Traffic engineering construction carbon emission prediction method and system based on digital twinning

Abstract

The invention relates to the technical field of traffic carbon emission prediction, in particular to a traffic engineering construction carbon emission prediction method and system based on digital twinning. According to the method, sites are used as nodes, carbon emission data sequences are used as node values, distances are used as edge weights between the sites, and an initial graph structure is constructed; obtaining interference factors corresponding to every two nodes; analyzing the nodes to obtain affected nodes and affected nodes, updating the initial graph structure to obtain a directed graph, and correspondingly updating interference factors of the two nodes of the directed graph; selecting influence components from IMF components of the carbon emission data sequences of the nodes in combination with interference factors between the nodes; and according to the carbon emission data sequence of the node and the difference in the influence component, determining the weight value of the corresponding site as an interpolation weight, and predicting to obtain predicted emission data of each site by using a digital twinning technology. According to the method, the carbon emission data of each unknown point can be accurately obtained, and a more reliable prediction result can be obtained according to digital twinning.

Description

Technical Field

[0001] The present invention relates to the technical field of transportation carbon emission prediction, and in particular to a method and system for predicting carbon emission from transportation engineering construction based on digital twins. Background Art

[0002] With the continuous acceleration of urbanization and the continuous growth of population, transportation engineering construction has become a key link in ensuring the normal operation of cities and improving the quality of life of residents. In urban development planning, transportation engineering construction has been given an important mission and responsibility. Advanced construction equipment and processes are required in transportation engineering construction, such as excavators, milling machines, and road levelers. However, when using these construction equipment, a certain amount of carbon emissions will be generated. In recent years, in the transportation industry, the control and prediction of carbon emissions are crucial. When the predicted carbon emissions exceed the standard, corresponding measures should be taken to avoid large amounts of carbon emissions during transportation engineering construction and cause adverse effects.

[0003] Currently, a common method for predicting carbon emissions from transportation engineering construction is to use digital twin technology to model and simulate the transportation engineering construction system in order to predict the carbon emissions of the transportation engineering construction system. In this process, the carbon emission data of unknown points needs to be predicted based on the carbon emission data of known construction sites, which is often done through interpolation calculations. However, the carbon emissions between different construction sites are fluid and influence each other, and the referenceability of a construction site is related to the stability of the carbon emissions of the construction site. Therefore, the results of the existing inverse distance weighted interpolation method usually have large prediction errors. Summary of the invention

[0004] In order to solve the technical problem that the results of the existing inverse distance weighted interpolation method usually lead to large prediction errors when analyzing carbon emissions, the purpose of the present invention is to provide a method and system for predicting carbon emissions from transportation engineering construction based on digital twins. The technical solutions adopted are as follows: In a first aspect, an embodiment of the present invention provides a method for predicting carbon emissions from traffic engineering construction based on digital twins, the method comprising the following steps: Obtain carbon emission data for each construction site in transportation engineering construction; The initial graph structure is constructed by taking each construction site as a node, the carbon emission data sequence of each construction site as the node value, and the distance between construction sites as the edge weight between construction sites; Obtain the embedding vector corresponding to each node; use the similarity of the embedding vectors of two nodes as the interference factor corresponding to the two nodes; The different trend sequences of the embedded vector of each node under different window sizes are calculated by the moving average method; based on the matching of different trend sequences between nodes, the trend distance between nodes is determined; based on the trend distance, the nodes are classified; based on the node value, the nodes of the same category are matched, and the time difference is obtained by the time difference between the matched node values, and the influencing nodes and the affected nodes are divided from the category based on the time difference; based on the influencing nodes and the affected nodes, the initial graph structure is updated to obtain a directed graph, and the interference factors corresponding to the two nodes of the directed graph are updated accordingly; Decompose the carbon emission data sequence of each node to obtain the IMF components corresponding to each node; match the IMF components corresponding to two nodes, and select the impact components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes; determine the weight value corresponding to each construction site based on the difference between the carbon emission data sequence of each node and the elements in the impact component; The weight value of each construction site is used as the interpolation weight in the inverse distance weighted interpolation method, and the digital twin technology is used to predict the predicted emission data of each construction site.

[0005] Preferably, matching the IMF components corresponding to the two nodes and selecting the influence components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes includes: Using the KM matching algorithm, multiple IMF components corresponding to two nodes are matched to obtain component matching pairs; wherein the cosine similarity of the two IMF components is used as the edge value corresponding to the two IMF components; When the edge value corresponding to two IMF components is greater than the interference factor corresponding to the two nodes, the ratio of the interference factor to the edge value is used as the influence value; when the edge value corresponding to two IMF components is less than or equal to the interference factor corresponding to the two nodes, the ratio of the edge value to the interference value is used as the influence value; the IMF component corresponding to the maximum influence value is used as the influence component corresponding to the two nodes.

[0006] Preferably, determining the trend distance between nodes based on the matching of different trend sequences between nodes includes: For any two nodes, KM matching is performed on the different trend sequences of the two nodes to obtain the maximum weight sum, which is used as the KM matching result value; the negative correlation normalized value of the KM matching result value is used as the trend distance between the nodes.

[0007] Preferably, matching nodes of the same category based on the node values, obtaining a time difference from the time difference between the matched node values, and dividing the influencing nodes and the affected nodes from the category based on the time difference includes: Any two nodes in the same category are respectively used as the first node and the second node; the node value of the first node and the node value of the second node are matched by a dynamic time warping algorithm to obtain multiple carbon emission data pairs; the time difference between the two carbon emission data in each carbon emission data pair is calculated, and the average of the time differences of all carbon emission data pairs corresponding to the first node and the second node is used as the time difference between the first node and the second node; For the first node, the average of the time differences between the first node and other nodes in the same category is taken as the influence value of the first node; The node corresponding to the largest influence value in the category is taken as the influencing node, and the other nodes in the category except the influencing node are taken as the affected nodes.

[0008] Preferably, the initial graph structure is updated based on the influencing nodes and the affected nodes to obtain a directed graph, including: A direction is added to the initial graph structure, where the direction is from the influencing node to the influenced node in each category, thereby obtaining a directed graph.

[0009] Preferably, determining the weight value corresponding to each construction site according to the carbon emission data sequence of each node and the difference between each element in the impact component includes: The elements at the same position in the influence component corresponding to each node are averaged to obtain the reference component corresponding to each node; The difference between the carbon emission data series of each node and each element in the reference component is calculated to obtain a difference series; the coefficient of variation of each difference series is used as the weight value corresponding to each construction site.

[0010] Preferably, the weight value of each construction site is used as the interpolation weight in the inverse distance weighted interpolation method, and the predicted emission data of each construction site is predicted by using the digital twin technology, including: Using the inverse distance weighted interpolation method, based on the known carbon emission data of each construction site and the corresponding interpolation weights, the carbon emission data of unknown construction sites in the same period are obtained; through digital twin technology, based on historical data, the predicted emission data of each known construction site and unknown construction site are obtained.

[0011] Preferably, obtaining the embedding vector corresponding to each node includes: The embedding vector corresponding to each node is obtained through the Node2Vec graph embedding method.

[0012] Preferably, the using the similarity of the embedding vectors of two nodes as the interference factors corresponding to the two nodes includes: The cosine similarity of the embedding sequence composed of the embedding vectors of the 5-neighborhood of any two nodes is used as the interference factor corresponding to the two nodes.

[0013] In the second aspect, an embodiment of the present invention provides a carbon emission prediction system for transportation engineering construction based on digital twins, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the above-mentioned carbon emission prediction method for transportation engineering construction based on digital twins.

[0014] The embodiments of the present invention have at least the following beneficial effects: The present invention first constructs a graph structure, determines the weight value of each node according to the interference factor of the node in the initial graph structure and the IMF component of the carbon emission data sequence corresponding to the node, and uses the weight value as the interpolation weight. The present invention removes the influence of other construction sites on the data monitored by each construction site, and then obtains the interpolation weight according to the volatility of the carbon emission data of each construction site, thereby obtaining more accurate carbon emission data of each unknown point, and then obtaining more reliable prediction results according to the digital twin. When the carbon emission data of the known construction site is interpolated and calculated for the carbon emission data of the unknown construction site, the traditional method only considers the distance factor, but for the data of each known construction site, if the degree of fluctuation of the carbon emission data of the point is greater, the instability of the point is greater, as a reference point for other points, that is, when other points refer to the data of the point, the smaller the weight should be given, and then a more reasonable interpolation weight is obtained. The interpolation weight of each construction site is obtained by the fluctuation of the real carbon emission value sequence of each construction site, and then more accurate carbon emission data of each unknown construction site is obtained, and then more reliable prediction results are obtained according to the digital twin. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 A method flow chart of a method for predicting carbon emissions from transportation engineering construction based on digital twins provided in one embodiment of the present invention. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, describes in detail the specific implementation methods, structures, features and effects of the carbon emission prediction method and system for transportation engineering construction based on digital twins proposed by the present invention. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0019] The embodiment of the present invention provides a specific implementation method of a method and system for predicting carbon emissions in transportation engineering construction based on digital twins, and the method is applicable to the scenario of carbon emission prediction in transportation engineering construction. In this scenario, except for the unknown carbon emission data of the unknown construction site, all other data are known. In order to solve the technical problem that the results of the existing inverse distance weighted interpolation method usually lead to large prediction errors when analyzing carbon emissions. The present invention proposes a method and system for predicting carbon emissions in transportation engineering construction based on digital twins, which removes the influence of other construction sites on the data monitored by each construction site, and then obtains the interpolation weight according to the volatility of the carbon emission data of each construction site, thereby obtaining more accurate carbon emission data for each unknown point, and then obtaining more reliable prediction results based on the digital twins.

[0020] The specific scheme of the carbon emission prediction method and system for transportation engineering construction based on digital twin provided by the present invention is described in detail below with reference to the accompanying drawings.

[0021] See also Figure 1 , which shows a flowchart of a method for predicting carbon emissions from traffic engineering construction based on digital twins provided by an embodiment of the present invention, the method comprising the following steps: Step S100, obtaining carbon emission data of each construction site in the transportation engineering construction.

[0022] The purpose of the present invention is to predict the carbon emissions of transportation engineering construction, so it is necessary to first obtain the carbon emission data of each construction site. It should be noted that in the subsequent steps, the construction site can also be directly referred to as a site. The carbon emission data sequence corresponding to each construction site is composed of the carbon emission data of each construction site at different times. In the embodiment of the present invention, the carbon emission data is detected once an hour. Among them, the carbon emission data includes: carbon emissions from raw materials for engineering construction, carbon emissions from transportation for engineering construction, and carbon emissions from land use. The corresponding carbon emission data detected every hour is the sum of carbon emissions from raw materials for engineering construction, carbon emissions from transportation for engineering construction, and carbon emissions from land use.

[0023] Step S200, constructing an initial graph structure with each construction site as a node, the carbon emission data sequence of each construction site as a node value, and the distance between the construction sites as the edge weight between the construction sites.

[0024] The purpose of this invention is to predict the carbon emissions of transportation engineering construction. In practice, each area with a large amount of construction will have a construction site for observation. Therefore, it is necessary to first obtain the location of the construction site. Since the carbon emissions between different construction sites are flowing and affect each other, the scene is described by constructing a graph structure.

[0025] Specifically: each construction site is used as a node of the initial graph structure, and the carbon emission data sequence collected from each construction site within one month is used as the node value; the distance between any two construction sites is used as the edge weight to construct the initial graph structure.

[0026] Step S300, obtaining an embedding vector corresponding to each node; and using the similarity of the embedding vectors of two nodes as interference factors corresponding to the two nodes.

[0027] When interpolating the carbon emission data of unknown points through the carbon emission data of known construction sites, the traditional method only considers the distance factor. However, for the carbon emission data of each known construction site, the greater the fluctuation of the carbon emission data of this point, the greater the instability of this point, and the smaller the weight should be given when it is used as a reference point for other points, that is, when other points refer to the data of this point.

[0028] It should be noted that since the carbon emissions between different construction sites are fluid and mutually influential, the carbon emission data monitored by each construction site is affected by other construction sites. Therefore, it is necessary to first remove this influence to obtain the pure data of each construction site, and then obtain the actual data of each construction site for further prediction. This influence can be obtained through the similarity in the node aggregation process.

[0029] First, obtain the embedding vector corresponding to each node. Specifically, obtain the embedding vector corresponding to each node through the Node2Vec graph embedding method.

[0030] The similarity of the embedding vectors of the two nodes is used as the interference factor corresponding to the two nodes. Specifically, for any two nodes, the cosine similarity of the embedding sequence composed of the embedding vectors of the 5 neighborhoods of any two nodes is used as the interference factor corresponding to the two nodes. In the embodiment of the present invention, the neighborhood node range of each embedding vector is set to 5, so that the embedding vector of each node including 1 neighborhood node, 2 neighborhood nodes, 3 neighborhood nodes, 4 neighborhood nodes, and 5 neighborhood nodes can be obtained. The embedding vectors of 1 neighborhood node, 2 neighborhood nodes, 3 neighborhood nodes, 4 neighborhood nodes, and 5 neighborhood nodes of each node are spliced ​​into a vector sequence, which is called an embedding sequence, and then the cosine similarity of any two embedding sequences is calculated and recorded as the interference factor.

[0031] Step S400, calculate different trend sequences of the embedded vector of each node under different window sizes by the moving average method; determine the trend distance between the nodes based on the matching of different trend sequences between the nodes; classify the nodes based on the trend distance; match the nodes of the same category based on the node value, obtain the time difference from the time difference between the matched node values, and divide the influencing nodes and the affected nodes from the category based on the time difference; update the initial graph structure based on the influencing nodes and the affected nodes to obtain a directed graph, and update the interference factors corresponding to the two nodes of the directed graph accordingly.

[0032] Among them, nodes are divided into influencing nodes and affected nodes. Influencing nodes refer to the carbon emissions of the construction site corresponding to the node that are easily transferred to other construction sites through air flow, thereby increasing the carbon emission data of other construction sites, resulting in inaccurate carbon emission data of other construction sites. The affected nodes refer to the carbon emission data of the affected nodes of the construction site corresponding to the node, resulting in inaccurate carbon emission data of the construction site. Therefore, by analyzing the nodes, the nodes in the initial graph structure are divided into influencing nodes and affected nodes. And the initial graph structure is further updated based on the influencing nodes and affected nodes.

[0033] The interference factor between any two nodes currently obtained is large, which only means that the changes in the carbon emission data of these two nodes are similar. They are not necessarily nodes that affect each other. They may be two construction sites that are far apart, but have similar terrain and construction conditions, resulting in similar carbon emission data, and they do not necessarily affect each other.

[0034] There must be an influencing node and an affected node in the two nodes that affect each other, but it is unknown which is the influencing node and which is the affected node. Since the influencing node affects multiple nodes, and the affected node becomes similar to the change of the influencing node due to being affected, the influencing node and the affected node are calculated by the influence centrality and the affected proximity. Since the impact proximity means that the change trend after being affected is similar, and the carbon emission data of the construction sites corresponding to these affected nodes are not necessarily similar, the trend proximity is used to represent the affected proximity, and the inverse proportional normalized value of the trend proximity is used as the trend distance.

[0035] First, the different trend sequences of the embedding vector of each node under different window sizes are calculated by the moving average algorithm. In the embodiment of the present invention, the default window sizes are [3, 5, 7, 9, 11, 15, 17, 19] as the window sizes of the moving average. According to each window size, a trend line can be obtained, that is, a trend sequence. According to eight different window sizes, eight trend sequences corresponding to each node can be obtained.

[0036] Based on the matching of trend sequences between nodes, the trend distance between nodes is determined. Specifically, for any two nodes, different trend sequences of the two nodes are KM matched to obtain the maximum weight sum, which is used as the KM matching result value, and the negative correlation normalized value of the KM matching result value is used as the trend distance between nodes. The KM matching result value is also the trend similarity.

[0037] In the embodiment of the present invention, the calculation formula of trend distance is: in, The trend distance between the two current nodes; is a natural constant; The maximum value of the sum of the maximum weights of the KM matching of any two nodes; It is the sum of the maximum weights of the KM matching between the current two nodes.

[0038] The calculation formula of the trend distance realizes the negative correlation normalization processing of the sum of the maximum weights of the KM matching of the current two nodes.

[0039] After the trend distance is obtained, the nodes are classified based on the trend distance. Specifically, the trend distance is used as a clustering basis, and multiple categories can be obtained through the hierarchical analysis method.

[0040] After obtaining nodes with similar trends, it means that these nodes are mutually influencing nodes, but the influencing and affected nodes cannot be determined yet. It is necessary to use time information to distinguish the influencing and affected nodes by the changes of the influencing nodes at an earlier time and causing similar changes of the affected nodes at a later timeline.

[0041] Therefore, further, based on the node value, the nodes of the same category are matched, and the time difference is obtained by checking the matched node values. The node value corresponding to each node is the carbon emission data sequence of the construction site corresponding to each node.

[0042] Any two nodes in the same category are respectively regarded as the first node and the second node; the node value of the first node and the node value of the second node are matched through a dynamic time warping algorithm to obtain multiple carbon emission data pairs; the time difference between the two carbon emission data in each carbon emission data pair is calculated, and the average of the time differences of all carbon emission data pairs corresponding to the first node and the second node is taken as the time difference between the first node and the second node.

[0043] Based on the time difference, influencing nodes and affected nodes are divided from the category. Specifically: for the first node, the average of the time difference between the first node and other nodes in the same category is taken as the influence value of the first node; the node corresponding to the largest influence value in the category is taken as the influencing node, and the other nodes in the category except the influencing node are taken as the affected nodes.

[0044] The larger the influence value, the earlier the change of the node occurs in the category with greater mutual influence, and then other nodes have corresponding change trends. Therefore, the node with a larger influence value has a greater probability of being an influential node.

[0045] After obtaining the influencing nodes and affected nodes of each category, the initial graph structure is updated based on the influencing nodes and affected nodes to obtain a directed graph. That is, the direction of each influencing node to the corresponding affected node is obtained to obtain a directed graph. More specifically, a direction is added to the initial graph structure, which is the direction from the influencing node to the affected node in each category, thereby obtaining a directed graph. Then, based on the graph embedding of the directed graph, the embedding vector is re-obtained, and the corresponding interference factor is further calculated to achieve the correction of the interference factor.

[0046] It should be noted that the interference factors used in the subsequent steps are the corrected interference factors; since the direction information is added, the degree of fit with the real scene is improved, and therefore, the accuracy of the corrected interference factors is higher than that of the interference factors before correction.

[0047] Step S500, decompose the carbon emission data sequence of each node to obtain the IMF component corresponding to each node; match the IMF components corresponding to the two nodes, and select the impact components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes; determine the weight value corresponding to each construction site based on the difference between the carbon emission data sequence of each node and the elements in the impact component.

[0048] Furthermore, the carbon emission data sequence of each node is decomposed to obtain the IMF component corresponding to each node. Specifically, the carbon emission value sequence of each node is decomposed by EMD to obtain the IMF component corresponding to each node. The EMD decomposition algorithm is a method for decomposing a sequence to obtain information of different frequency components. The similarity between the IMF components containing more interference factor information is the frequency information corresponding to the interference factor. After removing this information, the unaffected carbon emission data of each construction site can be obtained.

[0049] Similarity calculation between IMF components To ensure the maximum overall matching degree, a one-to-one match of IMF components is obtained through KM matching, and then the IMF component with the closest similarity to the interference is taken as the influencing component. Specifically, for any two nodes, the IMF components corresponding to the two nodes are matched, and combined with the interference factors of the two nodes, the influencing components corresponding to the two nodes are selected from the IMF components. More specifically: Using the KM matching algorithm, multiple IMF components corresponding to two nodes are matched one by one to obtain component matching pairs; the cosine similarity of the two IMF components is used as the edge value corresponding to the two IMF components. It should be noted that the similarity calculation between IMF components is to ensure the maximum overall matching degree, so the KM matching algorithm is used to obtain a one-to-one match of the IMF components, and then the IMF component whose one-to-one matching similarity is close to the interference factor is used as the influencing component.

[0050] The existing KM matching is to calculate the matching of the left node and the right node, the left node belongs to one category, the right node belongs to another category, and each left node is connected to all the right nodes. In this embodiment, all IMF components of a carbon emission sequence are used as left nodes, all IMF components of another carbon emission sequence are used as right nodes, and the cosine similarity of the IMF components of the left node and the right node is used as the edge value. Through KM matching, a one-to-one matching relationship between the left and right nodes is obtained.

[0051] When the edge value corresponding to two IMF components is greater than the interference factor corresponding to the two nodes, the ratio of the interference factor to the edge value is used as the influence value; when the edge value corresponding to two IMF components is less than or equal to the interference factor corresponding to the two nodes, the ratio of the edge value to the interference value is used as the influence value; according to the influence value, the influence components corresponding to the two nodes are selected from the IMF components.

[0052] According to the influence value, the influence components corresponding to the two nodes are selected from the IMF components. Specifically, the IMF component corresponding to the maximum influence value is used as the influence component corresponding to the two nodes.

[0053] Taking the carbon emission sequence of any two nodes, node A and node B, as an example, first obtain the interference factor a corresponding to the two nodes, and at the same time obtain the one-to-one relationship of the IMF components, each of which corresponds to a boundary value. The influence component corresponding to node A and node B is obtained by calculation, which can also be said to be the influence component of node A and the influence component of node B. Since an influence component can be calculated between any two nodes, each node corresponds to multiple influence components.

[0054] The elements at the same position in all the influencing components corresponding to each node are averaged to obtain the reference component corresponding to each node, and the difference between the carbon emission data sequence of each node and each element in the reference component is calculated to obtain the difference sequence. This difference sequence is used as the real carbon emission data sequence of each node. It should be noted that the greater the fluctuation of the difference sequence of each node, the smaller the corresponding reference weight.

[0055] Furthermore, the coefficient of variation of each difference sequence is used as the weight value corresponding to each construction site.

[0056] Step S600: The weight value of each construction site is used as the interpolation weight in the inverse distance weighted interpolation method, and the predicted emission data of each construction site is predicted using the digital twin technology.

[0057] Since the existing inverse distance weighted interpolation method usually has prediction errors, the interpolation weights of each construction site are obtained through analysis and calculation in the embodiment of the present invention, and the inverse distance weighted interpolation method is used to realize the prediction of carbon emission data of each construction site. Therefore, the inverse distance weighted interpolation method is finally used to obtain the carbon emission data of unknown construction sites in the same period based on the known carbon emission data of each construction site and the corresponding interpolation weights; the predicted emission data of each known construction site and unknown construction site are obtained based on historical data through digital twin technology. Among them, the interpolation weight takes into account the mobility of carbon emissions. The carbon emission data of each construction site obtained by the embodiment of the present invention is more accurate, which helps to improve the prediction accuracy of the digital twin model.

[0058] The carbon emission data of each construction site obtained through interpolation weights can be used as part of the digital twin model input to improve the accuracy of the prediction model. By combining the data of known construction sites with the data obtained through interpolation weights, the carbon emission levels of different regions can be more accurately reflected, thereby improving the credibility of the prediction results.

[0059] Model the real-world traffic engineering construction system, including various components of the traffic engineering construction system, such as construction equipment, construction time, etc.

[0060] Then, it is necessary to use digital twin technology to simulate the transportation engineering construction system to predict the carbon emissions of the transportation engineering construction system and obtain predicted emission data for known construction sites and unknown construction sites.

[0061] Finally, by comparing the digital twin simulation results with the actual carbon emission data in the real world, the carbon emissions of the transportation engineering construction system can be optimized and controlled.

[0062] In summary, the present invention relates to the technical field of transportation carbon emission prediction. The method first obtains the carbon emission data of each construction site; takes each construction site as a node, takes the carbon emission data sequence of each construction site as a node value, takes the distance between the construction sites as the edge weight between the construction sites, and constructs an initial graph structure; obtains the embedding vector corresponding to each node; takes the similarity of the embedding vectors of two nodes as the interference factor corresponding to the two nodes; analyzes the nodes to obtain the influencing nodes and the affected nodes; updates the initial graph structure based on the influencing nodes and the affected nodes to obtain a directed graph, and updates the interference factors corresponding to the two nodes of the directed graph accordingly; decomposes the carbon emission data sequence of each node to obtain the IMF component corresponding to each node; matches the IMF components corresponding to the two nodes, and selects the influencing components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes; determines the weight value corresponding to each construction site according to the difference between the carbon emission data sequence of each node and the elements in the influencing component; takes the weight value of each construction site as the interpolation weight in the inverse distance weighted interpolation method, and uses the digital twin technology to predict the predicted emission data of each construction site. The present invention obtains the actual carbon emission data of each construction site by removing the influence of other construction sites on the data monitored by each construction site, and then obtains the interpolation weight according to the volatility of the carbon emission data of each construction site, so as to obtain more accurate carbon emission data of each unknown point, and then obtains more reliable prediction results based on the digital twin.

[0063] The embodiment of the present invention also proposes a carbon emission prediction system for traffic engineering construction based on digital twins, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above method when executing the computer program. Since a carbon emission prediction method for traffic engineering construction based on digital twins has been described in detail above, it will not be repeated here.

[0064] It should be noted that the sequence of the above embodiments of the present invention is only for description and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0065] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for predicting carbon emissions from transportation engineering construction based on digital twins, characterized in that: The method comprises the following steps: Obtain carbon emission data for each construction site in transportation engineering construction; The initial graph structure is constructed by taking each construction site as a node, the carbon emission data sequence of each construction site as the node value, and the distance between construction sites as the edge weight between construction sites; Obtain the embedding vector corresponding to each node; use the similarity of the embedding vectors of two nodes as the interference factor corresponding to the two nodes; The different trend sequences of the embedded vector of each node under different window sizes are calculated by the moving average method; based on the matching of different trend sequences between nodes, the trend distance between nodes is determined; based on the trend distance, the nodes are classified; based on the node value, the nodes of the same category are matched, and the time difference is obtained by the time difference between the matched node values, and the influencing nodes and the affected nodes are divided from the category based on the time difference; based on the influencing nodes and the affected nodes, the initial graph structure is updated to obtain a directed graph, and the interference factors corresponding to the two nodes of the directed graph are updated accordingly; Decompose the carbon emission data sequence of each node to obtain the IMF components corresponding to each node; match the IMF components corresponding to two nodes, and select the impact components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes; determine the weight value corresponding to each construction site based on the difference between the carbon emission data sequence of each node and the elements in the impact component; The weight value of each construction site is used as the interpolation weight in the inverse distance weighted interpolation method, and the digital twin technology is used to predict the predicted emission data of each construction site.

2. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The matching of the IMF components corresponding to the two nodes and selecting the influence components corresponding to the two nodes from the IMF components in combination with the interference factors of the two nodes include: Using the KM matching algorithm, multiple IMF components corresponding to two nodes are matched to obtain component matching pairs; wherein the cosine similarity of the two IMF components is used as the edge value corresponding to the two IMF components; When the edge value corresponding to two IMF components is greater than the interference factor corresponding to the two nodes, the ratio of the interference factor to the edge value is used as the influence value; when the edge value corresponding to two IMF components is less than or equal to the interference factor corresponding to the two nodes, the ratio of the edge value to the interference value is used as the influence value; the IMF component corresponding to the maximum influence value is used as the influence component corresponding to the two nodes.

3. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The determining of the trend distance between nodes based on the matching of different trend sequences between nodes includes: For any two nodes, KM matching is performed on the different trend sequences of the two nodes to obtain the maximum weight sum, which is used as the KM matching result value; the negative correlation normalized value of the KM matching result value is used as the trend distance between the nodes.

4. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The nodes of the same category are matched based on the node values, a time difference is obtained from the time difference between the matched node values, and the influencing nodes and the affected nodes are divided from the category based on the time difference, including: Any two nodes in the same category are respectively used as the first node and the second node; the node value of the first node and the node value of the second node are matched by a dynamic time warping algorithm to obtain multiple carbon emission data pairs; the time difference between the two carbon emission data in each carbon emission data pair is calculated, and the average of the time differences of all carbon emission data pairs corresponding to the first node and the second node is used as the time difference between the first node and the second node; For the first node, the average of the time differences between the first node and other nodes in the same category is taken as the influence value of the first node; The node corresponding to the largest influence value in the category is taken as the influencing node, and the other nodes in the category except the influencing node are taken as the affected nodes.

5. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The updating of the initial graph structure based on the influencing nodes and the affected nodes to obtain a directed graph includes: A direction is added to the initial graph structure, where the direction is from the influencing node to the influenced node in each category, thereby obtaining a directed graph.

6. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: Determining the weight value corresponding to each construction site according to the carbon emission data sequence of each node and the difference between the elements in the impact component includes: The elements at the same position in the influence component corresponding to each node are averaged to obtain the reference component corresponding to each node; The difference between the carbon emission data series of each node and each element in the reference component is calculated to obtain a difference series; the coefficient of variation of each difference series is used as the weight value corresponding to each construction site.

7. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The weight value of each construction site is used as the interpolation weight in the inverse distance weighted interpolation method, and the digital twin technology is used to predict the predicted emission data of each construction site, including: Using the inverse distance weighted interpolation method, based on the known carbon emission data of each construction site and the corresponding interpolation weights, the carbon emission data of unknown construction sites in the same period are obtained; through digital twin technology, based on historical data, the predicted emission data of each known construction site and unknown construction site are obtained.

8. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The obtaining of the embedding vector corresponding to each node includes: The embedding vector corresponding to each node is obtained through the Node2Vec graph embedding method.

9. The method for predicting carbon emissions from transportation engineering construction based on digital twins according to claim 1 is characterized in that: The method of using the similarity of the embedding vectors of the two nodes as the interference factors corresponding to the two nodes includes: The cosine similarity of the embedding sequence composed of the embedding vectors of the 5-neighborhood of any two nodes is used as the interference factor corresponding to the two nodes.

10. A carbon emission prediction system for traffic engineering construction based on digital twins, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, it implements the steps of a method for predicting carbon emissions from transportation engineering construction based on digital twins as described in any one of claims 1 to 9.

Images