Autonomous Agent-Based Electricity Trading in a Grid-Connected Microgrid

Grishma Manandhar

doi:10.7273/000006346

The increased deployment of distributed energy technologies is transforming regular electricity consumers into active prosumers. This, along with advancements in Information and Communication Technology (ICT) devices, is setting the scene for a new paradigm where consumers can interact and trade their electricity with one another in community microgrids that are interconnected to the larger utility grid network. Local end-node microgid trading can make renewable energy more accessible in communities while also making better use of prosumer decentralized generation.Efficient community microgrids offer win-win possibilities for participants where prosumers can benefit from increased efficiency and gains from trade. At the same time, they enhance grid resiliency and reliability of local electricity delivery, potentially reducing the scope of large-scale power outages by providing isolated local power balancing capability, often refered to as islanding. This research proposes a novel approach to community-based micro-grid trade. Our approach is based on a simultaneous clock auction that is augmented with a cost minimization algorithm to match buyers and sellers and clear the market. Several competition-enhancing design elements are included in our auction to support welfare gains obtained through competitive pricing. An important feature of our auction is that it captures behavioral aspects of prosumers that can be easily implemented through AI agents to facilitate real-time pricing. In this research, we present an AI agent capable of bidding on behalf of the customer ensuring invisible to minimal effect to the residents. The bids are based on thermodynamic adjustments that are made by the AI so that a small change in internal temperature made to optimize the economy-comfort balance in the current period can be readjusted in later periods thereby maintaining an overall forward-looking optimal balance for the present period. Our study demonstrates how our AI agent can find an optimal HVAC load by adjusting the thermostat set points in response to price changes while ensuring that the consumer's preferred set point is restored in the third time period. Additionally, the study further highlights how the AI agent and automated trading can reduce peak load while better utilizing the local renewable resources.

Autonomous Agent-Based Electricity Trading in a Grid-Connected Microgrid

Files and links (1)

Abstract

Metrics

Details