This paper proposes a recurrent neural network based model to segment and classify multiple combined multiple power quality disturbances (PQDs) from the PQD voltage signal. A modified bi-directional long short-term memory (BI-LSTM) model with two different types of attention mechanisms is developed. Firstly, an attention gate is added to the basic LSTM cell to reduce the training time and focus the memory on important PQD signal part. Secondly, an attention layer is added to the BI-LSTM to obtain the more important part of the voltage signal by assigning weightage to the output of the BI-LSTM model. This attention gate applied in the LSTM cells improves effective and decisive key information extraction from future and past states of the PQD signal and the addition of attention layer improves the overall decision capability of the model, saving computation time, and increasing PQD classification accuracy. Finally, a SoftMax classifier is applied to classify the combined PQD signal in 96 different combinations. The proposed BI-LSTM model with attention gate and attention layer mechanism is compared to different baseline models based on recurrent neural network and convolution neural network (CNN). From the simulation study, it is inferred that with the proposed method, the multiple combined PQD signals are easily segmented from the voltage signal which makes the process of PQD classification more accurate with less computation complexity and in less time as compared to popular signal processing based approaches for which classification of multiple PQDs is computationally difficult.
Power quality disturbance signal segmentation and classification based on modified BI-LSTM with double attention mechanism
Siano P.;
2024-01-01
Abstract
This paper proposes a recurrent neural network based model to segment and classify multiple combined multiple power quality disturbances (PQDs) from the PQD voltage signal. A modified bi-directional long short-term memory (BI-LSTM) model with two different types of attention mechanisms is developed. Firstly, an attention gate is added to the basic LSTM cell to reduce the training time and focus the memory on important PQD signal part. Secondly, an attention layer is added to the BI-LSTM to obtain the more important part of the voltage signal by assigning weightage to the output of the BI-LSTM model. This attention gate applied in the LSTM cells improves effective and decisive key information extraction from future and past states of the PQD signal and the addition of attention layer improves the overall decision capability of the model, saving computation time, and increasing PQD classification accuracy. Finally, a SoftMax classifier is applied to classify the combined PQD signal in 96 different combinations. The proposed BI-LSTM model with attention gate and attention layer mechanism is compared to different baseline models based on recurrent neural network and convolution neural network (CNN). From the simulation study, it is inferred that with the proposed method, the multiple combined PQD signals are easily segmented from the voltage signal which makes the process of PQD classification more accurate with less computation complexity and in less time as compared to popular signal processing based approaches for which classification of multiple PQDs is computationally difficult.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.