论文标题

部分可观测时空混沌系统的无模型预测

The ASTAROTH project

论文作者

D'Angelo, Davide, Zani, Andrea, Alessandria, Franco, Andreani, Alessandro, Castoldi, Andrea, Coelli, Simone, Cortis, Daniele, Di Carlo, Giuseppe, Frontini, Luca, Gallice, Niccolò, Guazzoni, Chiara, Liberali, Valentino, Monti, Mauro, Orlandi, Donato, Sorbi, Massimo, Stabile, Alberto, Statera, Marco

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 most discussed topic in direct search for dark matter is arguably the verification of the DAMA claim. In fact, the observed annual modulation of the signal rate in an array of NaI(Tl) detectors can be interpreted as the awaited signature of dark matter interaction. Several experimental groups are currently engaged in the attempt to verify such a game-changing claim with the same target material. However, all present-day designs are based on a light readout via Photomultiplier Tubes, whose high noise makes it challenging to achieve a low background in the 1-6 keV energy region of the signal. Even harder it would be to break below 1 keV energy threshold, where a large fraction of the signal potentially awaits to be uncovered. ASTAROTH is an R\&D project to overcome these limitations by using Silicon Photomultipliers (SiPM) matrices to collect scintillation light from NaI(Tl). The all-active design based on cubic crystals is operating in the 87-150 K temperature range where SiPM noise can be even a hundred times lower with respect to PMTs. The cryostat was developed following an innovative design and is based on a copper chamber immersed in a liquid argon bath that can be instrumented as a veto detector. We have characterized separately the crystal and the SiPM response at low temperature and we have proceeded to the first operation of a NaI(Tl) crystal read by SiPM in cryogeny.

扫码加入交流群

加入微信交流群

微信交流群二维码

扫码加入学术交流群,获取更多资源