• DocumentCode
    3353651
  • Title

    Forecasting Short-Term Load of Southwestern Power Market in China by Chaotic BP Network

  • Author

    Yin, Kuang ; Gang, Luo

  • Author_Institution
    Key Lab. of Network Applic. Project, Neijiang Normal Univ., Neijiang
  • fYear
    2009
  • fDate
    27-31 March 2009
  • Firstpage
    1
  • Lastpage
    3
  • Abstract
    Power is important to modern society and national economy. To forecast short-term load more accurately, phase space of the complex nonlinear system was reestablished according to chaos theory and properties of short-term load were analyzed. It proves that forecasting short-term load is a classic decision-making process, full of chaos. Combining with chaos theory and traditional BP network, an improved BP network (chaotic BP network, CBP network) was presented in the chaotic phase space. Learning algorithm of traditional BP network was improved because of initial value sensitivity and good ergodicity of chaos operator. The forecasting system has been applied in the power market in southwestern China. The results show that the forecasting system based on CBP network is more accurate than traditional BP network and reliability and accuracy can be used as needed.
  • Keywords
    backpropagation; learning (artificial intelligence); load forecasting; neural nets; power engineering computing; power markets; chaotic BP network; chaotic phase space; complex nonlinear system; decision-making process; learning algorithm; power market; short-term load forecasting; southwestern China; Artificial neural networks; Chaos; Delay effects; Demand forecasting; Economic forecasting; Laboratories; Load forecasting; Power markets; Predictive models; Timing;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Power and Energy Engineering Conference, 2009. APPEEC 2009. Asia-Pacific
  • Conference_Location
    Wuhan
  • Print_ISBN
    978-1-4244-2486-3
  • Electronic_ISBN
    978-1-4244-2487-0
  • Type

    conf

  • DOI
    10.1109/APPEEC.2009.4918388
  • Filename
    4918388