Bottom-up Markov Chain Monte Carlo approach for scenario based residential load modelling with publicly available data

M. Nijhuis, M. Gibescu, S. Cobben

Research output: Contribution to journalArticleAcademicpeer-review

27 Citations (Scopus)
558 Downloads (Pure)

Abstract

In the residential sector, with the introduction of electric vehicles and photovoltaics, developments are taking place which have an impact on residential load curves. In order to assess the integration of these new types of technologies on both the generation and load side, as well as to develop mitigation strategies like demand side management, detailed information is required about the load curve of a household. To gain knowledge about this load curve a residential load model is developed based on publicly available data. The model utilises a Markov Chain Monte Carlo method to model the occupancy in a household based on time use surveys, which together with weather variables, neighbourhood characteristics and behavioural data are used to model the switching pattern of appliances. The modelling approach described in this paper is applied for the situation in the Netherlands. The resulting load curve probability distributions are validated with smart meter measurements for 100 Dutch households for a week. The validation shows that the model presented in this paper can be employed for further studies on demand side management approaches and integration issues of new appliances in distribution grids.

Original languageEnglish
Pages (from-to)121-129
Number of pages9
JournalEnergy and Buildings
Volume112
DOIs
Publication statusPublished - 15 Jan 2016

Fingerprint

Markov processes
Smart meters
Electric vehicles
Probability distributions
Monte Carlo methods
Demand side management

Keywords

  • Electricity consumption
  • Household appliance
  • Load curves
  • Markov Chain
  • Occupancy
  • Residential energy consumption modelling

Cite this

@article{b1019f169c22442c986f67ce87ce74d2,
title = "Bottom-up Markov Chain Monte Carlo approach for scenario based residential load modelling with publicly available data",
abstract = "In the residential sector, with the introduction of electric vehicles and photovoltaics, developments are taking place which have an impact on residential load curves. In order to assess the integration of these new types of technologies on both the generation and load side, as well as to develop mitigation strategies like demand side management, detailed information is required about the load curve of a household. To gain knowledge about this load curve a residential load model is developed based on publicly available data. The model utilises a Markov Chain Monte Carlo method to model the occupancy in a household based on time use surveys, which together with weather variables, neighbourhood characteristics and behavioural data are used to model the switching pattern of appliances. The modelling approach described in this paper is applied for the situation in the Netherlands. The resulting load curve probability distributions are validated with smart meter measurements for 100 Dutch households for a week. The validation shows that the model presented in this paper can be employed for further studies on demand side management approaches and integration issues of new appliances in distribution grids.",
keywords = "Electricity consumption, Household appliance, Load curves, Markov Chain, Occupancy, Residential energy consumption modelling",
author = "M. Nijhuis and M. Gibescu and S. Cobben",
year = "2016",
month = "1",
day = "15",
doi = "10.1016/j.enbuild.2015.12.004",
language = "English",
volume = "112",
pages = "121--129",
journal = "Energy and Buildings",
issn = "0378-7788",
publisher = "Elsevier",

}

Bottom-up Markov Chain Monte Carlo approach for scenario based residential load modelling with publicly available data. / Nijhuis, M.; Gibescu, M.; Cobben, S.

In: Energy and Buildings, Vol. 112, 15.01.2016, p. 121-129.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Bottom-up Markov Chain Monte Carlo approach for scenario based residential load modelling with publicly available data

AU - Nijhuis, M.

AU - Gibescu, M.

AU - Cobben, S.

PY - 2016/1/15

Y1 - 2016/1/15

N2 - In the residential sector, with the introduction of electric vehicles and photovoltaics, developments are taking place which have an impact on residential load curves. In order to assess the integration of these new types of technologies on both the generation and load side, as well as to develop mitigation strategies like demand side management, detailed information is required about the load curve of a household. To gain knowledge about this load curve a residential load model is developed based on publicly available data. The model utilises a Markov Chain Monte Carlo method to model the occupancy in a household based on time use surveys, which together with weather variables, neighbourhood characteristics and behavioural data are used to model the switching pattern of appliances. The modelling approach described in this paper is applied for the situation in the Netherlands. The resulting load curve probability distributions are validated with smart meter measurements for 100 Dutch households for a week. The validation shows that the model presented in this paper can be employed for further studies on demand side management approaches and integration issues of new appliances in distribution grids.

AB - In the residential sector, with the introduction of electric vehicles and photovoltaics, developments are taking place which have an impact on residential load curves. In order to assess the integration of these new types of technologies on both the generation and load side, as well as to develop mitigation strategies like demand side management, detailed information is required about the load curve of a household. To gain knowledge about this load curve a residential load model is developed based on publicly available data. The model utilises a Markov Chain Monte Carlo method to model the occupancy in a household based on time use surveys, which together with weather variables, neighbourhood characteristics and behavioural data are used to model the switching pattern of appliances. The modelling approach described in this paper is applied for the situation in the Netherlands. The resulting load curve probability distributions are validated with smart meter measurements for 100 Dutch households for a week. The validation shows that the model presented in this paper can be employed for further studies on demand side management approaches and integration issues of new appliances in distribution grids.

KW - Electricity consumption

KW - Household appliance

KW - Load curves

KW - Markov Chain

KW - Occupancy

KW - Residential energy consumption modelling

UR - http://www.scopus.com/inward/record.url?scp=84950996585&partnerID=8YFLogxK

U2 - 10.1016/j.enbuild.2015.12.004

DO - 10.1016/j.enbuild.2015.12.004

M3 - Article

AN - SCOPUS:84950996585

VL - 112

SP - 121

EP - 129

JO - Energy and Buildings

JF - Energy and Buildings

SN - 0378-7788

ER -