The (TM) machine learning model acts as an "engine" receiving as input sensor data form a network, like a sewer or a water distribution network (both stationary and mission based data) and generating stationary sensor data values for the same network, which can then be fed to a digital twin software.

The main functionality of is that of "filling in the voids" both in space and in time, generating a smart interpolation of the data based on the available information. This is particularly useful when the available info is in the form of mission based data collections, which typically are not well suited to be fed in a digital twin layer.

The software is also dynamic, so that if additional data (added sensors, additional missions etc.) become available, the generation of data is improved accordingly. The software can be setup to mark the generated data so that the digital twin is aware of which data is "raw" and which one is generated by the engine. is currently used in the NETWORKleak softare for localization of Inflow and Infiltration in sewer networks.