This project is done in the context of the Professional Education Lead Program of the Jheronimus Academy of Data Science [JADS] as a graduation project.
Enexis is a regulated regional distribution network operator (DNO) in the Netherlands, responsible for transporting electricity and gas to 2.6 million customers.
Enexis aims to keep the delivery of gas and electricity reliable, affordable and sustainable for all its customers.
Power is delivered via grid components such as cables and transformers (for electricity) to residential and industrial connections of our customers.
The capacity of these grid components should always be sufficient for the power flows in the grid. This is monitored by Grid Planners and if needed transformers are swapped for heavier ones, or cable connections are strengthened.
For decades this monitoring was done mainly by looking to historical yearly extremes.
However, the increasing growth rate of power demand and supply driven by emerging technologies as electrical vehicles (EV) and photovoltaics (PV), requires shorter monitoring periods and nowadays even forecasts. Otherwise, there is no time for mitigating actions and customers will be out of power.
Enexis started a few years ago with the measurement of its distribution transformer population (in the project “distribution automation light” (DALI)).
Power measurements are since then available which enable monitoring ad hoc by the grid planners.
This project takes the next step by forecasting on that data in autumn if a transformer overloads in spring.
This project and its documentation are set up around the CRoss Industry Standard Process for Data Mining [CRISP-DM].
It is an iterative process and this documentation focuses on the end result without ignoring the lessons learned along the way. The latter is generally noted at the end of every chapter.
I would like to thank Jeroen the Mast for the valuable feedback and supervision.
Special thanks also goes out to PDEng candidate Akshaya Ravi for her technical support. Together with my buddy David Rijlaarsdam she provided me with helpful insights and discussions on the project.