论文部分内容阅读
针对现代电力系统所呈现的强烈非线性特征,在工作点附近进行电力系统模型线性化方式效果不佳,因此提出一种基于ESD阻尼控制的深度学习网络启发式动态规划(DNNs-HDP)振荡模态识别控制方法。首先,针对双区四电机基准系统,基于残差方式并借鉴粒子群算法振荡阻尼控制器设计过程,实现振荡模态识别控制框架设计。同时针对粒子群算法在进行控制参数优化过程中易于早熟收敛问题,利用深度学习算法构建启发式动态规划算法,并结合储能装置,对双区四电机基准系统进行控制优化设计。仿真结果显示,所提算法相对传统残差方式和粒子群算法设计的振荡阻尼控制器,控制效果更佳。
In view of the strong nonlinear characteristics of modern power system, it is ineffective to linearize the power system model near the operating point. Therefore, a DNNs-HDP oscillation mode based on ESD damping control is proposed State recognition control method. First of all, for the two-zone four-motor reference system, based on the residual method and using particle swarm optimization to design the oscillation damping controller, the control framework of oscillation mode identification is designed. At the same time, aiming at the premature convergence problem of particle swarm optimization algorithm in optimization of control parameters, a heuristic dynamic programming algorithm based on depth learning algorithm is constructed. Combined with energy storage device, the control design of dual-zone four-motor reference system is optimized. The simulation results show that the proposed algorithm has better control effect than the traditional residual method and the oscillation damping controller designed by the particle swarm optimization algorithm.