Market Equilibrium in Active Distribution System With µVPPs: A Coevolutionary Approach

Hao Fu, Xiao-Ping ZhangIEEE Access 2017 · DOI

Abstract

The transition toward decentralized energy systems requires market mechanisms that empower small-scale Distributed Energy Resources (DERs). This paper proposes a novel Active Distribution System Market (ADSM) where micro Virtual Power Plants (µVPPs) act as strategic price-makers rather than passive price-takers. Using a bilevel Equilibrium Problem with Equilibrium Constraints (EPEC) modeled via a coevolutionary algorithm, the authors identify a pure strategy Nash Equilibrium. Results indicate that the ADSM significantly enhances µVPP profitability and distributed generation utilization compared to existing market models.

Introduction

Traditional electricity retail markets are often characterized by high entry barriers, dominated by large, vertically integrated suppliers. While the proliferation of DERs like solar photovoltaics and microgrids offers potential for a more resilient grid, these "prosumers" are typically relegated to passive roles. They often function as price-takers or are limited to self-consumption, leaving their full economic and operational potential untapped. This paper addresses the lack of efficient market mechanisms that allow small-scale entities to actively bid and offer energy and reserve services, thereby missing out on the benefits of localized competition.

The Proposed ADSM Framework

The researchers developed the ADSM to mimic wholesale auction dynamics at the distribution level. The market is structured as a bilevel EPEC:

  1. Upper Level (UL): Individual µVPPs seek to maximize their private profits through strategic bidding.
  2. Lower Level (LL): The Distribution System Operator (DSO) clears the market, aiming to maximize overall social welfare while respecting physical grid constraints.
Comparison of Current DSM, PDSM, and the proposed ADSM frameworks.

To solve this complex problem, the authors move away from traditional mathematical reformulations like Karush-Kuhn-Tucker (KKT) conditions, which often struggle with non-convexity and require restrictive assumptions. Instead, they employ a coevolutionary approach—an agent-based simulation where different µVPP "species" evolve strategies simultaneously using a real-coded Genetic Algorithm (GA). This method effectively identifies the pure strategy Nash Equilibrium (NE), a stable state where no player can improve their profit by unilaterally changing their strategy.

Experimental Setup

The model's validity was tested using real-world data and standard benchmarks:

  • Market Data: Retail energy and reserve prices were sourced from the 2016 Nord Pool data (UK).
  • Network Benchmark: A modified IEEE 33-bus system was used to simulate physical grid constraints.
  • µVPP Profiles: Four distinct profiles with varying Renewable Energy Source (RES) penetration (0% to 70%) were created to evaluate diverse competitive scenarios.

A key insight from the study is that a µVPP’s optimal strategy is highly context-dependent, relying not just on its internal costs but also on the regional penetration of renewables and the competitive behavior of rivals.

Results and Analysis

The paper evaluates three distinct market structures: the current Distribution System Market (DSM), a Passive DSM (PDSM), and the proposed Active DSM (ADSM).

  • Profitability: In the ADSM, a representative µVPP (µVPP3) saw a 14% increase in total profit compared to the current DSM.
  • Asset Utilization: Distributed Generation (DG) utilization rose from 5% in the current system to 31% in the ADSM. While the PDSM showed higher theoretical utilization (up to 65%), it was characterized by high volatility and lacked a stable equilibrium, making it an unattractive option for risk-averse investors.
  • Algorithmic Performance: The coevolutionary algorithm successfully converged to a Nash Equilibrium in 253 iterations. While the computation took approximately 14 hours, it demonstrated the ability to handle the non-linearities inherent in complex retail markets.

Conclusion and Future Directions

The ADSM framework offers a promising path toward a more democratic and efficient distribution-level energy market. By allowing µVPPs to act as strategic players, the system achieves higher stability and better returns on investment. However, several challenges remain. The computational intensity of the coevolutionary approach suggests a need for optimization to support real-time operations. Furthermore, future iterations of the model should incorporate stochastic variables to account for the inherent uncertainty of wind and solar generation, as well as more detailed physical network constraints like voltage limits and line losses.

Summary updated

BibTeX
@article{fu2017market,
  author = {Hao Fu and Xiao-Ping Zhang},
  title = {{Market Equilibrium in Active Distribution System With µVPPs: A Coevolutionary Approach}},
  journal = {IEEE Access},
  year = {2017},
  doi = {10.1109/access.2017.2691316},
}
Figure