多端直流电网架空线路故障概率高、设备耐受过电流能力有限,无选择性的重合将严重危害一次设备的安全性和电网的稳定性,因此提出了一种基于动态时间规整(dynamic time warping, DTW)的重构序列重合方法,实现重合前故障性质的预判。首先...多端直流电网架空线路故障概率高、设备耐受过电流能力有限,无选择性的重合将严重危害一次设备的安全性和电网的稳定性,因此提出了一种基于动态时间规整(dynamic time warping, DTW)的重构序列重合方法,实现重合前故障性质的预判。首先介绍了直流线路的贝杰龙时域等值建模方法,在此基础上,建立了各关键一次设备的贝杰龙等值模型,分析了瞬时性、永久性故障工况下模型等效时域波过程的差异性。然后以瞬时性故障为基准,对直流断路器注入信号所引发的电气分量暂态成分进行理论计算,并借助DTW重构序列量化整定门槛、进一步消除色散误差,最终以此为指标构造判据实现故障性质的识别。基于PSCAD/EMTDC平台的四端直流电网仿真算例显示,所提出的自适应重合方法仅需要单端量,不依赖通信,能够消除死区,且拥有600Ω的灵敏性。展开更多
This paper presents a new method for the calculation of the transient voltage distribution over a transformer winding——the Bergeron's method. Using the nodal admittance matrix, formulae are developed for the re...This paper presents a new method for the calculation of the transient voltage distribution over a transformer winding——the Bergeron's method. Using the nodal admittance matrix, formulae are developed for the resistances, and self and mutual inductances of the equivalent circuit. Calculated results are given. The method reduces internal memory requirement, allows easy formation of the nodal admittance matrix, reduces computational time, etc.展开更多
文摘多端直流电网架空线路故障概率高、设备耐受过电流能力有限,无选择性的重合将严重危害一次设备的安全性和电网的稳定性,因此提出了一种基于动态时间规整(dynamic time warping, DTW)的重构序列重合方法,实现重合前故障性质的预判。首先介绍了直流线路的贝杰龙时域等值建模方法,在此基础上,建立了各关键一次设备的贝杰龙等值模型,分析了瞬时性、永久性故障工况下模型等效时域波过程的差异性。然后以瞬时性故障为基准,对直流断路器注入信号所引发的电气分量暂态成分进行理论计算,并借助DTW重构序列量化整定门槛、进一步消除色散误差,最终以此为指标构造判据实现故障性质的识别。基于PSCAD/EMTDC平台的四端直流电网仿真算例显示,所提出的自适应重合方法仅需要单端量,不依赖通信,能够消除死区,且拥有600Ω的灵敏性。
基金国家自然科学基金(50907036)国家重点基础研究发展计划(973计划)(2011CB209403)+1 种基金Project supported by National Natural Science Foundation of China(50907036)National Basic Research Program of China(973 Program)(2011CB209403)
文摘This paper presents a new method for the calculation of the transient voltage distribution over a transformer winding——the Bergeron's method. Using the nodal admittance matrix, formulae are developed for the resistances, and self and mutual inductances of the equivalent circuit. Calculated results are given. The method reduces internal memory requirement, allows easy formation of the nodal admittance matrix, reduces computational time, etc.
