Source-linked AI summary
Optimal sizing of renewable energy storage: A comparative study of hydrogen and battery system considering degradation and seasonal storage
Son Tay Le, Tuan Ngoc Nguyen, Dac-Khuong Bui, Tuan Duc Ngo
TL;DR
The study addresses the need for energy storage under complex, dynamic sizing conditions by investigating an optimised long-duration strategy for hydrogen energy storage systems. MOMFA outperforms NSGA-II, while NSGA-II may require multiple runs to ensure the global Pareto front is identified.
Problem
Complex and dynamic sizing of battery and hydrogen energy-storage components creates a need for suitable optimisation methods.
Method
The study investigates an optimised long-duration strategy for hydrogen energy storage systems.
Results
MOMFA outperforms NSGA-II, whose optimisation results are inconsistent across SSR = 65% and SSR = 72%.
Takeaways & Limitations
NSGA-II may require multiple runs to ensure that the global Pareto front has been identified.
Takeaways & Limitations
Some NSGA-II runs may fall short of finding the local Pareto front, creating a risk in identifying the global Pareto front.
Abstract
from arXiv · showhide
Renewable energy storage (RES) is essential to address the intermittence issues of renewable energy systems, thereby enhancing the system stability and reliability. This study presents an optimisation study of sizing and operational strategy parameters of a grid-connected photovoltaic (PV)-hydrogen/battery systems using a Multi-Objective Modified Firefly Algorithm (MOMFA). An operational strategy that utilises the ability of hydrogen to store energy over a long time was also investigated. The proposed method was applied to a real-world distributed energy project located in the tropical climate zone. To further demonstrate the robustness and versatility of the method, another synthetic test case was examined for a location in the subtropical weather zone, which has a high seasonal mismatch. The performance of the proposed MOMFA method is compared with the NSGA-II method, which has been widely used to design renewable energy storage systems in the literature. The result shows that MOMFA is more accurate and robust than NSGA-II owing to the complex and dynamic nature of energy storage system. The optimisation results show that battery storage systems, as a mature technology, yield better economic performance than current hydrogen storage systems. However, it is proven that hydrogen storage systems provide better techno-economic performance and can be a viable long-term storage solution when high penetration of renewable energy is required. The study also proves that the proposed long-term operational strategy can lower component degradation, enhance efficiency, and increase the total economic performance of hydrogen storage systems. The findings of this study can support the implementation of energy storage systems for renewable energy.
Nomenclature
The nomenclature defines storage-system abbreviations, grid-energy timing variables, and storage limits for sunny and cloudy periods.
- Grid-energy variables distinguish one-hour intervals at off-peak, peak, and shoulder rates.
- LIMIT_CLOUDY and LIMIT_SUNNY denote storage limits during cloudy and sunny periods, bounded by the sunny-period start and end times.
- BESS denotes Battery Energy Storage System, while HESS denotes Hydrogen Energy Storage System.
- PEMEL and PEMFC denote the Proton Exchange Membrane Electrolyser and Proton Exchange Membrane Fuel Cell, respectively.
- SSR denotes Self Sufficiency Ratio.
1 Introduction
The introduction motivates storage for intermittent renewable energy and identifies a gap in jointly optimising system sizing, operation, degradation, and changing electricity prices. It presents a MOMFA-based framework and a hydrogen strategy for long-duration or seasonal storage, comparing battery and hydrogen options.
- Motivation: Renewable-energy adoption is challenged by geographic and weather dependence, uncontrollable resources, intermittency, supply-demand mismatches, power-quality issues, and network constraints.
- Motivation: Energy storage enables energy produced at one time to be used later and supports load shifting, PV-curtailment reduction, and peak shaving.
- Storage technologies: Batteries are mature, flexible, reliable, economical, responsive, and effective for smoothing load fluctuations and short-term changes.
- Storage technologies: Higher renewable penetration increases the importance of long-duration storage, for which hydrogen can provide a flexible option for backing up intermittent renewable sources.
- Research gap: Prior sizing studies commonly fixed the operational strategy, limiting suitability for long-term hydrogen storage; few jointly optimise storage capacity and operation.
- Research gap: Previous work examined coordinated sizing and operation, but did not account for future electricity-price variation or component degradation over the analysis period.
- Study contribution: This study develops a framework that simultaneously optimises renewable-system sizing and energy management using MOMFA while considering degradation, price variations, and operational strategies.
- Study contribution: A two-mode summer-winter operational strategy is introduced to use hydrogen for long-period storage, with analysis extending across a 25-year project lifetime.
2 Methodology
The methodology section introduces the system configurations and the proposed optimal-sizing framework.
- The study describes system configurations together with the proposed optimal sizing and operational strategies.
2.1 System configurations
The study models grid-connected PV systems with battery or hydrogen storage, comparing conventional operation with an optimised long-duration strategy for hydrogen. It also specifies cost scenarios, degradation assumptions, and component-level storage models.
- System configurations: Three configurations use rooftop PV, an electricity load, the external grid, and either a battery or hydrogen energy storage system.Battery and hydrogen systems follow conventional operation in Cases 1 and 2; Option 3 applies the optimised long-duration strategy to hydrogen storage.
- Cost assumptions: Two cost scenarios are evaluated: current costs from prior reviews and ultimate costs based on IEA and United States Department of Energy targets.The study notes substantial variation in BESS and HESS component costs across the reviewed literature.
- Hydrogen energy storage system: The hydrogen system combines a PEM electrolyser, hydrogen tank, and PEM fuel cell, allowing charge power, storage capacity, and discharge power to be sized independently.The electrolyser determines charging power, the tank determines storage capacity, and the fuel cell determines discharge power.
- Hydrogen energy storage system: Hydrogen storage has a flexible energy-to-power ratio, whereas lithium-battery ratios are usually under eight hours.This decoupling makes hydrogen suitable for applications requiring longer-duration storage.
- Operational strategies: The optimised long-duration strategy stores hydrogen during sunny months and consumes it during cloudy months to represent seasonal storage.It introduces sunny- and cloudy-period limits and time indicators, while modifying discharge decisions during low-price off-peak hours.
2.2 Optimization method and implementation
The study optimises PV-energy storage sizing and operation to maximise NPV and SSR. NPV captures economic performance, while SSR measures the share of demand supplied by renewable energy.
- The optimisation targets component sizing and operation variables for battery and hydrogen energy storage systems.
- NPV represents the economic incentives of the renewable energy system, while SSR represents renewable electricity supplied to demand.
- System revenue equals electricity-bill savings from using expanded PV and energy storage instead of the baseline grid-supplied system.
- PV and storage costs include capital, operation and maintenance, and replacement costs over the project period.
- SSR is the percentage of electricity demand supplied by renewable energy.
2.3 Optimisation Algorithm
The study uses MOMFA to search for Pareto-optimal PV-storage sizing and operating strategies. The algorithm combines modified firefly mechanisms with chaotic initialisation, adaptive randomness, and Lévy flights.
- MOMFA optimises BESS and HESS sizing and operational variables.
- The modified firefly algorithm uses firefly brightness to determine the objective function and attract other fireflies.
- The initial population is generated with a Logistic map, while chaotic and absorption parameters are set using sensitivity analysis and literature.
- Firefly movement includes attraction, randomisation, and Lévy-flight components, with randomisation reduced over iterations.
- For simultaneous NPV and SSR maximisation, MOMFA searches for non-dominated solutions forming a Pareto front.
3 Case study
The case study evaluates grid-connected PV-storage systems using real monitored data from a tropical warehouse in Ho Chi Minh City and a synthetic seasonal-mismatch case in Melbourne.
- The real-world distributed energy project is located in Ho Chi Minh City, Vietnam, in a tropical climate.
- The warehouse rooftop PV system produces 1.5 million kWh/year, and monitored PV and electricity-use data support the analysis.
- The study expands PV capacity and adds storage to increase renewable-energy penetration and system efficiency.
- PV output is modelled with linear expansion and a 0.55% annual depreciation rate; expanded solar costs $881/kW.
- The PV lifetime is 25 years, with annual operation and maintenance costs assumed at 1% of CAPEX.
- A hypothetical Melbourne case represents a subtropical location with high seasonal mismatch, using HOMER-generated solar output.
4 Result and discussion
MOMFA produces more consistent Pareto fronts than NSGA-II and is used to compare battery and hydrogen storage across climates, cost scenarios, and operational strategies. Batteries perform better economically for short-duration storage, whereas hydrogen becomes advantageous for very high renewable penetration and seasonal storage.
- Algorithm comparison: MOMFA outperforms NSGA-II by densely filling the Pareto-front region across repeated runs.
- Algorithm comparison: NSGA-II produces separate fronts around SSR = 65% and SSR = 72%, with results varying by initial random conditions.
- Current-cost comparison: At current costs, BESS achieves higher NPV and SSR than HESS, reflecting hydrogen’s current cost challenge.
- High renewable penetration: When SSR exceeds 95%, hydrogen becomes preferable, while BESS has no optimal solution approaching 100% renewable supply.
- Climate comparison: HESS dominates in Melbourne at SSR = 80%, whereas BESS performs well in Ho Chi Minh City up to 95%.
- Climate comparison: Subtropical systems are larger and have NPCs 10 to 80% greater than tropical systems under the same cases and cost scenarios.
- Operational strategy: The long-term operational strategy increases HESS NPV by 17% in Ho Chi Minh City and 32% in Melbourne.
5 Conclusion
The study optimises sizing and operation of grid-connected PV-hydrogen/battery systems using MOMFA, evaluating real and synthetic locations with seasonal variation. BESS performs better economically at current costs and low-to-medium SSR, while HESS is viable for high renewable penetration and benefits from OLDS.
- 5 Conclusion: MOMFA optimises sizing and operational strategy for grid-connected PV-hydrogen/battery storage systems.The framework considers electricity price forecasting, solar output depreciation, and storage-component degradation.
- 5 Conclusion: The study evaluates a real-world tropical project and a synthetic subtropical case with high seasonal variations.The tropical case uses actual monitored data, while the subtropical case examines performance across different geographic conditions.
- 5 Conclusion: MOMFA is more accurate and robust than the popular NSGA-II algorithm for this storage optimisation problem.The comparison addresses the complex and dynamic nature of energy storage systems.
- 5 Conclusion: BESS outperforms HESS economically under current costs and low-to-medium SSR, yielding a better Pareto front with higher SSR and NPV.The comparison concerns the solar system using BESS versus HESS.
- 5 Conclusion: HESS is viable for SSR>80 under ultimate costs, especially with OLDS in locations with high seasonal variations.This identifies a setting where hydrogen's long-term storage potential becomes advantageous.
- 5 Conclusion: OLDS lowers HESS component degradation and enhances storage efficiency and total economic performance over the project lifetime.The strategy controls HESS functioning through distinct seasonal operation modes.