Year: 2013
Type: dissertation
Abstract: Demand side management is regaining attention because of the integration of intermittent renewable energy sources such as wind and solar power. Accurate electrical load data are needed to estimate the potential and the impact of demand response. Data itself is hard to grasp. Analytical techniques are required to describe the data in a comprehensive way. Privacy laws prevent data usage in certain contexts. Data modelling gets around the privacy problem and allows for Monte Carlo simulations.First, residential electricity demand is described. The demographic properties describing the electricity demand in Belgium are determined with the use of machine learning techniques. Residential electricity consumers are grouped based on timing of and amount of electricity demand with clustering techniques. Measurement data from a project are scaled up by employing the groups found to get an impression of electricity demand of wet appliances.Then, residential electricity demand is modelled. The combination of electricity consumer groups together with Markov models allow for an electricity demand model. Wet appliance load cycles are detected and parametrised according to the previously defined consumer groups. The first is able to regenerate load data at the household connection point, the latter is able to regenerate load cycles of appliances.Finally, the models are applied to estimate potential and impact of direct appliance control and corresponding privacy issues.
Source: Lirias
Author: Wouter Labeeuw
Language: English
Topic: Smart Grid Energy Management
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2559759316