为了快速准确地计算静态电压稳定裕度，本文提出了两种鞍结分岔点快速求取算法，分别采用二分搜索和抛物线近似来进行计算。基于Levenberg-Marquardt算法在潮流方程的不可行域也能求得最小二乘解的特性，二分搜索算法利用解得的最小二乘值判断此算点是否处于潮流不可行域，通过二分搜索来快速逼近鞍结分岔点。抛物线近似算法对不可行域的最小二乘值-负荷裕度曲线进行抛物线近似，曲线的零点即为所求的鞍结分岔点。多个经典算例测试结果表明，相较于传统的连续潮流算法，二分搜索算法在保证计算准确地同时可以大幅度提升计算效率。而抛物线近似算法牺牲了一定的计算精度，在二分搜索算法的基础上进一步提升了效率。并且得益于Levenberg-Marquardt算法的强鲁棒性，本文两种算法即使在面对大型病态算例时也可以收敛，保证了计算的稳定性。

In order to calculate static voltage stability margins quickly and accurately, this paper proposes two algorithms for searching saddle-node bifurcation points based on bisection search and parabolic approximation respectively. The Levenberg-Marquardt algorithm is applied to obtain the least-squares solution in the infeasible region of the power flow equations. The bisection search algorithm uses the least-squares solution to determine whether the current trial step is in the power flow infeasible region, and finally approaches the saddle-node bifurcation point. The parabolic approximation algorithm performs a parabolic approximation to ‘the least-squares value’ - ‘load margin’ curve in the infeasible region, and the zero point of the curve is corresponding to the required saddle-node bifurcation point. The numerical experiments on several classical cases show that compared with the traditional continuous power flow algorithm, the bisection search algorithm can significantly improve the computational efficiency while ensuring the accuracy. The parabolic approximation algorithm sacrifices some accuracy, but further improves the efficiency based on the bisection search algorithm. Also, thanks to the robustness of the Levenberg-Marquardt algorithm, these two algorithms in this paper can well converge under large ill-conditioned cases, ensuring the numerical stability of these two algorithms.