ETP has developed a tool to quantify generator shaft fatigue caused by AI data center load oscillations over extended operating periods, from hours to months.
​
The key challenge with conventional PSCAD studies is that EMT simulations are typically run for short windows, often around 10 seconds. While these studies can identify torsional interactions and oscillatory behavior, they do not directly translate the results into a quantitative measure of generator shaft life degradation.
​
ETP’s tool bridges this gap by using EMT results from PSCAD studies and converting them into practical generator life-impact metrics. The assessment considers expected AI training duration, training algorithm characteristics, torsional response, generator mechanical parameters, and damping behavior. This provides decision-makers with a quantitative basis to understand how different AI training profiles may affect generator shaft fatigue over time
OEM Selection and Technology Comparison
Optimization of BESS, E-STATCOM, and Flywheel Sizing
Identifying When Mitigation is Required Before Equipment Life is Impacted
Long-duration Reliability Assessment for Colocated Generation and AI Data Centers
Generator Operating Limits During AI Training Events
