基于自适应幂级数初始点的电力系统全纯嵌入潮流并行计算

在利用全纯嵌入潮流计算方法求解电力系统潮流时，若平启动的电压幂级数初始点远离系统潮流解实际值，则易导致计算量增大，且在求解各节点电压的过程中，传统基于中央处理器(CPU)串行计算架构的计算耗时较长。提出一种基于自适应幂级数初始点的全纯嵌入潮流并行计算方法。根据系统节点导纳矩阵和平衡节点电压自适应调整系统全纯嵌入潮流的幂级数初始点，构建基于自适应幂级数初始点的全纯嵌入潮流计算模型，以加快电压幂级数的收敛。借助多核CPU计算平台实现各节点电压Padé近似的并行化，以提高系统潮流方程的求解效率。不同规模测试系统的仿真结果表明，所提方法可实现电力系统全纯嵌入潮流的高效、准确求解。

Adaptive power series initial points based holomorphic embedding load flow parallel calculation for power system

When the holomorphic embedding load flow calculation method is used to solve the load flow of power system, if the initial points of voltage power series of flat start are far away from the actual values of load flow solution of the system, the calculation amount is easy to increase. Moreover, the calculation time based on the traditional central processing unit(CPU) serial calculation architecture is too long in the process of solving the voltage of each node. A holomorphic embedding load flow parallel calculation method based on the adaptive power series initial points is proposed. The initial points of holomorphic embedding load flow of the system are adaptively adjusted according to the node admittance matrix of the system and the voltage of equilibrium node, a calculation model of holomorphic embedding load flow based on the initial points of adaptive power series is constructed to accelerate the convergence of voltage power series. Based on the multi-core CPU calculation platform, the parallelization of voltage Padé approximation for each node is realized to improve the solving efficiency of load flow equation of the system. The simulative results of different scale test systems show that the proposed method can realize efficient and accurate solution of holomorphic embedding load flow of power system.