Distributed Model Predictive Load Frequency Control for Virtual Power Plants with Novel Event-Based Low-Delay Technique Under Cloud-Edge-Terminal Framework

In this paper, the distributed model predictive load frequency control problem for virtual power plants (VPPs) under the cloud-edge-terminal framework is addressed, where the data packets are transmitted under a novel dynamic event-triggered mechanism (DETM) with hybrid variables. The proposed DETM has the ability to flexibly manage packet releases and reduce network congestion, thus decreasing the communication delay of the VPP. A method of the DETM-based distributed model predictive control (DMPC) is proposed, which can shorten the data processing time and further decrease the communication delay. The DMPC problem is described as a “min-max” optimization problem (OP) with hard constraints on the system state. By utilizing a Lyapunov function with an internal dynamic variable, an auxiliary OP with matrix inequalities constraints is proposed to optimize the controller gain and the weighting matrix of the DETM. The effectiveness and superiority of the designed DETM and dynamic event-based DMPC algorithm are demonstrated through a case study on two-area VPPs.