Transforming Electricity Price Forecasting with Advanced Computing 

SUMMARY

Developing long term electricity price scenarios and forecasts is crucial for market participants, including utilities, grid operators, energy traders, and policymakers. Accurate forecasting helps optimize bidding strategies, manage risks, and ensure system reliability. However, predicting electricity prices is inherently complex due to the high volatility of demand, renewable generation, fuel prices, and market conditions. Traditional models often rely on simplifications and heuristics that struggle to capture market dynamics accurately, especially with increasing renewable penetration.

Ceres, our state-of-the-art planning tool, improves long term electricity price modelling by leveraging full temporal resolution modeling, cloud computing, and advanced acceleration techniques. Unlike conventional models, Ceres processes all 8,760 hours per year over a 30-year horizon without resorting to simplifications like representative days when evaluating capacity expansion. This results in a more robust approach, able to more accurately assess network expansion and storage needs, which are crucial for electricity price forecasting, providing a competitive advantage to market participants.

Challenges in Electricity Price Forecasting Electricity prices fluctuate due to multiple factors, including:  

  • Variable Renewable Energy (VRE): Wind and solar generation are weather-dependent, introducing uncertainty in price forecasts. 
  • Reference System Expansion: Future storage and interconnectors capacity need to be correctly estimated as they have a key impact on electricity prices. 
  • Grid Congestion and Transmission Constraints: Localized congestion leads to price spikes and variations across regions. 
  • Fuel Price Fluctuations: Prices of natural gas, coal, and other fuels impact the marginal cost of generation. 
  • Market Rules and Regulations: Changing policies and market structures influence price formation. 

Traditional models often address these complexities using: 

  • Simplified time aggregation (e.g., representative days, load blocks) which distort system expansion and therefore price signals, especially with a high share of renewables.  
  • Regression models and heuristics that fail to capture non-linear market interactions resulting from structural changes of the market over time. 
  • Time-consuming simulation approaches that require hours or days to run, making real-time forecasting impractical. 

Figure 1: Range of errors incurred in capacity expansion plans from traditional models with large shares of renewable energy in storage and grid expansion

How Ceres Enhances Price Forecasting Ceres overcomes these limitations with a next-generation computational approach. Key innovations include: 

  1. High-Resolution Temporal Analysis: Unlike traditional models, Ceres considers all hourly price fluctuations over multi-decade horizons when planning new capacity, eliminating distortions caused by aggregation techniques that affect price formation. 
  2. Integrated Renewable Uncertainty Modeling: Ceres simulates the impact of renewable variability on market prices, helping stakeholders anticipate fluctuations and develop risk-mitigation strategies. 
  3. Cloud-Based Acceleration: Leveraging cloud computing, Ceres processes large-scale market datasets efficiently, reducing runtimes to under 30 minutes. 
  4. Advanced Market Simulation: Ceres integrates detailed power market structures, including locational marginal pricing (LMP), capacity markets, and ancillary services, to provide comprehensive forecasts. 

Use Case: Market Assessment of new Projects: A utility, developer or financial institution interested in evaluating the profitability of a new generation or storage project requires robust price forecasts to guide investment and project dimensioning decisions. Using conventional tools, the utility may: 

  • Misestimate the financial viability of a power generation, transmission or storage project due to inaccurate price volatility estimates caused by inaccurate storage and transmission expansion plans. 
  • Struggle cost recovery due to forecasting errors and or to sub-optimally dimensioned facilities. 
  • Face difficulties in assessing the value of flexible assets such as battery storage and demand response. 

With Ceres, the utility can: 

  • Evaluate long-term price trends considering solid network and storage expansion plans developed using full chronology minimizing forecast errors. 
  • Quantify the economic value of storage and flexibility investments based on accurate price and volatility projections. 

The Strategic Advantage of Ceres Ceres provides an unprecedented advantage for energy market participants by offering: 

  • Unmatched Accuracy: Full temporal resolution eliminates errors from simplified models. 
  • Rapid Computation: Cloud-based acceleration delivers results in minutes. 
  • Comprehensive Insights: Advanced modeling captures real-world price dynamics. 
  • Scalability and Flexibility: Suitable for long-term planning, and policy analysis. 

Figure 2: Forecasted scenario of average and captured electricity prices

Conclusion As electricity markets become more complex, traditional long term forecasting methods struggle to keep up with increasing renewable penetration and market volatility. Ceres revolutionizes electricity price forecasting by combining high-resolution modeling and advanced computational techniques. With Ceres, utilities, regulators, market analysts and financial institutions, can obtain faster, and more accurate long term price forecasts, enabling informed decision-making in the rapidly evolving energy landscape. 

About the authors 

This article was written by Diego Luca de Tena, Managing Director of Pharoes, who is based in Madrid. The author wants to thank Javier Andersen, Solutions Analyst of Pharoes, for his contribution to the development of Ceres.