Reserve market operation with the integration of distributed energy resources

Abstract

Integrating renewable energy resources (RERs) into power systems hasbeen developing very fast recently. The uncertainty and intermittence of RERscan easily result in the mismatch between generation and load, which maycause power systems to collapse. Operation reserve is necessary for the stableand reliable operation of power systems. Traditionally, the reserve capacity isprovided by generators. With the development of demand-side managementtechnology and battery technology, distributed energy resources, such as electricvehicles (EVs), prosumers, and so on, present a new opportunity to providereserve to the power systems. For example, EVs can inject power back to thegrid in vehicle-to-grid mode when necessary. This thesis focuses on the reservemarket operation with the integration of distributed energy resources.First, we study the reserve management problem of the EV aggregator. Abilevel optimization model is developed to describe the interaction between theaggregator and the EV owners. Unlike existing literature, binary variables areemployed to model the exclusive right constraint for accessing EV battery inthe lower-level model. An exact and nite algorithm is proposed to solve theproposed model.Second, based on the above work, we consider the uncertain EV connectivityto the grid due to the inevitable transportation randomness. We propose atrilevel prot maximization model for EV aggregators participating in the dayaheadreserve market. Firstly, total unimodularity property, primal-dual, andvalue-function methods convert this problem into a single-level mixed-integernonlinear program (MINLP). Then, a sample-based algorithm is developed to solve the single-level MINLP, and the convergence is proved. In addition, anacceleration strategy is proposed to facilitate the computation.Third, we investigate the day-ahead economic and secure managementproblem of the prosumer in the energy and reserve market, considering theuncertainties of renewable energy, market prices, and reserve activation rate.Due to the distributions of these uncertain parameters are ambiguous, Wassersteindistance-based distributionally robust optimization (DRO) approach isemployed to hedge against these uncertainties. The proposed optimizationmodel includes DRO objective function and chance constraints. We employinner approximation to convert the developed DRO model into a convenientlycomputable form. Dierent from existing literature, we analyze the optimalitygap in objective function approximation.Finally, we study the joint energy and reserve sharing problem betweenmassive prosumers considering renewable generation uncertainty and limitednetwork resources. Firstly, a data-driven distributionally robust energy andreserve sharing model between dierent agents in electricity markets is proposed.Then, considering the agents in the internet of things exchange informationby a resource-limited network, we develop a communication-censored consensusalternating direction method of multipliers to save the limited networkresources and solve the sharing problem fully decentralized. We also analyzethe convergence of the proposed algorithm. In addition, an adaptive penaltyparameter method is proposed to speed up the convergence.