英语翻译UNCERTAINTY OF THE SHORT-TERM ELECTRICAL LOAD FORECASTIN

英语翻译
UNCERTAINTY OF THE SHORT-TERM ELECTRICAL LOAD FORECASTING IN UTILITIES
Deregulation in an energy sector and the energy market origin needs an accurate of the Short-Term Load
Forecasting (STLF) method.One can find in bibliography number of methods used for the STLF in power systems
with discussion of their errors.Results of those researches cannot be easily applied for the STLF accuracy evaluation
in utilities.Capacity of an utility in Poland is nearly about one order lower than the power system capacity,moreover
the load curve in utilities are much more distorted than the load curve in power system.The aim of the paper is
presentation of the Authors investigation on the STLF in one of the Polish utilities.
At the beginning we will present the results of researches on the accuracy of the STLF method in power system
on a base of selected papers printed in 1998 and 1999.Pretty all of the papers applied one- or combination of the
Artificial Intelligence (AI) tools:Artificial Neural Networks (ANN),Expert Systems (ES) with Fuzzy Logic (FL)
comparing the traditional methods used in time series analysis.
The non parametric regression based STLF method comparing it’s accuracy with the nonparametric regression-static
and ANN receiving mean absolute percentage errors equal to 2.78,3.01 and 2.64 respectively has been describedin
PI.
The ANN with fuzzy-set based classification used for STLF of hourly load in 24,48,72,96 and 120 hours
ahead,following values of the Mean Absolute Percentage Errors (MAPE) 1.529,1.563,1.354,1.666 and 1.661
respectively.ANN based on the phase-space embedding of a load time-series STLF in two big US utilities and in
four months (January,April,July and October) receiving average (from number of different ANN used) MAPE 1.0 to
4.63 and maximum error 2.15 to 4.63 [ 121.Using the same method,the Authors investigated in [ 131 influence of the
different weather and consumption patterns in the two US utilities receiving following errors:

一个能源部门和能源市场起源中的不确定度的短期电 LOAD 预测的公用事业放松管制需要一个准确的短期负荷预测 (STLF) 方法.书目数的方法在电力系统的他们错误的讨论中,STLF 用于中可找到.这些研究的结果不能轻松地应用的 STLF 精度评定的实用程序.一个实用程序,在波兰的容量是近约一个订单比电源系统容量低、 电力系统亦更扭曲比负荷曲线实用程序中的负荷曲线.这份文件的目的,是在波兰的实用程序之一 STLF 作者调查的演示文稿.开始时我们会在电力系统的基础上的所选文件打印在一九九八年及一九九九年对 STLF 方法的准确性提出研究的结果.所有的报纸一或人工智能 (AI) 工具组合的应用相当：人工神经网络 (ANN)、 专家系统 (ES) 与模糊逻辑 (外语) 在时间序列分析中使用的传统方法比较.基于非参数回归的 STLF 方法比较与非参数回归-静态的准确性并接收等于 2.78 平均绝对百分比错误的神经网络,3.01 和 2.64 分别已 describedin PI.人工神经网络与模糊集的分类用于 STLF 的逐时负荷 24、 48、 72、 96 和 120 小时前面,分别后的平均绝对率错误 (MAPE) 1.529、 1.563、 1.354、 1.666 和 1.661 的值.基于负载的相空间嵌入的人工神经网络时间序列 STLF 两大美国实用程序中和四个月内 （一月、 四月、 七月和十月) 接收平均 （从不同的人工神经网络的多使用） MAPE 1.0 到 463 和最大错误 2.15 到 463 [121.在同一法作者调查 [中这两个美国实用程序收到下列错误模式的 131 不同天气和消费的影响：