Prespector is a dashboard based on Machine Learning algorithms for predictive maintenance in the petrochemical industry. Prespector is created by making use of merely open source software. The name Prespector is a contraction of Preventative (to prevent failures) and Prospector (a miner for gold and minerals).
Prospecting is the first stage of the geological analysis of a territory after which the exploration stage starts. This is exactly the stage in which many maintenance departments are in with the mining of their huge amounts of raw maintenance data leading to better preventative maintenance. The Prespector dashboard is the result of the first pioneering steps of the R&M Engineers into the world of Machine Learning in their own environment and every day work.
The advantage of making use of open source software is that is a fast and easy way to start with Machine Learning by Reliability and Maintenance Engineers (R&ME) without big investments in resources. Compare it to starting your finance with an Excel sheet before using ERP software like SAP or Oracle.
Because of the flexibility of open source software packages like Python and R you can make use of almost all data sources that come to mind like;
– Field measurements from the SCADA / DCS
– Failure data from the CMMS
– Asset utilisation analysis from the Process Engineers
The idea of Prespector is that the dashboards and algorithms are created by the Subject Matter Experts themselves (in this case the R&M Engineers) so that they understand their own models and can optimize them anytime they want.
If you want to know if your maintenance department is ready to start mining contact me to for the possibilities (bart@bastiaanengineering.nl).