BRIDGE-ESM

Context and Background

The BRIDGE-ESM project represents a strategic collaboration between Eurac Research (Italy) and the University of Victoria (Canada), focusing on advancing machine learning applications in energy systems modeling. This research builds upon a 2024 publication in Energy journal titled "Machine learning as a surrogate model for EnergyPLAN: speeding up energy system optimization at the country level," which demonstrated that ML-based surrogate models can reduce computational time by 64% while maintaining high accuracy.

Traditional energy system models require significant computational resources, making them impractical for real-time decision-making and extensive uncertainty analyses. This limitation hinders the ability of policymakers and planners to explore multiple scenarios during stakeholder meetings or planning sessions. By developing sophisticated surrogate models using deep learning techniques, this project aims to revolutionize how local and regional authorities interact with energy planning tools.

Research Methodology

The project employs a comparative research approach, examining two distinct surrogate modeling strategies:

Eurac Research's Approach: Focuses on operational optimization, creating surrogate models that can rapidly predict energy system operation under various scenarios.

University of Victoria's Approach: Centers on direct capacity expansion optimization, using deep learning to predict optimal infrastructure investments for decarbonization pathways.

Both methodologies were applied to the Canadian energy system, providing a comprehensive case study to evaluate their relative strengths, computational efficiency, and accuracy levels. This parallel implementation enables direct comparison and synthesis of approaches, generating insights that advance the entire field of ML applications in energy modeling.

Work Structure and Timeline

The research was organized into five overlapping work packages executed between June and August 2025:

WP1 - Methodology Integration: Established collaboration protocols, aligned computational environments, and exchanged detailed methodological approaches between research teams.

WP2 - Canadian Energy System Model Development: Adapted both surrogate modeling approaches to the Canadian context, including data collection, preprocessing, and initial implementation.

WP3 - Comparative Analysis: Parallel implementation of both approaches, running simulations, analyzing computational performance, and documenting differences in results and efficiency.

WP4 - Validation and Uncertainty Analysis: Comprehensive testing including sensitivity analyses, uncertainty quantification, and robustness assessment across different scenarios.

WP5 - Results Documentation: Statistical analysis of comparative results, preparation of visualizations, and collaborative writing of research paper for peer-reviewed publication.

Key Results and Outcomes

The project successfully achieved all planned objectives. The comparative analysis revealed significant insights into how different parameter inputs affect energy transition modeling outcomes. Comprehensive validation protocols were established, providing robust frameworks for future applications of surrogate models in energy system planning.

A major outcome was the development of methodologies that enable real-time decision support capabilities for energy planners. These tools can provide almost instantaneous results, making sophisticated energy system analysis accessible to non-technical stakeholders and enabling interactive exploration of scenarios during policy discussions.

The collaboration resulted in the preparation of a high-impact research paper documenting methodological comparisons and best practices, currently in the final writing phase before submission to a peer-reviewed journal.

Strategic Impact and Future Perspectives

Beyond the immediate research outcomes, this collaboration has generated substantial strategic benefits. The partnership was further strengthened through joint preparation and submission of a major research proposal to the Schmidt Sciences Foundation's Decarbonization & Energy Virtual Institute (DEVI) program. Matteo Prina served as Principal Investigator with Eurac Research as project coordinator for this comprehensive proposal, which includes University of Victoria as a key partner alongside 19 other international institutions.

If successful (results expected December 2025), this will result in a $10 million project coordinated by Eurac Research, establishing a global network for advancing decarbonization modeling methodologies. This represents a significant amplification of the initial research visit's impact, demonstrating how strategic international collaboration can catalyze major research initiatives.

The project positions both institutions at the forefront of an emerging field, creating opportunities for continued collaboration and establishing expertise that will be valuable as demand for computationally efficient modeling approaches continues to grow in the energy sector.