Forecasting electricity consumption by separating the periodic variable and decompositions the pattern

Electricity consumption pattern has been affected by some socio-economic and environment factors like consumers and environmental fluctuations. Due such parameters electricity consumption pattern will demonstrate various seasonal, monthly, daily and hourly variations. With this type of diversity, it is difficult to find the appropriate model to estimate. In this article it is attempted to decompose the pattern by clustering primary date and eliminating the periodic variance. The complicated pattern is changed to a set of simple patterns which easily could be used in point of view of macro decision making to determine the parameters affected on consumption pattern and forecast detailed executing consumption that is mainly used by middle management and technical engineers. Improved results with descriptive estimation for any part of consumption pattern are concluded in any time intervals.