基于FPGA的电磁瞬态仿真器研究

Research on Electromagnetic Transient Simulator Based on FPGA

为了提高配电网故障电磁瞬态特性的实时仿真效率,避免复杂的现场可编程门阵列(field-programmable gatearray,FPGA)编程,本文研究了一种基于FPGA的自动实时电磁瞬态仿真器。首先通过综合改进的增广节点分析(modified augmented nodal analysis, MANA)方法和固定的导纳矩阵节点方法(fixed admittance matrix nodal method,FAMNM)构建了求解器;然后基于开关电导参数的最优选择,建立了高效的稀疏矩阵-向量乘法器;最后通过对三相配电网络中的单相接地故障后电磁瞬态特性进行测试,验证了仿真器的性能。实验结果表明:MANA-FAMNM的特殊结构可以实现极低的积分时间步长,并且避免了仿真器重新设计FPGA代码,该应用程序可直接用于不同的拓扑结构;仿真器能够实时精确地再现配电网故障中发生的电磁瞬态特性以及在传输线中传播的电磁波,FPGA-RTS仿真与软件EMTP-RV离线仿真的C相相电压值的最大误差约为0.006pu,完全满足工程应用的要求。该仿真器可以推广到工业实时嵌入系统中。

In order to improve the real-time simulation efficiency of electromagnetic transient characteristics of distribution network faults, and avoid complex field-programmable gate array (FPGA) programming, an automatic real-time electromagnetic transient simulator based on FPGA was studied. First, the comprehensive modified augmented nodal analysis (MANA) method and the fixed admittance matrix nodal method (FAMNM) were used to build up a solver. Then an efficient sparse matrix-vector multiplier was established based on the optimal selection of the switching conductance parameters. Finally, the electromagnetic transient characteristics of the single-phase-to-ground fault were tested in the three-phase distribution network to verify the performance of the simulator. The results show that the special structure of MANA-FAMNM can achieve extremely low integration time step and avoid redesigning the FPGA code, and the applications program can be used directly in different topologies. The simulator can realize accurate reproduction of the electromagnetic transient characteristics occurring in the distribution network fault and the electromagnetic waves propagating in the transmission line in real time. The maximum error between the phase-phase voltage of the FPGA-RTS simulation and the software EMTP-RV offline simulation is about 0.006 pu which fully meet the requirements of engineering applications. The simulator can be extended to industrial real-time embedded systems.