基于粒子群优化的支持向量机在瓦斯浓度预测中的应用研究

Analysis of Mine Low-voltage Leakage Protection in Situation of Great Disparity between Long and Short Lines and Big Unbalanced Current

  • 摘要: 为了准确预测煤矿瓦斯浓度,基于从芦岭煤矿KJ98监控系统中提取的生产现场瓦斯浓度时间序列数据,对基于粒子群优化的支持向量机理论在瓦斯浓度短期预测中的应用进行了研究。首先对瓦斯浓度时间序列进行小波软阈值去噪和相空间重构等预处理,然后采用粒子群优化算法对支持向量机的惩罚因子、损失函数、核函数参数进行了优化,并基于最优参数建立了瓦斯浓度预测的支持向量机模型。仿真结果表明,采用粒子群优化的支持向量机理论进行煤矿瓦斯浓度预测,极大地提高了预测的准确性和精确度;误差分析结果表明,该方法预测结果的误差很小,且测试样本越小,误差越小。

     

    Abstract: The paper explained that traditional leakage protection principles such as zero-sequence direction relay principle and zero-sequence current's maximum principle are unable to implement selective leakage protection in situation of great disparity between long and short lines and big unbalanced current through analyzing an actual example, and proposed a scheme of selective leakage protection in the situation by use of zero-sequence current's break variable principle. The scheme judges leakage by comparing zero-sequence current with inherent capacitive current of each power line as follows: if zero-sequence current of one power line changes suddenly compared to inherent capacitive current, then the power line would be leakage line; if zero-sequence current of non power line changes suddenly compared to inherent capacitive currents, then main line leakage happens. The application showed the validity of the zero-sequence current's break variable principle.

     

/

返回文章
返回