Application in Power Load Forecasting Based on an Improved Combination Grey Model Weighted by Correlation

WANG Jing-min; LI Ya-kun

It is exceedingly difficult to get accurate predictions for single traditional prediction models because of the volatility, non-linear increment and complexity of power loads. In order to improve the accuracy for long-term power load forecasting, the multivariate exponential weighting grey prediction model, residual grey prediction model, dynamic and equal-dimensional information grey prediction model and equal time sequence grey prediction model were constructed and weighted by correlation so that an improved combination grey prediction model could be built to make a prediction and for empirical analysis. The example shows that the volatility can be effectively reduced by the multivariate exponential weighting model and dynamic equal-dimensional information model. Similarly, the residual model and equal time sequence model are suitable for the power load forecasting with non-linear increasing trend. Considering all kinds of features of power loads, the constructed combination model can improve the accuracy of power load forecasting effectively and ensure the economic and safe operation for power system.

Select Volume / Issues:

• Volume 6 Issue 1, Jan. 2015

Year:

2015

Type of Publication:

Article

Keywords:

Improved Grey Model; Power Load Forecasting; Correlation Method; Combination Forecasting

Journal:

IJECCE

Volume:

6

Number:

1

Pages:

44-49

Month:

Jan.-Feb.

ISSN:

2249-–0