By Finn Jagd Larsen and Marc Berry Bendtsen
Examiner
Søren Støvring-Hallsson (Ærø Elforsyning)
In cooperation with
DONG Energy
Supervisors
Pierre Pinson, CEE, DTU Elektro
Rikke Helbirk Kristensen, DONG Energy
Abstract
The main objective of DONG Energy Distribution is to operate the grid in an optimal technical-economical way. In order to do so an increased knowledge regarding the present and future load must be obtained. The solution is not to implement measuring equipment in the entire grid but to acquire methods for modelling of the energy consumption for customers in the distribution grid. Today the modelling is carried out using the Velander correlation. This method is not very precise and is only valid for larger groups of customers. The goal of this master is thesis is to investigate and formulate new methods for modelling of the energy consumption. The new method should not be limited by the same issues experienced with the Velander correlation.
Throughout the project available data at DONG Energy and Danish Meteorological Institute have been collected. Afterwards relations between the collected data and the energy consumption has been established. This investigation has shown that variables such as income and education have a significant impact on the energy consumption. Today DONG Energy uses a method for classification of their customers by grouping them in different customer categories depending on whether the customer is a traditional household or industry. This classification is challenged to see if customers can be grouped using other parameters. The gathered knowledge regarding the investigated relations and the new grouping of customers will be used to formulate statistical models by using the Generalized Additive Model. Using the Generalized Additive Model showed that it is very useful for estimating the energy consumption compared to the Velander correlation. Furthermore this model is very flexible and can be expanded with relative ease in regards to adding more explanatory variables. Lastly recommendations given to DONG Energy in regards to how a new classification of their customers can be initiated and how the results from the Generalized Additive Model can be used prospectively.