ANN Control for Damping Low-frequency Oscillation using Deep learning

If the Low-frequency oscillations in power systems are not properly damped out, they can cause critical effects, including wide-area outage. In 1996, Low-frequency oscillation incurred a wide-area outage and the economic, social loss from this phenomenon was estimated at 3 billion dollars. In order to damp out these oscillations, the PSS (Power System Stabilizer) is typically used. This, however, cannot damp out inter-area oscillations. Although research on damping inter-area oscillations is conducted to resolve the mentioned problems, the Lead Lag Controller for damping the oscillations used in FACTS (Flexible AC Transmission System) devices including VSC (Voltage Source Converter) has limitations in that it is a linear controller designed at a particular power system operating condition and there are some limitations incurred by the variable and unpredictable power system states, so that fixed controller parameters cannot properly damp out these oscillations. Meanwhile, the AI (Artificial Intelligence) techniques including supervised learning, unsupervised learning, and reinforcement learning are applied in various engineering fields in order to overcome a lot of nonlinear problems like those of power systems. In this paper, the ANN (Artificial Neural Network) controller, a kind of AI controller, was used to damp out inter-area oscillation at different power system operating conditions, after which the results were analyzed.