03.2021 - 10.2022

Price forecast of intraday auction and continuous intraday trading

FfE project with the Entrix GmbH

In the context of the energy transition, battery storage is crucial for ensuring system stability. The Entrix GmbH operates an optimization and trading platform for large battery storage systemssystems with the aim of accelerating the transition to a green energy future. A relevant use case is arbitrage trading on the spot market. Price forecasts are required for optimized trading in order to profitably market the large battery storage systems. In the project thus, intraday price forecasts for Germany and Great Britain were created in a joint collaboration, taking into account the relevant influencing factors with the implementation of the algorithms by the Entrix GmbH. The FfE had performed the basic energy economic analyzes and, based on this, the modeling of the algorithm.

Figure 1: Price forecast of the price characteristics of the intraday auction in Germany and historical prices

Two forecasts for the prices of the EPEX Spot were created:

  • Price forecast of the quarter-hourly (DE) / half-hourly (UK) intraday auction price:
    The price forecast is performed between the hourly day-ahead auction and the quarter-hourly intraday auction (DE) or half-hourly day-ahead auction (UK). The focus of the modeling is on the representation of the intraday characteristics based on fundamental influencing factors.
  • Forecast of continuous intraday price distribution after the intraday auction:
    The price uncertainty forecast takes place immediately after the quarter-hourly intraday auction (DE) or the half-hourly day-ahead auction (Great Britain). The forecast is performed once and not updated continuously based on new forecasts. In contrast to the previous price forecast, quarter-hourly (DE) or half-hourly (UK) products have already been traded at this point. The focus of the modeling is on the stochastic consideration of how big the uncertainty of the expectancy value of the continuous intraday price is.

The model develops its prediction from historical data, which is why the definition of the input period takes a crucial role. As further analysis indicated changing market conditions, historical data for both price forecasts is taken from a rolling input period to ensure results remain current.

The algorithms were then tested by using historical data. The day-ahead price and the price forecast of a commercial service provider served as a reference for the evaluation of the forecast quality of the price characteristics. In particular for the intraday auction in Germany, a stable price forecast could be achieved, which showed a significantly higher forecast quality than the two references. The algorithms were handed over to the Entrix GmbH for implementation. In addition, based on the algorithm developed jointly with the Entrix GmbH, the FfE publishes an ex-post demonstration of the forecast for the intraday auction in Germany on a daily basis.