infraSOFT is the Machine Learning-powered water pipes Replacement and Rehabilitation (R&R) management platform developed by infraPLAN.
infraSOFT helps water utilities and their consultants create the least expensive plan that meets utility goals. infraSOFT streamlines all the necessary analytical computations and tasks.
infraSOFT has 4 modules: CLEAN, STATS, PREDICT and PLAN.
infraSOFT is web-based with embedded mapping. Built outside of GIS to maximize computing capacity, there is no need to be GIS-literate. Users access the platform for different purposes based on their role in the planning process. Some users may generate specific results, others may simply consult those results, track data cleaning, or monitor the physical condition of the pipes.
The data needed is simply a pipes and breaks database, GIS shape or csv files.
Let's face it, Asset Management is about data. The better the data, the more specific, accurate, optimized and defensible your plan will be. We also learned through our many consulting projects that data cleaning can be very tedious, which is why we developed the CLEAN module.
The CLEAN module includes various innovative and powerful proprietary approaches, including Machine Learning-powered infraSOFT.ml, that identify missing and incoherent values as well as structural issues.
We believe that, even in the age of Machine Learning, running simple and rigorous descriptive statistics helps interpret black box forecasting results and planning recommendations.
Statistical results also allow the analyst to better assess data quality, understand the system and program of R&R, define the next analytical steps, and monitor changes in data quality, R&R and system performance.
The STATS module automatically generates results we know to be essential to achieve the above goals. The results come in the form of configurable charts and tables as well as in a dashboard. The platform is dynamic which means that all results get updated as data changes.
The PREDICT module generates a likelihood of failure score for each pipe for each year in the future with exceptional validation; it can be used to prioritize inspection or rehabilitation projects, generate aging curves, and draw the Useful Lives.
This is all done using physical data (pipes, breaks, pressure, etc.) as well as environmental data (traffic, soil, groundwater, landfill, zoning, etc.) analyzed in advanced statistical and Machine Learning models.
The PLAN module makes use of the output results from PREDICT. It allows the generation of multiple scenarios of R&R (which pipes to rehabilitate and when).
It is the culmination of all the steps that took place to that point.
Output results include charts that track the Break Rate, Business Risk Exposure, R&R rate and cost.
What do we mean with optimal plan? Doing more with less!