Source-linked AI summary
A Multiobjective Approach to Multimicrogrid System Design
Wei-Yu Chiu, Hongjian Sun, H. Vincent Poor
TL;DR
Existing aggregate-utility designs can neglect emergency storage and may favor particular participants. This paper formulates a multiobjective design problem and uses an MOIA to approximate Pareto-optimal prices and power distribution. Simulations report fair participant treatment and satisfaction of physical constraints during operation.
Problem
Existing aggregate-utility designs can neglect emergency-operation storage, seldom address relationships among participant utilities, and may favor particular participants.
Source: §I · p. 2Method
The paper formulates microgrid utility, power-grid revenue, and stored energy as objectives, then uses an MOIA to approximate a Pareto front and select a fair design.
Source: §I · p. 2Results
The simulated operation considered all participants' interests, satisfied physical constraints, and adjusted power demand and stored-energy levels properly.
Source: §V · p. 8Takeaways & Limitations
The approach produces Pareto-optimal market prices and power distribution for the MO and DNO while seeking a fair design across participants.
Source: §VI · p. 8The work addresses high-level multi-microgrid design through market pricing and power distribution.
Source: §I · p. 1
Abstract
from arXiv · showhide
The main goal of this paper is to design a market operator (MO) and a distribution network operator (DNO) for a network of microgrids in consideration of multiple objectives. This is a high-level design and only those microgrids with nondispatchable renewable energy sources are considered. For a power grid in the network, the net value derived from providing power to the network must be maximized. For a microgrid, it is desirable to maximize the net gain derived from consuming the received power. Finally, for an independent system operator, stored energy levels at microgrids must be maintained as close as possible to storage capacity to secure network emergency operation. To achieve these objectives, a multiobjective approach is proposed. The price signal generated by the MO and power distributed by the DNO are assigned based on a Pareto optimal solution of a multiobjective optimization problem. By using the proposed approach, a fair scheme that does not advantage one particular objective can be attained. Simulations are provided to validate the proposed methodology.
I. INTRODUCTION
The paper frames multi-microgrid design as a conflict among microgrid utility, power-grid revenue, and stored energy for emergency operation. It proposes a Pareto-based approach to generate prices and power distribution without favoring one participant.
I. INTRODUCTION
- The system includes an MO that generates price signals and a DNO that distributes power, with microgrids using nondispatchable renewables and some also using storage.Source: §I · p. 1
- The high-level design jointly optimizes microgrid and power-grid asset use and stored energy for ISO emergency operation.Source: §I · p. 1
- Existing aggregate-utility designs can neglect emergency-operation storage, leave relationships among participant utilities unaddressed, and favor participants depending on weighting and objective conflicts –1.Source: §I · pp. 1–2
- The proposed formulation maximizes microgrid net value, power-grid net revenue, and the sum of stored energy levels as three objectives.Source: §I · p. 2
- An MOIA searches feasible decision points for nondominated solutions, producing an approximate Pareto front that shows how objectives affect one another.Source: §I · p. 2
- The design selects the Pareto-front vector with the greatest minimum normalized improvement across all three dimensions to avoid favoring an extreme participant.Source: §I · p. 2
II. SYSTEM MODEL
The system model covers a network of microgrids and its high-level design. Its assumptions include no power exchange among distinct microgrids and information exchange among network participants.
II. SYSTEM MODEL
- The model assumes no power exchange between microgrids; linked microgrids that can exchange power are combined into one.Source: §II · pp. 2–3
- An ICT system exchanges network information, including prices and power demand, among microgrids, the power grid, and the ISO.Source: §II · p. 3
- The study models high-level system design and omits current- and voltage-balance equations because those are managed at a lower level.Source: §II · p. 3
A. Microgrids
The microgrid model represents demand, renewable generation, and optional local storage. It defines microgrid utility as the net value of consumed power at the market price.
A. Microgrids
- Microgrids are modeled with nondispatchable renewable sources and no dispatchable distributed-generation units.Source: A. Microgrids · p. 3
- For microgrids with storage, stored energy is bounded by maximum capacity, with charging and discharging rates subject to limits.Source: A. Microgrids · p. 3
- Demand includes shiftable loads and a known nominal base load, which the model assumes can be forecast accurately.Source: A. Microgrids · p. 3
- The microgrid objective is to maximize overall utility, defined as the net value of consumed power at price λ(k).Source: A. Microgrids · p. 3
B. Power Grid
The power-grid model evaluates the grid's utility from the power exchanged with microgrids at the market price.
B. Power Grid
- The power-grid utility function Ug(pg(k), λ(k)) represents net value derived from pg(k) at price λ(k), and the model seeks to maximize the grid's interest.Source: B. Power Grid · p. 3
C. ISO
The model uses stored energy in microgrids to support emergency operation, with price and distributed power determined as decision variables. A Taiwan dormitory scenario illustrates how time-varying prices and existing facilities could support this setup.
C. ISO
- The emergency-operation objective seeks to maximize total stored energy while maintaining a minimum energy level required for emergency operation.The model aims to store as much energy in the network as possible to achieve a safer secure level.Source: C. ISO · p. 4
- The price λ(k) and power distributed to each microgrid pgn(k) are decision variables because distributed power affects stored energy and depends on price.Source: C. ISO · p. 4
- A Taiwan dormitory scenario illustrates time-varying prices that may let students shift washing-machine and dryer use, benefiting both students and universities.Universities may level load and reduce costs and energy losses, while students may reduce expenses.Source: C. ISO · p. 4
- The scenario maps dormitories with batteries to microgrids and identifies rooftop solar, smartphones, and metering infrastructure as supporting facilities.The paper notes that Yuan Ze University already has an AMI system.Source: C. ISO · p. 4
III. PROBLEM FORMULATION AND RELATED ANALYSIS
The formulation jointly addresses microgrid utility, power-grid revenue, and stored energy as a multiobjective problem. It treats system stability as feasibility and uses Pareto optimality to identify candidate designs.
III. PROBLEM FORMULATION AND RELATED ANALYSIS
- The formulation assumes hourly real-time market settlement and treats forecast renewable output as known during each optimization.Source: §III · p. 4
- The design problem simultaneously addresses objectives (6), (8), and (9) through a multiobjective formulation.Source: §III · p. 4
- The decision variables are the MO's price λ(k) and DNO's distributed power pgn(k), while implementation costs for existing RES and storage infrastructure are excluded.The paper frames dynamic pricing as high-level energy management using infrastructure that already exists.Source: §III · p. 4
- The authors define system stability as feasibility at every time slot and prove stability by choosing distributed power so stored energy remains unchanged.The proof says this construction is feasible regardless of the assigned price.Source: §III · p. 4
- Pareto dominance compares objective vectors across all dimensions, requiring no-worse values everywhere and a strict improvement somewhere.Source: §III · p. 5
- Theorem 2 establishes equivalence between the constrained problem and its reformulation using Uc, provided Uc(p*) = 0; infeasible Pareto points are then removed.The remaining feasible points are Pareto optimal solutions of the original problem.Source: §III · p. 5
IV. PROPOSED ALGORITHM FOR MULTI-MICROGRID SYSTEM DESIGN
The proposed MOIA searches for feasible nondominated designs using an artificial immune system structure. It selects an output that maximizes the minimum normalized improvement across objectives.
IV. PROPOSED ALGORITHM FOR MULTI-MICROGRID SYSTEM DESIGN
- The method adapts artificial immune system structures for Pareto search and calls each decision-space point an antibody.Source: §IV · p. 5
- The algorithm maintains nondominated, feasible antibodies as an approximation of Pareto-optimal solutions, then selects the candidate maximizing minimum normalized improvement across dimensions.The selected solution determines price generation and power distribution.Source: §IV · p. 5
- The pseudocode tracks the current population, nominal and maximum population sizes, iteration counter, and maximum iteration number.Source: §IV · p. 5, Figure 2
- Dominated antibodies are removed so the population's objective vectors can approach the Pareto front as iterations proceed.Source: §IV · p. 5
- Gene operations preserve antibody diversity, while infeasible and low-quality candidates are removed to manage population size.The algorithm removes infeasible candidates with Uc(p) > 0 and prunes the population to its nominal size.Source: §IV · p. 6
V. NUMERICAL RESULTS
Simulations show that selected Pareto solutions supported operation considering all participants while meeting physical constraints and adjusting demand and stored energy. The results also show how power distribution responds to renewable input and how energy storage behaves under multiple objectives.
V. NUMERICAL RESULTS
- The three-microgrid simulation modeled demand as percentage shifts from base load, increasing at low prices and decreasing at high prices.Two microgrids had storage, with maximum capacities of 250 and 200 kWh and secure levels of 125 and 100 kWh.Source: §V · p. 6, Figure 3 · p. 7, Figure 4 · p. 7
- Normalized nondominated vectors lay in the middle of the sampled objective-front graphs, indicating a design that does not favor one participant.The figure samples times k = 12, 24, and 36.Source: §V · p. 7, Figure 5 · p. 8
- Distributed power increased as renewable input decreased and decreased as renewable input increased; demand was usually larger than grid supply.Source: §V · p. 7, Figure 6 · p. 8
- Stored energy levels vibrated with time-varying power flows, renewable input, and price, whereas an ISO-only objective could eventually attain total storage capacity.Source: §V · p. 7
- High prices need not produce storage discharge when only a small portion of demand is shiftable and the ISO prioritizes secure energy levels.Source: §V · pp. 7–8
- Selected Pareto solutions supported operation considering all participants, satisfied physical constraints, and adjusted power demand and stored energy levels properly.Source: §V · p. 8
VI. CONCLUSION
The paper formulates multi-microgrid design as a multiobjective problem and uses MOIA to produce Pareto-optimal prices and power distribution. The authors also argue that the approach can accommodate other objectives and constraints.
VI. CONCLUSION
- The MOP and MOIA simultaneously maximize microgrid, power-grid, and ISO utilities, producing Pareto-optimal market prices and power distribution for the MO and DNO.The ISO objective concerns stored energy at microgrids.Source: §VI · p. 8
- The conclusion’s figures depict microgrid stored-energy levels and the price signal generated by the MO.Source: Figure 7 · p. 8, Figure 8 · p. 9
- The approach can accommodate other objectives and constraints because MOIA does not rely on particular objective or constraint-function structures.The authors describe the methodology as general and flexible.Source: §VI · p. 8