Machine learning and energy storage - Visiting Speaker Series

Published on 26 November 2020

Part of our Visiting Speaker Series

On this page

Energy storage capacity is considered an important tool for aiding in energy transition, and significant developments pertaining to this tool are being debated both in the energy economics literature and by the policy makers

Jose Lanzagorta is a Senior Data Scientist at Arenko, where he analyses the data generated by the different markets for electricity in UK, including the wholesale electricity market and balancing services market, to develop automation software for optimising the performance of grid scale batteries based on machine learning programming languages. He was invited to talk about the use of machine learning for the energy storage under the CEPMLP’s Visiting Speaker Series.

Mr. Jose highlighted that value of battery asset is dependent upon control of the asset to respond to various opportunities presented by the market. This control could either be achieved through a dedicated trading desk or through automating the asset. Later is a cheaper option due to machine learning programming as it helps in combining historical data, weather forecasts and demand output to make relatively accurate predictions. Machine learning models for energy storage once developed can conduct training and tuning of the model automatically and speedily.

He explained that batteries facilitate in the flexibility and stability of an electricity system. A grid scale battery asset (say of capacity of 50MW) may get charged when the supplies on the system is more than the system demand due to a relatively higher output from the renewable energy sources; the assets can supply energy to the system when the system supplies are in deficit of system demand. However, he highlighted the traditional preference of Electricity System Operator (ESO) for gas-based assets to balance the grid, and licensing/grid connection conditions of traditional generating assets being imposed on the batteries as barriers to entry. This may change as ESO engaged with the battery operators in a three-week demonstration project at system level in September, and the results were positive. At the end of his presentation, he answered some follow up questions from the attendees in details. The Visiting Speaker Series adds knowledge to CEPMLP students.

Story category Talks