直流电压下两相体放电路径预测（英文）

Prediction of the Discharge Paths of Two-Phase Mixtures under DC Voltage

通过实验研究了两相体放电路径的预测问题,结果表明:放电路径中选择空气或两相体由被畸变的电场决定,而电场的畸变受两相体颗粒粒径大小的影响。为了解释实验现象,利用传统的流注理论和概论统计理论,以泊松方程求解的空间场强为流注发展的判据,并假设流注发展的击穿时间满足Weibull分布,将两相体空间电场畸变后电场值的变化决定流注的发展方向,建立了正极性的放电路径选择的物理模型。将直流电压下两相体放电路径发展问题的目标函数(即最短路径)与连续性Hopfield神经网络的能量函数相对应,将经过的节点顺序(局部电场值的影响大小)与网络的神经元状态相对应,此时对应的节点发展顺序就是待求的最佳路线。仿真和实验结果比较显示,基于该模型两相体直流放电路径选择概率分布的计算结果与实验所得规律一致。

The prediction of the Discharge Paths of Two-phase Mixtures under DC Voltage has been investigated by the experiment. From the experimental results the probability of selecting the air or the two-phase mixtures is governed by the local distorted electric field,which is highly correlated with the macroparticle sizes. In order to explain the phenomena,a new stochastic model has been put up. On the basis of the streamer theory and probability and statistics theory,the paper uses the strength of an electric field from the Poisson's equation as the criterion of the streamer development. The breakdown time of the streamer development meets Weibull distribution and the distorted field value decides the direction for the streamer development in the TPMD space. The development of the DC discharge path can be affected by the local electric field. The objective function( optimal path) corresponds to the energy function of the Hopfield neural network. The comparison of the simulation with the experiment shows that the stochastic model has given a good approximation.