Ceres: Revolutionizing Power System Planning for a Renewable Future 

SUMMARY

As the power sector transitions towards higher renewable energy penetration, traditional planning tools face significant challenges in delivering accurate and timely results. Many of these tools rely on simplifications such as load blocks or typical days, which can introduce errors of 50% in capacity expansion planning with high shares of variable renewables. Ceres, our advanced power system planning tool, addresses these shortcomings with state-of-the-art techniques, cloud computing, and cutting-edge acceleration technology. By considering every hour over a multidecade planning horizon, Ceres delivers precise and actionable results in under 30 minutes, facilitating the planning process for utilities, regulators, and policymakers.

The Challenges of Power System Planning with High Renewables Traditional power system planning tools were developed in an era when generation was predominantly dispatchable, making system modeling relatively straightforward. However, as more renewable energy sources like wind and solar are integrated, the consideration of the variability of these resources presents new challenges. Conventional planning methods typically employ: 

  • Load blocks and typical days: These simplifications aggregate data to reduce computational burden but fail to capture the variability of renewables accurately. 
  • Limited temporal granularity: Many tools analyze only a subset of representative days per year, which leads to significant errors in estimating system adequacy and flexibility requirements. 
  • Lengthy computation times: Running large-scale simulations with traditional models can take several hours or even days, making decision-making challenging and increasing the burden on analysts. 

These challenges can be easily explained with the following example, which in a simple way illustrate the limitations of traditional planning models. Imagine a situation with a large share of solar electricity, which during daytime replaces thermal generation. Under such a situation the value of electricity in those hours is very low, even zero or negative (see left chart of Figure 1). However, when load blocks or typical days are used, we obtain the chart on the right, where solar generation is smoothen out as different hours or different days are grouped together artificially reducing its variability.  

Figure 1: Exemplary representation of power generation on one day employing hourly values (left) and load blocks (right)

For the given case we can also estimate the impact on the economic value of solar generation (see Figure 2), which in the example considered exceeds 60%. 

Figure 2: Market revenue with hourly representation (left) and blocks (right)

Finally, we can also see what the impact on the economic value of solar would be in case of different shares of solar power generation. We can see in Figure 3, that whereas for solar penetrations below 10% the impact is limited, however, for solar penetration shares of 25% the distortion can reach 150%!! 

Figure 3: Deviation on estimated market revenues considering blocks vs hourly data for different solar power penetrations, %

Use Case: Planning for a High-Renewable Future Consider a utility tasked with designing a power system expansion plan for the next 30 years under a scenario with 50+% renewable energy penetration. Using a conventional tool, the utility might:

  • Approximate demand and renewables generation patterns using load blocks or representative days, leading to errors of 50% in the resulting capacity expansion plans as it fails to consider less probable but high impact scenarios.

  • Require lengthy simulations, making analysis more costly and slow, and thus forcing the user to make excessive simplifications that distort the quantitative analysis. 

  • Struggle to assess the impact of energy storage, flexible demand and network expansion needs with sufficient granularity as fluctuating solar and wind would appear to be less volatile in the model. 

With Ceres, the same analysis: 

  • Captures hourly variations in renewable generation and demand across all several decades, leading to precise system adequacy and expansion needs assessments. 
  • Produces least-cost investment and operational plans in under 30 minutes, facilitating decision-making and iterating the hypothesis with the key stakeholders. 
  • Identifies optimal energy storage and grid reinforcements needs with high accuracy, ensuring long-term system reliability and cost-effectiveness. 

The Impact of Ceres By eliminating the inaccuracies introduced by traditional planning tools, Ceres empowers utilities, regulators, and policymakers to make data-driven decisions with confidence. Its ability to deliver rapid, high-precision results makes it an indispensable tool for planning the future grid. As the world moves towards decarbonization and increased renewable integration, Ceres provides the technological edge needed to navigate these complex transitions efficiently and cost-effectively. 

Conclusion Ceres represents a transformative advancement in power system planning. By leveraging cloud computing, our proprietary acceleration technology, and full chronological modeling, it outperforms conventional tools in both accuracy and speed. As the energy landscape evolves, Ceres ensures that planners have the insights they need to build resilient, cost-effective, and sustainable power systems. 

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 Salvador Guerrero, former Solutions Analyst of Pharoes, for his contribution to the development of Ceres. 

Figure 4: Range of errors incurred in capacity expansion plans from traditional models

Acknowledgments 

We wish to thank the Centre for the Development of Industrial Technology, the European Union for their financial, and el Ministerio de Ciencia e Innovación y Universidades support for the development of this technology: