Optimal Power System Planning

Power systems planning is an intricate and critical endeavor, balancing the demands of future infrastructure needs with the realities of financial constraints. Infrastructure investments in this sector often span decades, making meticulous planning a necessity. Through our collaboration with Redeia and Elewit in the Siroco project, eRoots has developed a groundbreaking approach that addresses these challenges head-on: a genetic algorithm-based method for optimal planning.

The Challenges of Power Systems Planning

With large-scale investments in grid infrastructure expected to last approximately 50 years, it is crucial to select projects that maximize technical benefits while minimizing costs. This challenge is further complicated by long lead times for construction, as well as the need to replace aging infrastructure while integrating new technologies. Operators require a roadmap to prioritize investments, ensuring immediate and sustainable improvements.

Genetic Algorithms: A Transformative Solution

Our method employs the Non-Dominated Sorting Genetic Algorithm III (NSGA-III), an advanced evolutionary optimization technique tailored for multi-objective problems. The approach is designed to simultaneously minimize:

1. Investment Costs: The sum of capital expenditures (CAPEX) and operational expenditures (OPEX).
2. Technical Costs: A combination of monetary penalties applied to technical violations within the grid, such as power losses, voltage deviations, and branch overloading.

This dual-objective framework generates a Pareto front—a curve representing the trade-offs between these objectives. Operators can use this visualized curve to identify optimal investment strategies within their budget constraints, balancing cost and technical improvements.

How NSGA-III Revolutionizes Planning

NSGA-III offers a significant advantage over traditional methods by effectively exploring the Pareto front for complex, multi-dimensional problems. The algorithm operates through the following key mechanisms:

• Non-Dominated Sorting: Solutions are ranked based on their performance across objectives, ensuring only the most competitive configurations advance.
• Reference Directions: These guide the algorithm toward a balanced distribution of solutions, enhancing diversity and coverage along the Pareto front.
• Mutation and Crossover: Genetic operators introduce variations and recombine configurations, enabling the discovery of new, optimal solutions.

By employing these mechanisms, NSGA-III identifies a diverse set of investment strategies, offering grid operators a comprehensive view of possible improvements.

Key Insights and Results

When tested on a 130-bus grid with 389 investment candidates, our method demonstrated significant improvements:

• Improved Pareto Front Exploration: NSGA-III uncovered more optimal solutions compared to previous methods.
• Computational Efficiency: The algorithm is approximately 25 times faster than other algorithms, delivering results within minutes even for this large grid.
• Visual Clarity: Operators receive a clear and actionable roadmap, highlighting the most impactful investments.

‍

Integration with GridCal

This cutting-edge methodology is fully integrated into the open-source GridCal platform. Available on our website and the GridCal repository, this integration empowers utilities, researchers, and developers to conduct in-depth investment evaluations and experiment with various scenarios and constraints to refine planning strategies

Future Directions

Building on this robust foundation, we aim to:

1. Develop Surrogate Models: These will further reduce computation times, enabling faster evaluations of investment scenarios.
2. Expand Testing: Apply the methodology to diverse grid configurations, including smaller and larger systems, to ensure its robustness and adaptability.
3. Enhance Usability: Incorporate user-friendly features to make the tool more accessible to a broader audience.

Conclusion

Through the integration of genetic algorithms and open-source innovation, eRoots is redefining power system planning. Our method provides a clear roadmap for strategic investments, balancing cost efficiency with technical performance. By empowering stakeholders with actionable insights, we contribute to the development of resilient and future-ready power grids. Explore this transformative tool within GridCal and join us in driving the evolution of power systems engineering.

‍