Typical modal tracking algorithms make use of frequency, mode shape and environmental conditions, among other information to identify, classify and follow the modal response. To correctly characterize mode shapes, well distributed and located sensors are required. In civil strictures, the use of large number of sensors is limited by the size of the structure, the difficulty to install cables or wireless connections and limitations due to cost, technical, usage or aesthetics requirements. In this article, we explore the effects of a limited set of sensors to perform modal tracking in a building that has been monitored continuously for approximately 5 years and 8 months. We implement a methodology that relies mostly on the modal frequency and a poorly defined modal shape. The reduced number of sensors are used to discriminate between closely spaced modes and modes that are highly sensitive to environmental conditions. In order to increase the robustness of the tracking algorithm, the global environmental conditions, particularly temperature, are monitored and used. It is concluded that in order to obtain a reliable tracking, a minimum number of sensors is required if closely spaced modes are present. Also, the need of highly defined mode shapes can be reduced if the environmental conditions are considered. As expected, we have found that tracking is possible with very limited number of sensors, if the objective is to capture the global response contained in the lower modes of the system. In order to capture low excited higher modes, the number of sensors should be increased and location selected cautiously.
[1]
Randall J. Allemang,et al.
A Correlation Coefficient for Modal Vector Analysis
,
1982
.
[2]
Guido De Roeck,et al.
REFERENCE-BASED STOCHASTIC SUBSPACE IDENTIFICATION FOR OUTPUT-ONLY MODAL ANALYSIS
,
1999
.
[3]
Rubén Boroschek,et al.
Interpretation of stabilization diagrams using density-based clustering algorithm
,
2019,
Engineering Structures.
[4]
Hans-Peter Kriegel,et al.
OPTICS: ordering points to identify the clustering structure
,
1999,
SIGMOD '99.
[5]
Alessandro Cabboi,et al.
Automatic Operational Modal Analysis: Challenges and practical application to a historical bridge
,
2013
.
[6]
Rune Brincker,et al.
Modal identification of output-only systems using frequency domain decomposition
,
2001
.
[7]
Carlo Rainieri,et al.
Near real-time tracking of dynamic properties for standalone structural health monitoring systems
,
2011
.