When a big generator trips on the Australian energy network, grid frequency drops and other generators must immediately increase their output to compensate. AGL Energy required a system to automatically determine whether their large fleet of generators had responded appropriately to these events. To achieve this required a multi-disciplinary effort; this presentation will walk through the build process of the following system:
The SCADA system pushes high speed meter data (~20ms sampling rate) into Pi following each event, trigger event frames
An assessment tool coded in Python picks up new event frames and processes the high-speed data to assess the performance of each generator
PDF reports and performance KPIs are written back to the event frames with alerting for poor performance
A front-end built on Asset Intellect allows users to view reports at both generator and fleet level
AGL Energy owns the biggest power generation fleet in Australia including coal, gas turbines, wind, hydro, solar and batteries.
David Bowly leads AGL's Operational Analytics team. This small team combines engineering and data science experts to drive better understanding of the operation of AGL's power generation fleet, improving reliability and efficiency.
Dave started his career with ExxonMobil, then as an associate at Boston Consulting Group. He then completed his masters at Cambridge as the 2014 Menzies Scholar in Engineering and has been with AGL for the past 6 years.