The automatic generation control mechanism in power generators comes into operation whenever an over-supply or under-supply of energy occurs in the power grid. It has been shown that the automatic generation control mechanism is highly vulnerable to load altering attacks. In this type of attack, the power consumption of multiple electric loads in power distribution systems is remotely altered by cyber attackers in such a way that the automatic generation control mechanism is disrupted and is hindered from performing its pivotal role. The existing literature on load altering attacks has studied implementation, detection, and location identification of these attacks. However, no prior work has ever studied design of an attack-thwarting system that can counter load altering attacks, once they are detected in the power grid. This paper addresses the above shortcoming by proposing an attack-thwarting system for countering load altering attacks. The proposed system is based on provoking real-time adjustment in power consumption of the flexible loads in response to the frequency disturbances caused by the load altering attacks. To make the adjustments in-proportion to the frequency disturbances, the proposed attack-thwarting system uses a transactive energy framework to establish a coordination between the flexible loads and the power grid operator.
Distributed deep reinforcement learning for integrated generation‐control and power‐dispatch of interconnected power grid with various renewable units
