Study On The Grid Electric Short-term Load Forecasting Based On Distributed Data Warehouse

This paper designs and constructs the distributed data warehouse for power load analysis and forecast, according to the structure characteristic of state grid operation and management. The system framework and interior structure strategy of distributed data warehouse are described in detail, as well as on-line analytical processing (OLAP). Meanwhile, the paper presents an idea of gridding to analyze the influence of weather to power load more concretely, gives a depth studies on the relationship between weather and power load, characterizes the combined effects of different weather factors on the load in different areas with bio-meteorology index, such as sense temperature index, thermal humidity index, cold index and comfort index. This paper analyzes various weather indices and the level of load by Wavelet Denoising and Gray algorithms, chooses the index which better reflects the changes in the load as a variable of multivariate time series, forecasting the daily power load. the conclusion testifies the validity of the proposed method.