openLAB

Table of contents

1. 1. Motivation
   2. Research consortium and funding
   3. openLAB - A research bridge in Lusatia
      1. Test concept
      2. Structural Health Monitoring System
      3. Evaluation concepts
      4. As-maintained model
   4. Data access and license agreements
   5. Cooperation opportunities
   6. Contact
   7. Selected publications

Motivation

The goal of the IDA-KI (Infrastructure Data Analysis with Artificial Intelligence) research project is to develop methods for automated analysis and evaluation of monitoring data. The validation is carried out using real measurement data acquired on a 45 m long research bridge (openLAB). First, the reference condition of the bridge is measured with and without service loads over a period of one year. Subsequently, targeted load tests will be performed up to the severe damage state. At the same time, redundant sensors will be damaged or manipulated to create a real database with specific signal characteristics for metrological and structural anomalies. This will provide an important basis for the development of evaluation methods (also based on machine learning), so that measurement faults in monitoring data can be corrected in the future and differentiated from structural damage.

Research consortium and funding

The research consortium is made up of an equal number of partners from practice and research:

Technische Universität Dresden
Institut für Massivbau
August-Bebel-Str. 30/30a01062 Dresden

Technische Universität Hamburg
Institut für Digitales und Autonomes Bauen
Blohmstraße 1521079 Hamburg

MKP GmbH
Zum Hospitalgraben
299425 Weimar

Hentschke Bau GmbH
Zeppelinstraße 15
02625 Bautzen

The research project IDA-KI (Automated assessment of monitoring data for infrastructure constructions using AI and IoT) is funded by the Federal Ministry for Digital and Transport, Germany, within the innovation program mFUND (funding reference: 19FS2013A).

openLAB - A research bridge in Lusatia

The research bridge, known as openLAB, is a prestressed concrete bridge approximately 45 m long and 4.5 m wide, which is being built on the site of Hentschke Bau GmbH in Bautzen, Germany. The bridge provides a wide range of common construction types. Each of the three fields has its own research focus, see Figure 1. The openLAB will be used to test the suitability of the evaluation algorithms to be developed and to enable their practical transfer.

Fields 1 and 2 each consist of three precast elements (PE) with a T-beam cross-section, supplemented with cast-in-place concrete. Field 1 contains typical damage and damage mechanisms of early prestressed concrete constructions: Coupling joint problems in PE 1.1; stress corrosion cracking in PE 1.2; areas of reduced shear capacity in PE 1.3. In addition, gravel pockets and blowholes are present for diagnostic investigation. Further damage, e.g., to the tendons, is planned for a later stage.

Field 2 is built according to the current state of the art. A special feature is located in PE 2.1, in which a so-called "smart tendon" is installed – a cooperation with the mFUND project of the same name. Distributed fiber optic sensors (DFOS) are integrated into the smart tendon to measure the strain distribution along the tendon and thus detect any damage to the structure at an early stage.

In field 3, PE of a rapid construction system without in-situ concrete additions is provided, which is fully loadable immediately after installation. Load distribution in the transverse direction of the bridge is achieved by casting reinforced joints.

Further structural features, such as a movable bearing of the columns (axis 20), a monolithic connection between the superstructure and the columns (axis 20), and a mineral fiber-reinforced transition (axis 30), are highlighted in Figure 1. Details can be found in a corresponding publication describing the project idea, the bridge construction and the monitoring system [1].

In order to achieve relevant loading conditions on the structure up to the ultimate limit state (ULS) with a reasonable test effort, only 25% of the load model (LM) 1 according to DIN EN 1991-2 was used for the design.

Figure 2 shows the construction progress in December 2023. Completion of the openLAB is planned for early 2024.

Test concept

After completion of the bridge, the reference condition of the structure is first measured over a period of one year under climatic conditions and "traffic" (hereinafter also referred to as the "reference phase"). To simulate the effects of traffic, a rail-guided load vehicle is driven over the bridge several times a month. Additional loading runs are made on particularly hot and cold days.

After the one-year reference phase, the structure is ready for static and dynamic load tests. In the static tests, the force is applied using hydraulic jacks and a load traverse. In addition to the dynamic effects of the load vehicle, the structure can also be excited with a straightening exciter to study the dynamic behavior in different frequency ranges.

Structural Health Monitoring System

A comprehensive monitoring system is installed at openLAB to monitor the structural behavior and to enable reliable damage detection. The monitoring system initially installed at openLAB is described below.

During the reference phase, an accelerometer is installed in the center of each PE axis in fields 1 and 2, and an inclinometer is installed in the point of zero moment near axis 20 (both on the outside of the web), see Figure 3. The tests are used to analyze how different types of damage (e.g., partial cutting of the prestressing) affect the modal parameters or the deformation figure.

Sensors are also used to measure environmental conditions, in this case air temperature, relative humidity (RH), solar radiation, and precipitation. Strain gauges, displacement transducers, and component temperature sensors are used to monitor local effects.

DFOS offer great potential for structural health monitoring due to their ability to perform distributed measurements with high spatial resolution [1,2]. Robust DFOS were installed before concreting in the precast elements and in the in-situ addition to allow distributed strain and/or temperature measurements with high spatial resolution from "hour zero". With a total length of about 1 km, the DFOS create an artificial nervous system of the bridge, see Figure 4.

The exact arrangement of the sensors, including semantic information about the sensor type, can be taken from the as-maintained model. In general, the configuration of externally attached sensors can be temporarily adapted to the task at hand, e.g., during a load test.

Evaluation concepts

Automated evaluation and assessment of the large amounts of data is the most important prerequisite for near real-time condition assessment. In the project, the free software fosanalysis is developed for the DFOS measurement data [3,4]. It allows to correct anomalies in the measurement signal and to monitor cracks along the sensor.

For anomaly detection in the measurement data of "conventional" measurement technology, a fault diagnosis approach based on analytical redundancy is developed. Fault diagnosis consists of the subtasks (i) fault detection, (ii) fault isolation, (iii) fault identification and (iv) fault elimination [5,6].

It is planned to expand the monitoring system with additional measurement technology from external project participants to evaluate the potential of different sensor types for anomaly detection. All monitoring data will be integrated into the evaluation processes.

As-maintained model

All relevant information from the design, construction and operation of the bridge is provided centrally via an as-maintained model, see Figure 5. The range of stored data extends from planning documents and information on the monitoring system to recorded damage and its assessment, as well as sensor raw data. The as-maintained model thus serves as a virtual image of the real structure, allowing the current condition of the structure to be assessed of the basis of aggregated condition information from monitoring, structural inspections and diagnostic examinations, and thus providing a decision-making aid for predictive maintenance.

External partners are given read access via an EPLASS InfoClient. Further information can be found in the following instructions:

PDF-Download: Anleitung Nutzung EPLASS InfoClient

Data access and license agreements

The following measurement data is made available via the research data repository OpARA (Open Access Repository and Archive):

• Measurements in the reference phase
• Measurements during selected load tests (e.g., also for system identification)

The link to the data is provided via this website.

Data may be used under the following conditions:

• The data is used for non-commercial research purposes.
• The origin of the data must be stated in publications. Example of acknowledgements: "The measurement data were obtained at openLAB and provided by the Technische Universität Dresden as part of the IDA-KI research project (FKZ 19FS2013A-D)." The DOI of the data set must be specified.
• Relevant publications from the IDA-KI project (see below) must be cited.
• The IDA-KI research consortium must be informed of publications using the data provided at the time of publication at the latest.

• If own sensors are installed at openLAB, the measurement data must be made available to the consortium and the public (if necessary after a short latency period). The type and scope of data provision will be detailed in consultation with the consortium leader.

Cooperation opportunities

The large-scale demonstrator will also be made available to external research groups who wish to test and validate sensor and monitoring techniques. Collaboration with national and international industry partners and research institutions at openLAB is encouraged by both mFUND and the project partners. Interested parties are invited to contact the consortium lead or the participating project partners.

For example, there is active cooperation with the monitoring provider Statotest, which has supplemented the existing monitoring system with own measurement technology. A total of eight measuring boxes are used to dynamically record accelerations and inclinations. The data complements the measurements from the initial monitoring system and provides an opportunity to compare different sensor types and measurement principles in terms of their potential for anomaly detection. Statotest uses the monitoring data from the damaging load tests to verify and further develop existing evaluation procedures.

Contact

Selected publications

• Herbers, M.; Bartels, J.-H.; Richter, B.; Collin, F.; Ulbrich, L.; Al-Zuriqat, T.; Chillón Geck, C.; Naraniecki, H.; Hahn, O.; Jesse, F.; Smarsly, K.; Marx, S. (2024) openLAB – Eine Forschungsbrücke zur Entwicklung eines digitalen Brückenzwillings. Beton- und Stahlbetonbau. https://doi.org/10.1002/best.202300094
• Herbers, M.; Richter, B.; Gebauer, D.; Classen, M.; Marx, S. Crack Monitoring on Concrete Structures – Comparison of Various Distributed Fiber Optic Sensors with Digital Image Correlation Method. Structural Concrete. 2023, 24(5):6123–40. https://doi.org/10.1002/suco.202300062
• Richter, B.; Herbers, M.; Marx, S. (2023) Crack Monitoring on Concrete Structures with Distributed Fiber Optic Sensors – Toward Automated Data Evaluation and Assessment. Structural Concrete. https://doi.org/10.1002/suco.202300100
• fosanalyis – A framework to evaluate distributed fiber optic sensor data [Software]. https://github.com/TUD-IMB/fosanalysis.
• Al-Zuriqat, T.; Chillón Geck, C.; Dragos, K.; Smarsly, K. Adaptive Fault Diagnosis for Simultaneous Sensor Faults in Structural Health Monitoring Systems. Infrastructures. 2023; 8(3):39. https://doi.org/10.3390/infrastructures8030039
• Al-Zuriqat, T.; Peralta, P.; Chillón Geck, C.; Dragos, K.; Smarsly, K. (2023) Implementation and validation of a low-cost IoT-enabled shake table system. In: International Workshop on Structural Health Monitoring (IWSHM). Farhangdoust, S.; Guemes, A.; Chang, F.-K. [Hrsg.], Stanford: DEStech Publications, Inc., S. 1063–1070.