LEs based framework for transient instability prediction and mitigation using PMU data

Abstract

This study presents a Lyapunov exponent (LE)-based framework for transient instability prediction and mitigation using wide area measurement systems data. For transient instability prediction, maximum LEs calculated using post-fault voltage magnitudes are used as input features to support vector machines to predict the transient stability status. To ensure that the proposed transient instability prediction approach does not rely on full network observability and is effective even if limited numbers of phasor measurement units (PMUs) are installed across large power grids, an optimal PMU placement from the transient stability point of view is also proposed. The robustness of the proposed instability prediction scheme is verified with respect to network topology change and measurement errors. For transient instability mitigation, an LE-based controlled islanding scheme is proposed. The proposed controlled islanding scheme determines the coherent generators, identifies proper load buses that should remain inside each group of coherent generators, specifies the boundary circuits and finally partitions the system into different islands while placing the coherent generators in the same island, maintaining connectivity inside islands and minimising the generation–load imbalance of islands.