On Application of State Dependent Parameter Models in Electrical Demand Forecast
Electrical demand forecast is critical to power system operation since it serves as an input to the management and planning of such systems, such as power production, transmission and distribution, dispatch and pricing process as well as system security analysis. From the system’s point of view, this is a complex nonlinear dynamic system in which the power demand is a highly nonlinear function of the historical data and various external variables. This chapter describes an application of a type of State Dependent Parameter (SDP) models, two-dimensional wavelet based SDP model (2-DWSDP) to the modeling and forecast of daily peak electrical demand in the state of Victoria, Australia. Using the proposed approach, the essentials of such a system’s dynamics can be effectively captured by a compact mathematical formulation. The parsimonious structure of the identified model enhances the model’s generalization capability, and provides very descriptive views and interpretations about the interactions and relationships between various components which affect the system’s behaviours.
KeywordsElectrical Demand Peak Demand Daily Peak Nonlinear System Identification State Dependent Parameter
- 6.Young, P.C.: Stochastic, dynamic modelling and signal processing: time variable and state dependent parameter estimation. In: Fitzgerald, W.J., Walden, A., Smith, R., Young, P.C. (eds.) Nonlinear and Nonstationary Signal Processing, pp. 74–114. Cambridge University Press, Cambridge (2000) Google Scholar
- 7.Young, P.C.: The identification and estimation of nonlinear stochastic systems. In: Mees, A.I. (ed.) Nonlinear Dynamics and Statistics, pp. 127–166. Birkhäuser, Boston (2001a) Google Scholar
- 18.Niebur, D.: Artificial Neural Networks in the power industry, survey and applications. Neural Netw. World 6, 945–950 (1995) Google Scholar
- 23.Bunn, D.W., Farmer, E.D.: Comparative Models for Electrical Load Forecasting. Wiley, New York (1985) Google Scholar