Modal analysis can offer insight into civil structures, which can drive the design of life-saving buildings that are resilient to major catastrophe. Unfortunately, the enormity of civil structures, like bridges and skyscrapers, means traditional, wired data loggers are simply not fit for this purpose.
The University of Exeter’s Vibration Engineering Section used LabVIEW and CompactRIO to develop a set of bespoke vibration data loggers that can be widely distributed across huge structures, whilst maintaining sub-microsecond synchronisation without long, logistically challenging cables, RF transmissions, or GPS antennae. This distributed system was successfully used to measure ambient vibrations of the Jiangyin Suspension Bridge in China to estimate its modal properties.
James Brownjohn - University of Exeter - College of Engineering, Mathematics and Physical Sciences
James Bassitt - University of Exeter
Karen Faulkner - University of Exeter
Zhen Sun - Jiangsu Transportation Institute
Yichen Zhu - University of Liverpool
NI Products Used: LabVIEW, LabVIEW FPGA Module, cRIO-9064, NI 9234, NI 9402, NI 9912, NI 9977
The University of Exeter’s Vibration Engineering Section (VES) is home to an exceptional team of researchers with an international reputation in vibration serviceability, structural health monitoring, and active vibration control, and their application to real world problems. We also offer opportunities for talented engineers to study as undergraduates or post-graduates. We use the NI platform for education and research.
Modal analysis is the form of vibration testing that engineers use to identify the modal properties (including natural frequencies, damping ratios, and mode shapes) of all kinds of objects, from cars to aircrafts to civil structures.
In particular, the design of tall buildings and long-span bridges benefit from reliable damping information. For tall buildings, it governs wind-induced sway that affects occupant comfort and the integrity of façade panels. For long-span bridges, damping guards against instability due to flutter. For footbridges, damping establishes the limiting number of pedestrians before excessive sway occurs (for example, the infamous London Millennium Bridge wobble).
Ambient vibration testing (AVT) may be the only practical means of providing the vibration response data (output) without knowing the excitation force (input), such as due to wind or human excitation. Hence, we use AVT output only analysis, better known as operational modal analysis (OMA).
Unfortunately, AVT of super tall and long structures is particularly challenging. Their enormous scale prohibits the use of conventional cabled test systems, while the ultra-low natural frequencies and cross-talk from quasi-static deformations require extreme accuracy and synchronisation.
Modal analysis of long and tall structures relies on the correct phase information between vibration data recorded at distant locations. As such, tight synchronisation between the distributed data loggers is vital. Even slight time/phase shifts corrupt the critical modal analysis process.
Prior to beginning our research, we evaluated off-the-shelf data loggers from many vendors. Few met our demanding requirements for sensor compatibility and signal fidelity, and none offered the scalability required for our future research. Also, the data loggers we evaluated were limited to using wireless communications (limits range and does not work if some loggers are inside the structure) or GPS (which does not meet our synchronisation requirements).
Our experience applying AVT for large civil structures, such as dams, bridges, and skyscrapers, strongly indicated the need for an entirely new type of distributed, tightly synchronised embedded data logger. Simply put, we had no option but to develop our own distributed system.
Our research team has domain expertise within the fields of structural health and active vibration control, but lacks experience withlow-level electronic or embedded design. This made the NI platform an obvious choice for building our novel distributed data logging system. LabVIEW delivers intuitive graphical programming, even down to the hardware level, which accelerates the development process. Also, the CompactRIO platform not only offers powerful computational capabilities on both embedded and FPGA processors, but a range of I/O modules deliver instant connectivity and signal conditioning for a wide variety of sensors.
For each of the data loggers, we housed a cRIO-9064 embedded controller within a waterproof Peli storm case. For vibration acquisition, we coupled each CompactRIO with an NI-9234 vibration input module. This module’s wide dynamic range and 24-bit A/D converter ensures measurement accuracy, while its integrated IEPE signal conditioning simplifies sensor connectivity.
As previously highlighted, tight-synchronisation between the distributed data loggers is critical to the correlation of the data. To achieve this, we boxed each CompactRIO with an oven-controlled crystal oscillator (OCXO). This type of oscillator achieves high-frequency stability by precisely maintaining the quartz crystal’s temperature. OCXOs are commonly used to control the frequency of cellular base stations and military communications equipment. For convenience, we packaged the OCXOs inside blank C Series modules, and inserted them into each CompactRIO chassis.
To commence synchronisation, we briefly connect a ‘master’ OCXO to the ‘local’ OCXO inside each of the slave units. This syncs the two clocks and ensures that all units operate at the same time. Once we complete this synchronisation process, which takes only a few seconds at the start of the day, each slave unit can be detached from the master OCXO clock and roam the bridge freely.
We attached an NI-9402 high-speed digital I/O module to each CompactRIO to monitor the ‘local’ OCXO clock ticks at 120 MHz, whilst the CompactRIO FPGA (customized with LabVIEW) drives the vibration acquisition by overriding the native CompactRIO clock with the acquired OCXO ticks.
Using this novel approach, each of data logger is wirelessly synchronised within a fraction of a microsecond with negligible drift, even over extended test times of 12 hours.
Using high-speed direct memory access, we stream all acquired vibration/acceleration data to a Technical Data Management Streaming (TDMS) file. Once testing is complete, the CompactRIO allows us to access the stored data via Ethernet, FTP, or USB flash storage.
Funded by the Engineering and Physical Sciences Research Council (EPSRC), as part of the BAYOMALAW project (EN/N017897/1), our research team travelled to the Jiangsu province of China to test the capabilities of the new data logging system on the Jiangyin suspension bridge.
The Jiangyin bridge has a single suspended span of 1,385 m, with straight back stays. The concrete towers rise 191 m above ground level and each have three hollow portal beams. The mid portal is adorned by the bridge’s name in stylised Chinese characters created by the former Chinese premier, Jiang Zemin, who opened the bridge in 1999. The 32.5 m-wide deck carries three traffic lanes (plus a narrow emergency lane) in each direction. The 2.2 m-wide cantilevered walkways on either side are for maintenance as there is no pedestrian access to the bridge. The test team used these walkways for moving the loggers around.
For the experiment, we took four of the CompactRIO loggers along with a set of 12 force balance accelerometers. On the bridge, we brought the loggers together to be synchronised to the master OCXO clock to within fraction of a microsecond, before being distributed over the bridge deck and towers to acquire 14 separate ambient vibration measurements over a period of three days. For each measurement, we left the master logger recording vertical and lateral vibrations continuously on the east walkway, and “roved” the slave loggers to synchronously record an hour of vibration data at various locations on the east and west sides and inside the south tower.
On return to the United Kingdom, researchers performed full data analysis at Liverpool University, using their new Bayesian Operational Modal Analysis (BAYOMA) procedures. The data showed a rich set of vibration modes, with predominance in lateral (sway) and vertical directions, as well has some involving torsion and some involving the towers. Good synchronisation was evident from the analysis, which is very sensitive to phase drift. The character of the modes, including the extreme low frequencies of fundamental modes in lateral and vertical directions provided an excellent opportunity to challenge, evaluate, and enhance the BAYOMA procedure.
Following the successful trial in China, we now plan to reuse the wireless CompactRIO data loggers for testing landmark structures in the United Kingdom and overseas.
Additionally, we plan to scale our distributed system to identify the modal properties of tall buildings, such as skyscrapers or towers. Running the cables required by conventional systems through tall structures is even more problematic than long-span bridges, since the only access between levels is via stairwells. Even dropping cables between staircases, which might work on low-rise structures, is difficult on super tall structures, due to circular layouts, walls, and doors.
Ultimately, this highlights the real benefits of our choice of NI as a technology provider. Not only did the NI platform empower our team of domain experts to quickly develop a novel, distributed monitoring system, but the modular nature of LabVIEW and CompactRIO helped us scale the system to meet our ambitious future research plans.
James Brownjohn
University of Exeter - College of Engineering, Mathematics and Physical Sciences
North Park Road
Exeter EX4 4QF
Tel: +44 (0)1392 723698