Rail transit engineering low-carbon evaluation method

Abstract

The invention provides a rail transit engineering low-carbon evaluation method, and relates to the technical field of green low carbon, and the method comprises the steps: obtaining multi-source data of rail transit engineering, and determining a dynamic carbon emission factor of the rail transit engineering based on the multi-source data; based on the dynamic carbon emission factor, determining the total carbon emission of the full life cycle of the rail transit project, wherein the total carbon emission comprises a first carbon emission in a construction stage, a second carbon emission in an operation stage and a third carbon emission caused by the influence of the rail transit project on the social environment; and constructing a correlation model of the total carbon emission and the economic benefit, and determining the comprehensive benefit index of the rail transit project based on the correlation model. According to the method, the multi-source operation data of the rail transit project can be integrated to calculate the dynamic carbon emission factor, accurate quantification of the full-life-cycle carbon emission of the rail transit project is achieved, and therefore a decision basis is provided for low-carbon planning, operation and policy making of the rail transit project.

Description

Technical Field

[0001] The present application relates to the field of green and low-carbon technologies, and in particular to a method for low-carbon assessment and evaluation of rail transit projects. Background Art

[0002] With the acceleration of urbanization, rail transit projects have experienced rapid development. However, energy consumption and carbon emissions are becoming increasingly prominent issues during the construction and operation of rail transit projects. Traditional rail transit project evaluation methods often focus on technical performance, safety, and economic benefits, while paying relatively little attention to carbon emissions and environmental impacts.

[0003] Existing assessment systems typically use relatively simple indicators and methods, making it difficult to fully and accurately reflect the carbon emissions of rail transit projects throughout their lifecycle. For example, during the construction phase, carbon emissions from the production, transportation, and construction of building materials are not accurately calculated. During the operational phase, energy consumption and emissions assessments of train operations are not detailed enough, failing to fully consider the impact of different line conditions, train types, and operating modes on carbon emissions.

[0004] Furthermore, existing assessment methods lack carbon emission comparisons and optimization analysis across different technical solutions and measures. This fails to provide decision-makers with an effective basis for selecting low-carbon options during the planning and design phases, leading to potential energy conservation and emission reduction opportunities being overlooked. Furthermore, existing assessment systems face difficulties in data collection and processing, making it difficult to ensure data accuracy and reliability, impacting the scientific nature and practicality of the assessment results. Summary of the Invention

[0005] This application provides a low-carbon assessment and evaluation method for rail transit projects. By integrating multi-source operational data of rail transit projects to calculate dynamic carbon emission factors, it achieves accurate quantification of carbon emissions throughout the entire life cycle of rail transit projects, thereby providing a decision-making basis for low-carbon planning, operation and policy formulation of rail transit projects.

[0006] The low-carbon assessment method for rail transit projects provided in the embodiments of the present application includes:

[0007] Acquire multi-source data of a rail transit project, and determine a dynamic carbon emission factor of the rail transit project based on the multi-source data; wherein the multi-source data includes operation data of the rail transit project during the operation phase; determine the total carbon emissions of the rail transit project over its entire life cycle based on the dynamic carbon emission factor, the total carbon emissions including a first carbon emission during the construction phase, a second carbon emission during the operation phase, and a third carbon emission caused by the impact of the rail transit project on the social environment; construct a correlation model between the total carbon emissions and economic benefits, and determine a comprehensive benefit index of the rail transit project based on the correlation model, wherein the comprehensive benefit index is used to output a low-carbon assessment result of the rail transit project.

[0008] In some embodiments, the dynamic carbon emission factor has multiple updating mechanisms, including:

[0009] Performing a preliminary update based on a first preset period, wherein the preliminary update is performed by an automated script acquiring information data from multiple data sources according to preset rules and processing the information data before temporarily storing and updating the transitional database;

[0010] A deep update is performed based on a second preset period, wherein the deep update is to monitor the data in the transition database in combination with a domain expert experience model and an anomaly detection algorithm, and to update the dynamic carbon emission factor when an anomaly is detected.

[0011] In some embodiments, the first carbon emissions include carbon emissions from the production of building materials and carbon emissions from the operation of machinery during the construction process.

[0012] In some embodiments, the second carbon emission amount is obtained based on the energy consumption prediction model; and the process of constructing the energy consumption prediction model includes:

[0013] Acquiring operation data of the rail transit project, and preprocessing the operation data to use as sample data for training the energy consumption prediction model;

[0014] Based on the long short-term memory network as the model architecture of the energy consumption prediction model, the attention weight calculation formula of the energy consumption prediction model is:

[0015] e ij =v T tanh(W h h i +W x x j +b)

[0016] Among them, e ij is the attention weight, v, W h 、Wx , b is a learnable parameter; h i is the hidden layer state, x j As the input feature, e is processed by the softmax function ij Normalize the attention weights to get the context vector, and then perform weighted summation to get the context vector. This vector is then fed into the fully connected layer for energy consumption prediction. The output layer uses a fully connected layer, which contains one neuron node and outputs the predicted energy consumption value.

[0017] The energy consumption prediction model is trained based on the sample data, and the weight parameters of the energy consumption prediction model are adjusted by a back propagation algorithm to minimize the mean square error (MSE) loss function between the predicted energy consumption and the actual energy consumption:

[0018]

[0019] Among them, y i is the actual energy consumption value, To predict the energy consumption value, n is the number of sample data pairs; during the training process, the early stopping method is used to prevent overfitting, and the training process is stopped when the performance of the model on the validation set does not improve for five consecutive training rounds.

[0020] In some embodiments, during the training of the energy consumption prediction model, an Adam optimizer is used to dynamically adjust the learning rate of the model to accelerate the model convergence process; wherein the update rule of the Adam optimizer is:

[0021] m t =β1m t-1 +(1-β1)g t

[0022]

[0023]

[0024] Among them, m t and v t They are first-order moment estimation and second-order moment estimation, t is the number of iterations of the current training, β1 and β2 are the decay rates, and g t is the current gradient, is the first-order moment deviation correction, is the second-order moment deviation correction, θ t+1 is the new parameter after applying the update rule of Adam optimizer, θ t is the model parameter, α is the learning rate, and ∈ is a constant to prevent the denominator from being zero.

[0025] In some embodiments, the calculation formula for the third carbon emission amount is:

[0026] Esocial-env =E noise +E vibration +E land +E industry +E veg +E water

[0027] Among them, E social-env The third carbon emission, E noise Quantifying carbon emissions for noise pollution impacts, E vibration Quantifying carbon emissions for vibration impact, E land Quantifying carbon emissions for land impacts, E industry Quantifying carbon emissions for industrial agglomeration impacts, E veg Quantifying carbon emissions for vegetation impacts, E water Quantifying carbon emissions for water ecological impacts.

[0028] In some embodiments, the correlation model between total carbon emissions and economic benefits is:

[0029] I comprehensive =R operation -C construction -C carbon

[0030] Among them, I comprehensive is the comprehensive benefit index, R operation is the operating income of the rail transit project, C construction is the construction cost of the rail transit project, C carbon is the carbon emission cost of the rail transit project.

[0031] In some embodiments, the operating income is evaluated using the net present value method:

[0032]

[0033] Among them, R t is the operating income in year t, r is the discount rate, and T is the operating life.

[0034] In some embodiments, the carbon emission cost is obtained based on the carbon emissions and the carbon emission unit price. The carbon emissions are determined based on the dynamic carbon emission factor and the energy consumption prediction model. The carbon emission unit price is obtained by using the time series prediction ARIMA model to make a short-term prediction of the carbon emission unit price.

[0035] In some embodiments, the method may further include sending the low-carbon assessment results to a visual assessment platform for display, wherein the assessment platform is used to integrate and display the rail transit lines, stations, vehicle operation information and carbon emission data through three-dimensional maps and dynamic simulations.

[0036] Compared with the existing technology, the beneficial effects of this application are: by integrating multi-source operation data of rail transit projects to calculate dynamic carbon emission factors, accurate quantification of carbon emissions throughout the life cycle of project construction, operation and social impact is achieved; and an innovative correlation model between total carbon emissions and economic benefits is constructed, thereby outputting scientific comprehensive benefit indicators, providing a quantifiable decision-making basis with both environmental and economic benefits for low-carbon development planning, operation optimization and policy formulation of rail transit projects. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of the steps of a low-carbon assessment and evaluation method for rail transit projects provided in an embodiment of the present application.

[0038] Figure 2 A schematic diagram of the structure of the energy consumption prediction model provided in an embodiment of the present application. DETAILED DESCRIPTION

[0039] The present application is further described in detail below in conjunction with test examples and specific implementation methods. However, this should not be understood as limiting the scope of the above-mentioned subject matter of the present application to the following embodiments. All technologies implemented based on the content of the present application fall within the scope of protection of the present application.

[0040] Unless otherwise specified, in the description of the specific embodiments of this application, the terms indicating the orientation or position relationship such as "up", "down", "left", "right", "center", "inside", "outside", and "side" are based on the expression of the orientation or position relationship shown in the accompanying drawings, or the orientation or position relationship in which the product / device / apparatus is placed when it is usually used. These terms of orientation or position relationship are only for the convenience of describing the scheme of this application or simplifying the description in the specific embodiments to facilitate the technicians to quickly understand the scheme, and do not indicate or imply that a specific device / component / element must have a specific orientation, or be constructed and operated in a specific position relationship, and therefore should not be understood as limiting this application.

[0041] In the description of the embodiments of this application, the technical terms "first," "second," etc., merely distinguish one entity or operation from another and are not to be understood as indicating or implying relative importance or implicitly specifying the quantity, specific order, or primary-secondary relationship of the indicated technical features. In the description of the embodiments of this application, "plurality" means two or more, unless otherwise specifically defined.

[0042] References herein to "embodiments" mean that a particular feature, structure, or characteristic described in connection with the embodiments may be included in at least one embodiment of the present application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it constitute an independent or alternative embodiment that is mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0043] Please see Figure 1 , Figure 1 Schematic diagram of the steps of the low-carbon assessment method for rail transit projects provided in an embodiment of the present application. The steps of the low-carbon assessment method for rail transit projects may include:

[0044] S1. Obtain multi-source data on rail transit projects and determine the dynamic carbon emission factors of rail transit projects based on the multi-source data.

[0045] S2. Determine the total carbon emissions of the rail transit project over its entire life cycle based on the dynamic carbon emission factor. The total carbon emissions include the first carbon emissions during the construction phase, the second carbon emissions during the operation phase, and the third carbon emissions caused by the impact of the rail transit project on the social environment.

[0046] S3. Construct a correlation model between total carbon emissions and economic benefits, and determine the comprehensive benefit indicators of rail transit projects based on the correlation model. The comprehensive benefit indicators are used to output low-carbon assessment results of rail transit projects.

[0047] Before acquiring multi-source data in step S1, a comprehensive and highly accurate data acquisition system can be built. This data acquisition system can then be used to acquire multi-source data for rail transit projects. For example, for the production of building materials, a close data-sharing agreement can be established with major material suppliers. This data acquisition system not only captures general information such as energy type, consumption, and basic production processes for each batch of materials, but also provides in-depth insights into details such as the source and transportation method of raw materials, the energy efficiency rating of production equipment, and the efficiency of waste heat recovery during production. During construction, smart meters with two-decimal-point accuracy and highly sensitive sensors are installed on various types of construction machinery and equipment. These sensors can record equipment operating time, power profiles, and instantaneous energy consumption in real time and at high frequency. For train operations, critical and detailed data such as speed profiles accurate to the second, subtle acceleration fluctuations, and real-time traction power values ​​can be directly acquired from the train's advanced control system through a secure and reliable data interface. At the same time, smart meters and multi-parameter monitoring equipment with remote transmission and data storage functions are installed in stations and depots, which can accurately record the energy consumption details of facilities such as the correspondence between the brightness of station lighting and energy consumption, the relationship between the air volume and power of the ventilation system, and the temperature setting and energy consumption changes of the air-conditioning system.

[0048] In addition to tracking real-time data released by official authoritative organizations and information on carbon emission policies and regulations issued by various governments regarding rail transit, the data collection system can also automatically subscribe to professional industry carbon emission data journals to ensure that subtle industry trends are not missed. A deep data sharing mechanism is established with rail transit equipment manufacturers, not only to obtain carbon emission monitoring data from equipment operation, but also to require manufacturers to provide carbon emission estimates during the equipment development process. This data can provide early indications of carbon emission trends that may be brought about by new technologies. Leveraging intelligent sensor networks, high-precision carbon emission monitoring equipment is deployed at key nodes along rail transit lines (such as substations and depots), collecting real-time on-site data and supplementing it with data sources to further improve the timeliness and accuracy of data.

[0049] Data collection systems can use social media monitoring tools to monitor cutting-edge research results and technological breakthroughs shared by rail transit experts, scholars, and industry practitioners on social media platforms, exploring potential factors influencing carbon emissions and broadening the scope of data. For example, if an expert mentions research progress on new rail materials that reduce friction, timely follow-up assessments can be conducted to assess their impact on train energy consumption and carbon emissions.

[0050] During the research process, the applicant discovered that traditional calculation models usually assume that the carbon emission factor is constant, but the actual situation is that with the continuous advancement of technology, such as the application of new energy, the improvement of production processes, and the adjustment of energy structure, such as the increase in the proportion of renewable energy, and policy changes, such as restrictions on high-carbon emission industries and encouragement of low-carbon technologies, the carbon emission factor is in dynamic change. Therefore, in the embodiment of this application, a dynamic carbon emission factor is introduced, which is the carbon emission per unit energy consumption that changes with time, space or operating status. Compared with the traditional fixed factor, its core is that it can reflect the impact of energy transformation, changes in grid structure, and train operating conditions on carbon emission intensity in real time.

[0051] By establishing real-time connections with relevant databases and monitoring systems, dynamic carbon emission factors can be updated quarterly or even monthly, ensuring that the calculated results closely track changes in actual conditions. For example, when calculating carbon emissions from the production of building materials, the carbon emission factor for steel production will be adjusted based on the latest energy-saving and emission-reduction technologies used by steel mills to more accurately reflect actual carbon emissions levels.

[0052] The dynamic carbon emission factor is updated as time, space, or operating status changes. In the embodiment of the present application, the update mechanism of the dynamic carbon emission factor includes:

[0053] A preliminary update is performed based on a first preset period, wherein the preliminary update is performed by an automated script that obtains information data from multiple data sources according to preset rules and processes the information data before temporarily storing it in the transition database; and a deep update is performed based on a second preset period, wherein the deep update is performed by combining a domain expert experience model and an anomaly detection algorithm to monitor the data in the transition database, and updating the dynamic carbon emission factor when an anomaly is detected.

[0054] The first preset cycle could be one week, with a weekly preliminary data screening and update cycle. Automated scripts capture information from various data sources according to preset rules, perform preliminary cleaning to remove obvious errors or duplication, and then temporarily store it in a transitional database. The second preset cycle could be one month, with a monthly in-depth update. This in-depth update involves a detailed review of the data in the transitional database using domain expert experience models. Machine learning anomaly detection algorithms, such as the Isolation Forest algorithm, monitor data fluctuations in real time. Upon detection of anomalies, an in-depth analysis process is automatically initiated to determine whether an urgent update of the carbon emission factor is necessary. Immediate updates can be triggered when events occur, such as the large-scale deployment of new, energy-efficient train power systems (determined by monitoring industry news and equipment manufacturer production data) or policy adjustments (using policy and regulatory text analysis tools to capture key information in real time, such as significant changes in carbon emission trading prices), ensuring the timeliness of the assessment. Furthermore, a data version management system can be established to record detailed version logs for each update, including the update time, data source, reason for the update, and specific parameter changes involved, for traceability and auditability.

[0055] In an embodiment of the present application, the collected data can be classified and sorted in great detail according to different stages and links based on the carbon emission calculation model. The first carbon emissions include the carbon emissions of building materials during the production process and the carbon emissions of machinery operation during the construction process.

[0056] During the construction phase, the carbon emissions calculation of building materials not only considers the basic production energy consumption of common materials such as steel, cement, and wood, but also conducts an in-depth analysis of the carbon emissions during the production process of special materials such as high-performance concrete and new insulation materials.

[0057] For example, for steel, based on its specific production process, such as the type of steelmaking furnace (electric arc furnace, converter, etc.), energy source (traditional coal, natural gas, electricity, and their respective proportions), and the specific process and efficiency of waste recycling, combined with the dynamic carbon emission factor updated hourly, the following precise calculation formula is used to calculate the carbon emissions per unit weight of steel:

[0058]

[0059] Among them, C 钢材 MP is the carbon emission per unit weight of steel; 钢材 is the output of steel; EF i is the dynamic carbon emission factor of the corresponding energy i; is the energy consumption per unit steel produced by energy source i; RR is the waste utilization rate.

[0060] For the use of machinery during the construction process, we strictly distinguish different types and models of machinery (such as cranes, excavators, concrete mixers, etc.). Based on the accurate recording of their working time (accurate to the minute), the real-time changes in load (obtained through pressure sensors), and the dynamic monitoring of energy efficiency (taking into account equipment aging and maintenance), we calculate the energy consumption and corresponding carbon emissions of each machine in detail. Taking the crane as an example, the energy consumption calculation formula is:

[0061]

[0062] Where t1 and t2 are the start and end times of the crane's operation, respectively. The corresponding carbon emissions are:

[0063] Carbon emissions = energy consumption × dynamic carbon emission factor

[0064] Taking the calculation of train operation carbon emissions as an example, the original formula for calculating train operation carbon emissions is E train =∑ i F i ×D i , where E train is the total carbon emission of train operation, F i is the fixed carbon emission factor (carbon emission coefficient corresponding to electricity consumption per kilometer); D i For mileage.

[0065] Introducing dynamic carbon emission factor F d (t), that is, the carbon emission factors for different energy consumption types (such as electricity, diesel, etc.) are updated in real time according to time t, which can ensure that the calculation reflects the latest situation at all times. Furthermore, considering the differences in energy structure in different regions, the regional adjustment coefficient K is introduced in combination with the carbon emission factors of energy supply in different regions. r , the formula is updated to:

[0066]

[0067] In some embodiments, the second carbon emissions of the rail transit project may be determined based on an energy consumption prediction model. The process of constructing the energy consumption prediction model includes:

[0068] The operation data of the rail transit project is obtained and pre-processed as sample data for training the energy consumption prediction model; wherein, the data acquisition frequency is 1Hz, and the operation data of the rail transit project covers train speed, acceleration, load, line slope, power supply voltage, weather conditions and corresponding energy consumption data. The operation data is normalized using the Z-Score normalization method, and the formula is:

[0069]

[0070] Among them, x is the original data, μ is the data mean, σ is the data standard deviation, x norm To normalize the data and ensure that the data is in the same dimensional range, it is convenient for subsequent model training.

[0071] Using sliding window technology, we analyzed the autocorrelation and cross-correlation functions of historical data and determined that the primary cycle of energy consumption data was 15 minutes. We selected a 30-minute window size and divided the data for each continuous time period into multiple training samples. Each sample contains a period of historical operating data as input features and the corresponding energy consumption for the next period as the output label. To address noise interference in the data, we introduced a wavelet transform to denoise the raw data and improve data quality.

[0072] The model architecture based on the long short-term memory network as the energy consumption prediction model can effectively process the long-term and short-term dependencies in time series data. In the embodiment of the present application, a deep learning model combining a multi-layer convolutional neural network (CNN) and a long short-term memory network (LSTM) that integrates an attention mechanism is used to perform deep training on massive historical data. Train operation data for at least ten years are collected, covering energy consumption of different models (such as subway trains, light rail trains, trams, etc.) under various extremely complex line conditions (including detailed parameters such as minimum turning radius, maximum slope, slope length, etc.), different passenger flows (distinguishing peak, flat peak, and trough periods on weekdays, holidays, and special events), and different seasons (accurate to months) and weather conditions (including detailed parameters such as temperature, humidity, wind speed, and air pressure). During the training process, a composite loss function combining mean square error (MSE) and mean absolute error (MAE) is used to continuously adjust the weights and biases of the model through an optimization algorithm based on stochastic gradient descent (such as Adagrad, Adadelta, etc.) to minimize the error between predicted energy consumption and actual energy consumption. In actual applications, through high-speed, low-latency data interfaces, real-time access to advanced traffic management systems and high-precision meteorological data is achieved to obtain accurate hourly passenger flow forecasts for the day, detailed information on temporary route adjustments, and real-time weather conditions (temperature accurate to 0.1 degrees, humidity accurate to 1%, wind speed accurate to 0.1 m / s, etc.). These rich and new data are input into the trained model to obtain highly accurate energy consumption prediction results.

[0073] Please see Figure 2 , Figure 2A structural diagram of the energy consumption prediction model provided in an embodiment of the present application. The energy consumption prediction model provided in an embodiment of the present application includes an input layer, three hidden layers (LSTM unit layers) and an output layer. The input layer receives preprocessed train operation feature data, each feature corresponds to a neuron node, and there are 10 neuron nodes in total, corresponding to the 10 categories of key operation data collected above. The hidden layer learns and extracts features from the input data through the LSTM unit. The LSTM unit contains structures such as input gate, forget gate, and output gate, which can dynamically control the transmission and preservation of information. In order to further improve the performance of the model, the attention mechanism is introduced so that the model can automatically focus on the input features that have a greater impact on energy consumption prediction. The attention weight calculation formula of the energy consumption prediction model is:

[0074] e ij =v T tanh(W h h i +W x x j +b)

[0075] Among them, e ij is the attention weight, v, W h 、W x , b is a learnable parameter; h i is the hidden layer state, x j As the input feature, e is processed by the softmax function ij Normalize the attention weights to get the context vector, and then perform weighted summation to get the context vector. This vector is then fed into the fully connected layer for energy consumption prediction. The output layer uses a fully connected layer, which contains one neuron node and outputs the predicted energy consumption value.

[0076] The energy consumption prediction model is trained based on the sample data, and the preprocessed sample data is divided into a training set and a test set in a ratio of 8:2. The LSTM model is trained using the training set, and the weight parameters of the energy consumption prediction model are adjusted through the back propagation algorithm to minimize the mean square error (MSE) loss function between the predicted energy consumption and the actual energy consumption:

[0077]

[0078] Among them, y i is the actual energy consumption value, To predict the energy consumption value, n is the number of sample data pairs; during the training process, the early stopping method is used to prevent overfitting, and the training process is stopped when the performance of the model on the validation set does not improve for five consecutive training rounds.

[0079] Furthermore, during the training of the energy consumption prediction model, the Adam optimizer can be used to dynamically adjust the learning rate of the model to accelerate the model convergence process; wherein, the update rule of the Adam optimizer is:

[0080] m t =β1m t-1 +(1-β1)g t

[0081]

[0082] Among them, m t and v t They are first-order moment estimation and second-order moment estimation respectively, t is the number of iterations of the current training, β1 and β2 are decay rates, which can usually be taken as 0.9 and 0.999, g t is the current gradient, is the first-order moment deviation correction, is the second-order moment deviation correction, θ t+1 is the new parameter after applying the update rule of Adam optimizer, θ t is the model parameter, α is the learning rate, and ∈ is a constant to prevent the denominator from being zero.

[0083] During actual operation, real-time train operation data is fed into a trained LSTM model to quickly predict the energy consumption for the next phase, providing a basis for dispatchers to formulate energy-saving driving strategies and optimize the operation diagram in advance. For example, if energy consumption is predicted to be excessively high during a certain period, the train departure interval can be adjusted appropriately, such as from 5 minutes to 6 minutes, reducing energy consumption by reducing frequent train starts and stops. The speed curve can also be adjusted to appropriately reduce speeds on hilly sections to avoid high power output and reduce energy consumption. Simultaneously, the predicted results are compared with actual energy consumption in real time. If the deviation exceeds a set threshold of 10%, the model is automatically retrained and the parameters updated to ensure the accuracy of the predicted results.

[0084] For the quantification of social environmental factors, a variety of cutting-edge and scientifically rigorous methods are considered in the embodiments of the present application. Through large-scale, multi-level questionnaires, covering residents at different distances (accurate to ten meters) around rail transit stations, different age groups (every five years is an interval) and occupational types, a detailed understanding of the changes in their travel modes before and after the opening of rail transit, including travel frequency (accurate to the number of times per day), travel distance (accurate to kilometers), and the selected mode of transportation (distinguishing between walking, bicycles, buses, private cars, etc., and recording transfers), etc. At the same time, air quality monitoring stations with high-precision detection capabilities (capable of detecting fine particulate matter PM2.5, PM10 and the concentration of various harmful gases) and advanced traffic flow monitoring equipment (capable of distinguishing detailed information such as different vehicle models, traffic volume, speed, etc.) are set up in key areas to monitor the improvement of air quality and the relief of road traffic congestion before and after the opening of rail transit. These massive data are deeply mined and analyzed using statistical methods based on big data (such as Bayesian statistics, non-parametric statistics, etc.) and powerful geographic information system analysis tools. For example, to calculate the reduction in gasoline consumption and corresponding exhaust emissions due to rail transit attracting passengers who originally used private cars, the following quantitative model is used:

[0085]

[0086] Furthermore, the calculation formula for the third carbon emission amount provided in the embodiment of the present application is:

[0087] E social-env =E noise +E vibration +E land +E industry +E veg +E water

[0088] In the above formula, E social-env The third carbon emission, E noise Quantifying carbon emissions for noise pollution impacts, E vibration Quantifying carbon emissions for vibration impact, E land Quantifying carbon emissions for land impacts, E industry Quantifying carbon emissions for industrial agglomeration impacts, E veg Quantifying carbon emissions for vegetation impacts, E water Quantifying carbon emissions for water ecological impacts.

[0089] Among them, in order to quantify the carbon emissions of noise pollution, a noise monitoring point can be set up every 200 meters around the rail transit line to collect noise data when trains pass through at different times, with a sampling frequency of once per second. According to the environmental acoustic model, the decibel number of noise exceeding the standard is converted into an equivalent increase in carbon emissions. For every 1dB exceeding the environmental noise standard, it is assumed that the corresponding increase in E noise kg of CO2 emissions per kilometer (empirical values ​​were derived through a large-scale questionnaire survey of residents in the area to understand the impact of noise on energy-related behaviors such as air conditioning usage and window opening and closing frequency. This was combined with energy consumption statistics). Regression analysis was used to establish a relationship model between noise and energy-related behaviors such as air conditioning usage and window opening and closing frequency, further refining the quantification of noise's impact on carbon emissions. The analysis found that for every 3dB increase in noise, residents' average air conditioning usage time increased by 15 minutes, and the corresponding increase in carbon emissions was calculated.

[0090] To quantify the carbon emissions caused by vibration, high-precision vibration sensors can be used to set up monitoring points at 5m, 10m, and 15m from the center line of the track to monitor the ground vibration caused by train operation, with a sampling frequency of 10Hz. Similarly, based on the correlation model established by the research, the vibration intensity is converted into a carbon emission index based on the increase in energy consumption of surrounding buildings and the increase in health costs of residents. For example, if a certain intensity of vibration causes the air conditioning system of surrounding buildings to consume additional energy, the additional E per kilometer can be calculated. vibration kg of carbon dioxide emissions. Applying the principles of structural dynamics, we analyzed the impact of vibration on building structural stability. By integrating the relationship between increased building maintenance costs and carbon emissions, we refined the quantitative assessment of vibration impacts. Structural monitoring of surrounding buildings revealed that vibration increases the probability of cracks in building walls, leading to increased maintenance frequency. This increase in carbon emissions was quantified based on data such as maintenance material consumption and construction energy consumption.

[0091] Regarding the quantification of carbon emissions from land use impacts, if rail transit drives the improvement of land development and utilization efficiency in the surrounding areas, it reduces residents’ commuting distances, thereby reducing traffic congestion and corresponding energy consumption. By comparing the changes in the average commuting distances and traffic energy consumption of residents in the area before and after the construction of rail transit, the carbon emission reduction brought about by each kilometer of rail transit can be quantified as E land kilograms of carbon dioxide. Using the four-stage approach (trip generation, trip distribution, mode division, and traffic allocation) within a macro traffic flow model, we analyze the impact of rail transit on surrounding traffic flows and accurately calculate changes in commuting distances and traffic energy consumption. For example, we collect data on the starting points, destinations, and travel modes of residents in the surrounding area and input it into the four-stage model to simulate traffic flow distribution under different scenarios. We find that after the opening of rail transit, the average commuting distance for residents in the surrounding area has been shortened by 2.5 kilometers, and traffic energy consumption has been reduced by 15%. This quantifies the carbon reduction benefits.

[0092] The impact of industrial agglomeration on carbon emissions is quantified. The newly added economic benefits and corresponding energy consumption changes caused by industrial agglomeration along the rail transit line are analyzed. With a focus on positive promotion, the net carbon emission reduction benefit E per kilometer of line due to industrial development is calculated. industry kilograms of carbon dioxide (taking into account the difference between the increased energy consumption from industrial development and the reduction in transportation and communication energy consumption of traditional decentralized industries). Using an input-output model, we analyze the impact of rail transit investment on upstream and downstream industries, and, combined with industry energy consumption coefficients, quantify the impact of industrial agglomeration on carbon emissions. For example, through research on industries along the line, we determined that rail transit has driven the agglomeration of industries such as electronics and information technology and modern logistics. Based on the input-output table, we calculated the added output value and corresponding energy consumption of these industries. When compared with the transportation energy consumption of a traditional decentralized layout, we concluded that the net carbon emission reduction benefit from industrial development is 500 kilograms of carbon dioxide per kilometer of line.

[0093] To quantify the carbon emissions from vegetation impact, during rail transit construction, we used a combination of satellite remote sensing images and on-site mapping to accurately count the number of trees felled and the area of ​​grassland destroyed. Based on ecosystem carbon sink research, we calculated the carbon absorption loss caused by vegetation destruction, E. veg-loss At the same time, plan and monitor the greening restoration project along the route, and increase the carbon sink E corresponding to the newly added green area. veg-gain The difference between the two is the net carbon emission impact on the ecological environment. The ecological footprint model is used to assess the impact of vegetation destruction and restoration on the carrying capacity of the ecosystem and convert it into a carbon emission index. For example, if 100 mature trees are cut down, and each tree absorbs 18.3 kg of carbon dioxide per year, E veg-loss That is 1830 kg. If the new green area is 500 square meters, according to the carbon sequestration capacity of the local vegetation type, assuming that each square meter absorbs 0.7 kg of carbon dioxide per year, E veg-gain That is 350 kilograms, and the net carbon emission impact is an increase of 1,480 kilograms of carbon dioxide emissions.

[0094] For the quantification of carbon emissions from water ecological impacts, if rail transit projects involve crossing water bodies or have potential pollution risks to surrounding water bodies, factors such as the destruction of aquatic habitats and increased water treatment energy consumption due to changes in water quality are assessed and quantified as equivalent carbon emission changes E. water The water ecology QUAL2K model is used to simulate the process of water quality changes. The ecological impact of water bodies is quantified by combining the conversion relationship between water treatment costs and carbon emissions. For example, if the model simulation finds that the construction of a certain section of track has caused the dissolved oxygen content in the surrounding water to decrease, causing the living environment of aquatic organisms to deteriorate, in order to restore the water quality, the water treatment plant needs to increase energy consumption to treat sewage. Based on the energy consumption data, the corresponding E per kilometer of line is obtained. water 300 kg of carbon dioxide.

[0095] After determining the total carbon emissions of the rail transit project over its entire life cycle, a correlation model between the total carbon emissions and economic benefits is constructed. The correlation model between the total carbon emissions and economic benefits is:

[0096] I comprehensive =R operation -C construction -C carbon

[0097] Among them, I comprehensive is the comprehensive benefit index, R operation is the operating income of the rail transit project, C construction is the construction cost of the rail transit project, C carbon is the carbon emission cost of the rail transit project.

[0098] When establishing a model linking carbon emissions and economic benefits, the calculation methods and detailed parameters for each economic indicator were determined with extreme precision. Construction costs, in addition to direct construction investment (broken down to each project item) and land acquisition costs (including a detailed list of land appraisals and relocation compensation), also include detailed details such as design fees (including detailed costs for different design phases and design changes), supervision fees (calculated by work hours and difficulty), and relocation compensation (specifically, the amount paid to each household). Operating revenue includes not only precisely calculated ticketing revenue (differentiated by ticket tiers and sales channels) and advertising revenue (from different ad spaces and formats), but also government subsidies based on specific policies and project circumstances, detailed commercial leasing revenue (including store and venue rentals), and potential revenue from surrounding land development (calculated through detailed market research and evaluation models). By establishing a complex mathematical model based on quadratic programming, carbon emissions are closely linked to these economic indicators. For example, for a specific train energy-saving technology transformation plan, calculate its initial investment cost C0 (including equipment procurement, installation and commissioning, personnel training and other detailed expenses), annual energy savings E S Based on accurate calculations of energy price fluctuations and actual energy-saving effects), possible policy subsidies S for reducing carbon emissions (calculated based on the latest policy documents and actual emission reductions of the project), and increased ticket revenue R due to improved operational efficiency. t (calculated through passenger flow forecasting model and fare adjustment strategy), while fully considering the possible increase in maintenance costs C that may be caused by this solution m (including detailed costs such as equipment maintenance and parts replacement) and other factors. The evaluation index of economic benefits can be expressed as:

[0099]

[0100] Where n is the projected useful life of the solution (taking into account equipment lifespan and technology upgrade cycles), and r is the discount rate determined based on market conditions and project risk. Efficient numerical optimization algorithms (such as interior point methods and sequential quadratic programming) are used to solve multi-objective optimization problems, identifying the most economically efficient solution combination while meeting stringent carbon emission targets.

[0101] During the construction process, the cost quantification of carbon emissions is first considered. Based on the carbon emission market transaction price or the social cost of carbon emissions set by the government, the carbon emissions of rail transit projects are converted into economic costs. Assuming that the carbon emission cost is C carbon , carbon emissions are E total , the unit price of carbon emissions is P carbon , then C carbon =E total ×P carbon Here E total The above dynamic carbon emission factors and energy consumption prediction modules are used to comprehensively calculate and calculate the carbon emissions from various aspects including train operation, station equipment operation, construction, etc. carbon Real-time tracking of market dynamics or policy-driven values ​​ensures accurate cost accounting. A time series forecasting ARIMA model is used to make short-term forecasts of carbon emission unit prices, enabling more accurate estimates of future carbon emission costs. The ARIMA model modeling process includes data stationarity testing, using the Augmented Dickey-Fuller (ADF) test to determine whether the data is stationary. If not, differencing is performed; model order determination, using the tailing and truncation characteristics of the autocorrelation function (ACF) and partial autocorrelation function (PACF); and parameter estimation, using maximum likelihood estimation to estimate model parameters. For example, by analyzing carbon emission unit price data for the past year, an ARIMA (1,1,1) model is determined to predict carbon emission unit price trends for the next month, providing a reference for cost estimation.

[0102] The second is the economic benefit index consideration, construction cost C constructionThis includes all initial investments, including line laying, station construction, and rolling stock procurement, calculated through a detailed project budget list. Cost estimation models, such as analogical and parametric estimation, are used to quickly estimate construction costs in the early stages of a project and dynamically adjust them based on actual conditions during implementation. Analogous estimation references the costs of similar existing rail transit projects and incorporates factors such as the project's scale (line length accurate to the nearest 0.1 kilometer, number of stations accurate to the nearest 1) and technical difficulty (quantified by factors such as geological conditions and construction process complexity). Parametric estimation uses a cost function relationship established by linking key project parameters (line length, number of stations, etc.) with historical data. For example, if the construction cost per kilometer of a similar project is 150 million yuan, and the project's line length is 30 kilometers, taking into account the complex geological conditions and a technical difficulty coefficient of 1.2, the initial construction cost estimate is 1.5 × 30 × 1.2 = 5.4 billion yuan. This cost is then adjusted in real time based on actual bidding prices, project changes, and other factors.

[0103] Operating income R operation The main sources of revenue are ticket sales, advertising, and land value-added along the route. Future revenue streams are projected based on historical operating data and market research. Incorporating the time value of money, the net present value (NPV) method can be used to evaluate operating revenue:

[0104]

[0105] where R t is the operating income in year t, r is the discount rate, and T is the operating life. For ticket sales revenue, a passenger flow forecasting model is used to predict passenger flow over different time periods, incorporating factors such as population growth, regional development plans, and changes in surrounding competing transportation modes. This allows for an estimate of ticket revenue. Passenger flow forecasting can be modeled using a multivariate linear regression model, incorporating regional GDP, population density, distance to transportation hubs, and the number of newly built residential communities in the surrounding area as independent variables, and passenger flow as the dependent variable. For example, data analysis revealed that every 1% increase in regional GDP is associated with a 0.5% increase in passenger flow, enabling predictions of future passenger flow changes. Advertising revenue is estimated based on factors such as station tier (classified as extra-large, large, medium, and small, based on floor area and passenger flow), the number of advertising spaces, and the commercial prosperity of the area (quantified through indicators such as commercial rent levels and brand occupancy rates). Land value-added along the line is estimated based on land market assessment reports, regional planning adjustment documents, and expert judgment to estimate the extent of land value-added driven by rail transit, thus calculating revenue.

[0106] Through the analysis of I under different rail transit engineering schemes or operation strategies comprehensiveCompare and select the best solution to achieve a balanced decision between economy and environmental protection. For example, when planning a new line, compare the I of different line directions and station setting solutions. comprehensive value, select the solution with the largest positive value and the lowest carbon emissions.

[0107] A multi-objective optimization non-dominated sorting genetic algorithm (NSGA-II) is introduced to optimize comprehensive efficiency indicators. By simulating biological evolution, the NSGA-II algorithm simultaneously optimizes multiple objectives (such as maximizing economic efficiency and minimizing carbon emissions). The algorithm's workflow includes initializing the population and randomly generating a certain number (presumably 100) of initial route plans or operation strategy plans. Non-dominated sorting divides the population into different levels of non-dominated layers based on comprehensive efficiency indicators. Crowding calculation assesses the degree of crowding among individuals within the same non-dominated layer to maintain population diversity. Selection prioritizes individuals with high non-dominated levels and low crowding for the next generation. Crossover uses a simulated binary crossover operator to perform genetic crossover on selected individuals with a certain probability (presumably 0.8) to generate new individuals. Mutation involves randomly mutating individual genes with a small probability (presumably 0.05) to prevent regression into local optima. Ultimately, a set of Pareto optimal solutions is obtained, from which decision makers can select the appropriate solution based on their actual needs.

[0108] Furthermore, the method provided in the embodiment of the present application may also include: sending the low-carbon assessment results to a visual assessment platform for display, and the assessment platform is used to integrate and display the rail transit lines, stations, vehicle operation information and carbon emission data through three-dimensional maps and dynamic simulations.

[0109] In the embodiment of the present application, the visual evaluation platform is no longer a simple data table and chart display, but through intuitive forms such as three-dimensional maps and dynamic simulations, it integrates and displays information such as rail transit lines, stations, and vehicle operations with carbon emission data. Decision makers and relevant personnel can view carbon emissions in different areas and time periods by zooming and rotating the map. At the same time, virtual reality and augmented reality technologies are used to provide an immersive experience, allowing users to feel as if they are in an actual rail transit scene, and understand and grasp the distribution and changing trends of carbon emissions more clearly and quickly. For example, when displaying the carbon emissions of a large transfer station, virtual reality technology can be used to allow users to walk around the station from a first-person perspective, intuitively feel the differences in carbon emissions under different facilities and operating modes, and thus facilitate timely and accurate decision-making.

[0110] The visual assessment platform is developed based on state-of-the-art geographic information systems (GIS) and innovative data visualization technologies. Utilizing ultra-high-precision satellite maps (with decimeter-level resolution) and physically based rendering 3D modeling technology, a highly realistic model of rail transit lines and stations is constructed. Train operation data, station facility energy consumption data, carbon emissions data, and other data are seamlessly integrated with geospatial information through precise coordinate transformation and data matching algorithms. On the display interface, users can freely zoom and rotate the map using a multi-touch mouse wheel and smooth dragging to view detailed information for different regions and time periods. Carbon emissions data is presented in a variety of cutting-edge formats, including dynamic heat maps based on heat map principles, highly customizable bar charts (with selectable bar styles and colors), and real-time line charts (supporting multi-line comparison and data smoothing). Different carbon emission levels are clearly distinguished using automatically generated color gradients and brightness changes based on carbon emission levels. For example, by clicking on a station, a pop-up rich media window will display the detailed energy consumption data (accurate to kilowatt-hours) and carbon emissions percentage (in percentage form, accurate to two decimal places) of the station's lighting, air conditioning, elevators and other facilities; by sliding the timeline, the energy consumption and carbon emissions change curves at different times of the day (accurate to minutes) are dynamically displayed, and data export and historical data comparison are supported.

[0111] In the data access and integration of the visual evaluation platform, a data middle platform is built to access and integrate multi-source data such as the dynamic carbon emission factor database, energy consumption forecast model output, carbon emission and economic benefit correlation model calculation results, and quantitative data of social and environmental factors in real time. Ensure the consistency, accuracy and timeliness of the data to provide a solid data foundation for visual display. Use ETL (Extract-Transform-Load) tools to extract, clean, transform and load data from different data sources to meet the data requirements of the visualization platform. Use data lineage analysis technology to trace the source and conversion process of each data point so that the root cause of the problem can be quickly located when data anomalies occur. For example, when the visualization interface shows that the carbon emission data of a station fluctuates abnormally, through data lineage analysis, it is possible to quickly find out whether it is a problem with data source collection or a deviation in the data conversion and calculation links.

[0112] The visual assessment platform utilizes Geographic Information System (GIS) technology to construct a three-dimensional map model of rail transit lines and surrounding areas. This map visually displays train trajectories, station locations, and energy-intensive hotspots (such as high-energy-consumption uphill sections and frequent start-stop stations). Different colors and icons are used to identify carbon emission intensity, economic benefit indicators (such as a heat map of commercial revenue around stations), and social and environmental sensitive points (such as areas with excessive noise levels and ecological protection zones).

[0113] In addition, in the visual evaluation platform, users can interact with the mouse and click on specific objects to view detailed carbon emissions, energy consumption, economic data and other information, achieving a comprehensive understanding from macro to micro. For example, if you click on a station icon, a pop-up window will display detailed data such as the station's construction carbon emissions, operating energy consumption, advertising revenue, and noise complaints from surrounding residents. Using WebGL technology, efficient rendering of three-dimensional maps on the web page is achieved, improving the user experience. Setting interactive functions such as map zooming, panning, and rotation allows users to observe the relationship between rail transit and the surrounding environment from different angles. For example, users can zoom in and out to view the overall layout of multiple stations in a certain area and the corresponding carbon emission distribution, or rotate the map to observe the distance between the rail line and the surrounding ecological protection area from the side.

[0114] Finally, based on the comprehensive and in-depth evaluation results, specific, detailed, and highly targeted recommendations and improvement measures are provided for the planning, design, construction, and operation of rail transit projects. If the energy consumption of a particular line section is predicted to be too high, detailed recommendations may be made to optimize the line's alignment, precisely calculating the optimal angles and slopes to reduce curves. Alternatively, train operation strategies may be fine-tuned, such as appropriately reducing operating speeds by minutes during off-peak hours and optimizing the intervals between stops by seconds. For station facilities, specific recommendations may include adopting more energy-efficient lighting and air conditioning systems with intelligent dimming and temperature control capabilities, and optimizing ventilation designs through fluid dynamics simulations.

[0115] The following are two examples of a method for low-carbon assessment of rail transit projects provided in this application:

[0116] Example 1: Conduct a low-carbon assessment on an urban rail transit line that is under planning.

[0117] First, during the data collection phase, we obtained detailed production information from material suppliers for various construction materials that would be used to construct tracks, stations, and other facilities, including manufacturer, process, and energy consumption. For construction machinery, such as cranes, high-precision sensors were installed to record their operating hours, load conditions, and energy consumption at different times of the day throughout the week. We also obtained the planned train models and their technical specifications from the train manufacturer, including traction system efficiency and brake energy regeneration capabilities.

[0118] In the application of the carbon emission calculation model, let's take a certain type of steel used in rail construction as an example. The production process of this steel uses 60% electricity and 40% coal as energy. The carbon emission factor of electricity is 0.5 kg CO2 / kWh, and the carbon emission factor of coal is 2.5 kg CO2 / kg. The energy consumption for producing each ton of steel is 800 kWh of electricity and 300 kg of coal, and the waste recycling rate is 20%. The carbon emissions per ton of steel are calculated as:

[0119] C 钢材 =1×[(800×0.6×0.5+300×0.4×2.5)×(1-0.2)]=432kgco2 / kg

[0120] In terms of energy consumption prediction, a trained deep learning model was used to predict the energy consumption of trains on this line under different operating scenarios. During the morning rush hour, passenger flow is predicted to be high. The line includes a 2-kilometer uphill section with a 5% gradient. The model input feature vectors are: vehicle type code 001, line feature vector [2000, 5], passenger flow 1500, season code 03, and weather feature vector [18, 60, 3]. The model predicted the train's traction energy consumption to be 1500 kWh.

[0121] In the carbon emissions and economic benefit model, consider a new energy-saving train procurement plan. The initial investment cost is 50 million yuan, with an estimated annual energy savings of 3 million yuan. A policy subsidy of 500,000 yuan will be obtained due to reduced carbon emissions. Ticket revenue from improved operating efficiency will increase by 1 million yuan, while maintenance costs will increase by 800,000 yuan. The plan is expected to have a service life of 15 years and a discount rate of 8%. The economic benefits of this plan are calculated as follows:

[0122]

[0123] Through calculation, if the economic benefit is positive, the plan is economically feasible.

[0124] In the quantification of social and environmental factors, a questionnaire survey of residents within 500 meters of the station found that after the opening of the rail transit, residents who originally used private cars to travel five times a day, each trip covering 10 kilometers, with a unit mileage gasoline consumption of 0.1 liters / kilometer and a gasoline carbon emission factor of 2.3 kg of carbon dioxide per liter, now use rail transit instead, reducing their private car trips by three times per week. The indirect emission reduction is calculated as:

[0125] Indirect emissions = 3 × 5 × 10 × 0.1 × 2.3 = 345 kg CO2 / week

[0126] The visual assessment platform displays the energy consumption and carbon emissions of each station on the line. Clicking on a station displays a pop-up window showing that the lighting system consumes 200 kWh of energy and emits 100 kg of carbon dioxide per day; the air conditioning system consumes 500 kWh of energy and emits 250 kg of carbon dioxide per day.

[0127] Based on the evaluation results, recommendations were made for the line, such as optimizing the track laying angle in certain sections with larger slopes to reduce train traction energy consumption; and adopting natural lighting and ventilation designs in station design to reduce the usage time of lighting and air-conditioning systems.

[0128] Example 2: Conduct a low-carbon assessment on an urban rail transit line that has been in operation for five years.

[0129] During the data collection phase, detailed operational data for each train over the past year was collected from the train's operation control system, including per-second speed, acceleration, and traction power. Furthermore, monthly energy consumption data for lighting, ventilation, air conditioning, and other facilities was collected from the station's energy management system. Maintenance and replacement of building materials were also carefully reviewed from relevant maintenance records.

[0130] In the carbon emission calculation model, we take the use of a certain type of concrete in station construction as an example. The production of this concrete uses 70% natural gas and 30% electricity. The carbon emission factor for natural gas is 2 kg CO2 / m³, and the carbon emission factor for electricity is 0.6 kg CO2 / kWh. The energy consumption for producing each cubic meter of concrete is 120 kWh of electricity and 8 cubic meters of natural gas, and the waste recycling rate is 15%. The carbon emissions per cubic meter of concrete are calculated as:

[0131] C 混凝土 =1×[(120×0.3×0.6+8×0.7×2)×(1-0.15)]=28.56kgco2 / m 3

[0132] For energy consumption prediction, a deep learning model is used to predict train energy consumption under different weather conditions for the next month. For example, on a day predicted to be windy, the line feature vector is [wind speed: 15 m / s, wind direction: 30° angle with the line], the passenger flow is 1,200, the season code is 10, and the weather feature vector is [12, 40, 15]. The model predicts train energy consumption of 1,200 kWh.

[0133] In the carbon emissions and economic benefit model, consider the energy-saving renovation of the station lighting system. The initial investment cost is 800,000 yuan, with an estimated annual energy savings of 150,000 yuan. A policy subsidy of 80,000 yuan will be obtained for reducing carbon emissions. Maintenance costs remain unchanged. The projected service life of the project is 8 years, and the discount rate is 6%. The economic benefits of this project are calculated as:

[0134]

[0135] Through calculation, if the economic benefit is positive, then the transformation plan is economically reasonable.

[0136] In quantifying social and environmental factors, traffic flow monitoring equipment installed around the station revealed that after the opening of the rail transit, traffic volume on surrounding roads during peak hours decreased by 20%, while average vehicle speeds increased by 30%. Assuming that each vehicle originally idled for an average of 20 minutes during peak hours, with an idling fuel consumption of 0.2 liters per minute, and a gasoline carbon emission factor of 2.3 kg of CO2 per liter, the indirect emission reductions are calculated as:

[0137] Indirect emissions = 0.2 × 20 × 2.3 = 9.2 kg CO2 / peak hour

[0138] On the visual assessment platform, you can clearly see the energy consumption contribution of different facilities at each station. For example, the ventilation system in a certain station accounts for 30% of total energy consumption and 25% of carbon emissions.

[0139] Based on the evaluation results, it is recommended that in terms of train operation, the train departure frequency should be further optimized according to the passenger flow in different time periods; in terms of station facility management, the power of lighting and air conditioning should be intelligently adjusted according to the season and weather conditions.

[0140] During the implementation process described above, by establishing a comprehensive and accurate data collection system and advanced carbon emission calculation models, accurate carbon emission quantification results can be provided for rail transit projects. This helps relevant departments and enterprises clearly understand the carbon emissions profile of each project phase, enabling them to formulate targeted emission reduction strategies and reduce decision-making errors caused by inaccurate carbon emission calculations. Secondly, the deep learning-based energy consumption prediction module and multi-scenario comparative optimization analysis provide decision-makers with accurate and valuable references during the planning and design stages. This not only enables the selection of the most energy-efficient train models, route alignments, and operating modes, but also allows for the early prediction and avoidance of potential high energy consumption issues, effectively reducing project operating costs and improving energy efficiency. Furthermore, by closely linking carbon emissions with economic benefits, the economic feasibility of projects can be fully considered while pursuing low-carbon goals. This helps balance environmental protection and economic development, promotes the sustainable development of the rail transit industry, and attracts more investment and resources to low-carbon rail transit projects. Furthermore, the quantitative assessment and incorporation of social and environmental factors can more comprehensively reflect the comprehensive value of rail transit projects. Beyond the direct energy-saving and emission-reduction benefits, the project also benefits from improvements to residents' quality of life, relief from urban traffic congestion, and positive impacts on the ecological environment, enhancing public acceptance and support for the project. Finally, the development of a visual assessment platform significantly improves the efficiency of information transmission and the scientific nature of decision-making. This intuitive and clear presentation enables decision-makers, engineers, and the public to quickly understand the project's carbon emissions and potential impacts, fostering effective communication and collaboration among all parties and jointly driving rail transit projects towards a more low-carbon, efficient, and sustainable development.

[0141] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. A low-carbon assessment method for rail transit projects, characterized in that: include: Acquiring multi-source data of a rail transit project, and determining a dynamic carbon emission factor of the rail transit project based on the multi-source data; wherein the multi-source data includes operational data of the rail transit project during an operation phase; Determine the total carbon emissions of the rail transit project over its entire life cycle based on the dynamic carbon emission factor, wherein the total carbon emissions include first carbon emissions during the construction phase, second carbon emissions during the operation phase, and third carbon emissions caused by the impact of the rail transit project on the social environment; A correlation model between the total carbon emissions and economic benefits is constructed, and a comprehensive benefit index of the rail transit project is determined based on the correlation model. The comprehensive benefit index is used to output a low-carbon assessment result of the rail transit project.

2. The method according to claim 1, characterized in that The dynamic carbon emission factor has multiple updating mechanisms, including: Performing a preliminary update based on a first preset period, wherein the preliminary update is performed by an automated script acquiring information data from multiple data sources according to preset rules and processing the information data before temporarily storing and updating the transitional database; A deep update is performed based on a second preset period, wherein the deep update is to monitor the data in the transition database in combination with a domain expert experience model and an anomaly detection algorithm, and to update the dynamic carbon emission factor when an anomaly is detected.

3. The method according to claim 1, characterized in that The first carbon emissions include carbon emissions from the production of building materials and carbon emissions from the operation of machinery during the construction process.

4. The method according to claim 1, wherein The second carbon emission amount is obtained based on an energy consumption prediction model; the process of constructing the energy consumption prediction model includes: Acquiring operation data of the rail transit project, and preprocessing the operation data to use as sample data for training the energy consumption prediction model; Based on the long short-term memory network as the model architecture of the energy consumption prediction model, the attention weight calculation formula of the energy consumption prediction model is: e ij =v T fishy(W h h i +W x x j +b) Among them, e ij is the attention weight, v, W h 、W x , b is a learnable parameter; h i is the hidden layer state, x j As the input feature, e is processed by the softmax function ij Normalize the attention weights to get the context vector, and then perform weighted summation to get the context vector. This vector is then fed into the fully connected layer for energy consumption prediction. The output layer uses a fully connected layer containing one neuron node to output the predicted energy consumption value. The energy consumption prediction model is trained based on the sample data, and the weight parameters of the energy consumption prediction model are adjusted by a back propagation algorithm to minimize the mean square error (MSE) loss function between the predicted energy consumption and the actual energy consumption: Among them, y i is the actual energy consumption value, To predict the energy consumption value, n is the number of sample data pairs; during the training process, the early stopping method is used to prevent overfitting, and the training process is stopped when the performance of the model on the validation set does not improve for five consecutive training rounds.

5. The method according to claim 4, characterized in that During the training of the energy consumption prediction model, the Adam optimizer is used to dynamically adjust the learning rate of the model to accelerate the model convergence process. The update rule of the Adam optimizer is: m t =β1m t-1 +(1-β1)g t in,, t and v t They are first-order moment estimation and second-order moment estimation, t is the number of iterations of the current training, β1 and β2 are the decay rates, and g t is the current gradient, is the first-order moment deviation correction, is the second-order moment deviation correction, θ t+1 is the new parameter after applying the update rule of Adam optimizer, θ t is the model parameter, α is the learning rate, and ∈ is a constant to prevent the denominator from being zero.

6. The method according to claim 1, characterized in that The calculation formula for the third carbon emissions is: AND social-env =And noise +E vibration +E land +E industry +E veg +E water Among them, E social-env The third carbon emission, E noise Quantifying carbon emissions for noise pollution impacts, E vibration Quantifying carbon emissions for vibration impact, E land Quantifying carbon emissions for land impacts, E industry Quantifying carbon emissions for industrial agglomeration impacts, E veg Quantifying carbon emissions for vegetation impacts, E watet Quantifying carbon emissions for water ecological impacts.

7. The method according to claim 1, characterized in that The correlation model between total carbon emissions and economic benefits is: I comprehensive =R operation -C construction -C carbon Among them, I comprehensive is the comprehensive benefit index, R operation is the operating income of the rail transit project, C construction is the construction cost of the rail transit project, C carbon is the carbon emission cost of the rail transit project.

8. The method according to claim 7, characterized in that The operating income is estimated by the net present value method: Among them, R t is the operating income in year t, r is the discount rate, and T is the operating life.

9. The method according to claim 7, characterized in that The carbon emission cost is obtained based on the carbon emission amount and the carbon emission unit price. The carbon emission amount is determined based on the dynamic carbon emission factor and the energy consumption prediction model. The carbon emission unit price is obtained by using the time series prediction ARIMA model to make a short-term prediction of the carbon emission unit price.

10. The method according to any one of claims 1 to 9, characterized in that The method also includes sending the low-carbon assessment results to a visual assessment platform for display, wherein the assessment platform is used to integrate and display the rail transit lines, stations, vehicle operation information and carbon emission data through three-dimensional maps and dynamic simulation.

Images