Computational Analysis of Demand-side Management in Smart Grids

High tech software solutions will form a fundamental part of the future electricity grid. Software enables real time efficient communication and control of various devices for producing and consuming electrical energy and will be instrumental in analyzing the complex scenarios that arise when multiple parties with varying goals interact to create a safe and stable delivery system for electricity. With the increasing adoption of renewable energy sources (RES), maintaining grid stability becomes increasingly challenging because legacy power infrastructure was often not designed with decentralized generation in mind. The varying nature of RES also further complicates maintaining a balance between electricity production and consumption. Flexibility is needed in both production and consumption in order to achieve the European union's 20-20-20 climate and energy targets while maintaining a safely balanced electricity grid. Demand-side management (DSM) programs attempt to harness consumption flexibility for goals that include maintaining grid stability when dealing with RES. Successfully implementing these programs requires producers, consumers and system operators to deal with strategic choice situations concerning various aspects of DSM. Literature often focuses on singular technical aspects of consumption flexibility while the combination of various technical and economical aspects determine the actual long term efficacy of DSM programs. This dissertation focuses on various techniques from computer science (CS) literature for studying strategic choice situations in various aspects of using DSM to address problems related to the integration of RES into contemporary electricity grids. In this dissertation the choice in compensation payment structures for flexibility providers are computationally analyzed using evolutionary game theory (EGT) with replicator dynamics to gain insight into how financial compensation influences DSM participation willingness. Computational analysis with heuristic payoff matrices determined through microsimulation shows that users of flexibility can gain a higher market share of flexibility providers by favoring reservation payment structures. Besides financial compensation, users of flexibility must decide which provider is best suited at any given time to activate their flexibility. To this end, algorithms for the coordination of the use of consumption flexibility and the negotiation between users and providers of this flexibility are implemented from multi-agent and mechanism design literature. Both cooperative contract net protocol (CNP) and competitive qualitative Vickrey auction (QVA) mechanisms are compared and evaluated in terms of allocation efficiency. To solve grid balancing problems effectively, sizable investments in technological solutions for managing and activating consumption flexibility are necessary. Whether these investments are cost effective depends on the costs and benefits associated with these technology investments. An investment cost model is proposed and evaluated for an optimized end to end solution for excess wind power production using various active network management (ANM) techniques including dynamic line rating (DLR), DSM and battery storage. In addition to investment decisions faced by users of flexibility, flexibility providers also must decide on which flexibility program to take part in. As a last contribution, participation in different competing business cases is simulated and computationally analyzed using EGT. The choice dynamics of flexibility providers choosing business partners, are evaluated in terms of how flexibility users employ and compensate for the consumption flexibility offered by flexibility providers. This evaluation also includes a sensitivity analysis on the influence that contractual activation constraints have on the efficiency of the DSM program in use. Applying techniques from EGT literature to analyze strategic choice scenarios has shown that EGT can offer valuable insights into the complex interactions that occur between the multitude of parties in the energy domain.

Jaar van publicatie:2018

Toegankelijkheid:Open