Publicaties

Skip Navigation Links.
Recent verschenen
Expand per documenttypeper documenttype
Expand per Unitper Unit
Expand per Clusterper Cluster

Zoeken naar publicaties:
Beperk het zoeken tot de velden:

ECN publicatie:
Titel:
Identification and visualization of energy-consumption patters
 
Auteur(s):
 
Gepubliceerd door: Publicatie datum:
ECN 1-10-1998
 
ECN publicatienummer: Publicatie type:
ECN-RX--98-058 Artikel wetenschap tijdschrift
 
Aantal pagina's:
10  

Gepubliceerd in: Paper, presented at the Distributech DA/DSM Europe 1998 conference, London, United Kingdom, 27-29 October 1998 (), , , Vol., p.-.

Samenvatting:
The liberalisation of the energy markets will change the relationshipsbetween distribution company and customers. To take full advantage of future differentiation in tariff structures, a better understanding of the consumption patterns is required. The distribution company needs this insight to segment its market and improve load profiles, whereas the customer needs it to lower his electric bill by negotiation of a lower tariff, peak shaving, and reducing and leveling consumption. The aims of our project are: (1) to extract as much information as possible from data sets on electricity consumption, (2) to identify typical consumption patterns, and (3) to elucidate external factors that can explain these patterns. The development and evaluation of methods, tools, and techniques for load profiling is an important part thereof. Data sets were collected on the electricity consumption in our institute, which is a mixture of laboratories and offices, populated by 850 employees. Three different techniques were used: visualization, cluster analysis, and multiple regression analysis. The electricity consumption data of 1997, collected at 15-minute intervals, were visualized in a single image using software developed in-house. A prototype of an interactive time-series analysis tool with a calendar-based user interface was developed, which supports cluster analysis of the site consumption patterns. With this cluster analysis dominant patterns as well as typical outliers can be identified easily. Multiple linear regression analysis of the consumption data showed relations between the actual electricity consumption and the number of employees present at that time; the level of natural light; the day of the week; and the time of the day. Each technique provides its own perspective; together they lead to a better understanding of energy consumption patterns and effective demand side management measures. 1 ref.


Terug naar overzicht.