Leadership and engineers found making sense of the volume of data produced from 250,000 sensors extremely difficult.
We developed an analytics and data visualisation platform for monitoring a range of instrument and sensor data more efficiently.
We extended the traditional Gaussian models with spatio-temporal correlation derived from machine learning techniques to improve risk management on a £15b urban infrastructure project.
Our proprietary analytics are designed to help understand the relationship between observed movements and external factors in order to spot anomalies, forecast events and optimise the monitoring regime to improve risk management while minimising the cost. These analytics are embedded within a web interface that supports simple reporting and interpretation across the whole project organisation.
This ability to hunt for patterns between sensors is changing the industry by enabling real time anomaly detection, event forecasting and optimisation of the monitoring regime.
The analytics tool was successfully deployed operationally at two stations to reduce monitoring costs by 20% whilst improving risk management capability.
It enabled real time anomaly detection across the whole area rather than just within a limited ‘zone-of-influence’.
Enhanced the historical process with a predictive capability to forecast construction and monitoring events within a seven day window, enabling faster and better interventions.
This automation of the basic analysis enabled asset protection engineers to focus on value-add interpretation rather than spotting issues.