Leak detection technology has long been a staple across commercial buildings, data centres and critical infrastructure, helping organisations identify water ingress before it causes costly disruption. Traditional systems, particularly those using leak detection cables, have offered a relatively simple monitoring solution. However, these systems typically rely on fixed alert thresholds, which often fail to reflect the variability of real-world environments. As a result, while detection capabilities exist, their reliability has frequently been called into question.
According to Quensus, the challenge lies in how these signals are interpreted rather than how they are captured.
Leak detection cables generate numerical readings, typically on a scale from 0 to 5000, indicating moisture levels. Under ideal dry conditions, baseline readings are usually very low, often between 3 and 10.
When water is present, these values rise quickly, with even minor leaks pushing readings above 100. Despite this, most legacy systems rely on a static alert threshold—commonly set around 50—to trigger warnings, an approach that does not account for environmental fluctuations.
Environmental noise undermines reliability
In practice, leak detection is influenced by a wide range of external factors. Temperature shifts, humidity changes, condensation cycles, proximity to conductive materials, and even electrical interference can all affect readings.
Over time, these variables introduce natural fluctuations that may have nothing to do with actual leaks. Fixed thresholds therefore tend to generate false positives, creating unnecessary operational noise and gradually eroding trust in monitoring systems.
This erosion of confidence can have serious consequences. Facilities teams faced with frequent false alerts may begin to ignore warnings altogether, undermining the very purpose of leak detection. In environments where water damage can lead to significant operational or financial loss, unreliable alerts pose a substantial risk.
Adaptive analytics offer a smarter solution
To tackle this issue, Quensus has been developing a more intelligent approach centred on what it describes as a filtered average baseline. Instead of relying on a static threshold, this method continuously analyses historical data to establish a dynamic baseline that reflects normal environmental conditions. Alert sensitivity is then adjusted relative to this evolving benchmark.
This approach enables systems to filter out background noise, track gradual environmental changes, and more accurately identify genuine anomalies.
Early testing suggests the model can reduce false alerts by up to two orders of magnitude, while still maintaining sensitivity to real leak events. The result is a more reliable and actionable alert system that improves both efficiency and trust.
From detection to prevention
Improving alert accuracy is not just about reducing inconvenience—it plays a crucial role in enabling proactive risk management. More reliable signals allow for faster responses, greater confidence in automated systems, and better-informed operational decisions. Crucially, they also reduce alert fatigue among facilities teams, ensuring that genuine risks are taken seriously.
Accurate data also opens the door to preventative strategies. When systems can clearly distinguish between environmental variation and actual risk, automated interventions such as shut-off valves can be deployed with greater confidence. This marks a shift from reactive responses to a more preventative approach to water management.
Looking ahead, Quensus’ work reflects a broader trend in connected infrastructure, where the emphasis is moving from data collection to intelligent interpretation.
As analytics, adaptive thresholds and behavioural modelling become more advanced, they are set to play a central role in mitigating risk and improving resilience across modern building systems.
Read the full blog from Quensus here.
Copyright © 2026 FinTech Global
