Dynamic Risk-Averse Optimization of Distributed Energy Resource Aggregators

University Of Pittsburgh, Pittsburgh PA

Investigators

Abstract

Flexible distributed energy resources (DERs), such as energy storage devices, can provide diverse services to the nation's power grid that help mitigate variability and improve grid resilience. DER aggregators serve as mediators between electricity markets and consumers, seeking to incentivize DERs to provide multiple services to the grid while concurrently making commitments to the markets. However, current rules governing interactions with wholesale electricity markets are designed for large generators. These rules either discourage or do not allow for full utilization of the technological capabilities of small-scale energy storage and DER aggregations. The Federal Energy Regulatory Commission recently announced that independent system operators will soon be required to update their market rules in order to "remove barriers to the participation of electric storage resources and distributed energy resource aggregations in the capacity, energy, and ancillary markets..." Such rule changes will present opportunities for aggregators to effectively manage DER assets in a way that mutually benefits consumers and the grid. This research will create analytical and computational methods to devise optimal, or near optimal, strategies for DER aggregators who seek to incentivize DERs to provide multiple grid services, while making commitments to the market, in a risk-averse setting. Considered are multiple DERs and multiple grid services, uncertain interactions with the DERs and the market, and extendable notions of incentives and commitments to future rule implementations. It will also assess the value of aggregating DERs and co-optimizing multiple grid services and devise efficient techniques to compute risk-averse policies for problems that lack structure. A risk-averse Markov decision process (MDP) model with a dynamic risk measure is used to address the aggregator's sequential decision-making problem, and a suite of interrelated solution techniques for large-scale, risk-averse MDP models will be devised. The intelligent management of DERs for multiple uses promotes sustainable energy production and a reduction in environmental impacts by enabling deeper penetration of clean energy sources into the U.S. power grid while improving grid resilience. Furthermore, the techniques for risk-averse control are widely applicable to other large-scale operational problems that involve the joint management of several entities, including supply chain problems with multiple sources. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.