Robust Energy Management for Microgrids With High-Penetration Renewables (1207.4831v3)

Abstract: Due to its reduced communication overhead and robustness to failures, distributed energy management is of paramount importance in smart grids, especially in microgrids, which feature distributed generation (DG) and distributed storage (DS). Distributed economic dispatch for a microgrid with high renewable energy penetration and demand-side management operating in grid-connected mode is considered in this paper. To address the intrinsically stochastic availability of renewable energy sources (RES), a novel power scheduling approach is introduced. The approach involves the actual renewable energy as well as the energy traded with the main grid, so that the supply-demand balance is maintained. The optimal scheduling strategy minimizes the microgrid net cost, which includes DG and DS costs, utility of dispatchable loads, and worst-case transaction cost stemming from the uncertainty in RES. Leveraging the dual decomposition, the optimization problem formulated is solved in a distributed fashion by the local controllers of DG, DS, and dispatchable loads. Numerical results are reported to corroborate the effectiveness of the novel approach.

• The paper introduces a worst-case robust power scheduling model that minimizes transaction costs under high renewable variability.
• It employs dual decomposition for distributed economic dispatch, enabling local controllers to efficiently coordinate energy management.
• Extensive simulations validate that the convex optimization framework effectively balances supply and demand under varying market conditions.

Robust Energy Management for Microgrids With High-Penetration Renewables

The paper addresses the problem of optimal energy management in microgrids with significant penetration of renewable energy sources (RES). Key objectives include minimizing the net cost of the microgrid while balancing supply and demand amidst the intrinsic uncertainties of renewable power output. The problem considers distributed energy management with the added complexity of convexity through robust optimization and dual decomposition methods, making it not only relevant but also practical for large-scale implementation.

Key Contributions

1. Robust Formulation With Worst-Case Optimization: The paper introduces a novel approach to manage the intrinsically stochastic availability of RES by developing a robust power scheduling model. The novelty lies in addressing the power scheduling problem by considering both the actual renewable energy harvested and the energy traded with the main grid to maintain a supply-demand balance. This robust optimization strategy accounts for uncertainties in RES and minimizes the worst-case transaction cost.
2. Distributed Economic Dispatch: Leveraging the dual decomposition technique, the optimization problem is solvable in a distributed manner. This method allows local controllers of distributed generation (DG), distributed storage (DS), and dispatchable loads to operate cohesively yet independently, enhancing the system's robustness and reducing communication overheads.
3. Convexity and Efficiency: By formulating the transaction cost based on worst-case scenarios, the paper establishes a condition for the convexity of the overall optimization problem. This condition ensures that the combined problem of economic dispatch, demand-side management, and RES scheduling remains tractable and can be efficiently solved. The paper further refines the computational process using the proximal bundle method for improved convergence of the non-smooth subproblems.
4. Numerical Validation: Through extensive numerical simulations, the paper validates the effectiveness of the proposed approach. It compares scenarios with different transaction prices, demonstrating the impact of high transaction prices on the scheduling and overall costs. The results corroborate the merit of the robust and distributed energy management strategy, showing how it adapts to varying uncertainties and price conditions.

Methodology

Load Demand Model

The load demand comprises both inelastic and elastic loads. Elastic loads, further classified into class-1 and class-2, allow for dynamic scheduling. Class-1 loads have flexible power consumption within given limits, while class-2 loads, such as plug-in hybrid electric vehicles (PHEVs), have specific energy requirements over a predefined period.

Distributed Storage Model

The model accounts for distributed storage units (DS), including their (dis)charging constraints and efficiency. The optimization incorporates storage costs to maximize the lifetime of the DS units.

Worst-Case Transaction Cost

To handle RES uncertainties, the paper employs a robust optimization framework that maximizes the worst-case transaction cost, defined as the cost of power imbalance adjusted for the purchase and selling prices of energy. This worst-case scenario ensures resilience against the variability of RES.

Distributed Algorithm

The core of the algorithm is dual decomposition. By relaxing certain constraints, the problem is decomposed into smaller subproblems solvable by respective local controllers. The microgrid energy manager (MGEM) coordinates these local controllers, iterating on Lagrange multipliers to converge on the optimal solution.

Practical Implications and Future Directions

This paper has significant implications for the deployment and management of microgrids. The distributed and robust optimization framework ensures that microgrids can effectively integrate a high penetration of RES, enhancing both reliability and economic efficiency. The reduction in communication overheads and failure-resilience makes the approach suitable for real-world applications where infrastructure might be limited.

Future work could explore extensions to this model, including:

• Dynamic Pricing Models: Enhancing the framework to incorporate dynamic pricing strategies for demand response optimization.
• Advanced Storage Solutions: Evaluating new storage technologies and further refining the storage cost models.
• Integration of Machine Learning: Leveraging machine learning for more accurate forecasting of RES availability and demand patterns.

Additionally, classical power system problems such as optimal power flow (OPF) and unit commitment (UC) could be revisited under this robust and distributed framework, further contributing to the advancement of smart grid technologies.