LOF Overview

LOF Analysis forms the core of the Fracta solution. Calculated by Fracta machine learning algorithm, this section shows you an overall summary of your pipe asset risk and gives you access to all the details of the likelihood of failure values calculated for your system.

After your data has been uploaded, cleaned and normalized, the machine learning process generates around 200 additional variables to associate with the pipe asset data. The variables include variables from various environmental data, including soil properties, precipitation, population density and transportation features.

Each pipe segment is then assigned variables and a machine learning correlation is made based on the historical failures/breaks (if provided by your data). The algorithm builds a model of how the various pipe parameters and environmental variables correlate with main breaks. Using this data, the LOF application builds a model for future break events, resulting in a prediction for the probability of a failure for the next 1 year and 2 years.