Due to the increasing possibilities of information and computer technology, there is a strong tendency to collect and integrate more and more data from heterogeneous sources. To discern patterns in these complex data sources, special techniques have been developed, which are often referred to as machine learning, data-mining or pattern recognition.
Digital images can be regarded as a multivariate, highly structured sources of data, from which features and patterns need to be extracted (image analysis).
Water Distribution Monitoring
Monitoring water distribution systems can be challenging, due to the scale and maze-like nature of these networks, be they buried drinking or waste water pipes or meandering waterways. Insight in these systems is obtained based on a large numbers of sensors spread throughout the network, measuring flow, pressure and many other properties in real-time. Data collection alone does not yield the insight needed for maintenance strategies, thus a combination of black-box machine learning and white-box water dynamics modelling is required to gather information about the functioning of the whole system. With these tools, real-time algorithms are developed to serve as decision support and early warning tools, facilitate leakage detection and localization, temperature or contamination warning, and indicate optimal locations for placement of more sensors.
Watch the video to learn more about SMART detection and real-time learning in water distribution systems:
- Unfortunately, your cookie settings do not allow videos to be displayed. - check your settings