基于支持向量回归的配电网线损计算模型

Calculation of energy losses in distribution system based on a support vector regression model

提出了一种基于支持向量回归的计算配电网线损的可行方法,建立了配电网线损计算的支持向量回归模型。针对有代表性的配电线路的线损与特征参数的样本数据,利用支持向量回归的拟合特性映射线损与特征参数之间复杂的非线性关系,找出配电线路的线损随特征参数变化的规律。为了提高支持向量回归机的学习效率,采用样本分类处理的方法分别对其进行训练,使的计算结果更加符合实际。以配电线路数据为实例,仿真结果验证了所提的方法和模型的有效性和实用性。

Aiming at power system, a feasible method is presented for calculating the energy losses in distribution systems based on support vector regression （SVR） model. For representational samples of reflecting the relation between energy losses and characteristic parameter of distribution net, SVR method, due to its strong regression ability, is able to map complex non-linear relation between them, and find the rule of energy losses varying with characteristic parameter. In order to improve the learning accuracy of SVR model, a problem-specific scheme is performed to divide the samples into several clusters. Simulation verifies the validity and practicability of the proposed method.