Plateau railway new energy equipment carbon footprint prediction method and system
By using multi-source feature fusion and improved prediction models, the problems of feature loss and insufficient accuracy in predicting the carbon footprint of new energy equipment for plateau railways have been solved, achieving improved accuracy and generalization ability, and supporting the optimization and upgrading of green construction equipment.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- 川藏铁路技术创新中心有限公司
- Filing Date
- 2026-01-29
- Publication Date
- 2026-04-21
AI Technical Summary
Existing carbon footprint prediction methods for new energy equipment in plateau railways suffer from incomplete characteristic engineering systems, insufficient prediction model accuracy, and inadequate generalization ability, making it difficult to meet the refined requirements of green construction.
A carbon footprint prediction system for new energy equipment on plateau railways was constructed by employing multi-source feature fusion, the BiGRU-SA prediction model, and the ITSO intelligent optimization algorithm improved from the tuna swarm optimization algorithm. High-precision prediction was achieved through feature engineering, improved prediction model, and enhanced model generalization ability.
It has achieved high-precision prediction of the carbon footprint of new energy equipment for plateau railways, adapts to the actual needs of complex scenarios, and provides accurate carbon emission data to support the optimization and upgrading of green construction equipment.
Smart Images
Figure CN121903085A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of carbon footprint prediction technology for plateau railways, and in particular to a method and system for predicting the carbon footprint of new energy equipment for plateau railways. Background Technology
[0002] With the promotion of sustainable development strategies in my country, the sustainable development of the construction industry has received increasing attention and importance from all sectors of society. Green building design and green construction, as important means of implementing sustainable development strategies in the construction industry, have been widely understood and valued by many industry professionals. Green construction is the specific application of sustainable development strategies in the field of engineering construction and a comprehensive application of green construction technologies. Railways, as one of the main modes of transportation, play a significant role in promoting the overall socio-economic development of my country. my country is currently in a period of rapid railway construction development, with investment in railway infrastructure increasing year by year. However, actively developing railways should not only be a key focus for boosting local economies but also a measure to build a green and environmentally friendly transportation system. The construction phase is the main stage where railway engineering projects generate pollution and consume resources; without strong measures, it will cause significant environmental damage. Therefore, the green substitution of tunnel construction equipment is imperative. The strategic value of predicting the carbon footprint of green construction equipment for plateau railways lies in the deep integration of environmental, economic, and social dimensions. From an ecological protection perspective, the ecosystems in plateau regions are sensitive and fragile. Accurate prediction of carbon emissions from construction equipment can identify the intensity of pollution sources at each stage, providing data anchors for setting phased emission reduction targets and optimizing equipment combinations. From an engineering economics perspective, carbon footprint prediction can provide a reference for oil and electricity economics analysis. From an industry development perspective, carbon footprint prediction is conducive to building a green construction standard system for plateau railways and promoting the upgrading of construction equipment towards electrification and intelligence.
[0003] Current mainstream prediction methods are insufficient to meet the demands of green building for refined carbon emission management. Specifically, the main technical problems are as follows: (1) Incomplete feature engineering system leads to missing key environmental factors: Existing prediction methods do not fully consider features strongly correlated with carbon footprint in the feature input stage. For example, key parameters such as cumulative charging amount, cumulative discharging amount, total equipment working time, and cumulative carbon emission reduction are not included in the model. If the above-mentioned key parameters are ignored, the feature engineering may be seriously out of touch with the actual working conditions, thus affecting the prediction accuracy of carbon footprint. In addition, relevant features are not strongly correlated with carbon footprint in all scenarios. It is necessary to formulate a feature selection strategy to accelerate the model training speed and improve the prediction accuracy.
[0004] (2) The basic prediction algorithm has not been improved, and the model accuracy is limited: Although some methods have attempted to introduce many time series prediction models, the model structure has not been adapted. The carbon emission data of plateau construction equipment has strong nonlinear and non-stationary characteristics, and the construction cycle is long, with many features and a large amount of data. Traditional time series prediction models are prone to gradient vanishing when dealing with long time series dependencies, and the relevant models need to be improved.
[0005] (3) The prediction model has insufficient generalization ability and is difficult to adapt to the dynamic changes of actual business scenarios: Most data-driven models are only trained on datasets of specific regions or equipment. However, the construction of plateau railways has the characteristics of large spatial span, variable environment and diverse equipment types. When traditional models cross regions or face different types of construction equipment, the prediction error increases significantly.
[0006] In summary, existing carbon footprint prediction methods have many shortcomings in terms of feature selection, prediction accuracy, and generalization. There is an urgent need for a new carbon footprint prediction method for new energy equipment on plateau railways to meet the carbon footprint prediction needs in complex scenarios. Summary of the Invention
[0007] To address the aforementioned technical problems, this invention provides a method and system for predicting the carbon footprint of new energy equipment for plateau railways. The aim is to provide a method and system for predicting the carbon footprint of new energy equipment for plateau railways based on multi-source features. By feature engineering, improving the prediction model, and enhancing the model's generalization ability, this invention solves the problems of missing features, insufficient accuracy, and poor generalization in existing prediction methods, thereby achieving high-precision prediction of the carbon footprint of green construction equipment for plateau railways.
[0008] This invention is achieved using the following technical solution: Firstly, a method for predicting the carbon footprint of new energy equipment for plateau railways includes the following steps: Step S1: Collect multi-source data based on the data acquisition box of the plateau railway construction equipment to obtain data related to the equipment and carbon footprint; Step S2: Impute missing values in the original data, and use Person correlation analysis to obtain the correlation between equipment data and carbon footprint. Construct a correlation coefficient matrix, and select features with high correlation as feature variables for the prediction model according to a preset threshold. Step S3: A basic prediction model is constructed using BiGRU bidirectional gated recurrent units. An SA self-attention mechanism is introduced to capture long-term dependencies, and a basic BiGRU-SA prediction model is constructed. The dataset is divided and the original data is standardized. Carbon footprint and highly correlated features are input into the model. Step S4: Introduce the intelligent optimization algorithm ITSO, which is an improvement on the tuna swarm optimization algorithm, to globally optimize the hyperparameters of the basic BiGRU-SA prediction model, and obtain the BiGRU-SA-ITSO hybrid prediction model. Step S5: Use the optimal solution obtained from global optimization as hyperparameters to train the model and obtain the final model. Input historical data to predict the carbon footprint of new energy equipment for plateau railways.
[0009] Specifically, the acquisition of equipment and carbon footprint-related data in step S1 includes: cumulative equipment charging amount, cumulative equipment discharging amount, total equipment working time, cumulative equipment carbon emission reduction, and equipment carbon footprint.
[0010] Specifically, in step S2, missing value imputation of the original data is represented as follows: ; in, This refers to the carbon footprint data or related characteristic data at the current moment. This refers to carbon footprint data or related characteristic data from the previous moment. This refers to carbon footprint data or related characteristic data at a later point in time.
[0011] Specifically, the Person correlation analysis calculates the correlation coefficients between each feature and the carbon footprint, using the following formula: ; in, Pearson correlation coefficient between cumulative charging amount, cumulative equipment discharging amount, total equipment operating time, cumulative equipment carbon emission reduction, and equipment carbon footprint; and The first, representing different historical data One sampling point; and These represent the average values of different historical data.
[0012] Specifically, the BiGRU bidirectional gated recurrent unit in step S3 includes two GRU networks, and the calculation formula is expressed as follows: ; ; ; ; in, To reset the door; To update the door; This is the candidate hidden state; Output for hidden layer; Enter the current time; The hidden state from the previous moment; , , These are the weight matrices for the reset gate, update gate, and candidate hidden states, respectively. , , These are the biases for resetting the gate, updating the gate, and the candidate hidden state, respectively. This is the Sigmoid function.
[0013] Specifically, the SA self-attention mechanism captures long-distance dependencies as follows: ; in, It is a self-attention function; , and The input matrices represent the query matrix, key matrix, and value matrix, respectively. For vector dimensions.
[0014] Specifically, the intelligent optimization algorithm ITSO is based on the tuna swarm optimization algorithm, but is improved by introducing Logistic chaotic mapping and elite reverse learning strategies, including: Logistic chaotic mapping is used to generate population locations, enhancing population diversity. The computational formula is as follows: ; An elite reverse learning strategy is adopted to accelerate the convergence speed and accuracy of the tuna swarm optimization algorithm. The calculation formula is as follows: ; in, , , yes boundary values, The individual's position before implementing the strategy. The individual position after reverse learning from the elites.
[0015] Specifically, the foraging process of the intelligent optimization algorithm ITSO, which is based on the improved tuna swarm optimization algorithm, is represented as follows: ; ; ; in, for During the nth iteration The location of each individual; The optimal position for the individual; and These are the weighting coefficients; This represents the current iteration number; , Table of maximum number of iterations; It is a random number within (0,1); It is a constant; A reference point randomly generated in space; A random number that is either -1 or 1.
[0016] On the other hand, a carbon footprint prediction system for new energy equipment on plateau railways includes: Carbon emission data acquisition and calculation module: Collects and calculates carbon emission and related characteristic data, performs Pearson correlation analysis on the relevant data, and normalizes the selected carbon emission data and related characteristics; Model training and optimization module: Based on the selected raw data, a baseline prediction model is built, and the advanced tuna swarm optimization algorithm is used to optimize hyperparameters. The optimization process and results are stored, and the final prediction model is built using the optimal combination of hyperparameters. Carbon emission prediction module: Input the historical data matrix into the final prediction model to obtain the carbon emission prediction results.
[0017] The beneficial effects of this invention are as follows: (1) Key data from multiple sources, such as cumulative charging, discharging, working time, carbon emission reduction, and carbon footprint, were acquired through the data acquisition box of the plateau railway construction equipment. A correlation coefficient matrix was constructed using missing value imputation and Person correlation analysis, and highly correlated feature variables were selected based on thresholds. Compared with traditional single-source data prediction, multi-source feature fusion comprehensively depicts the equipment operation characteristics, missing value handling ensures data integrity, and correlation analysis eliminates redundant features, providing high-quality input for the prediction model and laying the foundation for accurate prediction. (2) A self-attention mechanism is introduced on the basis of the bidirectional gated recurrent unit to construct the BiGRU-SA core architecture. The bidirectional time series modeling captures the correlation features of equipment operation data, while SA focuses on key time steps and feature dimensions that have a significant impact on carbon footprint prediction through an adaptive weight allocation mechanism, effectively strengthening the model's ability to capture long-term dependencies and significantly improving the model's prediction accuracy; (3) The model hyperparameters are globally optimized using the Intelligent Optimization Algorithm (ITSO), which is an improvement on the tuna swarm optimization algorithm. This simulates the social behavior and search strategy of tuna swarms, overcoming the shortcomings of traditional optimization methods that are prone to getting trapped in local optima. The hyperparameter configuration optimized by ITSO enables the model to more efficiently approach the global optimum, ensuring stable and reliable prediction results and giving the prediction model good generalization ability. This method can meet the actual needs of accurate carbon footprint prediction for new energy equipment on plateau railways under various complex scenarios. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.
[0019] Figure 1 This is a flowchart of the carbon footprint prediction method for new energy equipment on plateau railways in an embodiment of the present invention. Detailed Implementation
[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.
[0021] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.
[0022] The following is in conjunction with the appendix Figure 1 The following describes some embodiments of the present invention in detail. Unless otherwise specified, the following embodiments and features can be combined with each other.
[0023] This invention proposes a method and system for predicting the carbon footprint of new energy equipment for plateau railways. In a preferred embodiment, the method is as follows: Figure 1 As shown, it includes: Step S1: Collect multi-source data based on the data acquisition box of the plateau railway construction equipment to obtain data related to the equipment and carbon footprint; Step S2: Impute missing values in the original data, and use Person correlation analysis to obtain the correlation between equipment data and carbon footprint. Construct a correlation coefficient matrix, and select features with high correlation as feature variables for the prediction model according to a preset threshold. Step S3: A basic prediction model is constructed using BiGRU bidirectional gated recurrent units. An SA self-attention mechanism is introduced to capture long-term dependencies, and a basic BiGRU-SA prediction model is constructed. The dataset is divided and the original data is standardized. Carbon footprint and highly correlated features are input into the model. Step S4: Introduce the intelligent optimization algorithm ITSO, which is an improvement on the tuna swarm optimization algorithm, to globally optimize the hyperparameters of the basic BiGRU-SA prediction model, and obtain the BiGRU-SA-ITSO hybrid prediction model. Step S5: Use the optimal solution obtained from global optimization as hyperparameters to train the model and obtain the final model. Input historical data to predict the carbon footprint of new energy equipment for plateau railways.
[0024] The following detailed description of this solution is provided in conjunction with specific embodiments: In one specific embodiment, the carbon footprint prediction of new energy equipment for plateau railways includes the following steps: Based on the multi-source data collected by the data acquisition box of the plateau railway construction equipment, important data such as the cumulative charging amount of the equipment, the cumulative discharging amount of the equipment, the total working time of the equipment, the cumulative carbon emission reduction of the equipment, and the carbon footprint of the equipment are obtained.
[0025] Missing values were imputed in the original data, and Pearson correlation analysis was used to obtain the correlation between cumulative equipment charging, cumulative equipment discharging, total equipment operating time, cumulative equipment carbon emission reduction, and carbon footprint. A correlation coefficient matrix was then constructed, and features with high correlation were selected as feature variables for the prediction model according to predefined thresholds. The dataset was divided into training, validation, and test sets in an 8:1:1 ratio, and the data was standardized.
[0026] A prediction module is constructed, using a bidirectional gated recurrent unit (BiGRU) to build a basic prediction model. A self-attention mechanism (SA) is introduced to capture long-term dependencies, and the hyperparameters of the prediction model are globally optimized using an intelligent optimization algorithm based on the tuna swarm optimization algorithm. Finally, a BiGRU-SA-ITSO hybrid prediction model is obtained.
[0027] The optimal solution obtained from ITSO optimization is used as the hyperparameters of BiGRU-SA and used for training. The final trained model is then tested on a test set.
[0028] In this embodiment, the formula for imputing missing values in the original dataset is as follows: ; In the formula, This refers to the carbon footprint data or related characteristic data at the current moment. This refers to carbon footprint data or related characteristic data from the previous moment. This refers to carbon footprint data or related characteristic data at a later point in time.
[0029] In this embodiment, Person correlation analysis needs to calculate the correlation coefficient between each feature and the carbon footprint. The calculation formula is as follows: ; In the formula, Pearson correlation coefficient between cumulative charging amount, cumulative equipment discharging amount, total equipment operating time, cumulative equipment carbon emission reduction, and equipment carbon footprint; and The first, representing different historical data One sampling point; and These represent the average values of different historical data.
[0030] In this embodiment, the bidirectional gated recurrent unit (BiGRU) is constructed from two GRU networks, and the calculation formula is as follows: In the formula, To reset the door; To update the door; This is the candidate hidden state; Output for hidden layer; Enter the current time; The hidden state from the previous moment; , , These are the weight matrices for the reset gate, update gate, and candidate hidden states, respectively. , , These are the biases for resetting the gate, updating the gate, and the candidate hidden state, respectively. This is the Sigmoid function.
[0031] In this embodiment, the self-attention mechanism (SA) can capture long-distance dependencies, and the calculation formula is as follows: ; In the formula, It is a self-attention function; , and The input matrices represent the query matrix, key matrix, and value matrix, respectively. For vector dimensions.
[0032] In this embodiment, the intelligent optimization algorithm ITSO, which is an improvement on the Tuna Swarm Optimization (TSO) algorithm, improves the TSO algorithm by employing Logistic chaotic mapping and an elite back-learning strategy. The TSO algorithm mimics the prey-chasing behavior of a swarm of tuna, performing optimization within a given space. It is used to optimize the learning rate, Dropout, and number of hidden layer nodes in the BiGRU-SA prediction model. The population initialization formula for the TSO algorithm is as follows: In the formula, Let i be the initial position of the i-th tuna; It is a random vector uniformly distributed within (0,1); and , where represents the upper and lower boundaries of the search space; N is the initial population size.
[0033] The improved intelligent optimization algorithm ITSO and the tuna shoal optimization algorithm (TSO) define the same foraging process of tuna shoals, and the calculation formula is as follows: ; ; ; In the formula, This represents the position of the i-th individual at iteration t+1. The optimal position for the individual; and t is the weight coefficient; t is the current iteration number; tmax is the maximum iteration number; b is a random number in (0,1); a is a constant. TF is a randomly generated reference point in space; TF is a random number of -1 or 1. .
[0034] In this embodiment, the intelligent optimization algorithm ITSO improves the tuna swarm optimization algorithm by employing Logistic chaotic mapping and elite back-learning strategies, specifically as follows: Logistic chaotic mapping is used to generate population locations, thereby enhancing population diversity. The formula is as follows: In the formula, These are the control parameters for the Logistic chaotic mapping. An individual within a school of tuna; An elite back-learning strategy is adopted to accelerate the convergence speed and accuracy of the tuna swarm optimization algorithm and avoid the algorithm getting trapped in local optima. The formula is expressed as follows: In the formula, , , yes boundary values, The individual's position before implementing the strategy. The individual position after reverse learning from the elites.
[0035] In this embodiment, the final model is the prediction model obtained by retraining on a specified dataset based on a combination of hyperparameters such as learning rate, Dropout, and the number of hidden layer nodes.
[0036] This invention also proposes a system for predicting the carbon footprint of new energy equipment for plateau railways based on multi-source characteristics, comprising: Carbon emission data acquisition and calculation module: This module is used to collect and calculate carbon emission and related characteristic data, perform Pearson correlation analysis on the relevant data, and normalize the selected carbon emission data and related characteristics; Model Training and Optimization Module: This module is used to build a baseline prediction model based on the filtered raw data, and to use the advanced tuna swarm optimization algorithm to optimize hyperparameters. The optimization process and results are stored, and the final prediction model is built using the optimal combination of hyperparameters. Carbon emission prediction module: Input the historical data matrix into the final prediction model to obtain the carbon emission prediction results.
[0037] The technical solution proposed in this invention solves the problems of missing features, insufficient accuracy, and poor generalization in existing prediction methods by feature engineering, improving the prediction model, and enhancing the model's generalization ability, thereby achieving high-precision prediction of the carbon footprint of green construction equipment for plateau railways.
[0038] For the foregoing embodiments, in order to simplify the description, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions involved are not necessarily essential to this application.
[0039] The above embodiments describe the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Modifications and variations made by those skilled in the art without departing from the spirit and scope of the invention should be within the protection scope of the appended claims.
Claims
1. A method for predicting the carbon footprint of new energy equipment for plateau railways, characterized in that, Includes the following steps: Step S1: Collect multi-source data based on the data acquisition box of the plateau railway construction equipment to obtain data related to the equipment and carbon footprint; Step S2: Impute missing values in the original data, and use Person correlation analysis to obtain the correlation between equipment data and carbon footprint. Construct a correlation coefficient matrix, and select features with high correlation as feature variables for the prediction model according to a preset threshold. Step S3: A basic prediction model is constructed using BiGRU bidirectional gated recurrent units. An SA self-attention mechanism is introduced to capture long-term dependencies, and a basic BiGRU-SA prediction model is constructed. The dataset is divided and the original data is standardized. Carbon footprint and highly correlated features are input into the model. Step S4: Introduce the intelligent optimization algorithm ITSO, which is an improvement on the tuna swarm optimization algorithm, to globally optimize the hyperparameters of the basic BiGRU-SA prediction model, and obtain the BiGRU-SA-ITSO hybrid prediction model. Step S5: Use the optimal solution obtained from global optimization as hyperparameters to train the model and obtain the final model. Input historical data to predict the carbon footprint of new energy equipment for plateau railways.
2. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 1, characterized in that, The specific data obtained in step S1 related to equipment and carbon footprint include: cumulative equipment charging amount, cumulative equipment discharging amount, total equipment working time, cumulative equipment carbon emission reduction, and equipment carbon footprint.
3. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 1, characterized in that, In step S2, missing value imputation is performed on the original data, which is represented as follows: ； in, This refers to the carbon footprint data or related characteristic data at the current moment. This refers to carbon footprint data or related characteristic data from the previous moment. This refers to carbon footprint data or related characteristic data at a later point in time.
4. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 3, characterized in that, The Person correlation analysis calculates the correlation coefficients between each feature and the carbon footprint, using the following formula: ； in, Pearson correlation coefficient between cumulative charging amount, cumulative equipment discharging amount, total equipment operating time, cumulative equipment carbon emission reduction, and equipment carbon footprint; and The first, representing different historical data One sampling point; and These represent the average values of different historical data.
5. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 1, characterized in that, The BiGRU bidirectional gated recurrent unit in step S3 includes two GRU networks, and the calculation formula is as follows: ； ； ； ； in, To reset the door; To update the door; This is the candidate hidden state; Output for hidden layer; Enter the current time; The hidden state from the previous moment; , , These are the weight matrices for the reset gate, update gate, and candidate hidden states, respectively. , , These are the biases for resetting the gate, updating the gate, and the candidate hidden state, respectively. This is the Sigmoid function.
6. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 5, characterized in that, The SA self-attention mechanism captures long-distance dependencies as follows: ； in, It is a self-attention function; , and The input matrices represent the query matrix, key matrix, and value matrix, respectively. For vector dimensions.
7. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 1, characterized in that, The intelligent optimization algorithm ITSO is based on the tuna swarm optimization algorithm, but is improved by introducing Logistic chaotic mapping and elite reverse learning strategies, specifically including: Logistic chaotic mapping is used to generate population locations, enhancing population diversity. The computational formula is as follows: ； An elite reverse learning strategy is adopted to accelerate the convergence speed and accuracy of the tuna swarm optimization algorithm. The calculation formula is as follows: ； in, , , yes boundary values, The individual's position before implementing the strategy. The individual position after reverse learning from the elites.
8. The carbon footprint prediction method for new energy equipment on plateau railways as described in claim 7, characterized in that, The foraging process of the intelligent optimization algorithm ITSO is represented as follows: ； ； ； in, for During the nth iteration The location of each individual; The optimal position for the individual; and These are the weighting coefficients; This represents the current iteration number; , Table of maximum number of iterations; It is a random number within (0,1); It is a constant; A reference point randomly generated in space; A random number that is either -1 or 1.
9. A carbon footprint prediction system for new energy equipment on plateau railways, used to implement the carbon footprint prediction method for new energy equipment on plateau railways as described in any one of claims 1 to 8, characterized in that, include: Carbon emission data acquisition and calculation module: Collects and calculates carbon emission and related characteristic data, performs Pearson correlation analysis on the relevant data, and normalizes the selected carbon emission data and related characteristics; Model training and optimization module: Based on the selected raw data, a baseline prediction model is built, and the advanced tuna swarm optimization algorithm is used to optimize hyperparameters. The optimization process and results are stored, and the final prediction model is built using the optimal combination of hyperparameters. Carbon emission prediction module: Input the historical data matrix into the final prediction model to obtain the carbon emission prediction results.