Energy systems planning faces serious challenges: many variables such as technology cost, resource availability, demand growth, policy changes, and emissions targets are highly uncertain. This Tech Xplore article explains a new framework developed by North Carolina State University researchers that integrates global sensitivity analysis into large-scale optimization models of energy systems.
Traditionally, energy system models, which aim to find least-cost pathways for building and operating resources such as renewables, storage, and transmission, run discrete “what-if” scenarios. But these don’t always identify, which input uncertainties matter most. The new method applies optimal-transport theory to map how changes in input distributions propagate to model outcomes like system cost or resource mix.
By ranking the influence of uncertain parameters, the technique allows planners to:
- Prioritize data collection and research efforts toward inputs that drive variability
- Explore investment and policy options that remain robust under a wide range of futures
- Speed up exploration of large “uncertainty spaces” by coupling the model with surrogate (machine-learning) systems to approximate responses.
The article emphasizes that the framework was applied using the open-source TEMOA model but can generalize across regions and scales. For planners designing infrastructure around climate and policy uncertainty, this means better insight into which assumptions really matter, and which decisions will hold up when the future shifts.
The key takeaway is that modeling is evolving from deterministic scenario snapshots toward structured sensitivity and robustness thinking. Tools such as this help reduce the risk of surprise, inform smarter investment in data, and build systems capable of coping with deep unknowns.