短期电力负荷预测有助于维持发电端和用电端的动态平衡,保障电力系统稳定且高效地运行。分布式能源的大规模并网以及气象和节假日等短期因素的影响,使得负荷序列呈现明显的波动性和非线性。为此,该文提出基于花授粉算法(flower pollinat...短期电力负荷预测有助于维持发电端和用电端的动态平衡,保障电力系统稳定且高效地运行。分布式能源的大规模并网以及气象和节假日等短期因素的影响,使得负荷序列呈现明显的波动性和非线性。为此,该文提出基于花授粉算法(flower pollination algorithm,FPA)优化变分模态分解(variational mode decomposition,VMD)和双向长短时记忆(bidirectional long and short time memory,BiLSTM)神经网络的新型两阶段短期电力负荷预测方法。第一阶段首先提出了一种关于分解损失的VMD评价标准,并采用FPA来寻找该标准下分解参数的最优组合,从而降低了经验设置参数的随机性并且减少了分解过程中的信号损失,提高了分解质量;其次针对分解所得的每个子序列分别建立具备双向处理和长期记忆的BiLSTM神经网络,从而可以更好地挖掘负荷数据的过去和未来的深度时序特征。第二阶段综合考虑模态分量以及气象和星期类型等短期因素的影响,建立基于BiLSTM神经网络的误差纠正模型,用以挖掘误差中所包含的隐含信息,从而降低了模型的固有误差。将该文方法应用于美国南部某地区的负荷数据集,最终的平均绝对误差(mean absolute error,MAE)、平均绝对百分比误差(mean absolute percentage error,MAPE)和均方根误差(root mean square error,RMSE)以及R2分别为108.03、1.19%、146.48以及0.9812。随后在冀北电网某供电公司的实际应用中,再次证明了该方法在区域性短期电力负荷预测中的有效性。展开更多
The estimate of dental caries among Chinese children at the microscale level using standard methodology remains unclear. In this study, we assessed and analyzed the disease burden of childhood dental caries in China b...The estimate of dental caries among Chinese children at the microscale level using standard methodology remains unclear. In this study, we assessed and analyzed the disease burden of childhood dental caries in China by extracting data from the Global Burden of Disease, Injuries, and Risk Factors Study 2016 （GBD 2016）. In 2016, the number of cases, prevalence, years lived with disability （YLD）, and age-standardized YLD rate of dental caries was 93.0 million, 43.0%, 32,200 person years, and 14.8 per 100,000, respectively. Across 33 provincial units, the disease burden was highest in Hubei （YLD rate 28.6 per 100,000）, lowest in Macao （9.1 per 100,000）, while geographical clustering was not observed. Compared with 1990, the prevalence in 2016 decreased from 46.8% to 43.0%, and the YLD rate decreased from 16.5 per 100,000 to 14.8 per 100,000. Given the slight decrease in dental caries burden, the prevalence and disease burden remained high among Chinese children. Strategies for addressing the spatial inequity of childhood dental caries require geographical targeting.展开更多
文摘短期电力负荷预测有助于维持发电端和用电端的动态平衡,保障电力系统稳定且高效地运行。分布式能源的大规模并网以及气象和节假日等短期因素的影响,使得负荷序列呈现明显的波动性和非线性。为此,该文提出基于花授粉算法(flower pollination algorithm,FPA)优化变分模态分解(variational mode decomposition,VMD)和双向长短时记忆(bidirectional long and short time memory,BiLSTM)神经网络的新型两阶段短期电力负荷预测方法。第一阶段首先提出了一种关于分解损失的VMD评价标准,并采用FPA来寻找该标准下分解参数的最优组合,从而降低了经验设置参数的随机性并且减少了分解过程中的信号损失,提高了分解质量;其次针对分解所得的每个子序列分别建立具备双向处理和长期记忆的BiLSTM神经网络,从而可以更好地挖掘负荷数据的过去和未来的深度时序特征。第二阶段综合考虑模态分量以及气象和星期类型等短期因素的影响,建立基于BiLSTM神经网络的误差纠正模型,用以挖掘误差中所包含的隐含信息,从而降低了模型的固有误差。将该文方法应用于美国南部某地区的负荷数据集,最终的平均绝对误差(mean absolute error,MAE)、平均绝对百分比误差(mean absolute percentage error,MAPE)和均方根误差(root mean square error,RMSE)以及R2分别为108.03、1.19%、146.48以及0.9812。随后在冀北电网某供电公司的实际应用中,再次证明了该方法在区域性短期电力负荷预测中的有效性。
文摘The estimate of dental caries among Chinese children at the microscale level using standard methodology remains unclear. In this study, we assessed and analyzed the disease burden of childhood dental caries in China by extracting data from the Global Burden of Disease, Injuries, and Risk Factors Study 2016 （GBD 2016）. In 2016, the number of cases, prevalence, years lived with disability （YLD）, and age-standardized YLD rate of dental caries was 93.0 million, 43.0%, 32,200 person years, and 14.8 per 100,000, respectively. Across 33 provincial units, the disease burden was highest in Hubei （YLD rate 28.6 per 100,000）, lowest in Macao （9.1 per 100,000）, while geographical clustering was not observed. Compared with 1990, the prevalence in 2016 decreased from 46.8% to 43.0%, and the YLD rate decreased from 16.5 per 100,000 to 14.8 per 100,000. Given the slight decrease in dental caries burden, the prevalence and disease burden remained high among Chinese children. Strategies for addressing the spatial inequity of childhood dental caries require geographical targeting.
