论文标题

部分可观测时空混沌系统的无模型预测

Probing thermalization and dynamics of high-energy quasiparticles in a superconducting nanowire by scanning critical current microscopy

论文作者

Jalabert, T., Driessen, E. F . C., Gustavo, F., Thomassin, J. L., Levy-Bertrand, F., Chapelier, C.

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。

Besides its fundamental interest, understanding the dynamics of pair breaking in superconducting nanostructures is a central issue to optimize the performances of superconducting devices such as qubits or photon detectors. However, despite substantial research efforts, these dynamics are still not well understood as this requires experiments in which quasiparticles are injected in a controlled fashion. Until now, such experiments have employed solid-state tunnel junctions with a fixed tunnel barrier. Here we use instead a cryogenic scanning tunnelling microscope to tune independently the energy and the rate of quasiparticle injection through, respectively, the bias voltage and the tunnelling current. For high energy quasiparticles, we observe the reduction of the critical current of a nanowire and show it is mainly controlled by the injected power and, marginally, by the injection rate. Our results prove a thermal mechanism for the reduction of the critical current and unveil the rapid dynamics of the generated hot spot.

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