Investigation of carbon emission in slurry shield tunnel construction based on modified process analysis method

Abstract

Underground engineering, including shield tunnel construction, is a significant contributor to carbon dioxide emissions in infrastructure engineering projects. To better predict and control the carbon emissions associated with shield tunnel construction, this paper presents a novel calculation method: the modified process analysis method based on input-output and process analysis methods. To evaluate the effectiveness of the proposed method, a specific shield tunnel construction project was selected as a case study. The modified process analysis method was used to analyze the various factors that influence carbon emissions during the project’s construction phase. In addition, a neural network approach was applied to validate the accuracy of the calculation using the LSTM and BP neural network. The results demonstrate that the proposed method not only combines the strengths of traditional methods but also offers high accuracy and acceptable error rates. Based on these findings, several measures to reduce carbon emissions during shield tunnel construction are suggested, providing valuable insights for reducing CO₂ emissions associated with infrastructure engineering projects. This study highlights the importance of adopting innovative approaches to reduce carbon emissions and promotes the implementation of sustainable practices in the construction industry. Through the use of advanced analytical methods, such as the proposed modified process analysis method, we can effectively mitigate the environmental impact of construction activities and make significant contributions to the global effort to combat climate change.

摘要

地下工程中的盾构隧道施工, 是基础设施工程中二氧化碳排放的重要贡献者。为了更好地预测和控制盾构隧道施工碳排放, 本文提出了一种新的计算方法, 即基于投入产出和过程分析方法的改进过程分析法。为评估所提方法的有效性, 选取某盾构隧道施工项目作为案例, 采用修正的过程分析法, 对项目施工阶段影响碳排放的各种因素进行剖析。此外, 采用神经网络方法, 验证了LSTM和BP神经网络的计算精度。结果表明, 该方法不仅结合了传统方法的优点, 而且具有较高的准确率和可接受的错误率。基于这些研究结果, 提出了减少盾构隧道施工过程中碳排放的若干措施, 为减少与基础设施工程项目相关的CO₂排放提供了有价值的见解。

1 Introduction

Sustainable development requires us to meet today’s needs without compromising our children and future generations and their needs [2]. For nearly a hundred years, the continued increase in greenhouse gases and the overall rise in global temperature are two objective facts, and an increasing number of researchers believe that the continued increase in greenhouse gases is a direct cause of global warming [9]. The infrastructure industry, particularly civil engineering, consumes a lot of raw materials, such as cement, steel, timber, and employs a variety of equipment, which is generating huge amounts of carbon dioxide emissions. Controlling the total carbon dioxide emissions of civil engineering has become the strategic goal of China’s sustainable development.

International mainstream carbon dioxide calculation methods are mainly IPCC (Intergovernmental Panel on Climate Change) emission coefficient and life cycle analysis (LCA). The IPCC Emission Factor Approach refers to a greenhouse gas emission calculation method proposed by the Intergovernmental Panel on Climate Change (IPCC). This method involves detailed categorization of potential greenhouse gas emissions from various economic activities, and provides corresponding emission factors based on data collection and analysis to estimate the greenhouse gas emissions of different sectors and regions. This approach is widely used internationally and can be used to assess the greenhouse gas emissions of different countries, industries, and products, providing a reference for the development of climate change-related policies. IPCC can have three calculation methods according to different sources of emission coefficient: IPCC default emission coefficient estimation method, country default emission coefficient estimation method and detailed emission calculation model. The LCA method includes the input-output method and the process analysis method. The input-output method is a final calculation method and a method for subsequent analysis. It takes the total amount of carbon dioxide as the main line, and the final result is the carbon emissions generated by all engineering activities, namely the total amount of carbon emissions. The process analysis method is a kind of analysis method, which takes engineering activities as the main line and is similar to the budget calculation method. Although the input-output method can obtain the total number, this method cannot fully reflect the proportion of each small part, which is not convenient for targeted control of carbon emissions. In the process of actual operation of process analysis, there will be deviations, resulting in a large gap between the final results and the actual results. Process analysis can solve the shortcomings of input-output method to some extent, but the results are not reliable.

Numerous scholars have conducted studies on carbon dioxide emissions in civil engineering. Some scholars have found that carbon dioxide emissions account for a large proportion of the total emissions in the global construction industry. Seo et al. used the LCA method to estimate the carbon dioxide emissions of residential buildings, and found that the emissions in the operation stage accounted for 87.5%~96.9% of the total, while the carbon dioxide emissions in the construction stage were 0.14–0.23 t/m² [14]. Huang et al. used the world environmental index to compare the carbon dioxide emissions caused by global construction activities, and found that the carbon dioxide emissions of emerging economies accounted for nearly 60 % of the total emissions of the global construction industry [5]. Lee et al. used inventory analysis to estimate carbon dioxide emissions from residential buildings during construction, and found that 84.9% of carbon dioxide emissions were from main buildings during construction [8]. Guo et al. used LCA method to propose five different functional units. Based on rock classification and pavement classification, and found that cement and electricity emissions accounted for more than 81% of the total emissions [4].

Some scholars have explored the main sources of carbon emissions in the construction industry. Itoh et al. used emission coefficient method to estimate the carbon dioxide emissions from bridge construction. It is found that most of the carbon dioxide emissions are mainly from building materials [6]. Seo et al. found that CO₂ emissions in the raw material production stage accounted for 93.4% of the total CO₂ emissions from the perspective of construction site, transportation and construction stages [13]. Zhang et al. proposed a process-based method to estimate CO₂ emissions in the whole life cycle of China’s construction industry from 2005 to 2012. It was found that carbon dioxide emissions were mainly from steel and cement in the material manufacturing stage, accounting for 87% [18].

Some scholars have given some suggestions and methods to reduce carbon emissions through exploratory research. Selving et al. studied the carbon dioxide emissions from residential buildings considering many factors, such as residential location, design. It was found that renovation and reuse of old buildings can reduce emissions by 60%~70%, and clean energy and energy-saving design can reduce emissions by at least 50% [12]. Zoubir et al. used the whole life cycle method to compare the carbon dioxide emissions of two same bridge decks constructed with high performance concrete and ordinary concrete in corrosive service environment. The results show that the carbon emissions of the bridge deck constructed with high performance concrete during construction and repair are 65% less than those of the bridge deck constructed with ordinary concrete [20]. Kim et al. proposed a quantitative method to evaluate the environmental impact of concrete CO₂ emissions [7]. Xu et al. extended the relationship between geological conditions construction parameters and greenhouse gas emissions, and found out the key indicators affecting tunnel construction emissions [16]. Zhou et al. based on the life cycle method, evaluated the CO₂ emissions at different stages of the building with life cycle-method, and provided some effective suggestions for reduction [19].

Compared with construction engineering and road and bridge engineering, the research on carbon emissions of underground engineering is relatively less, especially in tunnel, and other projects [11]. Piratla et al. used the whole life cycle method to analyze the carbon emissions of a water conveyance tunnel in the whole life cycle [10]. Xu et al. used the whole life cycle method to calculate the carbon emissions of multiple tunnels. And found out the influencing factors of greenhouse gas emissions, [15] was proposed. Chau et al. concluded that carbon emissions in underground engineering construction were greater than automobile emissions and personal’s daily life emissions [3]. Xu et al. used LCA method to evaluate and analyze greenhouse gases produced in five different tunnels, and put forward the importance of surrounding rock conditions for greenhouse gas emissions [16]. Ahn et al. studied carbon dioxide emissions during tunnel construction using input-output analysis. And found that the CO₂ produced by tunnel boring machine (TBM) is at least 35%, which is the main emission source in the process of construction, transportation and equipment operation [1]. Xu et al. realized the low carbon design of tunnel construction, discussed the relationship between tunnel design and greenhouse gas emission, and determined the marginal greenhouse gas emission caused by the change of tunnel lining design parameters [17]. Only by collecting a large number of energy consumption data of tunnel construction, can we deeply analyze the carbon dioxide emission characteristics of tunnel construction and propose corresponding carbon dioxide emission reduction measures.

LCA is a widely used methodology for evaluating the environmental impact of products and processes. However, traditional LCA approaches have limitations, including the difficulty in capturing complex relationships and patterns in large and diverse datasets. In recent years, the application of machine learning models has shown promising results in various fields, including environmental analysis. The incorporation of machine learning models in LCA analysis represents a significant innovation in the field. It has the potential to offer valuable insights into the environmental impact of products and processes, and to improve the accuracy and efficiency of the assessment.

This paper mainly studies the carbon emissions in the construction process of shield tunnels. Combined with a slurry shield construction project, a convenient and accurate calculation method for estimating the carbon emissions of shield tunnels is proposed on the basis of input-output and process analysis. In this study, we introduce the use of machine learning models in LCA analysis to improve the accuracy and efficiency of the assessment. By leveraging advanced algorithms, we aim to identify previously undetected patterns and relationships in the data, providing a more comprehensive understanding of the environmental impact of the studied product or process. Various factors affecting the carbon emissions are analyzed, and some measures to reduce the carbon emissions in the construction tunnel are put forward.

2 Process analysis method for carbon emission correction of shield tunnel

2.1 Principle of modified process analysis

Input-output method is used to calculate the amount of engineering carbon emission sources. Taking carbon emissions as the main line, one of the necessary conditions of input-output method is that the project must have very complete equipment records, material accounts and energy consumption accounts. In general, input-output method is a post-analysis method.

The process analysis method takes engineering activities as the main line, and divides the project of a system into subsystems according to the order of engineering activities. By calculating the carbon emissions generated by each subsystem, the total carbon emissions of engineering activities are finally obtained by summing up. The process analysis method is suitable for estimating the carbon emissions of the whole project before the beginning of engineering activities.

Although the process analysis method can calculate the proportion of carbon emissions in engineering activities, the reliability of the results is not high, the input-output method cannot reflect the proportion of carbon emissions in each part of the project, and it is not conducive to the targeted control of carbon emissions in engineering activities. Therefore, in order to accurately grasp the carbon emissions in the decision-making process of shield tunnel construction, the input-output and process analysis are combined to obtain the modified process analysis method.

2.2 Basic assumptions and calculation formulas of modified process analysis

In order to accurately grasp carbon emissions in the decision-making process of shield tunnel construction, a combination of input-output analysis and process analysis is used. The shield tunnel is divided into N standard sections, each section containing K standard rings. When calculating carbon emissions for the Jth standard section, one of the standard rings is taken as the calculation benchmark, and calculations are performed using both input-output analysis and process analysis. Since the calculation accuracy has been refined to one ring, theoretically, the two calculation results are not equal. Assuming that the accuracy of the process analysis method for fuel equipment can be accurately measured, and material consumption is basically the same, these two parts do not need to be corrected. Only the power consumption equipment needs to be corrected by obtaining the correction coefficient for carbon emissions based on a formula. Assuming that the carbon emission mode of the k-1 standard ring in the J interval is the same as that of the k ring, the carbon emission of the J interval can be obtained by summation, and the carbon emission of the whole shield tunnel can be obtained by summation again. Assume the consumption of fuel equipment and materials does not need to be corrected, only the carbon emission generated by power consumption equipment needs to be corrected. The equation of modified process analysis.

$$E^{RP} = \sum\limits_{N} {E_{J}^{RP} } \begin{array}{*{20}c} {} & {} \\ \end{array} \left( {J = 1 \cdot \cdot \cdot N} \right)$$

(1)

$$E_{J}^{RP} = k \cdot E_{s}^{RP}$$

(2)

$$E_{ES}^{I} = \eta_{S} \cdot E_{ES}^{P}$$

(3)

$$Q_{ES} \cdot e_{E} = \eta_{S} \cdot e_{E} \cdot \sum\limits_{j} {W_{Ej} \cdot } T_{j}$$

(4)

$$\eta_{S} = \frac{{Q_{ES} }}{{\sum\limits_{j} {W_{Ej} \cdot } T_{j} }}$$

(5)

$$E_{s}^{RP} = e_{D} \cdot \sum\limits_{S} {\beta_{DS} \cdot W_{DS} } + \eta_{J} \cdot e_{E} \cdot \sum\limits_{S} {\sum\limits_{j} {W_{ESj} } } \cdot T_{Sj} + \sum\limits_{S} {\sum\limits_{j} {q_{MSj} \cdot e_{Mj} } }$$

(6)

Where:

$E^{RP}$ is the total calculation that the carbon emission from shield tunnel construction.

$E_{J}^{RP}$: is the total calculation carbon emission from the standard shield subinterval

K: is the number of standard rings in standard interval that the calculate from the shield carbon emission

$E_{ES}^{I}$: is through the input-output method calculate to get carbon emission which the standard ring power equipment produce

$E_{ES}^{P}$: is through the process analysis calculate to get carbon emission which the standard ring power equipment pruduceη_S: is the modify coefficient of carbon emission that the standard ring power equipment

$Q_{ES}$: is the standard ring use the input-output method to calculate the energy consumption of power equipment

2.3 Carbon emission source of slurry shield

According to the characteristics of large-diameter slurry shield construction, the production process of construction materials is mainly carried out in the prefabricated field, the CO₂ emission at this stage is also mainly concentrated on the prefabricated site. The main emission sources at this stage are permanent construction materials such as cement and steel bars.

In the process of construction material transportation, carbon emission sources are mainly divided into three stages: building material production site transportation to prefabricated site, raw material transportation in prefabricated site, prefabricated components transportation from prefabricated site to construction site. The fuel equipment used in shield tunnel construction mainly refers to various vehicles transporting materials, including raw material transportation and prefabricated components transportation. The carbon emission produced by using fuel equipment is one of the sources of carbon emission in construction.

There are many kinds of carbon emission sources in the process of mechanical construction, and the construction technology of large diameter slurry shield is special, which involves many related processes, including shield tunneling and road construction, synchronous grouting, slurry circulation and many other auxiliary processes. Therefore, there are many kinds of emission sources in the process of mechanical construction. The shield construction process is mainly realized by shield tunneling, and the power consumption in the construction process is also one of the main emission sources of carbon emissions in shield tunnel construction.

2.4 Calculation boundary of carbon emission in shield tunnel construction

In order to accurately calculate the CO₂ emissions in the construction stage, it is also necessary to determine the calculation boundary, including the statistical calculation boundary of the engineering quantity in the construction stage and the calculation boundary of the main CO₂ emission sources.

The boundary of statistical calculation of engineering quantity includes the boundary of statistical calculation of the amount of building materials and mechanical equipment engineering quantity.

The calculation boundary of carbon emission source is analyzed from the use of construction material production, construction material transportation and construction equipment. Different construction materials are used in the construction stage of the project, and the dosage difference is also large. Cement, steel and sand aggregate are the three types of construction materials with a large amount of consumption, and CO₂ emissions are also large in the production process of three types.

The calcination process of clinker consumes a lot of energy. Therefore, the calculation boundary of cement carbon emissions mainly considers the production energy consumption in the cement factory and the CO₂ generated by the decomposition of limestone calcination process.

The calculation boundary of steel carbon emissions mainly considers the CO₂ produced by energy consumption in the production process of iron and steel smelting and rolling.

The CO₂ emission of sand aggregate mainly considers the CO₂ emission of sand or stone excavation, mining and screening process. For CO₂ emissions from the transport of sand and stone aggregates, which is included in numbers of mechanical equipment.

Concrete materials are used in abundance and consider the carbon emissions from concrete mixing. For CO₂ emissions from the transport of concrete, which is also included in numbers of mechanical equipment.

2.5 Determination of carbon emission coefficient in shield tunnel construction

When calculating the CO₂ emission coefficient of common energy sources, the CO₂ produced by energy combustion process is mainly considered, assume energy is fully burned, the specific conversion process is shown in Eq. (7).

$${\mathrm{e}}_{\mathrm{n}}={\mathrm{e}}_{\mathrm{nd}}{\mathrm{Q}}_{\mathrm{n}}$$

(7)

Where, ${\mathrm{e}}_{\mathrm{nd}}$ --is n^th energy source CO₂ emission factor (kg / kJ);

${\mathrm{Q}}_{\mathrm{n}}$—is the heating value of n^th energy source per unit mass (kJ / kg)

The CO₂ emission coefficient of commonly used energy-consuming working fluids is used to calculate, the CO₂ generated in the processing and production process of energy-consuming working fluids, which is determined according to the Eq. (8).

$$e_{{_{wl} }} = \sum\nolimits_{n = 1}^{N} {e_{n} q_{\ln } }$$

(8)

Where, e_n- is n^th energy source CO₂ emission coefficient (kg/kg) q_ln—is the consumption of type i energy in the production of type l energy-consuming working fluid with unit mass (kg)

Parameter factor of electric energy determined according to project location.

3 Carbon emission calculation of slurry shield tunnel construction

3.1 The basic situation of engineering

According to a tunnel project in Shanghai, this tunnel used large diameter slurry shield construction, and the length of tunnel is 5300 m. The starting point of the west tunnel is at WK15+091.500 m, and the end point is at WK11+700.010 m, with a total length of 3391.49 m. The starting point of the east tunnel is at EK15+088.590 m, and the end point is at EK11+700.037 m, with a total length of 3388.553 m. The outer diameter of the segment is 14,500 mm, and the inner diameter is 13,300 mm. Each ring is composed of 10 universal wedge-shaped segments, assembled by staggered joints. The thickness of the segment is 600 mm, and the width of the ring is 2000 mm. Tunnel buried sedimentary layers are mainly clay silt, silty clay, and silt layer. As in Fig. 1.

Fig. 1

Tunnel and segment lining

3.2 Carbon emission calculation results of shield tunnel construction

According to the calculation method and calculation formula mentioned above, the carbon emission from shield construction of a tunnel is calculated by using the modified process analysis method, and the calculation results are obtained in the following Table 1.

Table 1 Carbon emissions from tunnel construction (Unit: kg CO₂ / ring)

Table 1 is calculated based on a formula, with values calculated for each stage of the process. The production of building materials and material transformation occur within the same stage, resulting in consistent CO₂ emissions. The difference in carbon emissions arises during the material transportation stage and the tunnel construction stage, with the largest gap occurring during the latter. During the material transportation stage, the distance over which CO₂ is transported varies, resulting in differing amounts of CO₂ emissions. The carbon emissions during the tunnel construction stage are greatly influenced by the construction environment, such as the presence of underground water and the condition of surrounding rocks.

As shown in the Table 1, carbon emissions of one tunnel ring is composed of four parts: materials, material physicochemical process, in-field transportation of materials and shield construction. The carbon emissions of materials and material physicochemical process are almost constant. There are some changes in the field transportation of materials and shield construction. In particular, in the shield construction process, due to the influence of stratum conditions, tunneling distance, tunneling efficiency and others, the construction carbon emissions of each ring vary greatly, and the maximum difference reaches more than three times. The carbon emissions in the process of material production and preparation are the main emission sources, accounting for 93.8% of the total carbon emissions on average. Carbon emissions in the construction process account for only 6.2%.

3.3 Verification with machine learning model

This article uses BP neural network and LSTM for comparative prediction. During the operation of a machine learning model, there are input and output parameters. In this article, nine commonly used construction features in tunnels were selected as input parameters, and the output is the model’s predicted result. There is a certain degree of error between the predicted results and the actual carbon emissions. The five design parameters reflect the external environmental factors, while the four operational parameters reflect the internal environmental factors. The data is sourced from geological environmental investigation reports and daily monitoring system reports.

The learning process of BP network is composed of forward calculation process and error back propagation process. In the forward calculation process, the input is calculated layer by layer from the input layer through the hidden layer and transmitted to the output layer. The state of each layer of neurons will only affect the state of the next layer of neurons. However, if the value of the output layer does not reach the expected value, it will turn into the reverse propagation of the error until the output error reaches the acceptable range.

All the data are normalized, and all the sample data are randomly divided into training samples and test samples. There are 180 training samples and 30 test samples. The training samples are used to train the BP neural network, and the BP neural network obtained by the training samples is simulated and normalized. The average absolute error of the training samples is 0.157. As shown in Figs. 2 and 3.

Fig. 2

Error graph

Fig. 3

Prediction of carbon dioxide emissions

In addition to using the traditional BP neural network, the accuracy is verified by using the newer LSTM. Long short-term memory (LSTM) is a variant of RNN. RNN can only have short-term memory due to gradient disappearance, and LSTM networks combine short-term memory with long-term memory through subtle gate control and, to some extent, solve the problem of gradient disappearance to learn long-term dependent information.

The existing data is input into the neural network, including nine eigenvalues, and the output value is obtained. The loss is calculated according to the output value and the real value, and the loss is optimized by the optimizer. According to Fig. 4, the predicted results can be seen intuitively. The overall fitting degree is good and the error is within an acceptable range.

Fig. 4

LSTM’s combination of data fitting

Apply BP neural network and LSTM to predict the carbon dioxide emissions during the construction of shield tunnels, the carbon emissions of shield tunnels in the construction process of each ring section are quite different. Although BP neural network has strong nonlinear mapping ability and strong optimization computing ability, BP neural network has its own limitations. More training times lead to lower learning efficiency and lower speed. The choice of the number of neurons is also a key point. Too many neurons will cause the over-adaptability of the network, and too few neurons will cause the discomfort of the network. Through data prediction, the prediction results of lstm are more practical than the prediction of BP neural network, and the error is smaller. Although to a certain extent, LSTM solves the long-term dependence problem of RNN, due to the complex structure of LSTM itself, the training will be more time-consuming, the span will be larger, and the parallel processing data can not be better.

According to the prediction based on machine learning, it is found that there is little difference between the predicted results and the calculated results, which proves that it is feasible.

4 Discussion

4.1 Calculation results of carbon emissions in material physicochemical process

According to the calculation results and shown in Fig. 5, in the process of material materialization, the carbon emissions generated by the prefabricated segment account for the vast majority of the whole materialization process. The proportion of carbon emissions generated by the energy consumption of the pavement structure pouring on-site and the installation of the smoke plate in the materialization process is very small, which can be ignored. Therefore, the main carbon emission sources in the process of prefabricated segment should be considered.

Fig. 5

Carbon Emission Composition in Material Physicochemical Process of Shield Tunnel

4.2 Calculation of carbon emission during on-site transportation

During the shield construction process, various materials will produce carbon emissions when they are transported. The transportation of materials is mainly the transshipment of prefabricated components and the transportation of various materials in the tunnel. As shown in Fig. 6, the proportion of carbon emissions generated by the transshipment of segments reaches 95%. Some carbon emissions will also be generated, such as the transportation of flue plate and mouth parts. However, the carbon emissions generated by the transportation of smoke plate, mouth parts, segment under working well and synchronous slurry are small, the transit of segments should be optimized to reduce carbon emissions during on-site transport, such as combining prefabricated segments with construction sites where possible.

Fig. 6

Carbon emission composition of material transportation

4.3 Carbon emission calculation of materials in shield tunneling

The reinforced concrete such as tunnel segments and internal structure are the main components of the material. The carbon emission generated by using them accounts for 95.7% of the total carbon emission of the material. The contribution of slurry materials to the carbon emission of the whole ring tunnel material is particularly small. As shown in the Fig. 7, the segment accounts for the largest proportion of carbon emissions generated by materials in shield tunneling. The carbon emissions generated by segment bolts, synchronous grouting and flue plate are very small, which can be ignored. For reducing the carbon emissions generated by materials, some reasonable suggestions can be put forward for the segment.

Fig. 7

Carbon emission composition of material transportation

4.4 Calculation of carbon emission during shield tunneling

The shield propulsion process is characterized by high energy consumption, which is influenced by factors such as tunnel depth, stratum condition, and construction efficiency. As depicted in Fig. 8, the average carbon emissions resulting from shield tunneling account for 45% of the total emissions, with sludge transportation generating a relatively smaller amount. To analyze the carbon emissions associated with shield propulsion, the focus should be placed on the emissions generated by the energy consumption during the shield process.

Fig. 8

Carbon emission composition of shield tunneling

4.5 Impact of tunneling distance on carbon emissions

As shown in the Fig. 9, the carbon emission curve in the process of slurry transportation increases with the number of rings. It can be intuitively seen that the slurry transportation in shield tunneling has a significant increasing trend with the increase of propulsion distance. It can be seen that the change is relatively stable between 200 and 550 rings, and there is no obvious relationship. However, there is a mutation after 550 rings, and there is an upward trend after 600 rings, but the rise is relatively slow. The reason for the change is that a machine is added, resulting in an increase in carbon emissions. It can be seen that in terms of the slurry transportation process, with the advance of distance. Carbon emissions will increase with the increase, but there is no obvious regularity between the carbon emissions generated by slurry treatment, shield tunneling and ground auxiliary and the tunneling distance.

Fig. 9

Relationship between Carbon Emission and Distance of a 200–1000 Ring

4.6 Influence of shield depth on carbon emissions·

Referring to Fig. 10, the relationship between the total carbon emissions of tunnel construction and the depth at which the tunnel is buried is presented. Tunnel depth refers to the vertical distance between the top or side wall of a tunnel and the surface of the ground. Shield tunneling, which is an underground construction method, produces a certain amount of carbon emissions. It can be observed that the carbon emissions remain relatively stable between 34 and 36 m. Subsequently, the emissions begin to increase until reaching a peak at 39 m, before decreasing again. Beyond 39 m, the emissions exhibit an irregular pattern of increase, decrease, increase and decrease, with no clear trend. Consequently, it can be inferred that the depth at which the tunnel is buried does not significantly impact the total carbon emissions of shield construction. This is because an increase in tunnel depth can reduce the impact of construction on the surface environment, such as noise and vibration. At the same time, deeper tunnel depth can also improve construction safety, reduce construction interruptions and rework caused by geological disasters and other issues, and thus reduce the total carbon emissions.

Fig. 10

Relationship between Propulsion and Buried Depth of a 200–1000 Ring

Therefore, in shield tunneling construction, tunnel depth and other factors that affect the environment and construction should be considered comprehensively to develop a reasonable construction plan and achieve the goal of reducing carbon emissions.

4.7 Effect of shield propulsion efficiency on shield carbon emission

There is a certain relationship between the propulsion efficiency and the construction carbon emissions. In order to maintain the operation of the whole system, no matter whether the shield is advancing or not, there is always a certain amount of “basic emissions” in the whole shield tunneling system. As shown in the Fig. 11, with the increase in the number of daily propulsion rings, the average carbon emissions of the ring of the shield construction show an exponential downward trend. In other words, from this perspective, improving the propulsion efficiency of the shield is very effective to reduce the total carbon emissions. In order to avoid energy waste, we should try to avoid the abnormal pause of the shield machine, and the higher the daily progress, the smaller the average carbon emissions of the shield ring.

Fig. 11

Relationship between average daily propulsion cycles and carbon emissions

4.8 Influence of stratum conditions on carbon emission of shield construction

As shown in the Fig. 12, the average carbon emission of the four rings of shield tunneling in different strata. Figure a is carbon emissions from shield construction in sandy silt, figure b is carbon emissions from clay silt, figure c is carbon emissions from silt Compared with clayey silt and sandy silt, the carbon emission generated by shield tunneling in silty sand layer increases by 100%. The influence of different stratum conditions on the carbon emission of shield construction is different. The harder the soil layer is, the harder the shield travels in the soil layer, the more energy is consumed, and the more carbon emission is generated. When the soil strength is weakened, the energy consumption of shield tunneling and slurry system is reduced, and the carbon emission is also reduced. For reducing carbon emissions, in the stage of tunnel site selection, the buried length of the tunnel in the extremely hard stratum condition of silty sand layer is minimized.

Fig. 12

Carbon emission of shield construction in different strata

4.9 Proposals for emission reductions

• 1. On the basis of meeting the bearing capacity requirements, reducing the thickness of tunnel segments as much as possible is very effective to reduce the overall carbon emission of the tunnel.

• 2. The combination of prefabricated segment site and construction site can reduce the carbon emissions generated in the transportation process.

• 3. On the premise of maintaining the interface size of the segment unchanged, the reinforcement can be optimized and the steel content and reinforcement ratio of the segment can be reduced, which is beneficial to reduce the carbon emission in the process of material use.

• 4. On the site selection of the tunnel, as far as possible to avoid long-distance tunnel buried in silt layer, under conditions, as far as possible the tunnel buried in shallow silty clay layer, can effectively reduce the resistance of shield construction, reduce carbon emissions shield construction.

• 5. To improve the shield propulsion efficiency and avoid the long-term pause of the shield machine, under certain conditions, the daily average progress of the shield should be at least guaranteed to be greater than 3 rings.

5 Conclusions

Green and low-carbon tunnels are an important component of urban green transportation and are related to the modernization and sustainable development of the urban transportation industry. The modified process analysis method was used to calculate the carbon dioxide emissions of each stage of shield tunnel construction, combining machine learning with carbon emission prediction and the influencing factors of carbon dioxide emissions during shield tunnel construction were studied.

The modified process analysis method can help control carbon emissions during tunnel construction and achieve accurate calculations. Four points of analysis and corresponding recommendations are:

• 1. The segment thickness directly affects carbon emissions, so reducing reinforcement rates will effectively reduce emissions.

• 2. Prefabrication of segments generates the majority of emissions during materialization, but switching to electric boilers or other energy-saving methods can significantly reduce emissions.

• 3. Segment transfer generates the bulk of emissions during material transportation, which can be reduced by combining the prefabricated segment site with the construction site.

• 4. Material production generates most carbon emissions, and soil hardness affects emissions.

• 5. These findings emphasize the importance of reducing carbon emissions in tunnel construction and adopting sustainable practices to mitigate environmental impact.