Search for: All records

The paper proposes an approach for data generation and action recommender system development when small-signal stability assessment is performed using deep learning algorithms. These algorithms are trained on labeled time-series data solving a classification task. In this paper we propose an approach that includes automatically generated labeled data for the action recommender and the deep learning methodology to implement the recommender framework. The deep learning methodology is based on convolutional neural networks (ResNet, Encoder, Time-LeNet) that are tuned for time-series input data and shown the best performance in comparison to other architectures. The proposed approach is validated on synthetic but realistic measurement data from the IEEE 9-bus system as a reference and further applied to a 769-bus system representing a region in the U. S. Eastern Interconnection. The performance of the method is evaluated using accuracy as a most common machine learning metric, as well as precision and recall. We show that the evaluation of the methodology on the generated imbalanced data has to be treated with additional metrics other than accuracy. The training time and the classification performance time is evaluated. 

The paper proposes an approach for fast small signal stability assessment on a short data window using deep learning algorithms. This paper shows that the proposed deep convolutional neural networks (CNNs)-based assessment approach is faster than traditional methods (i.e. Prony’s method). The evaluated CNNs are fully convolutional network (FCN), CNN with sub-sampling steps performed through max pooling (Time LeNet), time CNN, fully convolutional network with attention mechanism (Encoder), and CNN with a shortcut residual connection (ResNet). The proposed approach is validated on different synthetic measurement data sets generated from the IEEE 9-bus system that is used as a reference, and further applied to a 769-bus system representing a region in the U. S. Eastern Interconnection. We show that precision and recall are more informative metrics than accuracy for the reliability of the stability assessment process using the proposed methodology. In addition, the method’s efficiency is compared to classical Prony method. 

Balanced data is required for deep neural networks (DNNs) when learning to perform power system stability assessment. However, power system measurement data contains relatively few events from where power system dynamics can be learnt. To mitigate this imbalance, we propose a novel data augmentation strategy preserving the dynamic characteristics to be learnt. The augmentation is performed using Variational Mode Decomposition. The detrended and the augmented data are tested for distributions similarity using Kernel Maximum Mean Discrepancy test. In addition, the effectiveness of the augmentation methodology is validated via training an Encoder DNN utilizing original data, testing using the augmented data, and evaluating the Encoder’s performance employing several metrics.