Adaptive Neural Data-Driven Learning Load Frequency Control for Nonlinear Power Systems with Repetitive Electric Loads

Energy-intensive (EI) enterprise power system is an important part of energy interconnection networks, and their short-term industrial electricity loads exhibit a typical cyclic behavior. In order to address the problem of frequency stability for EI enterprise power systems with the unknown model, a load frequency control (LFC) scheme based on an adaptive neural data-driven learning algorithm is proposed. Firstly, an equivalent iterative dynamic data model between frequency deviation and input/output information is established. Subsequently, a linear parameterized learning controller is developed based on dynamic linearization technology, which reduces the dependence of controller design on accurate models. Meanwhile, the pseudo-partial learning gain is continuously and adaptively updated by an RBF neural network approximator. In addition, the convergence of the system is proved by the Lyapunov stability theory. Finally, a simulation example is given.