Any activity in modern society is only possible with intensive and efficient use of energy. Among the various forms of energy the most important are those that are processed by companies and distributed to consumers, such as electricity. This work presents the development of two methodologies that will allow us to generate load capacity curves from (i) Random data obtained from power consumption of the PUC - Rio (ii) Bills (invoices) from the local utility company. With this data we will do an economic analysis of investment performed. This analysis can help us choose the optimum number of generators of electrical energy required to reduce the cost of the electricity supplied, all within the framework of the commercial sector and at peak hours. Both methods will be validated with the real power consumption, every fifteen minutes during four years from the energy profile of the PUC - Rio. The first method shows how we can generate load capacity curves making use of a statistic function called Probability Density, using data of any month. The second method shows how one can create the load capacity curves with only the invoices obtained from the local utility company. For this second methodology the program EES (Engineering Equation Solver) was used to solve systems of polynomial equations of the third degree.