Abstract
Decarbonizing urban environments requires sophisticated tools to manage the transition toward decentralized, sector-coupled energy systems. This paper introduces FOCUS, an open-source Python framework designed for multi-scale energy system optimization. By employing a modular, bottom-up approach based on Mixed-Integer Linear Programming (MILP), FOCUS enables the simultaneous optimization of sizing and operation across prosumer, district, and city levels. The framework supports multi-objective optimization, allowing decision-makers to evaluate trade-offs between economic costs and environmental impacts across multiple energy carriers, including electricity, heat, gas, cooling, and hydrogen.
Introduction
Urban areas are responsible for over 70% of global $CO_2$ emissions, making them central to climate change mitigation strategies. The shift toward renewable energy necessitates a move from centralized fossil-fuel-based systems to decentralized, "sector-coupled" architectures where electricity, heating, cooling, and transport are deeply interconnected. However, existing optimization tools often struggle with the computational complexity of city-scale modeling or fail to provide the multi-objective insights needed for nuanced policy decisions. FOCUS (Framework for Energy System Optimization) addresses these gaps by providing a scalable, modular framework for urban energy planning.
Methodology
The FOCUS framework is built upon a three-level bottom-up hierarchy designed to balance detail with computational efficiency:
- Prosumer Level: The smallest unit of optimization, representing individual entities such as residential buildings equipped with photovoltaics (PV), heat pumps, and battery storage.
- District Level: Models groups of interconnected prosumers sharing local infrastructure and energy resources.
- City/Coupled District Level: Addresses large-scale infrastructure, including high-voltage grids, gas networks, and centralized power plants.
To ensure computational tractability, FOCUS utilizes modularity. Instead of solving a monolithic city-wide optimization problem, it treats the results from lower levels—specifically the "residual load" (the net energy demand or surplus)—as inputs for higher-level optimizations. This hierarchical approach significantly reduces the mathematical complexity while maintaining high resolution. The core engine uses Mixed-Integer Linear Programming (MILP) to determine optimal component sizing (investment) and hourly operational strategies simultaneously.
Key Features and Innovations
Multi-Objective Optimization
Rather than providing a single "optimal" solution based solely on cost, FOCUS generates Pareto fronts. These visualizations allow planners to understand the trade-offs between competing criteria, such as the balance between total annual system costs and $CO_2$ emissions.
Sector Coupling and Flexibility
The framework includes an extensive library of mathematical models for diverse energy technologies. By integrating electricity, gas, heat, cooling, and hydrogen sectors, FOCUS can identify "flexibility potentials." This is critical for stabilizing the grid against the volatility of renewable energy sources through strategic use of storage and cross-sector energy conversion.
Open-Source Accessibility
FOCUS is developed in Python and released under the MIT license. This ensures that the tool is accessible to researchers and planners worldwide, facilitating collaborative development and the integration of new technology models as they emerge.
Results and Discussion
As a framework announcement, FOCUS demonstrates the successful integration of multiple energy sectors into a unified optimization workflow. The modular design effectively mitigates the risk of "unsolvable" large-scale problems by decomposing them into manageable sub-problems. Preliminary demonstrations show that the framework can handle complex interactions between various energy carriers while providing actionable insights into both infrastructure investment and daily operation.
Conclusion and Future Work
While FOCUS provides a robust foundation for urban energy optimization, ongoing research focuses on further reducing computational demands. The authors are investigating complexity reduction techniques, such as temporal aggregation (grouping hours into representative blocks), to speed up city-scale simulations. Future developments include expanding the component library to include industrial facilities and calibrating models against real-world urban energy data. The project is part of the "InEEd-DC" initiative, with full maturity expected by 2025.
Summary updated