Dynamic Self-Triggered Adaptive Control for Voltage Regulation of DC Microgrids

In this brief, a novel online near-optimal control scheme is investigated for the voltage tracking control problem of direct current (DC) microgrids under the dynamic self-triggered (DST) mechanism. First, the DC microgrid is modeled as an affine coupled system. The nonzero-sum game issue of the multi-input system is taken into account, where each distributed generator (DG) strives to minimize the individual performance index function and ensure the stability of the whole system. Subsequently, the value function with the non-quadratic utility function is established to seek the optimal constrained control pair. The critic neural network (NN) is devised to approximate the optimal value function. By virtue of experience replay technique, the persistent excitation condition is no more needed. What is more, dynamic event-triggered (DET) control can significantly reduce the waste of computation and communication resources by avoiding the redundant triggers. However, the continuous detection of the DET condition is dependent on dedicated hardware. To overcome the difficulties in hardware realization of DET control, a novel DST method with the dead-zone operation is proposed, in which design the next triggering instant is actively calculated by current data. Besides, the stability of the system and the minimum trigger interval are guaranteed. Finally, a simulation example validates the effectiveness of the algorithm.