地磁扰动对煤矿电子设备安全性的影响及其预测方法

Effects of Geomagnetic Disturbances on the Safety of Electronic Equipment in Coal Mines and their Prediction Methods

摘要: 随着煤矿无人化、智能化进程不断推进，电子设备在煤矿生产中的应用日益广泛，成为提高生产效率和保障人员安全的关键因素。然而，太阳风及其引发的地磁扰动对电子设备的安全运行构成了潜在威胁。太阳风Kp值作为衡量地磁扰动强度的关键指标，其预测的方法及准确性对于煤矿电子设备的安全性具有至关重要的意义。多层前馈神经网络模型用于预测Kp指数。该网络使用从太阳周期中提取的数据序列，然后通过输入已知的太阳风数据预测三小时后的Kp值。行星际磁场的Bz分量、太阳风密度n和太阳风速度V (一小时的平均值)是网络的输入参数。该网络采用误差反向传播算法，使用1982年至1992年间提取的数据序列进行训练，并使用遗传算法对网络进行优化，得出的模型的预测Kp值与实际Kp值均方根误差为0.9425 (三小时平均数据)，相关性为0.778 (三小时平均数据)。

Abstract: With the continuous advancement of unmanned and intelligent processes in coal mines, electronic equipment is increasingly widely used in coal mine production, becoming a key factor in improving production efficiency and ensuring personnel safety. However, solar wind and the geomagnetic perturbation it triggers pose a potential threat to the safe operation of electronic equipment. The method and accuracy of predicting the solar wind Kp value, as a key indicator of the intensity of geomagnetic perturbations, are of crucial significance for the safety of electronic equipment in coal mines. A multilayer feedforward neural network model is used to predict the Kp index. The network uses a sequence of data extracted from the solar cycle and then predicts the Kp value after three hours by inputting known solar wind data. The Bz component of the interplanetary magnetic field, the solar wind density n, and the solar wind velocity V (averaged over one hour) are the input parameters to the network. The network was trained using an error back-propagation algorithm using data series extracted between 1982 and 1992, and the network was optimized using a genetic algorithm, resulting in a model with a root mean square error of predicted Kp values versus actual Kp values of 0.9425 (for three-hourly averaged data) and a correlation of 0.778 (for three-hourly averaged data).