论文标题
部分可观测时空混沌系统的无模型预测
Data Driven Transfer Functions and Transmission Network Parameters for GIC Modelling
论文作者
论文摘要
典型的地磁诱导电流（GIC）建模假设传输网络中的节点处的诱导的Quasi-DC电流与一对网络参数与局部地理电场线性相关。给定测量的地磁和GIC数据的时间序列有限，提出了一种经验方法，该方法导致参数估计值的统计学意义，并在所识别的估计值中误差。该方法用于不同的传输网络和地磁风暴，并且在存在之前的建模时显示了GIC估计的改进。此外，在没有任何其他网络知识的情况下，可以在本地对建模的网络进行局部表征和探测。见解包括网络参数变化，有效的网络方向性和响应。合并网络参数和地球场估计，得出了传输函数，该函数提供了一种评估变形金刚暴露于GIC的替代方法。
Typical geomagnetically induced current (GIC) modelling assumes the induced quasi-DC current at a node in the transmission network is linearly related to the local geoelectric field by a pair of network parameters. Given a limited time-series of measured geomagnetic and GIC data, an empirical method is presented that results in a statistically significant generalised ensemble of parameter estimates with the error in the estimates identified. The method is showcased for different transmission networks and geomagnetic storms and, where prior modelling exists, shows improved GIC estimation. Furthermore, modelled networks can be locally characterised and probed without any further network knowledge. Insights include network parameter variation, effective network directionality and response. Merging the network parameters and geoelectric field estimation, a transfer function is derived which offers an alternative approach to assessing transformer exposure to GICs.
