Published Fri 26th May 2023
Wireless condition monitoring (WCM) provides engineers, surveyors and owners with a wide range of functionality and typically requires less expertise to install and operate than traditional wired or geodetic techniques. Installation can be performed by non-experts, often with remote support where required. This facilitates the use of WCM for spontaneous tasks and for small assignments where costs are far lower than observations using automated total stations (ATS).
Compared to manual observations there is an immense gain in terms of increased spatial and temporal sampling – and more data means more detail. Figure 1 above shows the comparison of occasional manual crack sensor observations vs. a continuously operating WCM crack sensor on the wing walls of a weir in Austria (Manual crack observation (red) vs automated crack sensor time series (blue)). Though the graphs coincide where data is available, it is obvious that the high frequency WCM observations reveal short-term structural behaviour that is missed by less frequent sampling of the manual operation.
Figure 2 shows the WCM constellation during the truss renovation of the baroque St.Oswald church in Regensburg, Germany [JANKA P.]. Paired tilt nodes were installed on the roof beams in order to determine whether beam movement took the form of tilt (if direction of tilt is directed equally), or bending (if tilt points in opposite direction). A bending motion would be characterised by vertical displacement with no tilt at the centre. To determine if this was the case the tilt nodes were supplemented by the use of optical displacement sensors installed at the centre of the beams pointing down to the floor.
Figure 2: Constellation of the QCM observation during the renovation of a historic church's roof truss
The data presented in Figure 3 shows the individual parameters indicated by colour coded alarm levels. In Figure 4 the time series of the tilt sensors, as well as the ambient temperature, is shown over an 18 month period. A correlation between temperature and movement is apparent, despite the relatively stable thermal environment inside the church. Furthermore, it is apparent that the walls moved less than the beams which provided confidence that the data indicates deformation of the beams rather than extraneous temperature effects on the sensors. The data also displayed a correlation between construction activity and structural behaviour throughout the monitoring programme.
Figure 3: Data visualisation platform of church roof truss observation
Figure 4: Time series of selected tilt sensors (truss beams: blue; wall: orange/yellow)
The higher sampling frequency of WCM is often more effective for measurement of relative movement than traditional geodetic surveys, although the latter will typically deliver more precision in terms of absolute measurement. There are therefore many situations where wireless monitoring can complement manual geodetic surveys rather then provide an alternative. A common approach is to conduct an initial geodetic survey, which can include the precise positioning of bolts or targets fixed to the structure. Subsequent WCM observations can characterise typical behaviour patterns, with any alarms triggered by the WCM followed up in a pre-determined way, which will often include verification by independent survey. Usually, activity induced deformation differs significantly from typical behaviour.
Where two elements of interest are physically separated (such as a pair of bridge piers) the use of 1D laser distance observations can be conducted using devices such as the Senceive Optical Displacement Sensor. These measurements are often conducted reflectorless straight onto the observed surface. Repeatability roughly equals those given for common total stations (±0.1mm), though achievable practical results depend on environmental conditions, surface structure and, of course, distance. With increasing distance the influence of refraction and visibility become greater. Even at long distances (>100m) however accuracies in the few millimetre range are realistic as the time series excerpt in Figure 5 shows for the observation of the distance between two bridge pillars.
Figure 5: Observation of pillars by tilt sensors. Temperature-related diurnal signal vs significant deviation on case of physical impact
It is quite common for monitoring to be required due to the influence of construction activity adjacent to existing buildings and structures. Typical examples include bridges, facades, sheet pile and retaining walls. In the case of the façade shown below the use of ATS observations was prevented by extensive scaffolding which blocked the line of sight between the total station and geodetic prism targets. An initial baseline survey was carried out manually. During demolition of the building core the remaining façade was observed using various wireless sensors. The system was configured so that breaches of threshold values defined by the designers would trigger alarms and further manual surveys would be initiated.
The ultra-long battery life of wireless nodes enables assessment of long-term behaviour trends such as those related to changing seasonal conditions. Dams and reservoir retaining walls, for example, may experience changing loads due to seasonal changes in water level, causing deformation, stress or even cracks. Bridges can suffer from increasing traffic load and from the effects of temperature and material deterioration. Data from wireless tilt nodes, strain gauges and dedicated temperature sensors can provide valuable insight that enables scarce maintenance resources to be effectively targeted.
During the renovation of the Spanish Martorell tunnel as part of the Mediterranean Rail Corridor project the invert of the tunnel was lowered and thus the tunnel height extended in order to allow future electrification. Designers expected this to result in deformation during the works and specified a monitoring programme to provide near real-time data on parameters including convergence of the lining.
Arrays of optical displacement sensors were installed at regular intervals around the arch to monitor convergence, and longitudinal chains of tilt beams were installed along the tunnel walls to monitor vertical displacements (Figure 6)
Figure 6: Example for real-life repeatability of automated distance observations at distance >130 m
Ease of installation is particularly significant when time is limited. Time is almost always limited when rail track monitoring is required, and returning to site for subsequent measurements or interaction with a monitoring system is expensive and complicated. Further challenges include the presence of obstructions such as masts, platforms, trains etc. that can complicate optical observations.
The option to utilize tilt sensors for rail track geometry monitoring has been documented in the past, but it is only since tilt sensors became wireless that the use of these methods become realistic, sustainable
and reliable. Rail and sleeper are considered as an interconnected unit, with the sleepers treated as rigid. Multi-axial tilt sensors observe lateral and longitudinal tilt simultaneously. Lateral tilt reveals changes in cant (crosslevel), while the differences in consecutive lateral tilts show track twist. Because the sleepers are connected by the rails it is possible to model virtual sensor chains longitudinally to provide indicative measurement of settlement. While this does not provide spirit level accuracies it represents a perfect compromise between the logistic constraints of ATS observations (interference with traffic, limited sampling etc..). Often both systems supplement each other and in synergy offer the best of best worlds.
In many cases the quantitative magnitude of movement is less significant than identification of events that are outside normal behaviour patterns. This is relevant in various structural and geotechnical applications, including the study of slopes and embankments. In their study of slope monitoring using tilt sensors UCHIMURA T. demonstrated the value of tilt sensors in the prediction of slope failure.
The authors not only pointed out the general usability but also evaluated the reliability of the prediction depending on the location of the nodes. Because earthworks such as embankments and cuttings extend over long distances and are especially difficult to observe in remote locations the autonomous operation of a WCM system proves to be specifically advantageous.
With over 30,000 wireless sensors installed on railway earthworks worldwide the methodology can be regarded as well-established and proven. Since 2019 rail operators have been able to deploy responsive wireless systems such as the Infraguard™ solution developed by Senceive to provide early warning of potentially disruptive and dangerous slope failures. In a typical deployment rows of tilt nodes are mounted on metal stakes Figure 13. Such systems are only effective if they can operate for long periods with minimal maintenance. This is made possible partly by improvements in battery performance, but more significantly by the network intelligence built into each node: this allows it operate in a semi-dormant mode for most its life, but to “wake-up” in the event of significant ground movement. Not only does such an event activate neighbouring tilt nodes to send updates, but also triggers cellular cameras to take and send photos of the site – enabling remote stakeholders to verify the significance of an event before they can get people to site.
Figure 13: InfraGuard system including tilt nodes, camera and cellular gateway
References
GOLSER J., (2010) „Fallbeispiele zur Bauwerksüberwachung mittels Online Neigungsmessung”; in 25. Christian Veder Kolloquium, Vortrag 9.4.2010 Graz
JANKA P.; Janka Ingenieurbüro für Vermessung und Geoinformation. 92421 Schwandorf, Deutschland
KADEN B., KALENJUK S., (RENNEN M.) (2022), Vortrag „Bauwerksmonitoring mit Senceive-Neigungssensoren Landpfeiler Verbund-Flusskraftwerk“ Geodätisches Kolloquium Universität
Innsbruck 16.03.2022, Innsbruck, Austria
KALENJUK S., LIENHART W. (2020) , Vortrag „Performance of Senceive Triaxial Tilt Sensor Nodes - Insights from lab tests and a field study“ Geodätisches Kolloquium Universität Graz
07.10.2020, Graz, Austria
PECK, R.B. (1969) “Advantages and Limitations of the Observational Method in Applied Soil Mechanics” 1969, Géotechnique, pp 171-187, V 19, N 2, 10.1680/geot.1969.19.2.171
UCHIMURA T. ET.AL. (2020) “Precaution and early warning of surface failure of slopes using tilt sensors”; in „Soils and Foundations”, Vol.55 Issue 5, Oct.2015 P.1086-1099 The Japanese
Geotechnical Society/Elsevier
