Important Parameters for Prediction of Power Loads - A Bottom-Up Approach Utilizing Measurements from an Automatic Meter Reading System

In Sweden the deregulation of the electricity market has increased the demand of efficient and accurate short term prognosis. Top-down models are today widely used for both planning and prognosis. The top-down models are known as accurate and effective when working with data in a well known domain, but can have problems estimating new trends in consumption patterns and in other situations with a limited access to reliable consumption statistics. In a bottom-up approach a stronger physical relationship can be implemented and individual parameter adaptations can be made. Increased knowledge on an individual consumption basis become more important to be able to increase the accuracy of such tool, and the extensive installations of automatic meter reading systems provides the necessary consumption data for these new models. In this paper an analysis method of individual customers is presented, a bottom-up approach is described and important parameters for forecasting and simulation are discussed.