In this work, a novel nonlinear approach is proposed for the stabilization of microgrids (MGs) with constant power loads (CPLs). The proposed method is constructed based on the incorporation of a pseudo-extended Kalman filter (EKF) into stochastic nonlinear model predictive control (MPC). In order to achieve high-performance and optimal control in dc MGs, estimating the instantaneous power flow of the uncertain CPLs and the available power units is essential. Thus, by utilizing the advantages of the stochastic MPC and the pseudo-EKF, an effective control solution for the stabilization of dc islanded MGs with CPLs is established. This technique develops a constrained controller for practical application to handle the states and control input constraints explicitly; furthermore, as it estimates the current by using the pseudo-EKF, it is a current-senseless approach. As noisy measurements are taken into account for the state estimation, it leads to a less conservative control action rather than the classical robust MPC, whereas it guarantees the global asymptotic stability in the presence of noisy measurements and parameter uncertainty. To validate the performance of the proposed controller, the attained results are compared with state-of-the-art controllers. Furthermore, the implementability of the proposed method is validated using real-time simulations on dSPACE hardware.

中文翻译：

---

用于DC微电网馈电CPL的新型随机预测稳定剂

在这项工作中，提出了一种新颖的非线性方法，用于稳定具有恒定功率负载（CPL）的微电网（MG）。该方法是基于将伪扩展卡尔曼滤波器（EKF）合并到随机非线性模型预测控制（MPC）中而构造的。为了在直流MG中实现高性能和最佳控制，估计不确定CPL和可用功率单元的瞬时功率流至关重要。因此，通过利用随机MPC和伪EKF的优势，建立了使用CPL稳定直流孤岛MG的有效控制解决方案。该技术为实际应用开发了一种约束控制器，以处理状态并显式控制输入约束。此外，由于它使用伪EKF估算电流，这是一种毫无意义的方法。由于将噪声测量值用于状态估计，因此与传统的鲁棒MPC相比，它导致的控制动作较不保守，而在存在噪声测量值和参数不确定性的情况下，它可以保证全局渐近稳定性。为了验证所提出的控制器的性能，将获得的结果与最新的控制器进行比较。此外，使用dSPACE硬件上的实时仿真验证了所提出方法的可实施性。为了验证所提出的控制器的性能，将获得的结果与最新的控制器进行比较。此外，使用dSPACE硬件上的实时仿真验证了所提出方法的可实施性。为了验证所提出的控制器的性能，将获得的结果与最新的控制器进行比较。此外，使用dSPACE硬件上的实时仿真验证了所提出方法的可实施性。