Evolving Disturbances Detection and Classification in Real-time for Grid-Connected System

In the past one decade, several signal processing techniques for detection and classification of single instant occurrence of both power quality (PQ) and islanding events have been reported. However, with the presence of several energy resources in the system, more critical conditions of PQ may arise due to transformer energizing, capacitor energizing, induction motor start, even fault at various locations of the network. As a result, evolving disturbances in a sequence may be observed. This study presents use of energy preserving ensemble empirical mode decomposition technique to detect evolving disturbances that occur in succession and classify them using least square support vector machine. The study is demonstrated on several cases of evolving disturbances as offline analysis for signals simulated in PSCAD/EMTDC and its analysis in MATLAB. Furthermore, novelty in study is established with real-time implementation of techniques in field programmable gate array blocks for detection and classification of evolving disturbances. Also, comparison with existing techniques is included in the discussion section.

Automated summary

This study introduces the use of energy preserving ensemble empirical mode decomposition technique to detect evolving disturbances that occur in succession and classify them using least square support vector machine. The real-time implementation of techniques in field programmable gate array blocks for detection and classification of evolving disturbances is demonstrated. Comparison with existing techniques is also included in the discussion section.