i-WASP—Intelligent and Wireless Technologies for Automatically Generated Shift Protocols in Conventional Tunnelling
Authors:
Clemens Shaw, Marlene Villeneuve, Fabian Dengg, Pascal Gruber, Christian Thienert, Christoph Klaproth, Lara Gutberlet, Claudia Ungers, Harald Bittmann, Stefan Auderer, Lukas Staggl
The research project i‑WASP, a collaboration of specialized German and Austrian companies and research institutions, looks at the question of how complex processes on a conventional tunneling construction site can be automatically recorded, analyzed, and classified through the development of innovative solutions for data acquisition, transmission, and processing. A trained artificial intelligence will be able to automatically produce accurate minute diagrams based on accelerometer and positional data gathered from the construction machines.
Notes
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
1 Introduction
Minute diagrams in conventional tunnelling visually represent the chronological sequence of construction processes (Fig. 1). The time and material costs for individual subprocesses can be clearly represented and easily understood. These diagrams are crucial for identifying and evaluating opportunities for optimization, which can significantly reduce construction time and costs, contributing to a more economically and environmentally sustainable construction model. They also provide valuable insights into machine performance (e.g. Which machine breaks down most frequently? Which achieves the highest performance?) and the construction cycle (e.g. Which processes are the most time-critical? How can the overall cycle be accelerated?).
However, minute diagrams are often completed manually, at a considerable time delay, and may not accurately reflect the actual construction cycle. Automating the generation of these diagrams in real time would not only reduce the workload but also deliver much more precise data for analysis.
Advertisement
To accomplish this, the project i‑WASP is tasked with the development of a system consisting of a wireless data transmission technology appropriate for underground conditions, an underground localization technology for determining the position of construction machinery, accelerometer sensor technology for recording machine activity, and an AI-algorithm for the reliable processing and classification of data.
2 Concept i-WASP
In conventional tunnelling, various machines are employed to perform each step of the cyclical process (Fig. 2). Given that these machines often differ in age and come from different manufacturers, it is essential to develop a system that is universally applicable and operates independently of each machine’s own control system.
The recognition and classification of activities rely on two primary data sets: vibrational signals and the machine’s position in the tunnel.
In the i‑Wasp concept, externally mounted accelerometers record the unique vibrational patterns generated by a machine during specific processes. This data is continuously fed into a trained artificial intelligence algorithm that classifies the construction process based on these distinct signals. To enhance the accuracy of process recognition, additional factors can be considered, such as which other machines are operating simultaneously, what process occurred immediately before, and the location of the active machine.
Advertisement
A key innovation for i‑WASP, and for the tunnelling industry as a whole, is the development and implementation of a localization technology capable of functioning underground (similar to GNSS, Global Navigation Satellite Systems). Knowing the location of every machine in the tunnel will help understand which process a particular machine is engaged in and will also provide valuable data for logistical purposes. Alongside this is the need for a robust and reliable data transmission system, which will serve as the central nervous system, transporting data out of the tunnel and enabling wireless connectivity for the sensors.
Given that no two conventional tunnelling sites are identical, the research and development process must emphasize flexibility, modularity, and ease of installation. These factors are crucial to ensuring that the final product is adaptable, practical, and marketable.
3 Project Partners
For the success of the project, experts have been gathered from all relevant fields. The i‑WASP project partners bring expertise in tunnel construction, communication technology, electrical engineering, digitalisation, and computer- and data science. The companies and research institutes all operate within the tunnelling industry and therefore understand the difficulties and complexities of subsurface engineering.
eguana GmbH: Founded in 2015, eguana develops, innovates, and integrates digitalisation solutions for specialist civil engineering and tunnelling construction projects. A deep understanding of the construction processes helps them design software and hardware for data collection and management, which is both robust, state of the art, and enables adaptable tailor-made solutions. Within the project i‑WASP, eguana is responsible for developing the accelerometer hardware as well as the relevant software for the preliminary data collection and management (Fig. 3).
BEMO Tunnelling GmbH: Based in Innsbruck, BEMO Tunnelling GmbH is a construction company with a long-standing history in the tunnelling industry. The company focuses on developing innovative solutions to complex underground construction challenges, covering a wide range of projects from long TBM tunnels to large conventionally excavated caverns. BEMO is involved in some of the most significant and complex projects in Europe. Their construction sites will be used by i‑WASP for field-testing systems and collecting real-world data. BEMO’s expertise and feedback will be essential in developing a marketable product (Fig. 4).
Studiengesellschaft für Tunnel und Verkehrsanlagen e.V. (STUVA): A prominent figure in the tunnelling sector, STUVA is dedicated to advancing underground construction through research and strategic networking. Leveraging a vast network of national and international collaborations, STUVA focuses on a broad spectrum of topics related to underground construction and urban transport infrastructure. For the i‑WASP project, STUVA will be responsible for developing the AI algorithm that classifies construction processes using data from accelerometers and machine positioning systems (Fig. 5).
Bechler Kommunikationstechnik GmbH: The communication and safety system HADES, developed by Bechler, is a fully-integrated and specialized system designed specifically for mechanised TBM tunneling. Its high-speed data networks are the central nervous system of a tunnel, connecting all areas and fields of the construction site. For conventional tunnelling Bechler is looking to develop an underground positioning technology, similar to GNSS. This technology coupled with their data communication expertise will be Bechler’s important contribution to project i‑WASP (Fig. 6).
Montanuniversität Leoben—Chair of Subsurface Engineering: The Chair of Subsurface Engineering at Montanuniversität Leoben stands as a leading academic and research institution specialising in geotechnics and underground construction. Under its initiative and management is the unique tunnelling research facility, Zentrum am Berg (ZaB), located in the scenic Austrian mountains. ZaB serves as a life-size tunnelling laboratory, providing a rare opportunity to conduct experiments and training in a controlled, yet highly realistic environment. For the i‑WASP project, ZaB offers an invaluable setting for preliminary testing. Additionally, the Chair of Subsurface Engineering will contribute by aiding in the collection of on-site data (Fig. 7).
The project kicked off in February 2024 with a meeting of all partners in Ettlingen, Germany (Fig. 8). This gave a chance for all to introduce and get to know each other and outline each other’s roles for the coming months. A construction site visit was also organised to give an idea of where this developed technology will be used and what its requirements will be.
Preliminary testing began in March 2024. These were mainly functional tests of the accelerometer prototypes developed by eguana. Three sensors were mounted onto various construction machinery at ZaB using strong magnets and connected to/controlled wirelessly from eguana’s headquarters in Vienna (130 km away) (Fig. 9). The success of the magnetic mounting system, data collection, and wireless data transfer paved the way for the second round of testing taking place on an active BeMo construction site.
For the second tests in May, the same three sensors were taken to the construction site at Angath in Tyrol. The aims of this second round of tests were many: How would the sensors function in a harsh tunnelling environment? Does the orientation and position of the sensor on the machine make a considerable difference to the collected data? Can reliable ‘labelled’ data be collected for training an AI? In order to answer these questions, ten days were spent collecting data from two machines; the drill-jumbo and the shotcrete machine.
All three sensors were mounted onto one machine simultaneously in three perpendicular planes, in this way it could be checked if the orientation of the sensor had any effect on the data (Fig. 10). The connection of each sensor to the WiFi in the tunnel meant that the state of the sensors could be checked and data downloaded at will. Whilst the machine was performing a particular construction process, its activity was logged manually. This activity log would later be overlaid onto the collected accelerometer signal to produce the labelled data necessary for AI training (Fig. 11).
Around the same time, Bechler began testing for its data transmission and underground positioning system. Again these preliminary tests were performed at ZaB, providing Bechler with an ideal and controlled underground environment. Varying WiFi bands (2.4 GHz and 5 GHz) were tested to evaluate their performance in a life-size tunnel system, taking into account the effects of reflections off the tunnel walls, diffraction around obstacles, and tunnel cross-sectional size.
Two positioning technologies were also tested. Here, importance was given to understanding not only the technical accuracy of each system but also its technical limitations. In order to achieve this, Bechler developed a simple experiment, first for the positioning of a human being and subsequently the positioning of a moving machine (Fig. 12). The results from each of the two technologies could then be compared.
The third and most recent test brought the accelerometers back to the construction site. This time six of them and in a new iteration. Eguana’s new sensors now had the capability of being connected directly to the machines’ power supply and therefore not having to run off batteries (Fig. 13). The benefits of this development are twofold: removing the hassle of charging and changing batteries, and ensuring that no unnecessary data was recorded (as the sensors were only turned on when the machine was also turned on).
Fig. 13
Wired accelerometer sensor connected to machine’s power supply
These six sensors, unlike in the previous test, were installed one per machine on six machines simultaneously (drill-jumbo, shotcrete machine, tunnel excavator, ITC-excavator, lifting platform, and dumper). The objective was to collect as much data from the whole construction cycle as possible. Again activity logs were manually taken to provide labelled data for the AI training. As with all previous tests the sensors functioned perfectly, their state and the recorded data could be wirelessly checked and downloaded in real time by eguana in Vienna.
5 i-WASP Comes Next?
A second project meeting over two days in Vienna at the end of August 2024 brought together again all project partners to discuss the results of the past months and plan further steps. On the first day, presentations were made by each partner summarizing their progress. Discussions were held to clarify what the results of the tests meant for the taking of future decisions.
For example, it is clear with the naked eye that, from the accelerometer signals, a machine process classification is possible. However, the questions remain; can an AI be trained to come to the same conclusions? What are the defining characteristics of each individual signal and how can they be programmed into an AI? These are questions that STUVA has taken upon themselves for the coming months.
For the underground positioning system, a decision was made with which of the two technologies to continue researching. The basic technical aspects of this system have, for the most part, been explored. Now comes the difficultly of figuring out how this system can be practically implemented on an active construction site. This is the next step for project partners BeMo and Bechler.
For eguana and the Chair of Subsurface Engineering MUL the goal remains to continue collecting valuable accurately labelled data for AI training.
As it stands, the project is progressing on the right track. Exciting and challenging times are ahead for project i‑WASP.
Funding
These results were compiled as part of the research and development project ‘i-WASP: Intelligent and wireless technologies for automatically generated shift protocols in conventional tunnelling’, funded from 2023 to 2026 by the German Federal Ministry of Economics and Climate Protection (BMWK) and by the Austrian Research Promotion Agency (FFG) under the umbrella of the IRASME initiative.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
i-WASP—Intelligent and Wireless Technologies for Automatically Generated Shift Protocols in Conventional Tunnelling
Authors
Clemens Shaw Marlene Villeneuve Fabian Dengg Pascal Gruber Christian Thienert Christoph Klaproth Lara Gutberlet Claudia Ungers Harald Bittmann Stefan Auderer Lukas Staggl
