Award No.
ARG02-0416-240192
A research collaboration developing machine-learning methods to detect and analyse sub- and super-synchronous oscillations (S2SO) in converter-dominated power systems — the hidden instabilities that emerge when inverter-based renewables replace conventional generation. Led by Qatar University with Hamad Bin Khalifa University and Iberdrola QSTP.
As Qatar's grid integrates more solar and wind, the machines that historically held the system steady are being replaced by power-electronic inverters — and that changes the physics of stability itself.
Conventional grids lean on the spinning inertia of synchronous generators to ride through disturbances. Inverter-based resources have none of that inertia. Instead, their fast controllers can interact with the network — and with each other — to produce sub- and super-synchronous oscillations (S2SO): sustained resonances that classical stability tools were never designed to catch.
This project builds the analysis and detection methods a renewable-heavy grid needs. It combines the established toolkit of power-system stability — eigenvalue analysis, frequency scanning, and impedance modelling — with modern machine learning and deep reinforcement learning trained on time-domain simulation data, so oscillation risk can be identified before it threatens the network.
From understanding how renewables reshape classical stability, through to a working data-driven detector for oscillation events.
Characterise how high renewable penetration reshapes classical power-system stability, and identify the new oscillation modes converter-dominated grids introduce.
Combine eigenvalue analysis, frequency scanning, and small-signal impedance modelling of inverter-based resources into a unified way to study S2SO.
Train machine-learning and deep-reinforcement-learning models on simulation data to flag S2SO events from measurements — even under weak-grid, series-compensated conditions.
Grow expertise in Qatar through graduate training, skills development, and research outputs that strengthen the region's grid-resilience knowledge base.
Each package runs on its own timeline of months (M1–M36) and deliverables. Year 1 focuses on WP2 and the early analysis methods.
Explore how growing renewable penetration affects classical stability, review transient-stability modelling for large-scale systems, benchmark ML algorithms for stability assessment, and generate a labelled dataset from IEEE bus-system simulations.
Apply eigenvalue analysis, frequency scanning, and combined approaches; use time-domain simulation and small-signal impedance models of inverter-based resources to study system stability.
Detect instability from IBR–grid and IBR–IBR interactions using DFT-based parameter identification, modal analysis, the open-loop modal proximity (OLMP) approach, and machine learning on measurement data.
Research-degree training, staff development, skills growth, and academic esteem — bibliometric impact, committee and editorial-board memberships built through the project.
Project reporting to the Research Office runs through a single guided form — no email attachments to chase, no version confusion. It walks the PI through six sections and captures everything the award requires.
Bringing together academic research, applied engineering, and industry to secure the grid of a renewable-powered Qatar.
Hosts the project through the College of Engineering. Leads the stability-analysis research, machine-learning model development, and graduate training across the work packages.
Contributes complementary research strength in advanced computing and energy systems, supporting the data-driven detection methods and analysis of converter-dominated grid behaviour.
Brings global utility and grid-operation expertise, grounding the project's methods in the real operational challenges of integrating renewables at scale.
Led from the College of Engineering at Qatar University, with graduate researchers across the work packages.
Electrical Engineering, College of Engineering, Qatar University. Leads the project's stability-analysis and machine-learning research direction.
Postgraduate students and research assistants driving dataset generation, modal analysis, and the machine-learning detection models across the work packages.
Partner researchers and engineers from Hamad Bin Khalifa University and Iberdrola QSTP, strengthening the project's methods in impedance modelling, frequency-domain analysis, and real-world grid-resilience assessment.
For research collaboration, data, or reporting queries, reach the principal investigator directly.
Email the PI