Demand response modelling in electricity markets

Abstract

In traditional power system operation, the power landscape comprises centralized generation utilities allocation where electricity is pushed one-way through transmission and distribution networks to customers. Once power imbalance occurs, generation sides are adjusted by system operator and consumption sides passively enjoy those adjustments where few interaction happens between demand sides and system. Although electricity has strong commodity attribute, its two special properties greatly shape the operation of real-time supply/demand balance of electric system. First, it cannot be economically stored. When contingency happens, system may face severe collapse even imbalance between supply and demand merely lasts within seconds. Second, the electric industry is capital intensive. These two properties determine that it costs a great amount of resources to balance system at all moments. However, the situation of one-way low-efficient power transmission has changed as the emergence of demand response.

In fact, the concept of demand response has existed for decades, but it was restricted by various challenges in the past, including policy support, technology accessibility, market-wide operability and client acceptance. Luckily, with the development of smart grid and power market deregulation, those four main obstacles to involving demand response into power system operation have been gradually tackled, which leads to a more efficient bi-directional information communication and power transmission.

The incorporation of demand response plays like an innovator in terms of power market operation and system dispatch. Benefits brought by it touch upon many dimensions, such as individual power consumption reliability benefits, economic benefits for both clients and entities, market-wide impact, energy efficiency, system reliability, environmental benefits, etc.

Due to the minimum entrance amount requirement, demand response aggregator, an agency acting as a broker between consumers and energy supplier, is introduced in markets. In this thesis, two models involving demand response aggregators are presented in day-ahead and real-time market respectively. To investigate the optimal market mechanism and bidding strategy with the participation of demand response aggregator, scenario-based stochastic programming and robust optimization are considered. Besides, considering the development of real-time pricing, the behavior of consumer is modelled and optimized through a dynamic rolling mechanism.