基于电流突变比和支持向量机的TN-C-S系统漏电故障二级判别方法

Two-stage leakage protection method of TN-C-S system based on ratio of current mutations and support vector machine

针对低压配电网TN-C-S接地系统固有剩余电流大,台区总保护和中级保护无法投运,线路漏电故障缺乏保护的问题,提出一种基于剩余电流-不平衡电流突变向量之比和粒子群优化支持向量机的漏电故障二级判别方法。首先,推导了TN-C-S接地系统漏电故障前后,剩余电流-不平衡电流突变比与零线阻抗、接地阻抗之比的关系,提出以突变比幅值为判据的故障一级判别方法。然后,以剩余电流、不平衡电流突变量的基波、3次谐波作为特征量,利用粒子群优化最小二乘支持向量机对判定结果进行二级判别,实现故障的可靠识别。最后,实际测试结果表明,在固有剩余电流幅值高于故障电流的情况下,所提方法的总识别率可达80%以上。所提方法计算量较小、成本低,使TN-C-S系统总保护和中级保护可靠投运成为可能。

Aiming at the problem that the general protection and intermediate protection cannot be put into operation in the low voltage TN-C-S systems, due to the large inherent residual current, this paper proposes a two-stage leakage protection method using the ratio of the residual and unbalanced current mutations and the particle swarm optimization support vector machine(PSO-LSSVM).The relationships between the ratio of the current mutations and the ratio of the neutral line impedance and the grounding impedance,before and after the leakage faults, are deduced and then a stage one fault detection method using the ratio of the current mutations is proposed. On this basis, the fundamental and the third harmonic waves of the residual and the unbalanced current mutations are used as the characteristic varieties, PSO-LSSVM is used to conduct secondary discrimination on the judgment results to achieve reliable fault identification. Field tests show that the total accuracy of the method is over 80% when the amplitude of inherent residual current is higher than the fault current. The proposed method requires less calculation and can be implemented at low cost, making the leakage protection possible in the TN-CS systems.