Demand response in a market environment

This thesis addresses the design, deployment and benefits of demand response in a market environment. Demand response is consumption that can be controlled by an external stimulus in the power system. Flexible consumption is a useful tool for absorbing volatile power from renewable sources like wind power and photovoltaics, and dealing with decentralised activity like electric vehicle charging.

Without flexible consumption or other new technologies like storage, there will be several occasions of surplus or deficit of generation to meet the demand of the future, sometimes expected and sometimes not, that will lead to power system failure. The type of demand response investigated is consumption controlled by indirect means, like an electricity price.

Initially, algorithms responding to real-time electricity prices are researched and benchmarked according to comfort and cost. After this simulation, real power system data from the Danish island of Bornholm is introduced and methods to quantify an aggregated load is developed. These methods can be used for real-time operation and to support investment decisions.

More specifically, they can be used to forecast the response to electricity pricing and to classify different types of customers. The proposed models are then embedded into new fiveminute electricity markets for system balancing and local congestion management. New market tools for exploiting and maintaining a degree of control over demand are developed, and the value of DR using indirect control is determined in terms of social welfare.

This thesis is written in the context of Danish and European power systems because the data used - and the data-driven models subsequently created - come from and were developed for the EcoGrid EU project. The demand forecasting models and electricity markets proposed in this thesis have been implemented on the Danish island of Bornholm in the EcoGrid EU project.

The real-time balancing market ran from October 2014 until May 2015, the congestion market operated from January 2015 onwards, and the demand forecast module operated from February 2015 onwards. EcoGrid EU is a large-scale smart grid demonstration with 1900 residential households and 100 industrial customers with a peak load above 5MW.

Customers are equipped with smart meters and a range of distributed energy resources with automated controllers that receive a new electricity price every five minutes and optimize consumption levels accordingly. DR from these customers is bid into the electricity market as balancing power and customer measurements are used in real-time to update the demand forecast.