Efficient Power Quality Disturbance Classification Using Teager-Kaiser Energy Operator and Fast Walsh-Hadamard Transform Features

Power Quality (PQ) disturbances are critically important for the reliability and efficiency of electrical systems. Such disturbances can negatively impact the performance of electrical devices, leading to malfunctions and significant energy losses. Accurate identification and classification of PQ disturbances are essential for maintaining system stability and optimizing energy usage. The proposed method begins with the application of the Fast Fourier Transform (FFT) to randomly generated PQ disturbance events from nine different types. Following this transformation, features are extracted focusing on Teager-Kaiser Energy Operator (TKEO) and Fast Walsh-Hadamard Transform (FWHT) outputs. These features provide a comprehensive representation of the disturbances, capturing both the energy distribution and structural patterns of the signals. These extracted features then feed into a Random Forest (RF) classification model. The performance of this model has proven to be highly effective, achieving a classification accuracy of 99.35% with pure signals. Additionally, the study investigates the effect of noise on classification performance. By assessing the robustness of the model with 40 dB noise, where it achieved 98.26% accuracy, its reliability in real-world scenarios, where noise is often a prevalent issue, has been demonstrated.