In this thesis, it is presented a statistical based model that allows the estimation of the load shape curve for appliances (end uses) for residential consumers that belong to low voltage group (BT clients), using information from two sources: PPHs (Energy Audit on Ownership and Usage of Electrical Appliances) and household measurements through specific devices that provide not only the total consumption (15 minute intervals) as well as the breakdown of this consumption for each device (end use) existing in the household. Through energy audits (PPHs) developed by Procel/Eletrobrás in 1996, one has a rough idea of the load shape curves by appliance. However, the curves obtained this way tend to be rather imprecise, as they are obtained by the consumer information of usage of the equipments on the surveys, which tend to be rather vague. In order to reduce such imprecision, an experiment was set where a subsample of the original survey sample is selected and the main appliances consumption measured by electronic meters at intervals of 15 minutes. The end use load shape obtained by these meters are then compared with the corresponding curve obtained by usage declaration. Using the linear regression model, the correction coefficients of the declarations are obtained for each hour of the day, for the main appliances. The results were applied to two distribution utilities of the ENDESA group: AMPLA and COELCE.