RESUMO
The application of virtual inertia (VI) has become a novel solution and new orientation for the frequency regulation of the high variable renewable energy (VRE) penetrated low inertia systems. How to provide a frequency regulation strategy for the system with VI is a key problem. Firstly, the power system is a multi inputs multi outputs (MIMO) system, in which the coupling characteristics of VI control and the primary and secondary frequency controls can't be ignored. Secondly, the system inertia constant is time-varying and the estimation of it is crucial for gaining a satisfying regulation performance. This paper intends to provide a frequency regulation strategy for this problem. In this work, an Elman neural network (ENN) based inertia estimation with a deadband method is proposed for the identification of the system's instant inertia. And the Laguerre function-based rate of change of frequency (ROCOF) constrained MIMO-MPC is constructed to provide the collaboration strategy. Then, the control parameters are tuned by the gravitational search algorithm (GSA) considering both the system performance indexes and the operation burden indexes. The proposed strategy is testified based on a three-area power system. The results prove not only the robustness of the controller but also the effectiveness of the ROCOF constraint, the operation burden index, and the instant inertia estimation method.