家庭负荷识别是实现需求侧精细化管理的关键。针对现有家庭负荷辨识研究中对所提取特征贡献度及相关性分析不足的问题,提出了基于ReliefF与互信息结合的特征评价、筛选的家庭负荷类型辨识方法。本文首先在现有研究基础上提取了16个家庭负荷运行暂、稳态特征,然后对其权重及特征间相关性进行分析,筛选了其中辨识效果最优的特征组合,最后利用基于粒子群优化的支持向量机(Support vector machine based on particle swarm optimization, PSO-SVM )分类模型对实测数据样本进行了辨识。算例结果验证了所提算法的准确性和优越性。
英文摘要:
Family load identification is the key to achieving demand side refined management. Aiming at the problem of insufficient analysis of the extracted feature contribution and correlation in the existing family load identification research, this paper proposes a family load type identification method based on the combination of ReliefF and mutual information. Based on the existing research, this paper firstly extracts the temporary and steady-state characteristics of 16 family load operations, then analyzes the weight and correlation between them, and selects the feature combination with the best identification effect. Finally, the support vector machine classification model based on particle swarm optimization is used to identify the measured data. The results of the example verify the accuracy and superiority of the proposed algorithm.