Displacement Monitoring of Subway Tracks and Tunnels According to Adjacent Construction

Abstract

In the Republic of Korea, large-scale deep excavation construction is being conducted adjacent to structures owing to overpopulation in urban areas. Securing the safety of earth-retaining and underground structures is crucial for adjacent excavation work in urban areas. Accordingly, an automated measurement system is used to monitor the behaviors of subway tunnels and track structures; however, the utilization of its results is extremely low. Existing techniques evaluate the safety of track and tunnel structures based on the measured maximum values. This study introduces an evaluation technique for improving behavior monitoring in subway tunnels and track structures. A substantial amount of long-term measurement data on subway tunnels and track deformation was quantitatively evaluated using the Gaussian probability density function. In addition, the results from the same location where the tunnel convergence meter (TL) and track bed settlement (RM) sensors were located were compared. The comparison results revealed a difference in the vertical displacement of the tunnel and track structure owing to adjacent excavation work. A technique to analyze the continuous behavior of the tunnel was presented by numerically analyzing the tunnel measurement results as input data. In addition, they emphasized the need for simultaneous monitoring of tracks and tunnels.

1. Introduction

In the Republic of Korea, the demand for social infrastructure, such as high-rise buildings, public transportation, and convenience facilities, is increasing owing to population growth in metropolitan areas [1,2,3,4]. In addition, owing to overcrowding in urban areas, large-scale, deep excavation construction is being conducted adjacent to structures [1,2,3,4]. To ensure the safety of subway tunnels and track structures, an automated measurement system was introduced for real-time safety management. However, current monitoring results of the automated measurement system compare the maximum value with safety management standards within a specific period [1,2,3,4]. This method evaluates the safety of subway tunnels and track structures based on whether they exceed safety management standards. In addition, the utilization of measurement results is extremely low, and research on reasonable and reliable standards and evaluation techniques is insufficient.

In Korea, subway structure deformation occurs owing to ground deterioration, an increase in groundwater level, and adjacent excavation work.

Accordingly, Choi et al. [1] evaluated structural damage caused by the uplift and subsidence of subway box structures owing to changes in ground conditions. They presented a technique for determining damage through stress analysis of subway box structures using survey data. Choi et al. [2] numerically analyzed ground settlement behavior when passing through a tunnel in a geologically vulnerable section. Choi et al. [3] numerically analyzed the behavioral characteristics of concrete tracks laid on a subway box structure during subsidence and uplift. In addition, they analyzed the level of damage to the concrete ballast resulting from concrete track settlement and uplift. Choi et al. [4] analyzed track deformation caused by excavation work close to urban and high-speed railways. They modeled the construction stage, ground conditions, and gravel track and performed a precise three-dimensional (3D) numerical analysis.

Park and Lee [5] investigated the lining behavior following the excavation of the upper part of an operating underground structure using laser scanning. Bae et al. [6] presented a technique for evaluating tunnel cracks using a tunnel scanner. Feng et al. [7] experimentally analyzed the feasibility of measurement methods (TDP&S, TDPS) used for microseismic monitoring. Manuello et al. [8] measured damage to precast concrete tunnels in real time using a multichannel acoustic emission system. Miliziano et al. [9] predicted the impact of tunneling on Roman architecture through numerical analysis using the Mohr–Coulomb model. Zhou et al. [10] monitored subway tunnel displacement based on a robotic total station (RTS) using a prism as a reflector. The monitoring range of the RTS was derived through tunnel experiments, and the accuracy of the long-term monitoring was verified. Xue et al. [11] accurately and automatically calculated the leakage area of a shielded tunnel lining using a deep-learning model. Wang et al. [12] analyzed the effects of foundation and excavation technology through numerical analyses using field measurement. Qin et al. [13] presented an automatic recognition system for tunnel lining elements using deep-learning-based ground-penetrating radar images. The lining thicknesses measured using the corresponding technique and field test were found to be 95.45% consistent.

Choi et al. [14] presented a plan for improving safety evaluation techniques for subway tunnel structures using an existing automated measurement system. They presented a Gaussian probability density function analysis technique for quantitatively evaluating a substantial amount of measurement data. Choi et al. [15] analyzed the dynamic wheel load using track deviation measurement results from adjacent excavation work.

This study presents a technique for monitoring the behavior of subway tunnels and track structures following actual adjacent excavation work. A substantial amount of long-term measurement data was quantitatively evaluated using the Gaussian probability density function. The vertical and horizontal behaviors of subway tunnels are monitored using a tunnel convergence meter sensor. The vertical behavior of subway tracks is monitored using a trackbed settlement sensor. In this study, data were collected from sites where adjacent excavation work was performed for 18 months. Subsequently, the safety of subway tunnels and track structures was evaluated using the collected long-term measurement data. In addition, the analysis results of the track bed settlement sensor and the tunnel convergence meter sensor were compared. Since the track bed settlement measurement results are like the tunnel invert displacement, the differences with the tunnel lining were analyzed. In this study, tunnel analysis results were used as input data for numerical analysis. A technique for analyzing the continuous behavior of tunnels was presented using numerical analysis. It is expected that existing cross-sectional measurement results can be efficiently analyzed through this technique.

2. Displacement Monitoring of Tunnel Structures

2.1. Field Measurement

In this study, tunnel convergence meter sensors were installed 25 m before and after the train direction based on the center of the adjacent excavation work area. Subway tunnel structures may deform owing to adjacent excavation work. Safety is evaluated using the measurement results of a tunnel convergence meter sensor. Before the adjacent excavation work, a tunnel convergence meter sensor is installed in the subway tunnel structure to secure the initial data. Subsequently, the deformation of the tunnel structure owing to adjacent excavation work is measured, and safety is assessed through monitoring until the construction is completed. In this study, long-term monitoring was conducted for 18 months, with data collected at 60 min intervals. Figure 1 shows a full view of the tunnel convergence meter and trackbed settlement sensors installed in the subway structure.

In this study, tunnel convergence meters and track bed settlement sensors were installed in a subway tunnel and track structure, as shown in Figure 1.

Table 1. Sensor specifications.

Category	Tunnel Convergence Meter (Tape Extensometer)		Track Bed Settlement Monitoring Sensor
	Angle	Length	Angle	Length
Measurement range	±60	0~50 mm	±15°	0~50 mm
Applied voltage/total resistance	5~15 V direct current	10 kΩ	5~15 V DC	(5 kΩ) ± 20%
Linearity	0.05°	±1%	±0.1°	±1%
Sensitivity	0.001°	0.01% maximum	0.001°	0.001 mm
Operating temperature of sensors	−20~+80 °C	−20~+80 °C	−30~+65 °C	−20~+80 °C

lists the specifications of the installed sensors [14,16].

Figure 2 shows the installation locations of the adjacent excavation work area, tunnel convergence meter sensor, and track bed settlement sensor.

As shown in Figure 2, tunnel convergence meters and track bed settlement sensors were installed to monitor subway tunnels and track structures close to the adjacent excavation construction site. The tunnel convergence meter sensor was installed based on a total of three tunnel front sections, from Section A to Section C. A tunnel convergence meter sensor was installed on the tunnel front section to analyze the behavior of the tunnel structure in the vertical and horizontal directions. Sixty-eight track bed settlement sensors were installed in the direction of the train movement at approximately 140 m. To analyze the behavior of the track structure, a track bed settlement sensor was installed to monitor the behavior in the vertical direction along the train length. The 18-month monitoring covered up to Step 7 of the adjacent construction process, as shown in

Table 2. Adjacent construction steps and measured months.

Construction Step	Construction Condition	Monitoring Months
Step—0	General condition	0
Step—1	Retaining wall construction	1–4
Step—2	Stage 1 excavation	5–7
Step—3	Stage 2 excavation and raker construction	8–11
Step—4	Stage 3 excavation and slab construction	12–13
Step—5	Stage 4 excavation and slab construction	14–15
Step—6	Stage 5 excavation and slab construction	16
Step—7	Stage 6 excavation and slab construction	17–18

.

The site under investigation was under construction, and the analysis was performed using the measurement results up to Step 7 of Table 2.

2.2. Measurement Results

Figure 3 shows the measurement results for the first and 18th months of the adjacent excavation work. Six sensors were installed in the cross-section of each tunnel: TL-1, left wall; TL-2, left shoulder; TL-3, left crown; TL-4, right crown; TL-5, right shoulder; and TL-6, light wall.

The existing subway tunnel structure monitoring system simultaneously monitored data from all sensors installed onsite, as shown in Figure 3. However, visually evaluating the measurement results proved challenging. In addition, the tunnel structure safety evaluation results were numerically evaluated based on the maximum value during the measurement period. In this study, time-series data were analyzed based on sensor TL-2 in Section C, where the maximum tunnel displacement occurred. Figure 4 shows the time-series data from sensor TL-2 in Section C, considering the adjacent excavation work.

An analysis of the vertical displacement of the tunnel structure (Section C, sensor TL-2) due to the adjacent excavation work showed a vertical displacement of (−)0.49 mm for the first month of measurement, as shown in Figure 4. After 10 months of monitoring, the vertical displacement for 1 month was (−)1.64 mm. The vertical displacement decreased by approximately 2.35 times after 10 months of construction compared with that in the first month of monitoring. The vertical displacement for 1 month after 18 months of construction was (−)1.81 mm. The vertical displacement 18 months after construction increased by approximately 10.37% compared with that at 10 months after construction. Therefore, the vertical displacement of the tunnel structure decreased by approximately 1.33 mm owing to the adjacent excavation work.

2.3. Gaussian Analysis Results

In this study, a gaussian probability density function analysis was performed using data obtained after 1, 10, and 18 months of monitoring. The 1-month monitoring data were used from the beginning to the 1st month of measurement. The 10-month data were used from the 9th to the 10th month. The 18-month data were used from the 17th to the 18th month. Figure 5 shows an example of the Gaussian probability density function analysis results for each section using the monitoring data.

The Gaussian analysis for each section revealed that in Section A (TL-1), the vertical displacement increased as the adjacent excavation work progressed, as shown in Figure 5a. Compared with that during the initial stage of the adjacent excavation work, the average vertical displacement decreased by approximately 0.022 mm 18 months after construction. As shown in Figure 5b, in Section A (TL-2), the displacement increase was greater than that measured by the TL-1 sensor.

Compared with the initial stage of the adjacent excavation work, the average vertical displacement decreased by approximately 0.792 mm 18 months after construction. Therefore, as shown in Figure 5, tunnel deformation was affected during construction at a location adjacent to the subway. The change in the vertical displacement of Section C (TL-2) was the largest, and the vertical displacement resulting from the adjacent excavation work decreased by approximately 1.634 mm.

Table 3. Safety evaluation results of the tunnel structure (mm).

Section	Sensor	Average (mm)			Standard Deviation (mm)			Field Maintenance Criteria (mm)
		1 Month	10 Months	18 Months	1 Month	10 Months	18 Months
A	TL-1	0.002	0.001	(−)0.020	0.044	0.010	0.014	3.0
	TL-2	0.009	(−)0.246	(−)0.783	0.191	0.127	0.154
B	TL-1	(−)0.049	(−)0.137	(−)0.286	0.114	0.041	0.068
	TL-2	0.073	0.018	(−)0.036	0.067	0.051	0.044
C	TL-1	0.015	(−)0.033	(−)0.159	0.107	0.119	0.089
	TL-2	(−)0.143	(−)1.672	(−)1.777	0.123	0.029	0.078

lists the safety evaluation results for the subway tunnel structure [14,16,17,18,19,20,21,22].

The safety evaluation results of the subway tunnel structure, shown in Table 3, revealed that the minimum value considering the standard deviation for sensor TL-2 in Section A was approximately (−)0.937 mm. The management standard for subway tunnel structures is 3.0 mm. The vertical displacement in Section A was approximately 31.23% of the management standard value. Considering the standard deviation of sensor TL-1 in Section B, the minimum value was approximately (−)0.354 mm. The vertical displacement of Section B was approximately 11.8% of the management standard. Considering the standard deviation of sensor TL-2 in Section C, the minimum value was approximately (−)1.855 mm. The vertical displacement of Section C was approximately 61.8% of the management standard. Figure 6 shows the analysis results of the maximum vertical displacement according to the measurement month for each section.

An analysis of the maximum vertical displacement by section revealed that in Section A, the TL-1 sensor measurement was close to 0, as shown in Figure 6a. However, the vertical displacement of sensor TL-2 increased significantly during the 2^nd to 3^rd months of measurement. In addition, the trend decreased until the 9^th to 10^th months of construction and then increased until the 18^th month. In Section B, as shown in Figure 6b, the change in vertical displacement was more significant for sensor TL-1 than for sensor TL-2. In Section C, no significant change in the displacement was observed for sensor TL-1, as shown in Figure 6c. However, the displacement continued to increase until the 6^th month of measurement with sensor TL-2; nonetheless, the change in displacement until the 18^th month of measurement was not significant.

Therefore, for the tunnel structure, the deformation of Section B, the center of the excavation area, was not substantial. However, notable tunnel structure deformations were observed in Sections A and C, located near the start and end points of the excavation area. Subway tunnel structures appear to be more vulnerable as they move from the center of the excavation area to the start and end points.

3. Displacement Monitoring of Track Structures

3.1. Measurement Results

In this study, a track bed settlement sensor was installed at a total distance of 140 m in the train direction, including the adjacent excavation work area. Adjacent excavation work may deform subway track structures. Safety is evaluated using a track bed settlement sensor. Before adjacent excavation work, initial data are obtained by installing a track bed settlement sensor on the subway track structure. Subsequently, the track structure deformation owing to the adjacent excavation work is measured, and safety is assessed through monitoring until the construction is completed. In this study, long-term measurements were taken for 18 months, with data collected at 60 min intervals.

Figure 7 shows the measurements during the 1^st and 18^th months of track bed settlement sensor monitoring.

As shown in Figure 7, the existing subway track structure was monitored simultaneously using monitoring data from all sensors installed onsite. However, the track bed settlement sensor was installed in the direction of train movement, and determining the results at the exact location using existing monitoring techniques is challenging. In addition, the tunnel structure safety evaluation results were numerically evaluated based on the maximum value during the measurement period. Figure 8 shows the results of sensor RM-41 among the track bed settlement sensors monitored during the 18 months of construction.

The vertical displacement measurements of track bed settlement sensor RM-41 revealed a 0.34 mm maximum vertical displacement at the beginning of the adjacent excavation work, as shown in Figure 8. The maximum vertical displacement at 10 months and 1 month after construction was 0.48 mm. The maximum vertical displacement increased by approximately 41.17% 10 months after construction compared with that during the initial construction period. The maximum vertical displacement after 18 months of construction was 0.59 mm. The maximum vertical displacement increased by approximately 22.92% 18 months after construction compared with that at 10 months after construction. Therefore, the vertical displacement of the track structure owing to the adjacent excavation work was approximately 0.25 mm.

3.2. Gaussian Analysis Results

Figure 9 shows the Gaussian analysis results using track bed settlement sensors (RM-01, RM-11, RM-21, RM-31, RM-41, and RM-52) and monitoring data obtained 1, 10, and 18 months after the adjacent excavation work.

The Gaussian analysis results of the track bed settlement sensor, shown in Figure 9, revealed that the vertical displacements measured by sensors RM-01, RM-11, RM-21, and RM-31 decreased as the adjacent excavation work progressed. However, for RM-41 and RM-52, the vertical displacement increased owing to the adjacent excavation work. The average vertical displacement measured by sensor RM-41 increased by approximately 0.513 mm 18 months after construction compared with that in the initial stage of the adjacent construction.

Table 4. Safety evaluation results of the track structure (mm).

Sensor	Location (m)	Average (mm)			Standard Deviation (mm)			Field Maintenance Criteria (mm)
		1 Month	10 Months	18 Months	1 Month	10 Months	18 Months
RM-01	0	(−)0.064	0.033	(−)0.115	0.079	0.074	0.076	3.6
RM-11	20	(−)0.086	(−)0.126	(−)0.167	0.066	0.058	0.073
RM-21	40	(−)0.049	0.059	(−)0.122	0.072	0.069	0.073
RM-31	60	(−)0.077	0.017	(−)0.213	0.109	0.149	0.119
RM-41	80	0.045	0.479	0.558	0.153	0.163	0.126
RM-52	102	(−)0.200	0.116	0.113	0.105	0.079	0.078
RM-61	120	0.074	0.071	0.082	0.150	0.043	0.063

lists the safety evaluation results for the subway track structures. The Seoul Metropolitan Rapid Transit Corporation track maintenance standards were applied to the field maintenance criteria of the track bed settlement sensor [14,16,17,18,19,20,21,22].

Based on the subway track structure safety evaluation listed in Table 4, the maximum value considering the standard deviation of sensor RM-41 measurements was approximately 0.684 mm. The management standard for subway track structures is 3.6 mm. The vertical displacement at the location of sensor RM-41 was approximately 19% of the management standard. By contrast, the vertical displacements at the locations of sensors RM-01, RM-11, RM-21, and RM-31 were close to 0.

4. Analysis and Discussion

In this study, a tunnel convergence meter sensor was installed in a subway tunnel structure, and monitoring was performed for 18 months. Gaussian probability density function analysis was performed for quantitative evaluation using the measurement results from long-term monitoring. Figure 10 shows the maximum vertical displacement analysis results from Sections A–C.

The Gaussian analysis results for each section according to the adjacent excavation work revealed that in Section A, the adjacent excavation work did not significantly affect the measurements of sensors TL-1, TL-3, TL-4, and TL-6. By contrast, for sensors TL-2 and TL-5, the displacement increased rapidly in the early stages of the excavation work. Therefore, the locations of sensors TL-2 and TL-5 in Section A were directly affected during the adjacent excavation work. Vertical displacement was observed in Section B; however, the impact of the excavation work was not significant. In Section C, the largest displacement occurred at the location of sensor TL-2. However, except for the location of sensor TL-2, most locations exhibited displacement behavior in the opposite direction. Therefore, the tunnel structure was observed and analyzed to have a more significant tunnel displacement at the start and end points than at the center of the excavation area. The tunnel maximum vertical displacement results using Gaussian analysis of 1-month and 18-month data are shown in

Table 5. Max vertical displacement results of the tunnel structure (mm).

Sensor	1 Month (Xc + SD)			18 Months (Xc + SD)
	Section A	Section B	Section C	Section A	Section B	Section C
TL-1	(−)0.024	(−)0.187	(−)0.092	(−)0.034	(−)0.354	(−)0.248
TL-2	0.000	0.006	(−)0.266	(−)0.937	(−)0.080	(−)1.855
TL-3	(−)0.236	(−)0.116	(−)0.265	0.000	(−)0.124	(−)0.037
TL-4	(−)0.070	(−)0.272	(−)0.211	(−)0.034	(−)0.397	(−)0.207
TL-5	(−)0.253	(−)0.175	(−)0.029	(−)0.937	(−)0.146	0.807
TL-6	(−)0.079	(−)0.090	0.198	0.000	(−)0.124	0.414

.

In this study, a continuous data analysis in the tunnel direction was performed using the analysis results of the tunnel convergence meter sensor, as shown in Figure 11. Using the data in Table 5, the range of tunnel vertical displacement influence was analyzed through displacement control for each location [22].

As a result of the analysis of the tunnel vertical deformation influence range, a large subsidence was found at the TL-3 location of Section C, as shown in Figure 11b. On the other hand, upward displacement was observed in TL-6 of Section C. In addition, it was found that similar displacement behavior occurred in Figure 11d, which is the 18th month data. Existing evaluations use data at nodes, making continuous data analysis difficult. It is believed that the continuous displacement behavior of the tunnel can be identified by using the technique presented in this study.

In this study, a track bed settlement sensor was installed on a subway track structure to quantitatively evaluate the monitoring data. Figure 12 shows the vertical displacement analysis results for sensors RM-01 to RM-68.

The per location analysis according to the adjacent excavation work, shown in Figure 12, revealed a vertical displacement of approximately (−)0.181 mm approximately 30 m before the adjacent excavation work area. However, the vertical displacement resulting from the adjacent excavation work was approximately (−)0.424 mm, corresponding to a 0.243 mm reduction. A vertical displacement of approximately 0.045 mm was observed approximately 80 m before the adjacent excavation work area. However, the vertical displacement resulting from the adjacent excavation work was approximately 0.555 mm, corresponding to a 0.51 mm increase. Decreasing and increasing sections are typically observed when using track bed settlement sensors. Therefore, because changes appear simultaneously when monitoring subway track structures, a safety evaluation of the minimum and maximum values is necessary. In this study, the analysis results of the measurements obtained by the tunnel convergence meter and track bed settlement sensors installed at the same location were compared. The vertical displacement analysis results of the subway track and tunnel structure revealed a significant difference in the vertical displacement between the tunnel and track in Sections A and C. By contrast, Section B, the center of the adjacent excavation work area, had a larger tunnel vertical displacement than the track. The results revealed a difference in the vertical displacement of the tunnel and track owing to the adjacent excavation work. Thus, in subways, the behavior of tunnels and tracks may differ; therefore, track and tunnel monitoring should be performed simultaneously.

5. Conclusions

This study performed safety evaluations using the monitoring data of subway tracks and tunnel structures, considering adjacent excavation work. In subway tunnel structures, vertical and horizontal behaviors were monitored using a tunnel convergence meter sensor. In subway track structures, vertical behaviors were monitored using a track bed settlement sensor. Displacement monitoring data were collected from sites where adjacent excavations were performed for over 18 months. The long-term measurement data collected were quantitatively evaluated using the Gaussian probability density function. The primary research results are summarized as follows:

(1) Based on the subway tunnel structure safety evaluation, the maximum vertical displacement occurred in the tunnel lining in Section A. However, the adjacent excavation work did not significantly affect the tunnel crown. In Section B, displacement occurred from the left lining to the left ceiling owing to the adjacent excavation work. In addition, the tunnel lining was the most affected in Section A. In Section C, the left and right linings of the tunnel were most affected by the adjacent excavation work. Therefore, the study site was determined to have the most vulnerable tunnel lining owing to the adjacent excavation work.

(2) Based on the subway track structure safety evaluation, the largest vertical displacement was measured at approximately 18 m, where sensor RM-10 was installed at the initial (standard) value of the adjacent excavation work. However, the location of the maximum vertical displacement changed as the adjacent excavation work progressed. The vertical displacement was significantly reduced to approximately 28 m owing to the adjacent excavation work. In addition, the vertical displacement increased significantly at approximately 60 and 80 m from the track bed settlement sensor installation location. Therefore, the track bed settlement sensor was directly affected at the start and end points and the center of the adjacent excavation area. In addition, the track vertical displacement increased and decreased from the center of the adjacent excavation work area to the start and end points, respectively.

(3) A comparison of the vertical displacements of the subway track and tunnel structure revealed that the vertical displacement of the track was larger at the start and end points of the excavation work area than at the tunnel. However, at the center of the excavation work area, the vertical displacement of the tunnel was larger than that of the track. This could affect the tunnel structure rather than the track at the center of the excavation work area. In addition, the start and end points of the excavation work area directly affect the track structure rather than the tunnel. Therefore, when performing adjacent excavation work, subway tunnels and track structures must be monitored simultaneously. In addition, periodically evaluating the safety of track and tunnel structures using long-term monitoring results is beneficial.

The purpose of this study is to process and analyze massive amounts of data. In addition, tunnel analysis results were used as input data for numerical analysis. A technique for analyzing the continuous behavior of tunnels was presented using numerical analysis. The implications of the study’s findings extend beyond the immediate context of subway construction in the Republic of Korea. They offer valuable insights and recommendations for improving safety practices, optimizing monitoring systems, and guiding future urban development initiatives.