Electricity price forecasting is a highly complex process that depends on numerous variables including demand, transmission conditions, weather forecasts, government regulations, market conditions, financial speculation, and more. Especially in deregulated, competitive markets, the supply and demand for electricity have numerous outside factors and variables influencing them. Demand is one of the key factors that decide how much electricity is generated at a point in time as large-scale storage of electricity without loss is still not practical. Development of efficient & cost-effective electricity storage systems would have a significant impact on electricity pricing as it would make it possible to generate electricity during low-demand, lower-cost sources and store and supply it to the grid when there is a higher demand.
The quality and accuracy of weather data are better than ever, and the energy industry is embracing digitalization, making available more data to base forecasting models on. We have also made tremendous advancements in our ability to manage, process and gain insights from big data, and this has contributed to increased accuracy in electricity price forecasting. With all the newly available data and technologies, electricity price forecasting is ready for disruption.
Munib Amin- Head of Market Analysis at Innogy says, “The electricity sector has developed to a complex environment exhibiting typical characteristics of volatility and uncertainty. Classical fundamental forecasting is not sufficient anymore to cope with a large number of assets, the variety of players and their diverse strategic behaviour, new applications and potentially flexible demand. That is why I expect fundamental forecasting to be developed further and new complementary methods to become more important. Parsimonious and econometrical models, as well as Monte Carlo methods, could be used to quantify a large number of possible future developments. Refined technology forecasting and technology scenarios could show the impact on business strategies. Multi-agent models could help to investigate future strategic behaviour of market participants and the feed-back on electricity prices. Overall, I expect electricity price forecasting to become a more strategic and less static tool.”
Diana Bacila- Senior Coal Market Analyst at Alpiq is of the opinion that better availability of fundamental and weather data will continue helping the forecasting process by reducing uncertainty and errors. She says, “This will increase the accuracy of power price forecasting in the short-term, urging players to strengthen their data management skills in order to be able to process and efficiently use the latest volume of information. For long-term price forecasting, the improved renewable forecasting process will help to better assess the “new normal” and therefore the need for residual power generation, potentially reducing the uncertainty for thermal asset owners. On the other hand, regulatory risks focused on reducing emissions will remain a significant driver in power price forecasting.”
Increased digitalization by power utility corporations has contributed to making more and more data available for forecasts and projections. Technology has helped refine the pricing forecasting process.
“Data transparency combined with artificial intelligence has played a significant role in increasing the accuracy of power price forecasting in the past years. The smart metering technology will help to improve demand forecasts, therefore reducing the imbalance costs in the system. Discovering trends and patterns in load and weather data series will give players a competitive advantage in the market, helping them make better trading decisions,” says Diana Bacila.
While experts agree that improvements in technology and increased availability and accuracy of data will improve price forecasting accuracy, some are also of the opinion that the fundamental electricity price forecasting processes are likely to remain the same.
Piergiacomo Sabino- Quantitative Risk Modeling Expert at E.ON says, “I am of the opinion that the price forecasting techniques will not change very much in the next coming years. I think that the existing statistical toolbox will be complemented with modern machine learning techniques even more. However, the real challenge is more from the economical-environmental point of view since the political choices will affect the shape and the level of prices. The digital transformation is shaping the energy sector like the other industry sectors. Having large databases and the application of Machine learning and AI techniques has made forecasts more accurate. On the other hand, these approaches heavily depend on the size of past available data while instead, the future situation will very likely exhibit a disruption with past behaviour. The real advantage will materialize combining technological tools with fundamental and political views and scenarios.”
According to Pietro Rabassi- Director of Central European Markets at Nord Pool, “Power markets in Europe are facing an increasingly higher share of iRES (intermittent Renewable Energy Sources). More iRES and other reasons explain the 6 trends that we see in power markets from now until 2030 (and beyond). These trends are happening and will take place in markets that are more and more integrated and more and more open to competition.”
The energy industry is now at a point in time when it is better equipped technologically to tackle electricity price forecasting than ever before. To discover what is new in the world of electricity price forecasting and to be in the know about the latest trends from the experts, join us in Amsterdam for the 8th Electricity Price Forecasting & Market Arrangements 2020 (26-28 February 2020).