Fast Power System Event Identification using Enhanced LSTM Network with Renewable Energy Integration
Abstract : Accurate and fast event identification in power systems is critical for taking timely controls to avoid instability. In this paper, a synchrophasor measurementbased fast and robust event identification method is proposed considering different penetration levels of renewable energy. A difference Teager-Kaiser energy operator (dTKEO)-based algorithm is first proposed to improve multiple-events detection accuracy. Then, feature extractions via the integrated additive angular margin (AAM) loss and the long short-term memory (LSTM) network are developed. This allows us to deal with intraclass similarity and inter-class variance of events when high penetration renewable energy occurs. With the extracted features, a multi-stage weighted summing (MSWS) lossbased criterion is developed for adaptive data window determination and fast event pre-classification. Finally, the re-identification model based on feature similarity is established to identify unknown events, a challenge for existing machine learning algorithms. Simulation results on the IEEE 39-bus, Kundur 2-area, and an actual large-scale power grid system are used to demonstrate the advantages of the proposed method over others.
? Although onshore wind turbines have reached maturity due to the sooner evolution and, currently, most of the existing wind farms are onshore, offshore harvesting has drawn attention in the past years as well due to its desirable features.
? Forecasting models based on deep learning are able to overcome the barriers and limitations of the existing statistical forecasting models, which are mostly formulated as a linear model in dealing with longer forecasting time horizons.
? Among other existing FCs, Proton Exchange Membrane Fuel Cell (PEMFC), Direct Methanol Fuel Cell (DMFC), Alkaline Fuel Cells (AFC), and Solid Oxide Fuel Cells (SOFC) are the most known ones.
? To handle the intra-class similarity caused by DERs integration, the diffusion kernel density estimation (DKDE) and a deep neural network (DNN) are combined .
? However, it does not handle the intra-class similarity issue with the increased penetration of DERs.
? Furthermore, existing algorithms do not address unknown events.It is worth noting that this paper mainly focuses on the large disturbance events detection and identification, such as faults, generation loss, or load switch-off/on, etc. The remainder of this paper is organized as follows shows the problem formulations.
• In , a wind power trading model, including a regulation strategy, is proposed for the day-ahead California electricity market.
• Moreover, a state-space dynamic model is proposed in for PV systems with full ancillary services support.
• In , a reactive power regulation method is suggested to provide ancillary services by PV inverters using three-phase control strategies.
• The authors of proposed a new Repeated Wavelet Transform (WT) based Auto Regressive Integrated Moving Average (ARIMA) (WT-ARIMA) model to predict the wind speed on very short time intervals.
? A residual CNN is proposed in which improves the network training convergence speed and detection accuracy.
? In , maximum overlap discrete wavelet transform and stacked sparse denoising autoencoder are combined to improve the anti-noise performance in disturbance detection.
? However, the Softmax loss only learns separable features that are not sufficiently discriminative, which may result in poor performance for some cases, such as the events with only slight differences.
? The overall accuracy of the test data is 97.99%, which indicates that the proposed method achieves good performances for event identification with renewable energy integrations
|