电压暂降作为电能质量的重要问题,严重影响电力系统稳定性与设备运行安全,因此亟须对电压暂降全面、精准的评估。现有方法多指标单一,缺乏权重差异的合理考虑。为此,文章提出一种基于主客观组合赋权多维指标的电压暂降评估方法。首先,...电压暂降作为电能质量的重要问题,严重影响电力系统稳定性与设备运行安全,因此亟须对电压暂降全面、精准的评估。现有方法多指标单一,缺乏权重差异的合理考虑。为此,文章提出一种基于主客观组合赋权多维指标的电压暂降评估方法。首先,构建涵盖系统侧、设备侧与用户侧的多维评估指标体系,包括电压幅值严重性、暂降能量、设备响应和经济损失等指标;采用序关系法与熵权法进行主客观赋权,并通过博弈论融合优化权重,增强权重分配的合理性;引入灰色关联度与欧氏距离改进传统优劣距离法TOPSIS(technique for order preference by similarity to ideal solution),提升模型判别力。最后,在IEEE 33节点系统上验证方法的有效性,结果表明该方法可准确反映各节点暂降严重程度,具备良好应用前景。展开更多
This study propose a new robust method to rank the performances of multi-assets （portfolios）, based purely on their return time series. This method makes no assumption on the distributions. Topsoe distance is symmet...This study propose a new robust method to rank the performances of multi-assets （portfolios）, based purely on their return time series. This method makes no assumption on the distributions. Topsoe distance is symmetrized Kullback-Leibler divergence by average of the probabilities. The square root of Topsoe distance is a metric. We extend this metric from probability density functions to real number series on （0, 1 ]. We call it ST-metric. We show the consistency of ST-metric with mean-variance theory and stochastic dominance method of order one and two. We demonstrate the advantages of ST-metric over mean-variance rule and stochastic dominance method of order one and two.展开更多
文摘电压暂降作为电能质量的重要问题,严重影响电力系统稳定性与设备运行安全,因此亟须对电压暂降全面、精准的评估。现有方法多指标单一,缺乏权重差异的合理考虑。为此,文章提出一种基于主客观组合赋权多维指标的电压暂降评估方法。首先,构建涵盖系统侧、设备侧与用户侧的多维评估指标体系,包括电压幅值严重性、暂降能量、设备响应和经济损失等指标;采用序关系法与熵权法进行主客观赋权,并通过博弈论融合优化权重,增强权重分配的合理性;引入灰色关联度与欧氏距离改进传统优劣距离法TOPSIS(technique for order preference by similarity to ideal solution),提升模型判别力。最后,在IEEE 33节点系统上验证方法的有效性,结果表明该方法可准确反映各节点暂降严重程度,具备良好应用前景。
文摘This study propose a new robust method to rank the performances of multi-assets （portfolios）, based purely on their return time series. This method makes no assumption on the distributions. Topsoe distance is symmetrized Kullback-Leibler divergence by average of the probabilities. The square root of Topsoe distance is a metric. We extend this metric from probability density functions to real number series on （0, 1 ]. We call it ST-metric. We show the consistency of ST-metric with mean-variance theory and stochastic dominance method of order one and two. We demonstrate the advantages of ST-metric over mean-variance rule and stochastic dominance method of order one and two.
