TL;DR
This paper introduces the CNU-TS-1, a novel mobile tunnel monitoring system that uses a combination of laser scanning and odometry to efficiently capture high-density, orthogonal cross-sections of tunnels. This system significantly reduces costs and improves efficiency compared to traditional methods, enabling rapid and accurate deformation detection for tunnel maintenance.
What Problem Does This Paper Address?
The rapid expansion of railway and urban transit infrastructure in China has led to a substantial increase in the number and length of tunnels. Ensuring the safety and usability of these tunnels requires regular inspection and maintenance, particularly monitoring for structural deformation. Traditional methods like total stations and terrestrial laser scanners are either inefficient, low-density, or require complex multi-station setups and stitching, making them unsuitable for the demanding environment of tunnels (poor lighting, lack of GPS). There's a clear need for a rapid, precise, and automated system that can efficiently capture detailed tunnel geometry for deformation analysis.
Methods
The core of the CNU-TS-1 system is its hardware integration and software development.
- Hardware: The system is built on a trolley that carries a Faro Focus3D 120 laser scanner, two odometers (ZSP3806-003), a displacement sensor (GH type MTS), a notebook computer, and a power supply. The laser scanner is operated in a "helical mode" (2D profiling) to capture cross-sections. Two odometers are used to improve mileage accuracy. The rigid structure of the trolley ensures that the captured cross-sections are orthogonal to the tunnel axis.
- Software: Custom software was developed for data acquisition and processing.
- Data Acquisition: This module controls the sensors, fuses their data, and records it. It ensures synchronized operation and aims for a uniform walking speed to achieve a high density of cross-sectional points (1-2 cm interval along the tunnel, 2-3 mm density within a section).
- Data Processing: This module fuses multi-sensor data, filters noise, generates cross-sections, and analyzes deformation. A key feature is the correction of mileage data using longitudinal joints of tunnel segments, which significantly reduces system cost by eliminating the need for absolute positioning.
Datasets
The paper describes experimental validation using data collected from a recently constructed subway tunnel in Chengdu, China. This tunnel has a standard gauge of 1435 mm and an interior diameter of 5400 mm. The authors collected data over a 100 m section of this tunnel. They also performed a simulation experiment using an artificial deformation sign.
Key Insights and Ideas
- Cost Reduction through Relative Positioning: By using odometers for initial positioning and then correcting mileage with longitudinal tunnel segment joints, the system bypasses the need for expensive absolute positioning systems (like DGPS or IMUs), making it significantly more cost-effective.
- Orthogonal Cross-Sections by Design: The rigid trolley design inherently captures cross-sections orthogonal to the tunnel axis, eliminating the need for complex post-processing algorithms to orient the data.
- High-Density Data Acquisition: The system achieves a very high density of cross-sectional points (1-2 cm interval along the tunnel), providing a much more detailed view of the tunnel's geometry than traditional methods.
- Efficiency Gains: The mobile, automated nature of the system leads to substantial time savings compared to traditional methods like total stations.
Key Results
- Efficiency: The CNU-TS-1 system is reported to be 15 times more efficient than a total station for tunnel monitoring. Measuring a 100m tunnel section took approximately 50 minutes with the CNU-TS-1, compared to an estimated 750 minutes with a total station.
- Measurement Accuracy:
- The measurement accuracy of TS1 for cross-sections is stated to be 1.1 mm.
- External coincidence accuracy (comparison with total station for convergent diameter) showed deviations between -3.1 mm and +2.2 mm, with absolute mean deviations of 1.6 mm (direct) and 1.4 mm (reverse). This meets the "Standard for monitoring and measurement of urban rail transit structure" requirement of ±3 mm.
- Internal coincidence accuracy (stability of measurements) showed an average deviation of 0.6 mm and a standard deviation of 0.3 mm, indicating high stability.
- Point Density: The system achieves a density of cross-sectional points of 2-3 mm and an interval between adjacent cross-sections of 1-2 cm.
Limitations and Future Work
- Tunnel Type Limitation: The current system is only suitable for tunnels with a standard track gauge of 1435 mm and is not applicable to other tunnel types.
- Odometer Accuracy: While corrected, odometer mileage data still has accumulated errors, and location accuracy needs further improvement.
- Absolute vs. Relative Accuracy: The current verification focuses on relative deformation accuracy, which is practical for operational tunnels but doesn't fully address absolute point location accuracy.
- Future Work: The authors plan to develop a second generation of the TS1 with electric automation and modular processing. They also aim to develop algorithms to transform dynamic cross-sections into an absolute coordinate system to restore the true tunnel line type and provide more comprehensive data for detection.
Summary updated