欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
随着电力行业改革的进一步深化，如何预测将来的电力需求是供电企业关注的问题之一。把预测误差看作供电企业的一种风险来源，研究风险价值在电力需求预测中的应用，是一种新方法。该方法利用三层BP神经网络根据电力需求历史数据建立预测模型：计算BP神经网络对历史数据的拟合残差（预测误差）；对计算得到的预测误差利用风险价值概念与方法计算给定置信水平下电力需求的置信区间；利用南京市电力需求历史数据进行案例分析。研究结果表明，该电力需求预测方法，充分考虑到供电企业的风险偏好．得到的电力需求区间能够在概率意义下反应电力需求的不确定性变化。
英文摘要：
With deepening the reform in power industry, how to predict the future power demand has become one of important problems, which catches the power supply enterprise＇s attention. Taking the predicting residual as the risk resource, this paper focused on the application of value at risk to the power enterprises. A predicting model was build based on the history data of power demand by utilizing a three-layer BP neural network. The confidence interval of the power demand of the given confidence level was obtained by using the conception and methods of value at risk. A ease was illustrated based on the power-demand history data of the Nanjing city. The results show that the power demand predicting method imposed in this paper takes sufficiently account of risk preference of the power enterprise, and the power demand prediction interval reflects the uncertainty of the power demand in the sense of probability.
同期刊论文项目
同项目期刊论文
期刊信息
位置：