Sourav Sudhakaran
PhD
Flow Over Complex or Forested Terrain: Physical versus Data-based Modelling
Host Organisation
University of Porto, Faculty of Engineering
Company
Inductiva Research Labs
Project Description
The PhD project aims to model airflow over complex and forested terrains using numerical simulations and machine learning (ML). Public datasets like Perdigão-2017 will be analyzed to classify wind patterns. High-quality simulation datasets will be generated to train and validate ML models. The project will develop physically informed ML architectures for wind field prediction and super-resolution, with models designed to generalize across terrains. Furthermore, data-driven models will be compared to classical simulators to assess their limitations, estimate errors, and identify failure cases. Data-driven methods will also be employed to predict calibration parameters for classical simulators, improving their accuracy. The ultimate goal is to create faster and computationally efficient models for wind flow prediction.
Supervisors
Hugo Penedones, Inductiva Research Labs
José Laginha Palma, Faculty of Engineering University of Porto
João Viana Lopes, Faculty of Sciences University of Porto
Background
I’m from the southwest state of Kerala in India. I have a bachelor’s in mechanical engineering (College of Engineering Trivandrum) and a master's in Thermal Engineering from NIT Rourkela. During my master’s, I developed a strong interest in Computational Fluid Dynamics (CFD), which led me to pursue various CFD roles after graduation. Over the next two years, I worked in academia and industry, focusing on simulation-based projects, finally leading to my current role at AptWind. Outside of research, I enjoy making art, travelling, cooking, and watching movies.
​
I’ve always been intrigued by the potential of machine learning (ML) to enhance technologies like CFD. The AptWind PhD project offered me a unique opportunity to work at the intersection of these fields. By applying my CFD background, I can explore how ML can drive advancements in wind energy and weather forecasting—areas with significant societal impact, especially in addressing global energy challenges and improving climate predictions. The doctoral network’s blend of academic and industrial experiences also provides invaluable opportunities for collaboration and networking, further enhancing my ability to contribute to impactful technologies.