y is the number of construction shifts of the i-th machine in the j-th construction module; ji is the i-th type of transport equipment shift for the j-th construction module; R ji The jth construction module and the i-th construction staff team.

[0127] According to the actual situation of each module, the module consumption sequence is integrated to establish the consumption matrix of the module. This matrix expresses the building materials (M ji ), Construction machinery (J ji ), transportation equipment (Y ji ), and construction teams (R ji ) is shown in formula (19).

[0128]

[0129] In order to establish a unified quantification method, the carbon emission factors involved in each module are used to establish a carbon emission factor sequence similar to the consumption sequence. The sequence includes the carbon emission factors of building material consumption, building material transportation, construction machinery and equipment, and personnel teams. The establishment process is shown in equations (20) to (21). According to the actual consumption of a single module, the carbon emission factor sequence is expanded into a carbon emission factor matrix. The size of the matrix can be changed according to the actual carbon source consumption in the project, and the part that does not appear on the construction site is replaced by 0.

[0130]

[0131]

[0132] Where: F j is the j-th construction module matrix; F mji is the carbon emission factor of the i-th material of the j-th construction module; F Jji is the carbon emission factor of the i-th construction machinery in the j-th construction module; F Yji is the carbon emission factor of the i-th transportation vehicle in the j-th construction module; F Rji is the carbon emission factor of the i-th construction worker team in the j-th construction module.

[0133] Based on the quantitative methods of process analysis and carbon emission factor method, the basic formula of carbon emission factor method is expanded into matrix product. Carbon source consumption and carbon emission factor generate consumption matrix and carbon emission factor matrix, and establish carbon emission quantification matrix of prefabricated building construction period. This matrix can quantify carbon emission of prefabricated building in more detail and comprehensively. The carbon emission quantification matrix of prefabricated building construction period is shown in formula (22).

[0134]

[0135] The data required for carbon emission quantification can be extracted from the corresponding results in the quantization matrix. In order to convert a matrix into a matrix containing only its diagonal data, matrix transformation can be used to achieve this. Specifically, the Kronecker delta function and the Hadamard product operation of a given matrix are used to achieve this. The Kronecker delta function is a function widely used in multiple mathematical and engineering fields. In solving linear equations, especially in matrix operations and vector spaces, the Kronecker delta function is used to represent the unit matrix, which is the unit element of matrix multiplication. Its mathematical formula is shown in formula (23); the Hadamard product is a binary operation, which is widely used in mathematics, engineering and scientific computing. Its mathematical formula is shown in formula (24). Through the above operation, the final quantification result is shown in formula (25).

[0136]

[0137] C zij =C ij ·δ ij (twenty four),

[0138]

[0139] In the formula, C z is the total carbon emissions, C1, C2, …, C i Represents the carbon emissions of each construction module.

[0140] Based on the established carbon emission quantification model for prefabricated buildings during construction, the matrix product quantification method is used to develop a quantification program in the Python programming language. In the actual quantification code operation, the data source and code parameters can be modified according to the construction module to quantify the carbon emissions during the construction period of prefabricated buildings. In the code example, only six types of building materials, construction machinery, and transportation equipment construction personnel are assigned values, and the corresponding data sources are set. This method simplifies the quantification calculation process while improving the accuracy and precision of quantification.

[0141] S3. Case analysis of carbon emissions during the construction period of prefabricated buildings:

[0142] A prefabricated public parking building in Chongqing was selected to conduct a case analysis on the established carbon emission quantification model during the construction period of prefabricated buildings.

[0143] The carbon sources at the construction site are identified by three methods: construction site carbon source identification, engineering data carbon source identification, and drone carbon source identification. The carbon emission quantification model for the prefabricated building construction period established based on the S2 step collects carbon emission data from the earthwork and foundation pit construction module, foundation and pile foundation engineering construction module, reinforced concrete engineering construction module, prefabricated construction module, decoration and renovation engineering module, electromechanical and equipment installation construction module, and garbage disposal engineering module.

[0144] Based on the matrix product quantification method in the carbon emission quantification model, the carbon sources identified on site and the collected carbon emission data are converted into a consumption matrix. Based on the analysis of the carbon emission factors during the construction period of prefabricated buildings, the carbon emission factors corresponding to the carbon sources in the case are selected to generate a carbon emission factor matrix. Based on the matrix product quantification method, the carbon emissions of the seven modules during the construction period of prefabricated buildings are quantified, and the carbon emissions during the construction period of prefabricated buildings are calculated.

[0145] According to the quantification results of carbon emissions of prefabricated parking buildings, the visualization analysis method is used to analyze the overall characteristics of the quantification data of carbon emissions during the construction period of prefabricated public parking buildings. The analysis content includes the overall carbon emission characteristics analysis of carbon sources and the individual carbon emission characteristics analysis of carbon sources, and the various sources and proportions of carbon emissions during the construction period of prefabricated public parking buildings are obtained. For example, the carbon emissions from material consumption are the largest, accounting for 92.6%, reaching 1,270 tons. The second largest source is carbon emissions from transportation equipment and construction machinery, accounting for 4.4% and 2.4% respectively. Transportation equipment includes transportation and on-site transportation, and construction personnel account for the smallest proportion, which is 0.6%.

[0146] Then analyze the individual carbon sources in the carbon emission quantification data of prefabricated public parking buildings, and analyze the carbon emission characteristics of individual independent carbon sources in the carbon sources. For example: the largest carbon emission of the material consumption of prefabricated public parking buildings is steel bars, followed by concrete. The carbon emissions of the two carbon sources account for more than 80% in total, which is the main source of carbon emissions from material consumption. The top five construction machinery with the highest emissions include AC arc welding machines, crawler excavators, concrete truck pumps, steel bar processing machinery, and rotary drilling rigs. The carbon emissions of the five types of construction machinery are all above 30t, accounting for more than 70% in total. The top five transportation equipment with the highest carbon emissions include dump trucks, transfer trucks, concrete mixer trucks, trucks, and self-elevating tower cranes. Among them, dump trucks account for the largest proportion, reaching 32.9%, and have the highest emission reduction potential.

[0147] Based on visualization analysis and time series analysis methods, the overall characteristics of carbon emissions and construction module characteristics during the construction period of prefabricated buildings are analyzed, and targeted emission reduction measures are formulated according to the analysis of carbon emission characteristics.

[0148] S4. Study on prediction of carbon emissions during the construction period of prefabricated buildings:

[0149] S41. Comparative analysis of carbon emission prediction models: Through comparative analysis of building carbon emission prediction methods, a machine learning algorithm is selected to construct a carbon emission prediction model for the construction period of prefabricated buildings.

[0150] Traditional carbon emission forecasting methods mainly include trend extrapolation forecasting method, multivariate linear regression method, grey forecasting model method, system dynamics forecasting method and econometric forecasting method. When applied in the field of carbon emission forecasting, they have various defects and therefore cannot achieve the expected results.

[0151] By comparing and analyzing the common carbon emission prediction methods in the research of carbon emission prediction for traditional buildings and highway traffic projects, we selected machine learning algorithms including support vector regression, BP neural network, extreme learning machine, random forest algorithm, etc. to construct a carbon emission prediction model for the construction period of prefabricated buildings, which is suitable for carbon emission prediction during the construction period of prefabricated buildings.

[0152] S42. Learning sample data sources and correlation analysis of the prediction model: Combine knowledge graph visualization and carbon emission characteristic analysis results to identify key influencing factors of carbon emissions in prefabricated buildings, and apply the Pearson correlation test to test the degree of correlation between influencing factors and carbon emissions per unit area.

[0153] Carbon emission influencing factors are generally divided into two categories: one is building characteristic influencing factors, and the other is carbon source influencing factors. Building characteristic factors mainly include building area, number of building floors, building height, etc. In the knowledge graph visualization research in step S1, the results show that the assembly rate and the number of prefabricated components are hot topics in current research and are also important carbon emission influencing factors for prefabricated buildings. In the quantitative characteristic analysis of carbon emissions during the construction period of prefabricated buildings in step S3, building materials have always been the dominant carbon emission influencing factors for each construction module, among which the use of concrete and steel bars is the main factor.

[0154] In summary, the correlation analysis of influencing factors selected the above seven influencing factors of carbon emissions during the construction period of prefabricated buildings, namely: building area, number of building floors, building height, assembly rate, number of prefabricated components, concrete consumption, and steel consumption. Since the influencing factor parameters of different prefabricated buildings are different, the carbon emissions per unit area are added to verify the correlation of the influencing factors on carbon emissions.

[0155] The commonly used correlation test method is the Pearson correlation test. The Pearson correlation coefficient (r) is an important indicator for measuring the correlation between two random variables. When the value of r is between 0.8 and 1.0, it is a very strong correlation; when it is between 0.6 and 0.8, it is a strong correlation; when it is between 0.4 and 0.6, it is a moderate correlation; when it is between 0.2 and 0.4, it is a weak correlation; when it is between 0.0 and 0.2, it is a very weak correlation or no correlation. The positive and negative signs of the r value represent the significance of the correlation direction. A positive value indicates a positive correlation, and a negative value indicates a negative correlation. The calculation formula for the r value is as follows:

[0156]

[0157] To discuss whether two variables are related, we must discuss the significance level p. The correlation between the two may be caused by accidental factors. Taking 5% as the threshold, if p<0.05, it means that the linear relationship between the two variables is significant. Figure 3 The results of the Pearson correlation test are shown in Table 7.

[0158] Table 7. Pearson correlation test results of influencing factors

[0159]

[0160] Depend on Figure 6 From the results of the Pearson correlation test in Table 7, we can see that the r values ​​are all greater than 0.4, and the p values ​​are all less than 0.05. The seven variables are significantly correlated with carbon emissions per unit area, and the correlations are all above medium. Therefore, these seven influencing factors can become significant influencing factors restricting carbon emissions during the construction period of prefabricated buildings.

[0161] S43. Expansion and normalization of prediction model learning samples: expand the learning samples based on the thin plate spline interpolation algorithm, and normalize the learning samples.

[0162] In order to ensure accurate distinction between different prediction models, the scale of training learning samples should be as large as possible. Since the number of learning samples in this study is small, it may lead to poor generalization ability of the prediction model. Therefore, the data interpolation method is used to expand the learning samples, and the original 60 groups of learning samples are expanded to 120 groups through the data interpolation method.

[0163] Thin plate spline interpolation is a relatively new two-dimensional interpolation method. The basic principle of the thin plate spline interpolation algorithm is similar to placing a very thin metal plate on a set of points and then letting it bend naturally to adapt to the positions of these points. The core principle of this method is to find a smooth function that takes a specific value at a given data point. Suppose there is a set of data points (x i y i ,z i ), x i and i is the coordinate of the point, z i is the function value at these points. The formula of thin plate spline interpolation function f(x,y) is as follows:

[0164]

[0165] In the formula, a1, a x ,a y and wi represents the parameter to be solved; U represents the radial basis function, and U(r)=r2log(r) is usually selected.

[0166] The differences in the dimensions of the various influencing factors of carbon emissions will affect the prediction accuracy of carbon emissions, such as the different units of building area and assembly rate, and the differences in the construction sites of different prefabricated building projects. Therefore, the learning sample data is normalized before training to reduce the impact of the differences in influencing factors, eliminate the impact of dimensions, and improve the fitting effect of the carbon emission prediction model. In this method, Min-Max normalization is used, and the carbon emission learning samples are imported into the Matlab program and normalized using the mapminmax function. The mapping range of the characteristic parameters is set to 0-1. The normalized data results of the learning sample data are shown in Table 8.

[0167] Table 8. Normalized data results of prediction model

[0168]

[0169]

[0170] S44. Construction and evaluation of carbon emission prediction model:

[0171] S441, Carbon emission prediction model construction:

[0172] Based on the machine learning algorithm, the prediction model is established. First, the hyperparameters of a single model and its optimization range are determined, and the optimal hyperparameters of the model are found through grid optimization and k-fold cross validation. Based on the optimal hyperparameters, a prediction model for carbon emissions during the construction period of prefabricated buildings is established.

[0173] Compare the prediction effects of the four carbon emission prediction models and select the optimal carbon emission prediction model. The model evaluation indicators are root mean square error (RMSE) and goodness of fit coefficient (R2). The prediction model is run in the Matlab2022b environment. The prediction model construction process is as follows Figure 4 shown.

[0174] S442, Carbon emission prediction model parameter optimization:

[0175] Grid optimization and k-fold cross validation were used to optimize the parameters of support vector regression, BP neural network, extreme learning machine and random forest models. The cross validation folds (k) were set to 3, 4, 5 and 6 respectively, and the root mean square error (RMSE) was used as the evaluation index. In theory, the closer the RMSE value is to 0, the smaller the model prediction error is and the more ideal the selected hyperparameter settings are.

[0176] (1) Optimization of support vector regression prediction model parameters:

[0177] The key hyperparameters of support vector regression include penalty factor C and kernel function scale parameter γ. In order to determine the optimal combination of hyperparameters, the optimization range of penalty factor C is set to 0.1~10, with a step size of 0.1; the optimization range of kernel function scale parameter γ is set to 0.01~1, with a step size of 0.01. The grid optimization method is used to statistically analyze the optimal parameter settings under different cross-validation folds, and the following results are obtained: Under the four cross-validation conditions, the RMSE values ​​first decrease and then increase with the increase of penalty factor C and kernel function scale parameter γ; with the increase of cross-validation folds, the RMSE shows a trend of first decreasing and then increasing; the optimal parameter optimization results are: penalty factor C is 0.7, kernel function scale parameter γ is 0.1, at this time, RMSE is the smallest, which is 0.407, and the cross-validation fold k is 4.

[0178] (2) Optimization of BP neural network prediction model parameters:

[0179] For BP neural network, the number of hidden layers, the number of hidden layer nodes and the learning rate are the key parameters that determine its complexity and performance, and directly affect the prediction ability and convergence speed of BP neural network. Due to the limited number of learning samples, the hidden layer configuration of BP neural network only considers single layer and double layer. The number of nodes in single hidden layer and double hidden layer and the learning rate of BP neural network are optimized by grid optimization and k-fold cross validation method. The optimization range of the number of hidden layer nodes is set to 1-50, with a step size of 1; the optimization range of the learning rate is set to 0.0002-0.01, with a step size of 0.0002. The number of network training is set to 1000 times.

[0180] The optimal parameter settings under different cross-validation were classified and statistically analyzed, and the optimization results are as follows: After comparison, the prediction effect of the double hidden layer is better than that of the single hidden layer. Finally, the parameters of the BP neural network are selected as a double hidden layer, with the number of hidden layer nodes being 5 and 32 respectively, and the learning rate being 0.0098. At this time, the cross-validation fold is 4, and the RMSE is 0.327.

[0181] (3) Optimizing the parameters of the extreme learning machine prediction model:

[0182] The main parameter of the extreme learning machine that needs to be adjusted is the number of hidden nodes. In order to find the optimal parameter setting, the number of hidden layer nodes n is set to 1-100, with a step size of 1. The optimal parameter settings under different cross-validation are classified and statistically analyzed, and the optimization results are as follows: Under the four cross-validation conditions, as the number of hidden layer nodes increases, the RMSE value shows a trend of first stabilizing and then increasing with oscillation; as the number of cross-validation folds increases, the RMSE has no obvious trend of change; The optimal parameter optimization results are: the number of hidden layer nodes is 22, the RMSE at this time is 0.430, and the number of cross-validation folds is 5.

[0183] (4) Optimizing parameters of random forest prediction model

[0184] In the random forest prediction model, the number of decision trees (n) and the minimum number of leaves (leaf) are two important parameters. A larger n can improve the accuracy of the model, but it will also increase the calculation time. A larger leaf can prevent overfitting, but it may also cause the model to be underfitted. Therefore, it is necessary to optimize the parameters n and leaf of the random forest carbon emission prediction model. Since the number of learning samples is 120, the range of n is set to 1 to 30, the step size is set to 1, and the range of leaf is set to 1 to 30, and the step size is set to 1 for optimization.

[0185] The optimization results are as follows: RMSE decreases as the number of leaves increases, and increases as n increases; when n is set to 6 and leaf is set to 5, the model has the best prediction effect, the minimum RMSE of the model is 0.463, and the cross-validation fold is 4.

[0186] S443. Analysis of prediction results of carbon emission prediction model:

[0187] When the cross-validation fold k=4, the four machine learning models of support vector regression, BP neural network, extreme learning machine, and random forest have the smallest prediction model error for carbon emissions during the construction period of prefabricated buildings. In order to compare the prediction effects of the four models on carbon emissions during the construction period under the same conditions, the learning samples are divided into training sets and test sets at a ratio of 3:1. The number of learning samples is 120 groups, so among the model learning samples, 90 groups are training sets and 30 groups are test sets. The model parameters are set according to the parameters that achieve the minimum RMSE when the cross-validation fold k=4. The evaluation indicators of the carbon emission prediction model are the root mean square error (RMSE) and the goodness of fit coefficient (R2). The smaller the RMSE, the smaller the model error; the closer R2 is to 1, the higher the prediction accuracy of the model. The carbon emission prediction results are shown in the figure. Figure 5 As shown in Figure 2, the comparison between the predicted carbon emission value and the actual value is as follows: Figure 6 As shown in Table 9, the performance index comparison of carbon emission prediction model is shown in Table 9.

[0188] Table 9. Comparison of performance indicators of carbon emission prediction models

[0189]

[0190]

[0191] Depend on Figure 5It can be seen that the RMSE of the support vector regression prediction model is 0.385, the RMSE of the BP neural network prediction model is 0.364, the RMSE of the extreme learning machine prediction model is 0.404, and the RMSE of the random forest prediction model is 0.413. Judging from the RMSE, the BP neural network prediction model has the best carbon emission prediction effect and the smallest prediction result error. The random forest prediction model has the worst carbon emission prediction effect, which is consistent with the results of the carbon emission prediction model parameter optimization in step S442. Figure 6 From Table 9, we can see that the R2 of the four prediction models are 0.859, 0.878, 0.848 and 0.841 respectively. The comparison points between the carbon emission prediction value and the actual value of the BP neural network prediction model are more concentrated on the reference line, the goodness of fit coefficient is lower, and its carbon emission prediction effect is the best. Figure 6 (b) It can be seen that the comparison points between the carbon emission prediction value and the actual value of the BP neural network prediction model are evenly distributed on both sides of the reference line, indicating that the size of carbon emissions has no effect on the results of the prediction model. The model is stable in predicting carbon emissions and has good applicability.

[0192] In summary, the carbon emission prediction model for prefabricated buildings during construction period based on BP neural network has the best prediction effect, and the prediction effect of this model is relatively stable.

[0193] The above is an exemplary description of the invention. Obviously, the specific implementation of the present invention is not limited to the above-mentioned method. As long as such non-substantial improvements are made by adopting the method concept and technical solution of the present invention, or the concept and technical solution of the invention are directly applied to other occasions without improvement, they are all within the protection scope of the present invention.

Claims

1. A method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph, characterized in that: The following steps are involved: S1. Visualization of knowledge graph of carbon emissions during the construction period of prefabricated buildings: Retrieve literature on carbon emissions during the construction period of prefabricated buildings from the CNKI database and the WoS core collection database, and select and construct a knowledge graph visualization dataset of carbon emissions during the construction period of prefabricated buildings; Based on the CiteSpace knowledge graph visualization software, deeply analyze the research hotspot keywords in the field of carbon emissions during the construction period of prefabricated buildings, obtain the research dynamics in the field of carbon emissions during the construction period of prefabricated buildings, the evolution path of keywords and their mutual relationships, and identify the research hotspots and future trends in this field; S2. Quantification of carbon emissions during the construction period of prefabricated buildings: Based on the comparative analysis of carbon emission quantification methods during the construction period, a carbon emission quantification model for prefabricated building construction period based on the combination of process analysis method and carbon emission factor method is established; carbon emission factors in current specifications and literature are analyzed and quantified, and a carbon emission factor library for prefabricated building construction period is constructed; a carbon emission quantification formula for prefabricated building construction period is constructed, and a matrix product quantification program is developed using Python programming language to improve the efficiency of carbon emission quantification; S3. Case analysis of carbon emissions during the construction period of prefabricated buildings: A case study of a prefabricated public parking building in Chongqing was conducted to analyze the established carbon emission quantification model for prefabricated buildings during construction. The carbon sources at the construction site were identified by three methods: construction site carbon source identification, engineering data carbon source identification, and drone carbon source identification. The carbon emission data during the construction period were collected, and the carbon emissions during the construction period of prefabricated buildings were calculated using the matrix product quantification method in different modules. The overall characteristics of carbon emissions during the construction period of prefabricated buildings and the characteristics of construction modules were analyzed based on visualization analysis and time series analysis methods, and emission reduction measures were formulated according to the analysis of carbon emission characteristics. S4. Study on prediction of carbon emissions during the construction period of prefabricated buildings: Select a machine learning algorithm to build a carbon emission prediction model for the construction period of prefabricated buildings; combine knowledge graph visualization and carbon emission feature analysis results to identify key influencing factors of carbon emissions of prefabricated buildings; expand learning samples based on the thin plate spline interpolation algorithm, use grid optimization and k-fold cross validation to find the optimal hyperparameters in the model, and use this to establish a carbon emission prediction model; use root mean square error and goodness of fit coefficient as evaluation indicators to compare the prediction effects of machine learning prediction models, and determine the carbon emission prediction model with the best prediction performance, thereby realizing the entire process of quantification and prediction of carbon emissions during the construction period of prefabricated buildings based on knowledge graphs.

2. According to claim 1, the method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph is characterized in that: The S2 step is as follows: S21. Carbon emission quantification basis: Establish the basic framework of the carbon emission quantification model during the construction period of prefabricated buildings, and clarify the basis for quantification and the system boundaries; S22. Improvement of carbon emission quantification method: Carry out modular quantitative research on carbon emissions during the construction period of prefabricated buildings. Based on the sub-projects of prefabricated buildings, comprehensively use the process analysis method and the carbon emission factor method to improve the carbon emission quantitative model during the construction period of traditional buildings, and establish a carbon emission quantitative model for the construction period of prefabricated buildings based on the combination of the process analysis method and the carbon emission factor method. S23. Carbon emission quantitative indicator system: Conduct quantitative index analysis on the carbon sources at the construction site of prefabricated buildings during construction, identify the specific sources of carbon sources, and then introduce the Pareto analysis method to classify and analyze the carbon sources at the construction site; S24. Selection and quantification of carbon emission factors: The carbon emissions during the construction period of prefabricated buildings are quantified by combining the process analysis method with the carbon emission factor method, and the values ​​of all carbon emission factors in the carbon emission quantification model are determined to construct a carbon emission factor library during the construction period of prefabricated buildings. S25. Construction of carbon emission quantification model: Construct a formula for quantifying carbon emissions during the construction period of prefabricated buildings, and use the Python programming language to develop a matrix multiplication quantification program to improve the efficiency of carbon emissions quantification.

3. The method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph according to claim 2 is characterized in that: In step S22, the carbon emission quantification model for the construction period of prefabricated buildings includes: earthwork and foundation pit engineering module, foundation and pile foundation engineering module, reinforced concrete engineering module, prefabricated engineering module, decoration and renovation engineering module, electromechanical and equipment installation engineering module, and construction waste treatment engineering module.

4. The method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph according to claim 2 is characterized in that: In step S24, the carbon emission factors include electricity, fossil energy, building materials, construction machinery, transportation and construction personnel.

5. The method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph according to claim 2 is characterized in that: In step S25, the carbon emission quantification matrix of the prefabricated building construction period is shown in formula (22): In the formula, E j is the consumption matrix of the jth construction module, M ji is the consumption of the i-th building material in the j-th construction module, J ji Y is the construction shift of the i-th machine in the j-th construction module, ji is the i-th type of transport equipment shift in the j-th construction module, R ji The i-th construction worker team of the j-th construction module; F j is the j-th construction module matrix, F mji is the carbon emission factor of the i-th material of the j-th construction module, F Jji is the carbon emission factor of the i-th construction machinery in the j-th construction module, F Yji is the carbon emission factor of the i-th transportation vehicle in the j-th construction module, F Rji is the carbon emission factor of the i-th construction worker team in the j-th construction module; i is a, b, c…g respectively; Convert a matrix into a matrix containing only its diagonal data using the Kronecker delta function and the given matrix Hadamard product operation. The final quantization result is shown in formula (25): In the formula, C z is the total carbon emissions, C1, C2, …, C i Represents the carbon emissions of each construction module.

6. The method for quantifying and predicting carbon emissions during the construction period of prefabricated buildings based on knowledge graph according to claim 1 is characterized in that: The S4 step is specifically as follows: S41. Comparative analysis of carbon emission prediction models: Through comparative analysis of building carbon emission prediction methods, a machine learning algorithm is selected to construct a carbon emission prediction model for prefabricated building construction period. The machine learning algorithm includes support vector regression, BP neural network, extreme learning machine, and random forest algorithm. S42. Sources and correlation analysis of learning sample data for prediction model: Combining the knowledge graph visualization and carbon emission characteristic analysis results, the key influencing factors of carbon emissions of prefabricated buildings are identified, and the Pearson correlation test is applied to test the degree of correlation between the influencing factors and carbon emissions per unit area; S43. Expansion and normalization of prediction model learning samples: Expand the learning samples based on the thin plate spline interpolation algorithm and normalize the learning samples; S44. Construction and evaluation of carbon emission prediction model: A prediction model is established based on a machine learning algorithm. First, the hyperparameters of a single model and their optimization range are determined. The optimal hyperparameters of the model are found through grid optimization and k-fold cross-validation methods to establish a prediction model for carbon emissions during the construction period of prefabricated buildings. Taking the root mean square error and goodness of fit coefficient as evaluation indicators, the prediction effects of four machine learning prediction models are compared, and the carbon emission prediction model for the construction period of prefabricated buildings with the best prediction performance is determined, thereby realizing the entire process of quantification and prediction of carbon emissions during the construction period of prefabricated buildings based on knowledge graphs.

Images