基于混合粒子滤波的电力设备载流故障预测

Current-carrying fault prediction of electric equipment based on hybrid particle filtering

为有效预防设备载流故障,提高设备稳定运行能力和延长使用寿命,提出基于混合粒子滤波的滚动式电力设备载流故障趋势预测方法。首先,借助传热学理论,分析电力设备触点在载流故障状态下的温度变化,建立触点温升模型;然后,利用降维思想,根据温升模型给定参数估计方程,初始化温升模型参数。最后,基于在线更新的温度,利用粒子滤波修正模型参数,滚动式优化温度预测曲线,实现载流故障发展趋势的精确预测。基于某电站实际温度数据的验证表明,所提方法能够快速精准地预测载流故障的发展趋势,有效保护电力系统安全运行。

In order to effectively predict current carrying fault of electric equipment, enhance equipment operation sta- bility,and prolong equipment service life time, this paper proposes a current-carrying fault trend rolling prediction method based on hybrid particle filtering. Firstly, under the guidance of heat transfer theory, the temperature change of electric equipment contacts under current-carrying fault condition is analyzed and the temperature rising model of the contacts is built. Then, with the idea of dimension reduction, the parameters of the temperature rising model are initial- ized based on the parameter estimation equation. At last ,based on the real-time data ,particle filtering is used to update the parameters of the temperature rising model, and the temperature prediction curve is optimized. The above method was verified with the actual temperature data of a certain power station, and the result shows that the method can pre- dict the development trend of current-carrvin~ fault accurately and protect the safetv of the nower svstem effectiwlv_