Donato Leo is the author of a study on the relationship between solar PV, batteries, and wholesale energy prices in Italy. His deep learning simulations predict how energy prices will change with increasing installed battery capacity.
pv magazine from ESS news site
As an energy professional working in utility operations, Leo uses deep learning and machine learning codes for analysis and forecasting to analyze and simulate market scenarios to optimize portfolio management strategies. In his latest analysis, Leo used deep learning techniques to simulate PUN's trends for utility-scale battery installations. PUN (Italian acronym for Prezzo Unico Nazionale, “National Single Price”) is the wholesale reference price for electricity purchased on the Italian Electricity Exchange (IPEX – Italian Electricity Exchange) and represents the national weighted average of hourly and daily zonal selling prices for electricity.
The graph of electricity demand on the power grid looks a bit like a duck and a camel (we'll call it the camel's back curve here), with demand highest in the morning and evening when people rely on the power grid, and then dropping off significantly during the day, when many people turn to solar power instead, resulting in less demand from the power grid.
According to Leo, BESS lowers the maximum electricity price, raises the minimum price, and has a seasonal effect on the average price, meaning that on days with less sunshine the average price drops, and on days with more PV generation the average price rises slightly.
BESS allows PV to avoid feeding power to the grid during daytime hours when PUN is low, and feed power to the grid during dark hours when PUN is high. This allows you to recover the higher cost of BESS and increase your revenue, but it depends on the situation, right? Can you explain?
Donato Leo: The shape of the PUN curve is closely related to the characteristics of the generating equipment. In the case of PV, generating equipment is not yet equipped with BESS on a large scale, so it can only generate and supply during sunshine hours. The gradual introduction of BESS (and the development of storage services offered by third parties) may encourage PV operators to store energy during the currently low-reward sunshine hours and supply it to the grid during PUN peaks, flattening the current camelback curve. If this is a plausible scenario, it is clear that in such a situation, PV utility-scale operators who use BESS first will enjoy greater benefits for some time, since they will initially notice that the current PUN “camelback curve” remains (or barely) unchanged, and there will be a significant daily difference between PUNmax and PUNmin during the day.
Using deep learning techniques, we created an algorithm to understand how the PUN curve, which is currently very much a camelback shape, will change. On closer inspection, it appears that the curve will lose its second peak in the evening, leaving only the first peak during periods of no energy input. Or will both peaks smooth out anyway? What does this mean for battery reward?
DL: First of all, let me state that the algorithm results you saw in my LinkedIn post are the result of an initial Convolutional Neural Network (CNN) training based on historical hourly stock market and energy balance data from 2023 to early 2024. The related forecasts should therefore be taken with great caution, taking into account that in the example I posted, I also assumed a significant large-scale shift of PV production from daytime to evening, but in the meantime, with the impact of unpredictable strategies of operators on the strategies adopted by the first movers.
To continue reading, please visit the new ESS news site.
This content is copyrighted and cannot be reused. If you would like to collaborate with us and reuse any part of our content, please contact us at editors@pv-magazine.com.
