Syrinix uses machine learning techniques to extract actionable insights from rising main pressure data

To begin, Syrinix worked with one utility and manually studied behaviour patterns within the daily workings of the rising main.

These manual studies fed the development of an algorithm, where performance patterns could be scored and plotted within a calibrated 'zone' on a graph.

A score was attributed to the following performance outcomes -

Good pump/bad pump
Low static head
High static head
Low delivery pressure
High delivery pressure

Any behaviour that 'scored' outside of the calibrated 'zone',  is immediately flagged to the utilty in the form of a 'burst alarm'.

In addition, pressure monitors were deployed at pump stations, and that data, containing burst and performance events, was also manually analysed to design algorithms.

Graph showing burst event
The graph shows a potential burst event.




This method has identified performance patterns, typical bursts, worn pumps, passing NRV's and blocked mains.


Interested to hear more? Enter details for further info or to book a demo and we will show you.

    How we work with global water utilities

    City of Galveston

    Case Study – City of Galveston uses PIPEMINDER-ONE pressure monitors to revitalise its system and improve resilience

    Read more