A Machine Learning Approach for Fast Future Grid Small-signal Stability Scanning

In this paper, we propose a novel fast scanning approach to perform small signal stability study of future power systems with high penetration of renewable generation. Stability assessment is an important component of power system planning. Due to generation technology diversity and inherent intermittent availability of most renewable sources in future grids, the conventional method to conduct stability analysis based on choosing a limited number of worst case operating points becomes infeasible. One way to capture the stability profile of a future grid scenario is to scan a large number of possible operating conditions. However, to achieve fast scanning and make the time consuming numerical study computationally affordable, the simulation burden has to be reduced. To that end, we propose a fast scanning approach based on feature selection and weighted clustering to reduce simulation burden when conducting stability scanning over a long period of time. We propose a novel algorithm based on conventional Relief-F and K-means techniques. Simulation results show the effectiveness of the proposed approach.