Power system voltage control by multiple STATCOMs based on learning humoral immune response

Abstract

An approach for power system voltage control is proposed based on the learning humoral immune response which can not only fulfil the conventional task of a voltage controller in a power system to keep the voltage at the controller's busbar to a setting reference value at the normal operation mode, but also minimise or eliminate voltage violations at its adjacent busbars in power system contingencies. This dual voltage control function is the major advantage of the approach proposed and is achieved by the capability of the humoral immune voltage control to identify self and nonself antigen intrusion (voltage variations and violations) to the power system. It is proposed that the parameters of a humoral immune voltage controller are trained by learning algorithms via offline simulation. The training is assisted by the injections of artificial reactive loads at relevant busbars in the power system and by the provision of a learning reference signal. An example 10-machine 39-node power system installed with two STATCOMs (static synchronous compensators) is presented. The success of applying the proposed humoral immune voltage control and associated learning algorithms to the design of STATCOMs is demonstrated.