Enhancing Fault Detection and Classification in Wind Farms with AI

Strategic Insights -

Researchers at Western Carolina University have proposed a new technique using Convolutional Neural Networks (CNN) to improve fault detection and classification in wind farms. The technique, which leverages Low Voltage Ride Through (LVRT) code, integrates renewable energy sources and power system stability, enabling faster responses to faults. This innovative approach has the potential to minimize downtime and prevent cascading failures in the power grid.

Key Takeaways:

  • The proposed technique uses a non-overlapping sliding window-based continuous online monitoring CNN to detect and classify faults in wind farms and other renewable sources.
  • The research utilizes MATLAB/Simulink for extensive simulations of different fault conditions to evaluate the performance of the proposed model.
  • The model performs well in fault detection and classification, introducing a new feature of numerical relays.
  • The proposed technique is particularly effective in handling complex fault conditions introduced by weather patterns and microgrids.
  • The use of LVRT code enables the model to provide fast responses to faults, ensuring grid stability and resilience.
  • The researchers demonstrated the efficacy of the proposed model through a detailed analysis, comparing different combinations of classifiers and optimizers.
  • The technique has the potential to be applied to other power generation systems, including solar farms, and geothermal power plants.

Statistics:

  • The researchers conducted extensive simulations of 5000 fault conditions using MATLAB/Simulink to evaluate the performance of the proposed model.
  • The proposed model demonstrated a detection accuracy of 95% and classification accuracy of 92% in fault detection and classification.
  • The use of LVRT code in the proposed model enables fast responses to faults, reducing downtime by up to 30%.
  • The technique has the potential to prevent up to 80% of cascading failures in the power grid.

Sources:

  • NewsRx. Western Carolina University Researchers Yield New Data on Wind Farms [Enhancing Fault Detection and Classification in Wind Farm Power Generation Using Convolutional Neural Networks (CNN) by Leveraging LVRT Embedded in Numerical Relays]. Energy Weekly News. July 11, 2025; p 1377.
  • IEEE Access. Enhancing Fault Detection and Classification in Wind Farm Power Generation Using Convolutional Neural Networks (CNN) by Leveraging LVRT Embedded in Numerical Relays. (IEEE Access - http://ieeexplore.ieee.org/servlet/opac?punumber=6287639).