Deep Learning for Dynamic Security Assessment of Power Systems with Adaptive Synthetic Sampling-Based Imbalanced Database: A Case Study

Recently, deep learning-based techniques in dynamic security assessment (DSA) have exhibited significant advancements, assuming a pivotal role in ensuring the secure operation of smart grids. However, imbalanced samples are a fundamental challenge for effective training of data-driven methods. In the DSA problem, especially in real-world power systems, the database is usually imbalanced and the number of secure cases is more than the number of insecure cases. This imbalance can lead to a loss of fit and generalization in insecure cases since the DSA model tends to focus too much on secure cases. The aim of this chapter is to address the data imbalance in DSA using the adaptive synthetic sampling (ADASYN) method to generate synthetic data resembling the original data. After addressing the data imbalance, this chapter presents a DSA model based on a long short-term memory (LSTM) network. The proposed model was implemented and tested on the IEEE 39 bus system. The test results show that solving the data imbalance problem has improved the performance of the proposed DSA model.