文章以电压无功优化的二次规划模型为基础，对原对偶内点法进行了扩展，使之能处理电压无功优化控制中大量的不等式约束。文章提出了一种壁垒参数的选取方法和控制计算步长的策略，并采用了一种有效的预测校正方法来提高算法的收敛性。 Based on the quadratic programming model of voltage and reactive power optimization, the paper extends the original dual dual inference method to deal with a large number of inequality constraints in voltage and reactive power optimization control. In this paper, we propose a method of choosing barrier parameters and a strategy of controlling the step size, and adopt an effective predictive correction method to improve the convergence of the algorithm.