Learning for Feasible Region on Coal Mine Virtual Power Plants with Imperfect Information

TL;DR

The paper addresses feasible region assessment (FRA) for coal mine industrial energy systems (CMIES) aggregated into a coal mine VPP under imperfect information and privacy constraints. It develops a data-driven inverse-optimization framework that learns a surrogate FRA ${\\tilde\\Omega}_{\\text{V}}({{\\tilde \\Xi}}_{|{\\rm I}|})$ from historical dispatch data by solving an inverse optimization problem constrained by the CMIES energy-transportation model. The proposed Learning-based FRA Algorithm (LFRA) iteratively updates parameters to improve approximation accuracy while reducing computational burden relative to NP-hard Minkowski-sum approaches. Case studies on an IEEE 33-bus system show LFRA achieving lower RMSE/MAE than a neural baseline for key bounds and dispatch variables, enabling reliable market participation with privacy-preserving aggregation. This work advances scalable, data-driven FRA for large-scale CMIES VPPs and lays groundwork for real-time, nonlinear FRA extensions.

Abstract

The feasible region assessment (FRA) in industrial virtual power plants (VPPs) is driven by the need to activate large-scale latent industrial loads for demand response, making it essential to aggregate these flexible resources for peak regulation. However, the large number of devices and the need for privacy preservation in coal mines pose challenges to accurately aggregating these resources into a cohesive coal mine VPP. In this paper, we propose an efficient and reliable data-driven approach for FRA in the coal mine VPP that can manage incomplete information. Our data-driven FRA algorithm approximates equipment and FRA parameters based on historical energy dispatch data, effectively addressing the challenges of imperfect information. Simulation results illustrate that our method approximates the accurate feasible operational boundaries under dynamic and imperfect information conditions.

Figures (4)

• Figure 1: Coal Mine VPP Aggregation with CMIESs
• Figure 2: System configuration
• Figure 3: The parameter identification results of 14 BCs
• Figure 4: The FRA and error results

Theorems & Definitions (1)

• Definition 1