A crack-monitoring time series becomes useful when you know where, when and under which conditions each value was obtained. Two isolated numbers may hide an interruption, a changed reference or an environmentally influenced variation. Monitoring therefore requires an organised approach to both data collection and review.
The aim is to document behaviour for the professionals responsible for analysis. A graph, even a smooth and regular one, is not a diagnosis of the structure on its own.
01 Define the question and responsibility
State what the monitoring should help examine: behaviour during a construction stage, variations over a given period or movement at an already identified point. Decide who reviews the data, how often, and who should receive an unusual observation.
The protocol must distinguish technical acquisition from decision-making. Where thresholds are used, defining them and the associated actions is the responsibility of the competent project professionals. They cannot be selected from a universal value found in an article.
02 Keep an understandable reference
Give every point a stable identifier and record its location, measurement axis and sign convention. Archive the initial reading and commissioning conditions. A zero selected in software does not necessarily represent a closed crack.
Record every intervention, including moved fixings, replacement sensors and changed settings. To establish this record from the outset, consult the checklist for installing a crack sensor. Do not silently join time series that use different references.
03 Select suitable intervals
The useful measurement frequency depends on the phenomenon and the project’s requirements. Widely spaced readings may miss variations between observations. More frequent acquisition produces more data, but cannot compensate for poor mounting or an unsuitable measurement method.
Separate measurement frequency from transmission frequency. Check whether intermediate readings are retained and retrievable. An average received at regular intervals does not necessarily reveal the extremes that occurred within that interval.
04 Check the record before interpreting it
Start with units, timestamps, sign conventions and continuity. Missing data should be shown as missing, not as an absence of movement. Preserve the original readings before filtering or making corrections.
A sudden change can have several causes: actual displacement, work on the device or a measurement problem. Review the site log and the measurement chain before deleting a value or assigning a cause. Depending on the situation, an independent check may be needed.
05 Compare measurements with their context
Document construction activity, interventions and relevant environmental conditions. Temperature may help put some variations into context. Instrumentation projects may therefore combine crack measurements with temperature and humidity observations. S10
However, similar-looking curves do not prove causation. Temperature measured near an enclosure is not necessarily representative of the whole structure. Any thermal correction must use a justified method and retain a record of the uncorrected data.
06 Review trends without jumping to conclusions
Look for recurring variations, persistent offsets and changes in rate while considering the observation period and measurement quality. An apparently stable period does not demonstrate that behaviour will remain unchanged under different conditions.
For an illustrative example, suppose a fictional series changes from 0.10 mm to 0.18 mm relative to the same reference. The difference is 0.08 mm. That calculation alone says nothing about severity. Relevant considerations include elapsed time, measurement performance, mounting condition and context. These values are neither thresholds nor ADVARNA results.
07 Produce a usable monitoring report
Present the objective, monitoring points, period, method, site events and limitations encountered. Graphs should show units, dates, references and interruptions. Separate measured observations from possible explanations and the decisions taken.
For a better understanding of small changes, read our guide to resolution and measurement uncertainty. To prepare the system, explore the remote crack monitoring solutions and describe your data-access requirements.