论文标题
部分可观测时空混沌系统的无模型预测
What are the spectroscopic binaries with high mass functions near the Gaia DR3 main sequence?
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。
The 3rd data release of the Gaia mission includes orbital solutions for $> 10^5$ single-lined spectroscopic binaries, representing more than an order of magnitude increase in sample size over all previous studies. This dataset is a treasure trove for searches for quiescent black hole + normal star binaries. We investigate one population of black hole candidate binaries highlighted in the data release: sources near the main sequence in the color-magnitude diagram (CMD) with dynamically-inferred companion masses $M_2$ larger than the CMD-inferred mass of the luminous star. We model light curves, spectral energy distributions, and archival spectra of the 14 such objects in DR3 with high-significance orbital solutions and inferred $M_2 > 3\,M_{\odot}$. We find that 100\% of these sources are mass-transfer binaries containing a highly stripped lower giant donor ($0.2 \lesssim M/M_{\odot} \lesssim 0.4$) and much more massive ($2 \lesssim M/M_{\odot} \lesssim 2.5$) main-sequence accretor. The Gaia orbital solutions are for the donors, which contribute about half the light in the Gaia RVS bandpass but only $\lesssim 20\%$ in the $g-$band. The accretors' broad spectral features likely prevented the sources from being classified as double-lined. The donors are all close to Roche lobe-filling ($R/R_{\rm Roche\,lobe}>0.8$), but modeling suggests that a majority are detached ($R/R_{\rm Roche\,lobe}<1$). Binary evolution models predict that these systems will soon become detached helium white dwarf + main sequence "EL CVn" binaries. Our investigation highlights both the power of Gaia data for selecting interesting sub-populations of binaries and the ways in which binary evolution can bamboozle standard CMD-based stellar mass estimates.