高级检索

基于XGBoost决策优化网络的矿井35kV供电防越级跳闸技术研究

Research on the Technology of Preventing Over-level Trip of 35kV Power Supply in Mines Based on XGBoost Decision Optimization Network

  • 摘要: 矿井35kV供电网络负荷波动、短路电流多级传递条件下易越级跳闸,传统整定依赖经验。提出XGBoost决策优化网络,以地面35kV变电所—井下中央变电所—采区配电点三级供电链为对象,构建电流突变量、零序分量、动作时差、负荷率、通信时延和CT状态特征集,联合保护定值复核、CT/VT改造和监控闭环防治。越级跳闸次数由26次降至10次,降幅61.5%;平均动作时间缩短约8.4%,高负荷区间越级跳闸概率下降。XGBoost决策优化网络提升保护配合精度与动作选择性。

     

    Abstract: To address over-tripping in mine 35 kV power supply networks caused by load fluctuations, multistage short-circuit current transfer, and experience-based relay settings, an XGBoost-based decision optimization network was proposed. Using the three-level supply chain from the ground substation to the underground central substation and mining distribution point, key features were extracted for coordinated protection. Over-tripping incidents fell from 26 to 10, and average operating time decreased by 8.4%, demonstrating improved coordination accuracy, selectivity, and engineering value.

     

/

返回文章
返回