The Open Automation and Control Systems Journal

2014, 6 : 1912-1918
Published online 2014 December 31. DOI: 10.2174/1874444301406011912
Publisher ID: TOAUTOCJ-6-1912

The Application of Neural Network and Logistic Regression in the Operation Performance Warning of Power Grid

Yixin Sun , Xiaobao Yu , Qingyou Yan and Zhongfu Tan
School of economics and management, North China Electric Power University, Beijing 102206, P.R. China.

ABSTRACT

In this paper, for a given initially analyses the operation performance of grid enterprises, we divide index factors affecting operation performance of grid enterprise into two types: investment index and result index. For the the method of warning analysis, we establish models of neural network and logistics regression, using Matlab mathematical software to calculate and analyse the above-mentioned two types of indexes, concludes their relationship of correlation coefficient and regression coefficient. For scenarios analysis, we use models with 3 scenarios to study on operation performance risks of grid enterprises herein. Eventually we analyses actual data from various places empirically and arrives at the related conclusion. For conclusion, we hold the opinion that risks of operation performance may occur in the future if present investment cannot satisfy requirement for power grid development of various regions; increasing investment ratio may be likely to decrease occurrence rate of risks of power grid operation performance.

Keywords:

Logistic regression, multi-scenario analysis, neural network, operation performance, warning.