![]() | Only 14 pages are availabe for public view |
Abstract Data mmmg means extraction of interesting information or patterns, which is non-trivial, implicit, previously unknown and potentially useful from data in large databases. Data mining is used to specify the kind of patterns to be found in data mining tasks. Data mining tasks can characterize the general properties of the data and perform inference on the current data in order to make predictions. Accurate models for electrical load forecast are essential. Knowing the load behavior in advance is very important in planning, analysis and operation of power systems to maintain reliable, secure and economic electric energy supply. Forecasting of electricity has always been the essential part of an efficient power system planning and operation, especially long term forecasts as it has become increasingly important since the rise of the competitive energy markets. The aim of long term load forecast is to predict future electricity demands based on historical data of some independent parameters such as total energy sale, total energy generation, GDP etc. This thesis investigates the suitable model using Data Mining and Knowledge Discovery in Databases (KDD) to increase the accuracy and revenue. AJI the implemented algorithms were written with MATLAB. The data was collecting from the different sites of the Egyptian Electricity Sector. Knowledge discovery process steps are implemented to the time series data, preprocessing the data in order to detect the missinvalue, odd value, outliers, and normalizing the data. The output from the preprocessing step is then fed into multiple regression to predict the coefficient parameters, or neural network for training. To obtain the forecast, the prediction data of the independent parameters must be entered into any of the two techniques (regression, neural network). These steps (phases) have been carried out for all parameters for ten models selected for implementations in order to reach the best model that fulfills good result with the Egyptian Power System. The simplest and more accurate model is the model that contains the following algorithms: natural cluster based interpolation algorithm for detecting the missing value, the histogram algorithm for detecting the odd value, and the regression method for forecasting the long term load forecast. Also the neural network it seen to be more accurate but more costly. The data results showed that the data mining with KDD process was capable of producing a reasonable forecasting accuracy in long term load forecast |