Abstract:

The paper shows the main types of consumption profiles used by distribution companies and the methodology by which these profiles can be determined.

A consumption profile means a function that depends on its daily temperature to estimate the volume of gas (or power) consumed by a customer.

In order to identify and extract the relevant values from the consumption data, a weighted average of the volume of gas consumed was calculated for all customers of a sample over a temperature range of one degree.

Various mathematical methods can define functions that approximate consumption trends on the basis of the points in the graph.

For this paper a case study was carried out in this work on 8 localities of a gas distributor in Romania. Consumption profiles have been built using non-linear regression (grade 3 and grade 5 polynomial and sigmoid) and neural networks to calculate profiling and nomination errors, as well as the financial impact of using a consumption profile created with neural networks.