Method and system for predicting carbon emission in railway infrastructure construction process

Abstract

The invention relates to the technical field of carbon emission prediction, and particularly discloses a railway infrastructure construction process carbon emission prediction method and system.The method comprises the steps that firstly, carbon emission basic data in the railway infrastructure construction process are extracted from railway infrastructure related IFC files, and meanwhile a related carbon emission feature database is constructed; on the basis of the Hash similarity between each carbon emission basic data and each carbon emission characteristic data element in the database, querying a corresponding carbon emission factor to calculate the carbon emission, and further introducing a deep learning algorithm; a plurality of candidate carbon emission characteristic data elements with relatively high similarity are subjected to semantic analysis and dynamic query matching fitting to deduce corresponding carbon emission factors, so that carbon emission calculation is realized. According to the method, the carbon emission characteristics of novel composite materials and construction activities can be effectively handled, and the generalization ability of carbon emission prediction is improved, so that powerful support is provided for carbon emission management in the railway industry.

Description

Technical Field

[0001] The present application relates to the technical field of carbon emission prediction, and more specifically, to a method and system for predicting carbon emissions during railway infrastructure construction. Background Art

[0002] As an important large-scale engineering field, the carbon emission problem of railway infrastructure construction has received increasing attention. As the scale of railway construction continues to expand, the impact of carbon emissions generated during the construction of railway infrastructure on the environment has become increasingly significant. Therefore, accurately monitoring carbon emissions during the construction of railway infrastructure is of vital importance to promoting energy conservation and emission reduction in the railway industry.

[0003] At present, the accounting of carbon emissions in the construction process of railway infrastructure is mainly based on the empirical formula method of static database, and carbon emissions are calculated by manually matching construction activities with the carbon emission factor library. However, its generalization ability is limited by the coverage of the database, and it is difficult to cope with the carbon emission characteristics of new composite materials and non-standard activities in railway projects. When encountering new building materials, carbon emission factor matching failure is prone to occur.

[0004] Therefore, it is necessary to provide an optimized carbon emission prediction method and system for the railway infrastructure construction process to solve the above technical problems. Summary of the invention

[0005] In order to solve the above technical problems, this application is proposed.

[0006] According to one aspect of the present application, a method for predicting carbon emissions during railway infrastructure construction is provided, which comprises: Extracting basic carbon emission data during the construction of railway infrastructure from relevant IFC files of railway infrastructure, including the type and amount of materials, transportation tools and distance of transportation activities, and energy type and consumption of construction machinery; Constructing a carbon emission characteristic database, wherein the carbon emission characteristic database stores carbon emission factors of various materials, energy sources and transportation tools in a table structure; Based on the carbon emission characteristic database, fuzzy matching and dynamic fitting derivation are performed on each data item in the carbon emission basic data to determine its carbon emission factor, which includes: Extracting a data item to be queried from the carbon emission basic data, wherein the data item to be queried is one of a material type, a transportation tool type, and an energy type; Calculating hash values ​​of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities; If the maximum value in the set of hash similarities is greater than or equal to a preset threshold, the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value is used as the carbon emission factor of the data item to be queried; If the maximum value in the set of hash similarities is less than the preset threshold, based on the descending order of the set of hash similarities, extracting the carbon emission characteristic data elements corresponding to the first N hash similarities to construct a set of candidate carbon emission characteristic data elements; Inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into a carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried; Based on each data item in the carbon emission basic data and its corresponding carbon emission factor, the total carbon emissions during the railway infrastructure construction process are calculated.

[0007] According to another aspect of the present application, a railway infrastructure construction process carbon emission prediction system is provided, which includes: A carbon emission data extraction module is used to extract basic carbon emission data during the construction of railway infrastructure from relevant IFC files of railway infrastructure. The basic carbon emission data includes the type and amount of materials, the means of transport and distance of transportation activities, and the energy type and consumption of construction machinery; A carbon emission database construction module is used to construct a carbon emission characteristic database, which stores carbon emission factors of various materials, energy sources and transportation tools in a table structure; The data matching and fitting module is used to perform fuzzy matching and dynamic fitting derivation on each data item in the carbon emission basic data based on the carbon emission characteristic database to determine its carbon emission factor, which includes: Extracting a data item to be queried from the carbon emission basic data, wherein the data item to be queried is one of a material type, a transportation tool type, and an energy type; Calculating hash values ​​of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities; If the maximum value in the set of hash similarities is greater than or equal to a preset threshold, the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value is used as the carbon emission factor of the data item to be queried; If the maximum value in the set of hash similarities is less than the preset threshold, based on the descending order of the set of hash similarities, extracting the carbon emission characteristic data elements corresponding to the first N hash similarities to construct a set of candidate carbon emission characteristic data elements; Inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into a carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried; The carbon emission calculation module is used to calculate the total carbon emissions during the construction of railway infrastructure based on each data item in the carbon emission basic data and its corresponding carbon emission factor.

[0008] The present application has at least the following technical effects: Compared with the prior art, the present application provides a method and system for predicting carbon emissions during the construction of railway infrastructure. It first extracts the basic carbon emission data during the construction of railway infrastructure from the relevant IFC files of railway infrastructure, and at the same time constructs a relevant carbon emission feature database. Based on the hash similarity between each carbon emission basic data and each carbon emission feature data element in the database, the corresponding carbon emission factor is queried to calculate the carbon emissions. At the same time, when the data elements in the carbon emission feature database are not sufficient to meet the query requirements, a deep learning algorithm is further introduced to perform semantic parsing and dynamic query matching and fitting of multiple candidate carbon emission feature data elements with high similarity to infer the corresponding carbon emission factors and thus realize carbon emission calculation. The present application can effectively respond to the carbon emission characteristics of new composite materials and construction activities, improve the generalization ability of carbon emission prediction, and thus provide strong support for carbon emission management in the railway industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0010] Figure 1 It is a flowchart of a method for predicting carbon emissions during railway infrastructure construction according to an embodiment of the present application.

[0011] Figure 2 Schematic diagram of data flow of a method for predicting carbon emissions during railway infrastructure construction according to an embodiment of the present application.

[0012] Figure 3 This is a flowchart of sub-step S3 of the method for predicting carbon emissions during railway infrastructure construction according to an embodiment of the present application.

[0013] Figure 4 This is a flowchart of sub-step S35 of the method for predicting carbon emissions during the railway infrastructure construction process according to an embodiment of the present application.

[0014] Figure 5 This is a flowchart of sub-step S352 of the method for predicting carbon emissions during the railway infrastructure construction process according to an embodiment of the present application.

[0015] Figure 6 This is a flowchart of sub-step S3521 of the method for predicting carbon emissions during the railway infrastructure construction process according to an embodiment of the present application.

[0016] Figure 7 This is a flowchart of sub-step S3522 of the method for predicting carbon emissions during the railway infrastructure construction process according to an embodiment of the present application.

[0017] Figure 8 It is a block diagram of a carbon emission prediction system for a railway infrastructure construction process according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] As shown in this application and claims, unless the context clearly indicates an exception, the words "a", "an", "an" and / or "the" do not refer to the singular and may also include the plural. Generally speaking, the terms "include" and "comprise" only indicate the inclusion of the steps and elements that have been clearly identified, and these steps and elements do not constitute an exclusive list. The method or device may also include other steps or elements.

[0019] Although the present application makes various references to certain modules in the system according to the embodiments of the present application, any number of different modules can be used and run on the user terminal and / or server. The modules are only illustrative, and different aspects of the system and method can use different modules.

[0020] Flowcharts are used in the present application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps may be processed in reverse order or simultaneously as required. Meanwhile, other operations may also be added to these processes, or a certain step or several steps of operations may be removed from these processes.

[0021] Below, the exemplary embodiments according to the present application will be described in detail with reference to the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application, and it should be understood that the present application is not limited to the exemplary embodiments described here.

[0022] It should be noted in advance that the acquisition and processing of all information or data in this application are carried out in compliance with the relevant data protection laws and policies of the place where you are located, and with the authorization of the authority administrator.

[0023] Figure 1 It is a flowchart of a method for predicting carbon emissions during railway infrastructure construction according to an embodiment of the present application. Figure 2 Schematic diagram of data flow of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 1 and Figure 2 As shown, the method for predicting carbon emissions during the construction process of railway infrastructure includes the following steps: S1, extracting basic carbon emission data during the construction process of railway infrastructure from relevant IFC files of railway infrastructure, wherein the basic carbon emission data include the type and amount of materials, the means of transport and distance of transportation activities, and the energy type and consumption of construction machinery; S2, constructing a carbon emission characteristic database, wherein the carbon emission characteristic database stores carbon emission factors of various materials, energy and means of transport in a table structure; S3, based on the carbon emission characteristic database, performing fuzzy matching and dynamic fitting deduction on each data item in the carbon emission basic data to determine its carbon emission factor; S4, calculating the total carbon emissions during the construction process of railway infrastructure based on each data item in the carbon emission basic data and its corresponding carbon emission factor.

[0024] In the above-mentioned method for predicting carbon emissions during the construction of railway infrastructure, the step S1 extracts the basic data of carbon emissions during the construction of railway infrastructure from the relevant IFC files of railway infrastructure, and the basic data of carbon emissions include the type and amount of materials, the means of transportation and distance of transportation activities, and the energy type and consumption of construction machinery. It should be understood that railway construction is an extremely complex systematic project with a long construction period and a wide range of involvement. During the construction process, the procurement and use of materials, the allocation of different modes of transportation, and the operation of various types of construction machinery will all generate carbon emissions. As an international standard format for digital information exchange in the construction industry, the IFC (Industry Foundation Classes) file comprehensively and normatively covers all types of information in all stages of railway infrastructure construction from planning and design to construction, providing an ideal data carrier for accurately extracting basic carbon emission data. Specifically, IFC files follow a strict and standardized information model structure, which comprehensively describes construction project information by clearly defining various entities (such as building materials, construction equipment, transportation activities, etc.), the attributes of entities (such as the material and amount of materials, the model and energy consumption of equipment, the type of transportation tools, the travel distance, etc.), and the relationship between entities (such as the association between materials and construction sites, the correspondence between equipment and construction tasks, etc.). By using IFC parsing engines (such as IfcOpenShell, etc.) to parse IFC files related to railway infrastructure and extract various entities and their attributes in the files, specific carbon emission basic data can be provided for carbon emission accounting during the construction process.

[0025] Specifically, when processing IFC files to obtain material types and quantities, it is necessary to dig deep into the file to find all the relevant information about building materials. Specific entities can be found in IFC files, which represent all the different types of materials used on the construction site. Each material has its own unique identifier, including but not limited to the material name, specification parameters, etc. These identifiers help identify the specific type of material. At the same time, for each material, the IFC file also contains detailed usage data, which is usually reflected in which parts or components the material is applied to. By analyzing these data, the material usage accurate to each construction link can be obtained, thereby providing accurate data support for subsequent carbon emission accounting.

[0026] Transportation activities are another important source of carbon emissions, so it is particularly important to extract relevant data from IFC files. Information on transportation activities not only covers the type of transportation tools, but also includes important parameters such as the distance traveled. Through clear entity definitions, IFC files closely link transportation activities with their corresponding construction tasks to ensure that every transportation activity is accurately recorded. In actual operation, by parsing these entities, we can clearly understand the basis for the selection of transportation tools and their route planning. This is of great significance for evaluating carbon emissions caused by transportation. In addition, transportation distance, as an important factor affecting carbon emissions, is also recorded in detail in IFC files, making it possible to calculate carbon emissions under different modes of transportation.

[0027] For construction machinery, its energy type and consumption is one of the key factors that determine carbon emissions. The information about construction machinery in the IFC file is rich and detailed, including not only basic information about the equipment such as model and power, but also detailed energy consumption data. This information helps to accurately calculate the carbon emissions generated by the operation of mechanical equipment during the construction process. Specifically, by analyzing the equipment energy consumption information provided in the IFC file and combining it with the actual working time, the total energy consumption of a certain type of construction machinery in a certain period of time can be estimated. This process requires careful consideration of the actual use of the equipment, including but not limited to factors such as the equipment's working efficiency and load changes, in order to more accurately reflect the energy consumption status in real conditions.

[0028] In the specific implementation process, open source libraries such as IfcOpenShell can be used to effectively parse IFC files and extract the required data. It supports multiple programming language interfaces and is easy to integrate into different system environments. Using such tools, specific information about materials, transportation, and mechanical equipment can be extracted from IFC files and converted into a data format for further processing. In this process, attention should also be paid to data quality control to ensure that the extracted data is accurate. This may involve multiple steps such as data cleaning and verification, aiming to improve the reliability and accuracy of the data.

[0029] In the above-mentioned method for predicting carbon emissions during the construction of railway infrastructure, the step S2 constructs a carbon emission feature database, and the carbon emission feature database stores the carbon emission factors of various materials, energy sources and transportation tools in a table structure. It should be understood that the carbon emission factor is used to characterize the carbon emissions generated by a unit amount of materials, energy sources or transportation tools during use, and is a key parameter for calculating carbon emissions. In this application, a table structure is used to store basic attribute information and carbon emission factors of various materials, energy sources and transportation tools, which is helpful for fast query and matching. Specifically, each row in the carbon emission feature database represents a specific material, energy source or transportation tool, and each column corresponds to its different attributes. For example, for materials, column attributes may include material name, material type (further subdivided into metal materials, inorganic non-metallic materials, organic polymer materials, new composite materials, etc.), material production process (carbon emissions of different processes vary greatly), carbon emission factor per unit dosage, etc. For energy, column attributes can include energy name (such as diesel, gasoline, electricity, coal, hydropower, wind power, etc.), energy source (thermal power may have different carbon emissions due to different power generation methods, such as coal-fired power and gas-fired power; hydropower needs to consider carbon emissions during the construction process, etc.), and carbon emission factor per unit of energy consumption. For transportation tools, column attributes can include transportation tool name, type (trucks can be subdivided into different loads and power types, trains can be divided into different models, and ships also have multiple types), load, carbon emission factor per unit of transportation distance, etc. By establishing this structured data storage method and using the powerful indexing and query functions of the database management system, it is possible to quickly locate and extract the required carbon emission factor data based on the input material, energy or transportation tool related information.

[0030] Specifically, for materials, a separate row record is created in the database for each material, and each column represents a different property. The material name serves as a basic identifier to help accurately identify each material. Next, the material type is classified in detail, such as metal materials, inorganic non-metallic materials, organic polymer materials, new composite materials, etc., which helps to define the characteristics of each material more accurately. In addition, the material production process is also an important attribute. Different process flows have a significant impact on carbon emissions, so this information is crucial for calculating the carbon emission factor per unit of use. In this structured way, a detailed description can be provided for each material, including not only basic properties, but also parameters directly related to carbon emissions. For example, when it comes to new composite materials, due to their complex composition and production process, special attention needs to be paid to their unique carbon emission characteristics so that the actual situation can be accurately reflected in subsequent calculations.

[0031] In terms of energy, the database also needs to be finely divided into different types and sources, using energy names as basic identifiers, such as diesel, gasoline, electricity, coal, hydropower, wind power, etc., to clearly distinguish various forms of energy. Then, it is further refined to the source of energy. For example, thermal power may vary due to different power generation methods. Coal-fired power and gas-fired power each have different carbon emission characteristics; hydropower needs to consider factors such as carbon emissions during the construction process. The purpose of this is to ensure that the carbon emission factor under each form of energy can accurately reflect its actual environmental impact. At the same time, the carbon emission factor per unit of energy consumption is also an indispensable part of the database. This part of the data is directly related to the calculation of carbon emissions based on the use of specific energy. By recording these attributes in detail, the corresponding carbon emission factors can be quickly located and extracted according to the specific energy consumption during use, thereby improving calculation efficiency and accuracy.

[0032] In addition, for the handling of transportation tools, the database design needs to take into account a variety of factors. First, the name of the transportation tool is used as the basic identifier to facilitate the quick identification of each tool. Then it is subdivided according to the type. For example, trucks can be classified according to different loads and power types, trains can be distinguished according to different models, and ships also have multiple types. In addition to the above basic information, key attributes such as load and carbon emission factor per unit transportation distance must also be recorded. This is because the carbon emissions generated by transportation activities depend not only on the type of transportation tool used, but also on the load and distance traveled. Therefore, accurately recording this information in the database is crucial to assessing the actual carbon emissions of each transportation activity. In addition, emerging green transportation modes, such as electric vehicles or hybrid ships, should also be reflected in the database to reflect the positive contribution of these technological advances to reducing carbon emissions.

[0033] In the entire database construction process, it is particularly important to use the powerful indexing and query functions of the database management system. By establishing appropriate indexes, the required carbon emission factor data can be quickly located and extracted based on the relevant information of the input materials, energy or transportation tools. This means that in actual operations, whether it is the application of new materials, the introduction of new energy, or the adoption of new transportation tools, they can be updated to the database in a timely manner to ensure the timeliness and accuracy of the data. At the same time, in order to support dynamic query and matching, the database design must also have good scalability so that it can flexibly adapt to changing needs as new data is added or existing data is adjusted.

[0034] In the specific implementation process, in the material part, in addition to the basic name and type, it is also necessary to record in detail the specific application areas of the materials and the carbon emission characteristics during their life cycle. For materials used in some special application scenarios, such as high-strength steel for bridge construction, or high-performance concrete for tunnel engineering, the carbon emission performance of these materials under different working conditions may be different. Therefore, this information should be recorded as detailed as possible in the database to facilitate accurate carbon emissions calculation in the complex and changeable railway infrastructure construction environment. Similarly, for the energy part, we should not only pay attention to common fossil fuels, but also pay attention to the development trend of clean energy, such as the gradual popularization of new energy such as solar energy and wind energy and its impact on carbon emission factors. This requires the database to continuously track the latest developments in the energy field and ensure that all relevant data are updated in a timely manner.

[0035] In the above-mentioned method for predicting carbon emissions during the construction of railway infrastructure, the step S3, based on the carbon emission feature database, performs fuzzy matching and dynamic fitting deduction on each data item in the carbon emission basic data to determine its carbon emission factor. Specifically, the present application takes into account that the carbon emission feature database may not be updated in a timely manner or the data is incomplete, and it is impossible to fully cover all carbon emission features in actual construction. Therefore, the present application adopts a dynamic fitting deduction method that combines similarity query and deep learning algorithm to improve the accuracy and generalization ability of carbon emission prediction. Among them, Figure 3 FIG. 4 is a flowchart of sub-step S3 of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 3 As shown, the step S3 includes the steps of: S31, extracting the data item to be queried from the carbon emission basic data, wherein the data item to be queried is one of material type, transportation tool type and energy type; S32, calculating the hash value of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities; S33, if the maximum value in the set of hash similarities is greater than or equal to a preset threshold, then using the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value as the carbon emission factor of the data item to be queried; S34, if the maximum value in the set of hash similarities is less than the preset threshold, then based on the descending order of the set of hash similarities, extracting the carbon emission characteristic data elements corresponding to the first N hash similarities to construct a set of candidate carbon emission characteristic data elements; S35, inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into the carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried.

[0036] Specifically, in step S31, the data item to be queried is extracted from the carbon emission basic data, and the data item to be queried is one of material type, transportation tool type and energy type. That is, by extracting the data item to be queried, the specific object that needs to be queried for carbon emission factor is determined, and input features are provided for subsequent intelligent matching of carbon emission factors.

[0037] Specifically, the step S32 calculates the hash value of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities. It should be understood that the hash algorithm is a technology that converts input data into a numerical summary of a fixed size through a specific function, and has the characteristics of efficient calculation and low collision probability. In this application, the hash algorithm is used to convert the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database into a hash value of a fixed length. By comparing the similarity between the hash values, the similarity between the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database can be effectively quantified, thereby providing a basis for subsequent carbon emission factor matching.

[0038] Specifically, in step S33, if the maximum value in the set of hash similarities is greater than or equal to a preset threshold, the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value is used as the carbon emission factor of the data item to be queried. Specifically, when the hash similarity is high, it means that the data item to be queried is closer to a certain carbon emission characteristic data element in terms of attribute characteristics. If the maximum hash similarity reaches or exceeds the preset threshold, it can be considered that the data item to be queried has a high degree of similarity with the carbon emission characteristic data element, and therefore the carbon emission factor stored in the carbon emission characteristic data element can be directly used as the carbon emission factor of the data item to be queried.

[0039] Specifically, in step S34, if the maximum value in the set of hash similarities is less than the preset threshold, then based on the descending arrangement of the set of hash similarities, the carbon emission characteristic data elements corresponding to the first N hash similarities are extracted to construct a set of candidate carbon emission characteristic data elements. That is, if the maximum value in the set of hash similarities is less than the preset threshold, it indicates that there are certain differences in attribute characteristics between the data item to be queried and the existing carbon emission characteristic data elements in the carbon emission characteristic database. If the carbon emission factors of the existing carbon emission characteristic data elements are directly used, the accuracy of carbon emission prediction may decrease. In this regard, the present application further narrows the candidate range by extracting the carbon emission characteristic data elements corresponding to the first N higher hash similarities as the set of candidate carbon emission characteristic data elements, and on this basis, performs deep learning derivation of the carbon emission factors of the data item to be queried.

[0040] Figure 4FIG. 4 is a flowchart of sub-step S35 of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 4 As shown, the step S35 includes the steps of: S351, performing semantic embedding coding on the data item to be queried and each candidate carbon emission feature data element in the set of candidate carbon emission feature data elements to obtain a semantic embedding coding vector of the data item to be queried and a set of semantic embedding coding vectors of the candidate carbon emission feature data elements; S352, performing a candidate carbon emission feature search based on semantic query on the semantic embedding coding vector of the data item to be queried and the set of semantic embedding coding vectors of the candidate carbon emission feature data elements to obtain a query response semantic coding vector of the data item to be queried; S353, performing feature decoding on the query response semantic coding vector of the data item to be queried to obtain the carbon emission factor of the data item to be queried.

[0041] More specifically, in a specific example of the present application, the step S351 includes: inputting the data item to be queried and each candidate carbon emission feature data element in the set of candidate carbon emission feature data elements into a semantic embedding encoder based on the Bert model to obtain a semantic embedding coding vector of the data item to be queried and a set of semantic embedding coding vectors of the candidate carbon emission feature data elements. It should be understood that since the traditional text matching algorithm mainly relies on the similarity calculation at the character level, the similarity at the semantic level is not sufficiently considered. In this regard, in order to improve the accuracy of the association analysis between the data item to be queried and the candidate carbon emission feature data element, the present application adopts the Bert model widely used in the field of natural language processing to perform semantic embedding coding on the data item to be queried and each candidate carbon emission feature data element. Specifically, the Bert model has a strong semantic representation ability by being pre-trained on a large-scale corpus, and can convert text into a vector in a high-dimensional semantic vector space, so that semantically similar texts are closer in the vector space. Based on this, in the present application, the Bert model is used to perform semantic embedding encoding on the data items to be queried and each candidate carbon emission feature data element, which can effectively capture the deep contextual semantic features of the data items to be queried and each candidate carbon emission feature data element, convert them into vector representations in the same feature space, and generate a set of semantic embedding encoding vectors of the data items to be queried and the semantic embedding encoding vectors of the candidate carbon emission feature data elements, thereby providing more accurate input features for the subsequent derivation of carbon emission factors.

[0042] More specifically, the step S352 performs a candidate carbon emission feature search based on semantic query on the set of the semantic embedding coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element to obtain a query response semantic coding vector of the data item to be queried. Specifically, in order to query the feature information that is most relevant to the data item to be queried at the semantic level from a large number of candidate carbon emission feature data elements as the basis for deriving the final carbon emission factor, the present application proposes a candidate carbon emission feature search method based on semantic query, which mines the deep semantic association between the semantic embedding coding vector of the data item to be queried and the semantic embedding coding vector of each candidate carbon emission feature data element by performing semantic query response encoding respectively, and dynamically adjusts the semantic contribution weight of each candidate carbon emission feature data element to the data item to be queried by introducing a semantic association gating agent mechanism, so as to aggregate all semantic query interaction information to obtain a more comprehensive and accurate query response semantic coding vector of the data item to be queried. Among them, Figure 5 FIG. 4 is a flowchart of sub-step S352 of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 5 As shown, the step S352 includes the steps of: S3521, performing semantic query response encoding on the semantic feature enhanced coding vector of the data item to be queried and each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors to obtain a set of candidate carbon emission feature data element semantic query score coding vectors; S3522, based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score coding vectors, performing semantic matching gated aggregation on the set of candidate carbon emission feature data element semantic query score coding vectors to obtain the query response semantic coding vector of the data item to be queried.

[0043] Figure 6 FIG. 4 is a flowchart of sub-step S3521 of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 6 As shown, the step S3521 includes the steps of: S35211, performing feature enhancement based on deconvolution coding on the semantic embedding coding vector of the data item to be queried to obtain a semantic feature enhanced coding vector of the data item to be queried, wherein the semantic feature enhanced coding vector of the data item to be queried has the same feature scale as the semantic embedding coding vectors of each candidate carbon emission feature data element in the set of candidate carbon emission feature data element semantic embedding coding vectors; S35212, respectively inputting the semantic feature enhanced coding vector of the data item to be queried and each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors into a monomer semantic query unit to obtain a set of candidate carbon emission feature data element semantic query score coding vectors.

[0044] In a specific example of the present application, the step S35211 is expressed by the formula: ; ; in, represents the semantic embedding encoding vector of the data item to be queried, represents the semantic feature enhanced encoding vector of the data item to be queried, represents the deconvolution weight matrix, represents the deconvolution coding process, represents the calculation norm, represents the set of semantic embedding encoding vectors of candidate carbon emission feature data elements, , , and They represent the first, second, and third sets of semantic embedding encoding vectors of candidate carbon emission feature data elements. and The semantic embedding encoding vector of candidate carbon emission feature data elements, is the number of semantic embedding coding vectors of the candidate carbon emission feature data element.

[0045] That is, through the adaptive feature learning mechanism of deconvolution coding, the model can parse the potential spatial topological relationship in the semantic embedding coding vector of the query data item, gradually restore the high-dimensional feature details through multi-layer transposed convolution, and solve the feature sparsity problem caused by semantic heterogeneity. In this way, through this feature enhancement method based on deconvolution coding, the semantic embedding coding vector of the query data item can dynamically adapt to the feature distribution of different candidate data elements. Specifically, the use of deconvolution coding instead of simple interpolation is essentially to establish a multi-scale feature space with engineering semantic consistency. Through the residual connection in the iterative upsampling process, the key semantic features in the original semantic embedding coding vector of the query data item are retained, and a high-dimensional compact representation that matches the set of semantic embedding coding vectors of candidate carbon emission feature data elements is generated, that is, the semantic feature enhanced coding vector of the query data item.

[0046] In a specific example of the present application, the step S35212 includes: cascading and fusing the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element, and inputting the tanh function-based neural network layer to obtain the candidate carbon emission feature data element semantic query score coding vector, which is expressed as: ; in, represents the tanh function, represents the bias term, represents the weight matrix of the neural network layer, Indicates cascade operation, Represents the semantic query score encoding vector of the candidate carbon emission feature data element.

[0047] That is, this application introduces a cascade fusion mechanism, which forms a joint representation with an engineering semantic topological structure by aligning and superimposing the high-order feature representation of the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element in the feature space. And through the tanh-based neural network layer, it not only realizes the dynamic weight allocation of bimodal features, but also ensures that the output semantic query score forms a standardized interpretability evaluation system in the range of [-1,1] through the value range constraint of the hyperbolic tangent function, so that the final generated candidate carbon emission feature data element semantic query score coding vector can synchronously represent the semantic similarity and feature matching degree between the data item to be queried and the candidate carbon emission feature data element.

[0048] Figure 7 FIG. 4 is a flowchart of sub-step S3522 of the method for predicting carbon emissions during the construction of railway infrastructure according to an embodiment of the present application. Figure 7 As shown, the step S3522 includes the steps of: S35221, based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score encoding vectors, determining the monomer semantic matching degree of each candidate carbon emission feature data element semantic query score encoding vector in the set of candidate carbon emission feature data element semantic query score encoding vectors to obtain a set of candidate carbon emission feature data element semantic matching degrees; S35222, inputting the set of candidate carbon emission feature data element semantic matching degrees into a relational gated proxy module to obtain a set of candidate carbon emission feature data element query semantic self-attention weights; S35223, aggregating the set of candidate carbon emission feature data element semantic query score encoding vectors based on the set of candidate carbon emission feature data element query semantic self-attention weights to obtain the query response semantic encoding vector of the data item to be queried.

[0049] In particular, in a preferred example of the present application, the step S35221 includes: first, based on the semantic feature interaction between the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vectors of each candidate carbon emission feature data element in the set of candidate carbon emission feature data element semantic embedding coding vectors, symmetry constraint optimization is performed on each corresponding candidate carbon emission feature data element semantic query score coding vector to obtain a set of optimized candidate carbon emission feature data element semantic query score coding vectors, which is expressed as follows: ; ; ; ; ; in, represents the cascade vector of query data-candidate carbon emission features, represents the linear mapping matrix, express and The interaction potential vector between is the coupling constant, represents the covariance matrix, express The corresponding optimized candidate carbon emission feature data element semantic query score encoding vector.

[0050] That is, when the monomer semantic query unit performs semantic query score encoding through direct feature splicing between the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element, it is expected to improve the representation accuracy of the semantic query score based on the improvement of the inherent alignment characteristics between the original feature space and the semantic query coding space. Based on this, the present application further maps the cascade features between the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element, generates the interaction potential between the two, and as an optimized generator, realizes the inherent alignment characteristics between the feature space and the semantic query coding space through the adjustment of the covariance matrix, thereby ensuring that the semantic query score coding vector of the candidate carbon emission feature data element can truly reflect the semantic similarity between the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element. Through this interactive optimization method, the semantic matching accuracy of the semantic query score encoding vector of the candidate carbon emission feature data element is improved. Through dynamic mapping and symmetry constraints, the model can more accurately identify subtle differences in semantic features, avoid matching failures caused by semantic gaps, and enhance the generalization ability of the model, so that the generated optimized candidate carbon emission feature data element semantic query score encoding vector can provide a more reliable basis for carbon emission prediction of railway projects.

[0051] Then, the contextual semantic relevance of each optimized candidate carbon emission feature data element semantic query score encoding vector in the set of the optimized candidate carbon emission feature data element semantic query score encoding vector relative to other optimized candidate carbon emission feature data element semantic query score encoding vectors is calculated as the monomer semantic matching degree to obtain the set of candidate carbon emission feature data element semantic matching degrees, which is expressed by the formula: ; in, The set of semantic query score encoding vectors representing the optimized candidate carbon emission feature data element is Different optimized candidate carbon emission feature data element semantic query score encoding vectors, represents the number of optimized candidate carbon emission feature data element semantic query score encoding vectors in multiple optimized candidate carbon emission feature data element semantic query score encoding vectors, represents the transpose of a vector, Indicates The exponential function operation with base , represents the softmax function, express The corresponding semantic matching degree of candidate carbon emission feature data element.

[0052] That is, by constructing a self-distributed feature analysis mechanism for a set of semantic query score encoding vectors that optimizes candidate carbon emission feature data elements, the potential correlation strength distribution between heterogeneous data can be quantitatively revealed, thereby breaking through the static comparison limitations of traditional cosine similarity and realizing dynamic calibration of semantic associations. For example, if the semantic query score encoding vector of the optimized candidate carbon emission feature data element is significantly higher than the average level of the set, its corresponding semantic matching degree is stronger and should be given a higher weight. On the contrary, if it is close to or even lower than the mean, it may represent a weaker semantic association or noise. In this way, the generated semantic matching degree of the candidate carbon emission feature data element can more effectively characterize the semantic association between each candidate carbon emission feature and the data item to be queried.

[0053] In a specific example of the present application, the step S35222 is expressed by the formula: ; in, represents the gating threshold, Represents a mask operation, express The corresponding candidate carbon emission feature data element query semantic self-attention weight.

[0054] That is, through the threshold control of the gating unit, the model can selectively retain the semantic association strength of the semantic query score encoding vector of the candidate carbon emission feature data element according to the physical constraints of the specific engineering scenario, thereby eliminating semantic associations below the preset importance level and ensuring that only highly relevant carbon emission feature data elements are retained, so as to improve the accuracy and efficiency of the model in the carbon emission factor fitting and derivation tasks.

[0055] In a specific example of the present application, step S35223 is expressed by the formula: ; in, Represents the query response semantic encoding vector of the data item to be queried.

[0056] That is, by weighting and modulating the set of semantic query score encoding vectors of the candidate carbon emission feature data element through the query semantic self-attention weights of the candidate carbon emission feature data element, the influence of the core features is automatically strengthened, while the interference of minor factors is weakened. This feature aggregation mechanism based on engineering semantic importance can transform the discrete candidate data element information flow into a query response semantic encoding vector of the query data item with physical interpretability, which helps to focus on the key candidate carbon emission features in the carbon emission factor fitting process, thereby improving the accuracy and reliability of the prediction results.

[0057] More specifically, in step S353, feature decoding is performed on the query response semantic coding vector of the data item to be queried to obtain the carbon emission factor of the data item to be queried. That is, the present application further uses a feature decoder to decode the query response semantic coding vector of the data item to be queried, so as to map it from the high-dimensional semantic vector space back to the specific value of the carbon emission factor. Specifically, the feature decoder is a neural network model built based on a deep learning framework. In the training stage, it uses known carbon emission data and its corresponding carbon emission factor as training samples, and continuously adjusts the network parameters through the back propagation algorithm, so that the feature decoder can learn the effective mapping relationship from the high-dimensional semantic vector space to the specific value of the carbon emission factor. In the derivation stage, the feature decoder receives the query response semantic coding vector of the data item to be queried as input, and outputs the carbon emission factor value corresponding to the new material through forward propagation calculation according to the learned mapping relationship. In this way, even if there is a certain difference between the data item to be queried and the existing data element in the carbon emission feature database, the carbon emission factor can be accurately derived based on its deep semantic features, thereby improving the accuracy and generalization ability of carbon emission prediction.

[0058] In the above-mentioned method for predicting carbon emissions during the construction of railway infrastructure, the step S4 calculates the total carbon emissions during the construction of railway infrastructure based on each data item in the carbon emission basic data and its corresponding carbon emission factor. Specifically, according to the basic calculation principle of carbon emissions, for different types of carbon emission sources, the corresponding calculation method is adopted. For the material part, since the carbon emissions generated by different materials in the production process are closely related to the material usage, the carbon emissions are equal to the material usage multiplied by the carbon emission factor per unit usage. For example, the carbon emission factor for producing 1 ton of steel is X kilograms of carbon dioxide / ton. If Y tons of steel are used in the construction, the carbon emissions of the steel production link are X*Y kilograms of carbon dioxide. For transportation activities, the carbon emissions during transportation mainly depend on the transportation distance and the energy consumption characteristics of the transportation tool. The carbon emissions are equal to the transportation distance multiplied by the carbon emission factor per unit transportation distance. For example, the carbon emission factor of a certain type of truck transporting per ton of goods per kilometer is Z kilograms of carbon dioxide / (ton・kilometer). M tons of goods are transported and N kilometers are traveled. The carbon emissions of the transportation activity are Z*M*N kilograms of carbon dioxide. For construction machinery, carbon emissions are generated due to energy consumption during operation, and the carbon emissions are equal to the energy consumption multiplied by the carbon emission factor per unit of energy consumption. For example, the carbon emission factor of a diesel construction machine is W kg CO2 / L for every liter of diesel consumed, and P liters of diesel are consumed during the construction process, so the carbon emissions of the machine are W*P kg CO2. By adding up the carbon emissions of all materials, transportation activities and construction machinery, the total carbon emissions during the construction of railway infrastructure can be obtained.

[0059] In summary, a method for predicting carbon emissions during the construction of railway infrastructure based on an embodiment of the present application is explained, which first extracts basic carbon emission data during the construction of railway infrastructure from relevant IFC files of railway infrastructure, and at the same time constructs a relevant carbon emission feature database, based on the hash similarity between each carbon emission basic data and each carbon emission feature data element in the database, queries the corresponding carbon emission factor to calculate carbon emissions, and at the same time, when the data elements in the carbon emission feature database are insufficient to meet the query requirements, a deep learning algorithm is further introduced, and semantic parsing and dynamic query matching and fitting of multiple candidate carbon emission feature data elements with high similarity are performed to infer the corresponding carbon emission factors to achieve carbon emission calculation. In this way, the carbon emission characteristics of new composite materials and construction activities can be effectively dealt with, and the generalization ability of carbon emission prediction can be improved, thereby providing strong support for carbon emission management in the railway industry.

[0060] Furthermore, a carbon emission prediction system for railway infrastructure construction process is also provided.

[0061] Figure 8FIG. 1 is a block diagram of a carbon emission prediction system for a railway infrastructure construction process according to an embodiment of the present application. Figure 8 As shown, according to the carbon emission prediction system 100 for the railway infrastructure construction process of the embodiment of the present application, it includes: a time series data acquisition module 110, which is used to obtain the actual power consumption during the operation of the train from the train traction power supply system based on a predetermined time period to obtain a time series of the actual power consumption, and at the same time obtain the power grid carbon emission factor corresponding to the train operation area from the power grid energy management system based on the predetermined time period to obtain a time series of the power grid carbon emission factor; a carbon emission calculation module 120, which is used to calculate the time series of the train carbon emissions based on the time series of the actual power consumption and the time series of the power grid carbon emission factor; a high emission identification module 130, which is used to mark the data items in the time series of the train carbon emissions that are greater than a preset carbon emission threshold as high emission time intervals, and obtain the train operation data in the high emission time interval to obtain a set of train operation status information; a cluster analysis module 140, which is used to perform cluster analysis on the set of train operation status information to determine the main influencing factors leading to high carbon emissions of the train.

[0062] The basic principle of the present invention is described above in conjunction with specific embodiments. However, it should be pointed out that the advantages, strengths, effects, etc. mentioned in the present invention are only examples and not limitations, and it cannot be considered that these advantages, strengths, effects, etc. must be possessed by each embodiment of the present invention. In addition, the specific details of the above embodiments are only for the purpose of illustration and facilitation of understanding, rather than limitation, and the above details do not limit the present invention to being implemented by adopting the above specific details.

[0063] In the above embodiments, the description of each embodiment has its own emphasis. For the parts that are not described or recorded in detail in a certain embodiment, please refer to the relevant description of other embodiments. In the several embodiments provided by the present invention, it should be understood that the disclosed system and method can be implemented in other ways. For example, the system embodiment described above is only schematic. For example, the unit division is only a logical function division, and there may be other division methods in actual implementation. The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the scheme of this embodiment.

[0064] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference to a figure in a claim should not be considered as limiting the claim to which it relates.

[0065] In addition, it is obvious that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units stated in the system claims can also be implemented by one unit through software or hardware.

[0066] Finally, it should be noted that the above description has been given for the purpose of illustration and description. In addition, the above embodiments are only used to illustrate the technical solution of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, it should be understood by those skilled in the art that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A method for predicting carbon emissions during railway infrastructure construction, characterized in that: include: Extracting basic carbon emission data during the construction of railway infrastructure from relevant IFC files of railway infrastructure, including the type and amount of materials, transportation tools and distance of transportation activities, and energy type and consumption of construction machinery; Constructing a carbon emission characteristic database, wherein the carbon emission characteristic database stores carbon emission factors of various materials, energy sources and transportation tools in a table structure; Based on the carbon emission characteristic database, fuzzy matching and dynamic fitting derivation are performed on each data item in the carbon emission basic data to determine its carbon emission factor, which includes: Extracting a data item to be queried from the carbon emission basic data, wherein the data item to be queried is one of a material type, a transportation tool type, and an energy type; Calculating hash values ​​of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities; If the maximum value in the set of hash similarities is greater than or equal to a preset threshold, the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value is used as the carbon emission factor of the data item to be queried; If the maximum value in the set of hash similarities is less than the preset threshold, based on the descending order of the set of hash similarities, extracting the carbon emission characteristic data elements corresponding to the first N hash similarities to construct a set of candidate carbon emission characteristic data elements; Inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into a carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried; Based on each data item in the carbon emission basic data and its corresponding carbon emission factor, the total carbon emissions during the railway infrastructure construction process are calculated.

2. The method for predicting carbon emissions during railway infrastructure construction according to claim 1 is characterized in that: Inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into a carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried, including: Performing semantic embedding coding on the data item to be queried and each candidate carbon emission characteristic data element in the set of candidate carbon emission characteristic data elements to obtain a set of semantic embedding coding vectors of the data item to be queried and semantic embedding coding vectors of the candidate carbon emission characteristic data elements; Performing a candidate carbon emission feature search based on a semantic query on the set of the semantic embedding coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element to obtain a query response semantic coding vector of the data item to be queried; Feature decoding is performed on the query response semantic encoding vector of the data item to be queried to obtain the carbon emission factor of the data item to be queried.

3. The method for predicting carbon emissions during railway infrastructure construction according to claim 2 is characterized in that: The method of performing semantic embedding coding on the data item to be queried and each candidate carbon emission characteristic data element in the set of candidate carbon emission characteristic data elements to obtain a set of semantic embedding coding vectors of the data item to be queried and semantic embedding coding vectors of the candidate carbon emission characteristic data elements includes: The data item to be queried and each candidate carbon emission feature data element in the set of candidate carbon emission feature data elements are input into a semantic embedding encoder based on the Bert model to obtain a set of semantic embedding coding vectors of the data item to be queried and semantic embedding coding vectors of the candidate carbon emission feature data elements.

4. The method for predicting carbon emissions during railway infrastructure construction according to claim 3 is characterized in that: The method of performing a candidate carbon emission feature search based on a semantic query on the set of the semantic embedding coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element to obtain a query response semantic coding vector of the data item to be queried includes: Performing semantic query response encoding on the semantic feature enhancement encoding vector of the data item to be queried and each candidate carbon emission feature data element semantic embedding encoding vector in the set of candidate carbon emission feature data element semantic embedding encoding vectors to obtain a set of candidate carbon emission feature data element semantic query score encoding vectors; Based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score encoding vectors, the set of candidate carbon emission feature data element semantic query score encoding vectors is subjected to semantic matching gated aggregation to obtain the query response semantic encoding vector of the data item to be queried.

5. The method for predicting carbon emissions during railway infrastructure construction according to claim 4, characterized in that: The semantic feature enhancement coding vector of the data item to be queried is respectively subjected to semantic query response coding with each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors to obtain a set of candidate carbon emission feature data element semantic query score coding vectors, including: Performing feature enhancement based on deconvolution coding on the semantic embedding coding vector of the data item to be queried to obtain a semantic feature enhanced coding vector of the data item to be queried, wherein the semantic feature enhanced coding vector of the data item to be queried has the same feature scale as each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors; The semantic feature enhancement coding vector of the data item to be queried and each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors are respectively input into the monomer semantic query unit to obtain a set of candidate carbon emission feature data element semantic query score coding vectors.

6. The method for predicting carbon emissions during railway infrastructure construction according to claim 5, characterized in that: The semantic feature enhancement coding vector of the data item to be queried and each candidate carbon emission feature data element semantic embedding coding vector in the set of candidate carbon emission feature data element semantic embedding coding vectors are respectively input into a monomer semantic query unit to obtain a set of candidate carbon emission feature data element semantic query score coding vectors, including: The semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vector of the candidate carbon emission feature data element are cascaded and fused, and then input into a neural network layer based on a tanh function to obtain a semantic query score coding vector of the candidate carbon emission feature data element.

7. The method for predicting carbon emissions during railway infrastructure construction according to claim 6, characterized in that: Based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score encoding vectors, the set of candidate carbon emission feature data element semantic query score encoding vectors is subjected to semantic matching gated aggregation to obtain the query response semantic encoding vector of the data item to be queried, including: Based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score encoding vectors, determining the monomer semantic matching degree of each candidate carbon emission feature data element semantic query score encoding vector in the set of candidate carbon emission feature data element semantic query score encoding vectors to obtain a set of candidate carbon emission feature data element semantic matching degrees; Inputting the set of semantic matching degrees of the candidate carbon emission feature data elements into the relational gating proxy module to obtain a set of query semantic self-attention weights of the candidate carbon emission feature data elements; The set of candidate carbon emission feature data element semantic query score encoding vectors is aggregated based on the set of candidate carbon emission feature data element query semantic self-attention weights to obtain the query response semantic encoding vector of the data item to be queried.

8. The method for predicting carbon emissions during railway infrastructure construction according to claim 7, characterized in that: Based on the feature set self-distribution characteristics of the set of candidate carbon emission feature data element semantic query score encoding vectors, determining the monomer semantic matching degree of each candidate carbon emission feature data element semantic query score encoding vector in the set of candidate carbon emission feature data element semantic query score encoding vectors to obtain a set of candidate carbon emission feature data element semantic matching degrees, including: Based on the semantic feature interaction between the semantic feature enhancement coding vector of the data item to be queried and the semantic embedding coding vectors of each candidate carbon emission feature data element in the set of candidate carbon emission feature data element semantic embedding coding vectors, symmetry constraint optimization is performed on each corresponding candidate carbon emission feature data element semantic query score coding vector to obtain a set of optimized candidate carbon emission feature data element semantic query score coding vectors; The contextual semantic association of each optimized candidate carbon emission feature data element semantic query score encoding vector in the set of optimized candidate carbon emission feature data element semantic query score encoding vectors relative to other optimized candidate carbon emission feature data element semantic query score encoding vectors is calculated as the monomer semantic matching degree to obtain the set of candidate carbon emission feature data element semantic matching degrees.

9. A carbon emission prediction system for railway infrastructure construction process, characterized in that: include: A carbon emission data extraction module is used to extract basic carbon emission data during the construction of railway infrastructure from relevant IFC files of railway infrastructure. The basic carbon emission data includes the type and amount of materials, the means of transport and distance of transportation activities, and the energy type and consumption of construction machinery; A carbon emission database construction module is used to construct a carbon emission characteristic database, which stores carbon emission factors of various materials, energy sources and transportation tools in a table structure; The data matching and fitting module is used to perform fuzzy matching and dynamic fitting derivation on each data item in the carbon emission basic data based on the carbon emission characteristic database to determine its carbon emission factor, which includes: Extracting a data item to be queried from the carbon emission basic data, wherein the data item to be queried is one of a material type, a transportation tool type, and an energy type; Calculating hash values ​​of the data item to be queried and each carbon emission characteristic data element in the carbon emission characteristic database to obtain a set of hash similarities; If the maximum value in the set of hash similarities is greater than or equal to a preset threshold, the carbon emission factor stored in the carbon emission characteristic data element corresponding to the maximum value is used as the carbon emission factor of the data item to be queried; If the maximum value in the set of hash similarities is less than the preset threshold, based on the descending order of the set of hash similarities, extracting the carbon emission characteristic data elements corresponding to the first N hash similarities to construct a set of candidate carbon emission characteristic data elements; Inputting the set of candidate carbon emission characteristic data elements and the data item to be queried into a carbon emission factor dynamic fitting model to obtain the carbon emission factor of the data item to be queried; The carbon emission calculation module is used to calculate the total carbon emissions during the construction of railway infrastructure based on each data item in the carbon emission basic data and its corresponding carbon emission factor.

Images