论文标题
部分可观测时空混沌系统的无模型预测
Discovery of double BSS sequences in the old Galactic open cluster Berkeley 17
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。
Blue straggler stars (BSS) are peculiar objects which normally appear as a single broad sequence along the extension of the main sequence. Only four globular clusters (GCs) have been observed to have two distinct and parallel BSS sequences. For the first time for any open cluster (OC), we report double BSS sequences in Berkeley 17. Using the machine-learning based membership algorithm ML-MOC on Gaia EDR3 data, we identify 627 cluster members, including 21 BSS candidates out to 15 arcmin from the cluster center. Both the BSS sequences are almost equally populated and parallel to one another in Gaia as well as in Pan-STARRS colour-magnitude diagram (CMD). We statistically confirm their presence and report that both BSS sequences are highly segregated compared to the reference population out to $\sim$5.5 arcmin and not segregated thereafter. The lower densities of OCs make BSS formation impossible via the collisional channel. Therefore, mass transfer seems to be the only viable channel for forming candidates of both sequences. The gap between the red and blue BSS sequences, on the other hand, is significant and presents a great opportunity to understand the connection between BSS formation and internal as well as external dynamics of the parent clusters.
