Public defence: Swaechchha Dahal

Swaechchha Dahal is defending her thesis for the degree philosophiae doctor (PhD) at the University of South-Eastern Norway.

She has followed the PhD programme in technology at the Faculty of Technology, Natural Sciences and Maritime Sciences.

All interested are welcome to attend the trial lecture and the dissertation defence.

• Read the thesis here (link to be updated).

Summary 

Renewable energy is transforming electricity generation, but they are weather dependent which makes the power grids harder to manage reliably. Swaechchha Dahal’s doctoral research has developed a new framework called WINDS (Weather-Integrated Neural-Symbolic Dispatch System) that enables grid operators to make smarter, weather-aware decisions in real time, reducing energy dispatch costs by up to 12.5% and eliminating security constraint violations even in adverse weather.

Modern power systems face growing challenges from the variability of renewable sources. Dahal’s research embeds detailed weather forecasts directly into the mathematical calculations that govern how electricity is distributed across the grid. WINDS combines neural networks which learn from weather and grid data to predict future conditions, with symbolic optimisation that enforces the physical rules the grid must obey. This hybrid approach reduces dispatch costs by 2.1–12.5% compared to conventional methods, while keeping the grid within safe operating limits at all times.

The research began in Nepal, where rainfall patterns were shown to predict run-of-river hydropower output with 97.7% accuracy. Building on this, an advanced forecasting model was developed for the Nordic grid, delivering reliable 36-hour ahead electricity forecasts by incorporating temperature, wind, and precipitation data. Further work found that combining modern optimisation algorithms with multiple-slack-bus operation reduces total power losses by 13.5%, while a spatial-temporal machine learning model achieved an 11-fold improvement in computational speed for power flow prediction.

For grid operators and policymakers, WINDS offers a practical pathway: as renewable energy’s share of generation continues to grow, weather-integrated tools can maintain grid reliability without costly infrastructure upgrades. The research was conducted under a cotutelle agreement between the University of South-Eastern Norway and Kathmandu University, drawing on field experience in Nepal’s transmission and hydropower sector.