论文标题
部分可观测时空混沌系统的无模型预测
Correlation effects in Barkhausen noise and magnetic attenuation in soft ribbons
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 propagation of the effects connected with the occurrence of magnetization reversals in an amorphous ribbon of Fe$_{63}$B$_{64}$Si$_{8}$Ni$_{15}$ has been investigated using a method based on two pickup coils separated by a distance variable between 2 an 40 mm. The ratio $R$ between the voltage signals induced in the coils contains information on the location where the magnetization reversal took place. This information can be extracted by knowing the function $V(x)$ which represents the attenuation of a signal generated by a reversal that took place at a distance x from the coil. A mathematical model for extracting this function starting from the histogram of the experimentally measured values of $R$ is presented. The attenuation function obtained in this way is relatively independent on the distance between the coils, and this is in strong support of the correctness of the model adopted.
