论文标题
部分可观测时空混沌系统的无模型预测
Understanding Impedance Ratio Criteria for Converter-Based AC Power System
论文作者
论文摘要
储层计算是预测湍流的有力工具，其简单的架构具有处理大型系统的计算效率。然而，其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间，然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用，该系统模拟了湍流的若干特征。我们表明，使用径向基函数作为非线性投影器，即使只有部分观测并且不知道控制方程，也能稳健地捕捉复杂的系统非线性。最后，我们表明，当测量稀疏、不完整且带有噪声，甚至控制方程变得不准确时，我们的网络仍然可以产生相当准确的预测，从而为实际湍流系统的无模型预测铺平了道路。
Nyquist criterion-based impedance ratio criteria (IRCs) have been widely applied for inspecting the risk of small-signal instability among converter-based AC power systems. Aided by a comparative study on voltage source converter, including the single-input single-output (SISO) and multiple input multiple output (MIMO) analyses in both the dq and the sequence domain, two aspects are emphasized in this paper: 1) the sufficiency of SISO analysis when the mapping function (MF) is observable to potentially unstable modes, and 2) the inconvenience of IRCs with an unintended right-half plane pole emergence of MF due to the source-load partition. The strictness of analyses is proved by a systematical deduction of explicit analytical impedance models using the state space. Moreover, a novel criterion that relies on the logarithmic derivative of MFs is proposed, which can identify the system modes directly, serve as an alternative to IRCs, and be extended to other transfer function-based stability analyses.
