Sparse Dictionary Learning for Transient Stability Assessment

Transient stability assessment (TSA) has always been a fundamental and challenging problem for ensuring the security operation of power systems. With more electronic interface resources integrated into grid large renewable energies, system is jeopardized. Therefore, TSA should be considered in advance to keep running stable. In recent years, with development artificial intelligence (AI) technologies such as neural network (ANN), support vector machine (SVM), Markov decision process, improved dramatically. this study, sparse dictionary learning approach proposed improve precision classification accuracy transient Case studies using multi-layer (ML-SVM) long short-term memory network–based recurrent (LSTM-RNN) are discussed benchmarks validate method. The stable unstable learnings designed based on datasets obtained by simulating thousands different time-domain simulation (TDS) scenarios performed New-England 39-bus PSAT (power analysis toolbox) toolbox. Stable dictionaries developed K-SVD approach. testing dataset contains both samples which steps coding process obtain indexes. Compared indexes, system’s final targeted. method exhibits satisfactory prediction provides ability reduce false alarms positives negatives system.