Addressing Grid Impact Challenges With Data Science

A critical challenge in modern grid stability is that inverter-based resources (IBRs) are often “black boxes” to utilities and system operators. Inverter manufacturers and plant developers understandably hesitate to disclose proprietary control strategies, leaving operators with limited visibility into internal dynamics. The problem is further compounded by the fact that IBRs can switch among multiple control modes, which are typically unknown to operators yet can exhibit dramatically different dynamic behaviors. In the final days of 2025, we were excited to learn that our paper on black-box IBR modeling was accepted by IEEE Transactions on Smart Grid. In this work, we develop a comprehensive data-driven framework that uses only terminal measurements to discover unknown control modes and learn continuous-time models that accurately capture IBR dynamics under each mode. By leveraging physics-inspired deep learning, the proposed approach addresses four major challenges in a unified way: 🚀 High-Order Nonlinear Representation Using only terminal measurements, the framework provides a general learning approach for characterizing arbitrary high-order nonlinear dynamics of IBRs. It is not tied to any specific control paradigm and can cover anything from power/voltage/current control loops to virtual synchronous machines (VSMs) and phase-locked loops (PLLs). 🚀 Continuous-Time Modeling Unlike most data-driven methods built on discrete-time models (e.g., RNNs, LSTMs, Transformers), our approach learns continuous-time state-space models (differential-algebraic equations). This enables seamless integration of the learned IBR models into standard power-system time-domain simulations with arbitrary numerical integration schemes and step sizes. 🚀 Discovery of Unknown Control Modes A physics-inspired deep unsupervised learning mechanism automatically identifies distinct control modes from historical disturbance data and learns separate state-space models that represent the dynamics associated with each mode. 🚀 Robustness to Noise and Uncertainty Inspired by Kalman filtering, the learning architecture explicitly accounts for system uncertainties and measurement noise, both of which are ubiquitous in real-world grid systems and data. It ensures the method’s robust performance in practical settings. The examples in the paper demonstrate how the proposed framework can learn accurate time-domain models of fully black-box IBRs and deliver highly accurate long-horizon predictions of their responses to grid disturbances, e.g., subsynchronous oscillations caused by PLL interactions in weak grids. See details here: https://lnkd.in/eFd5CU4e #PowerSystem #SmartGrid #InverterBasedResources #RenewableEnergy #PowerElectronics #Control #PowerSystemStability #PowerSystemModeling #PowerSystemSimulation #SystemIdentification #DataDriven #MachineLearning #DeepLearning #ArtificialIntelligence #PhysicsInformed #IEEETransactionsOnSmartGrid

Grid stability and security are becoming data + control problems. Utilities and large energy operators are already using Artificial Intelligence (AI) to move from reactive alarms to predictive, resilient, and cyber-aware operations—especially as renewables increase volatility. Here’s where Machine Learning (ML) and Deep Learning (DL) deliver real impact: ✅ Anomaly Detection: clustering + autoencoders to flag abnormal grid states and potential cyber events ✅ Fault Detection & Classification: Decision Trees, Random Forests, Support Vector Machine (SVM) models using voltage/current/frequency features ✅ Predictive Maintenance: Remaining Useful Life (RUL) forecasting to reduce unplanned outages (breakers, transformers, lines) ✅ Voltage Stability: Recurrent Neural Network (RNN) + Long Short-Term Memory (LSTM) models to anticipate instability and corrective actions ✅ Cybersecurity: Intrusion Detection System (IDS) + Anomaly Detection System (ADS) using supervised and unsupervised Machine Learning (ML) ✅ Optimal Power Flow (OPF): faster optimization with Machine Learning (ML) surrogates + Linear Programming (LP), Quadratic Programming (QP), Interior Point Method (IPM) constraint handling ✅ Forecasting: Autoregressive Integrated Moving Average (ARIMA) + Seasonal Autoregressive Integrated Moving Average (SARIMA) for load and generation inputs ✅ Uncertainty: Monte Carlo simulation + stochastic programming for renewables and market variability ✅ Autonomous control (next wave): Reinforcement Learning (RL) + Multi-Agent Reinforcement Learning (MARL), plus Federated Learning for privacy-preserving training What’s your biggest grid pain right now: false alarms, asset failures, voltage events, congestion, or cybersecurity? #ArtificialIntelligence #MachineLearning #DeepLearning #PowerSystems #GridReliability #Cybersecurity #PredictiveMaintenance #EnergyTransition

Data centers are emerging as large, geographically distributed, and increasingly controllable loads; and as market-driven workload scheduling becomes the norm, their interaction with the power grid is becoming a critical system-level concern. While most grid studies treat data centers as static loads, price-responsive scheduling can, if left unmitigated, concentrate demand sharply in space and time, inducing voltage stress and congestion the grid is not designed to anticipate. Our recent work, led by Shijie Pan & Zaint Alexakis addresses this challenge through three contributions: (1) A job-level mixed-integer framework capturing the full set of data center control actions (temporal deferral, inter-site transfer, runtime resource reallocation, and service termination) coupled to AC power-flow-based grid-security assessment. (2) A systematic comparison showing that runtime resource reallocation is the dominant driver of both economic gains and grid stress, robust across site placements and operating conditions. (3) A grid-facing ramping charge that smooths schedule-induced load profiles while preserving the data center's economic incentive. https://lnkd.in/ebKtRSdT #PowerSystems #DataCenters #GridSecurity #EnergyMarkets #SmartGrid

The Critical Link for Energy Transition in the AI Era After two weeks at Climate Week forums in New York, engaging with energy sector companies, one reality is clear: while media focuses on renewable generation, the immediate challenge is the infrastructure connecting that energy with demand, now radically transformed by AI. The energy sector faces a perfect storm: decarbonization urgency meets an unprecedented demand explosion from AI data centers. The IEA estimates a ~500 TWh increase by 2030 (twice Spain's current consumption!). Three fundamental conclusions from my conversations: 𝟭. 𝗚𝗿𝗶𝗱 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗶𝘀 𝗮𝘀 𝗰𝗿𝘂𝗰𝗶𝗮𝗹 𝗮𝘀 𝗰𝗹𝗲𝗮𝗻 𝗴𝗲𝗻𝗲𝗿𝗮𝘁𝗶𝗼𝗻 Modernizing existing grids can unlock significant capacity short-term, though it doesn't replace the need for new transmission lines. A Duke University study suggests that flexible curtailment in data centers (just 0.25% of the time) could add up to 76 GW of new capacity to the U.S. grid. 𝟮. 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗳𝗹𝗲𝘅𝗶𝗯𝗶𝗹𝗶𝘁𝘆 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝘀 𝗿𝗲𝗴𝘂𝗹𝗮𝘁𝗼𝗿𝘆 𝗺𝗼𝗱𝗲𝗿𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻 AI's electrical demands are highly concentrated geographically, creating unique grid challenges. Large GPU clusters can cause power fluctuations of hundreds of MW rapidly, stressing systems not designed for such volatility. ERCOT has documented over 2.6 GW of load trips from data centers, causing grid frequency spikes and risking cascading outages. These little-known effects show the immediate need for updated grid protocols. 𝟯. 𝗦𝗵𝗼𝗿𝘁-𝘁𝗲𝗿𝗺 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗲𝗻𝗮𝗯𝗹𝗲𝘀 𝗹𝗼𝗻𝗴-𝘁𝗲𝗿𝗺 𝘁𝗿𝗮𝗻𝘀𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 Grid intelligence solutions provide immediate capacity and buy critical time to execute the necessary 10-15 year transmission development plans. This dual approach is essential: maximizing current infrastructure while building the grid of tomorrow. The smart technologies of today will inform and optimize the transmission of the future. The most revealing insight: we have the technology. From large corporations to startups, companies are developing AI-based solutions for smart grids. What we need is an updated vision for our infrastructure and its regulatory framework. Regions that don't prioritize grid modernization will lose in both the energy transition and economic competitiveness. Territories with agile infrastructure and adaptive regulations will attract the next wave of tech investment, creating a virtuous circle of innovation.

Edited to add the link to Update 2: https://lnkd.in/gzqTfm6U 🚨 New White Paper Release 🚨 Datacenters are no longer just energy consumers—they’re becoming one of the biggest challenges for grid reliability. Unlike traditional loads, their rapid, sustained power fluctuations are already causing measurable disturbances across U.S. grids (ERCOT, WECC, PJM). As AI workloads scale, the urgency to address this problem has never been greater. Relyion’s latest white paper, AI-Enabled Intelligence for Datacenter Flexibility, demonstrates a game-changing solution: ⚡ AI-based forecasting of demand fluctuations ⚡ Intelligent optimization of Battery Energy Storage Systems (BESS) ⚡ Coordinated management of datacenter cooling and workloads The results? ✅ Up to 40–50% reduction in power fluctuation at the grid interconnection point ✅ Improved resilience and flexibility ✅ A pathway for datacenters to become good citizens of the grid 📄 Read the full white paper here: https://lnkd.in/gA-3Wtsw Together, we can transform datacenters from grid stressors into active partners for stability. #Datacenter #EnergyFlexibility #AI #GridStability #WhitePaper

Navigating the New Grid Reality: How DERs and Data Centers are Challenging T&D Infrastructure In today's rapidly evolving energy landscape, the widespread adoption of Distributed Energy Resources (DERs) and the explosive growth of power-hungry AI data centers are creating unprecedented challenges for our Transmission and Distribution (T&D) infrastructure. As someone who has spent years helping utilities adapt to these changes, I've seen firsthand how traditional grid equipment—designed for one-way power flow and predictable loads—is increasingly vulnerable to new failure modes. Transformers overheating from harmonic distortion, protection systems confused by bidirectional power flows, and capacitor banks damaged by resonance issues are just a few examples of what our industry now faces. I'm excited to share a comprehensive investigation framework that my team has developed specifically for identifying, analyzing, and addressing T&D equipment failures related to DER and data center integration. This approach combines rigorous data collection, advanced analytics, and targeted mitigation strategies to help utilities maintain reliability while supporting grid modernization. In the attached article, I explore how these modern grid constituents affect different types of equipment and outline practical steps for protecting your infrastructure investments. Whether you're a utility engineer, a grid operations manager, or an energy policy professional, you'll find actionable insights to help navigate this new grid reality. Looking forward to your thoughts and experiences with these challenges! #GridReliability #DERIntegration #DataCenters #EnergyTransition #UtilityInfrastructure

Last year Sagnik Basumallik and I wrote a paper on the challenges large loads pose to grid reliability and some potential solutions to mitigate these challenges. Our paper - “Reliability Challenges and Solutions for Large Load Integration in Bulk Power Systems,” was accepted for IEEE T&D 2026! We started this effort after working on the first NERC LLTF white paper and this paper built on our experience there. In this paper we expanded on that work with event reviews and identified possible mitigation options for the risks these loads pose to the bulk power system. In the paper we analyzed the impact to the grid from several events where large loads tripped in response to normal system faults, and oscillations originating from large loads across the AEP, Dominion, EirGrid, and ERCOT systems. Then we identified the following causes of events that have been seen and developed a taxonomy of root causes per their source - hardware or software. These causes included: ⚡️Fault-Induced Customer Initiated Load Reduction/Tripping ⚡️Oscillations due to Instability in Electronic Controllers ⚡️Oscillations due to Outdated Firmware Settings ⚡️Transients due to Regular, Cyclical Fluctuations in Data Center Digital Processes ⚡️Coordinated Customer Initiated Load Reduction After the event reviews we looked at what possible mitigations could address the reliability challenges that we identified. Facility side mitigations included: UPS and power supply controller changes to manage oscillations along with hardware updates for voltage ride-through support, coordination with transmission protection schemes, and grid forming loads. Grid side mitigations included E-STATCOMs, better dynamic modeling, improved monitoring capabilities, and market services. Future work is still needed however on large load dynamic modeling, improved monitoring such as point on wave monitoring, and large load characterization. You can read the preprint version of the paper here: https://lnkd.in/gKsJTRz6