Hydrogen and Battery Storage Technologies for Low Cost Energy Decarbonization in Distribution Networks (2202.02711v1)

Abstract: Deep energy decarbonization cannot be achieved without high penetration of renewables. At higher renewable energy penetrations, the variability and intermittent nature of solar photovoltaic (PV) electricity can cause ramping issues with existing fossil fuel generation, requiring longer term energy storage to increase the reliability of grid operation. A proton exchange membrane electrolyzer can produce H2and serves as a utility controllable load. The produced H2 can then be stored and converted back into electricity, or mixed with natural gas, or used as transportation fuel, or chemical feedstock. This paper considers the perspective of the distribution system operator that operates the distributed energy resources on a standard IEEE 33-node distribution network considering the technical and physical constraints with the goal of minimizing total investment and operation cost. Different case studies, at very high PV penetrations are considered to show the challenges and path to net-zero emission energy production using H2 energy. Sensitivity of utility PV costs and electrolyzer capital costs on producing H2 at $1/kg are presented showing that the distribution network could produce 100% renewable electricity and H2 could be produced with the same price by 2050 with conservative cost estimates and by 2030 with accelerated cost declines.

• The paper presents an optimization-based framework integrating PV, batteries, and hydrogen systems to minimize costs and emissions in distribution networks.
• It uses techno-economic analysis to evaluate DER scheduling, showing that hydrogen enables viable long-duration storage despite lower round-trip efficiency.
• The study demonstrates that combining short- and long-duration storage reduces PV curtailment and DG ramping issues while enhancing voltage support.

The paper "Hydrogen and Battery Storage Technologies for Low Cost Energy Decarbonization in Distribution Networks" (2202.02711) explores how distribution system operators (DSOs) can integrate high levels of renewable energy, specifically solar photovoltaic (PV), into their networks using battery and hydrogen storage technologies to achieve net-zero emissions while minimizing costs. The core idea is to use an optimization framework to schedule distributed energy resources (DERs) – including traditional natural gas generators (DGs), PV arrays, Li-ion batteries, Vanadium Redox Flow batteries, and hydrogen systems – on a standard distribution network model (IEEE 33-node) considering technical and physical constraints.

From a practical implementation perspective, the paper presents a techno-economic analysis guided by an optimization model. This model aims to determine the optimal dispatch of available resources and, in some cases, the optimal sizing of new assets (batteries, hydrogen system components) over a two-week operational horizon. The objective is to minimize the sum of investment costs (annualized capital expenditure) and operational costs (fuel, maintenance, power purchase from the grid) while meeting load demand and adhering to network physics (power flow, voltage limits) and operational constraints of each DER.

Key Technologies and Their Role in Implementation:

1. Solar PV (PV): Assumed to be utility-scale or aggregated distributed PV. The challenge is its intermittent nature, leading to potential oversupply (curtailment) or undersupply. The paper models PV generation based on a two-week profile, highlighting the need for storage to utilize excess generation.
2. Natural Gas Distributed Generators (DGs): Represent existing fossil fuel assets. The model includes their operational costs, capacity limits, and crucially, ramp rate constraints. High PV penetration exacerbates ramping issues for DGs, making them less flexible for grid balancing. Operational costs are assumed based on future projections (NREL ATB 2050), demonstrating the diminishing role of expensive DGs in a decarbonized future.
3. Battery Energy Storage Systems (BES): Li-ion and Vanadium Redox Flow batteries are considered.
   • Li-ion (4-hour duration): Modeled for shorter-term energy shifting, primarily storing excess PV for use during evening peaks or cloudy periods. High round-trip efficiency (81%) makes it suitable for daily cycling.
   • Vanadium Redox Flow (10-hour duration): Modeled for longer-duration storage compared to Li-ion, offering more flexibility for multi-hour energy shifts. Lower round-trip efficiency (67%) is a trade-off for longer duration. Implementation requires modeling charging/discharging power limits, energy capacity limits, state-of-charge dynamics, and round-trip efficiency losses.
4. Hydrogen System (HS): Composed of a Proton Exchange Membrane (PEM) electrolyzer, compressor, storage tank, and stationary PEM Fuel Cell (FC).
   • Electrolyzer: Consumes electricity (ideally excess renewable) to produce hydrogen. Modeled with efficiency (60%). Acts as a controllable load, absorbing excess PV to avoid curtailment.
   • Compressor: Increases hydrogen pressure for storage. Adds to the energy consumption and cost.
   • Storage Tank: Stores compressed hydrogen. Offers long-duration storage capability (seasonal or multi-day), which batteries typically cannot provide cost-effectively. Modeled by tracking the mass of hydrogen and its capacity limits.
   • Fuel Cell (FC): Converts stored hydrogen back into electricity. Modeled with efficiency (70%). Acts as a dispatchable generator during periods of low renewable generation. The hydrogen system's lower overall round-trip efficiency (electrolyzer + FC, approximately 42%) is a key consideration compared to batteries, but its advantage lies in cost-effective long-duration storage at scale. Implementing this requires modeling the conversion processes, efficiencies, storage dynamics (mass in tank), and power limits of the electrolyzer and fuel cell.

Implementation Architecture (Conceptual):

The paper implicitly suggests an optimization-based control architecture for the DSO:

+---------------------+ +-------------------+ | DSO Control Center | <--> | Grid Monitoring & | | (Optimization Engine)| | Control Systems | +---------------------+ +-------------------+ ^ ^ ^ ^ | | | | +-------+------+------+------+-------+ | Load | PV | DG | BES | HS | | Forecasts | Forecasts| Assets| Assets| Assets| +-------+------+------+------+-------+ | | | | v v v v +-------------------------------------+ | Distribution Network (IEEE 33-node)| | (Power Flow, Voltage Constraints) | +-------------------------------------+

The "Optimization Engine" would solve a complex mathematical program (likely a Mixed-Integer Linear Program or Non-linear Program, depending on how power flow is modeled) periodically (e.g., hourly or daily) for a look-ahead horizon (e.g., two weeks). Inputs include:

• Load demand forecasts.
• PV generation forecasts.
• Operational state of existing DGs.
• Current state of charge (batteries) and hydrogen level (tank).
• Technical parameters and constraints of all DERs (power limits, ramp rates, efficiencies, capacities).
• Network parameters (line impedances).
• Cost parameters (fuel costs, O&M costs, potential market prices, investment costs).

The output of the optimization would be a schedule for each asset: DG dispatch, PV curtailment, battery charge/discharge power, electrolyzer power consumption, and fuel cell power output. These schedules would then be sent as setpoints to the DER controllers via the grid control system.

Practical Applications and Insights from Case Studies:

The case studies provide concrete examples of how different DER combinations impact network operation and decarbonization goals:

• Without storage (Case 2): High PV penetration leads to significant curtailment (58.4%) and reliance on DGs for balancing, showing the necessity of storage.
• With short/long-duration batteries (Cases 3 & 4): Batteries effectively reduce DG usage and PV curtailment compared to Case 2, with longer duration (Redox Flow) offering more benefits in reducing peaking DG operation (DG30).
• With H2 system only (Case 5a, 5b): The H2 system, despite lower efficiency, enables significant reductions in fossil fuel use and PV curtailment, particularly when considering updated DG costs based on lower utilization (Case 5b). This highlights H2's role in absorbing large amounts of excess PV for long-term storage.
• With both batteries and H2 (Cases 6a, 6b, 7a, 7b): Combining storage technologies allows for near or 100% green energy supply (Case 7b). Batteries handle short-term fluctuations, while the H2 system provides long-duration storage. Penalizing curtailment (Case 6b) demonstrates how grid operators can incentivize storage usage for green energy maximization. Case 7b shows that starting with adequate hydrogen in the tank (50%) is crucial for achieving 100% fossil-free operation from the outset of the paper period.
• Voltage Support: The paper notes that inverters associated with PV, batteries, and fuel cells improve the voltage profile across the network compared to fossil fuel generation alone (Fig 11 vs 12). This is a critical technical benefit in distribution networks, where voltage can be a limiting factor for DER hosting capacity. Inverters providing voltage support (e.g., reactive power injection/absorption) would need to be included in a full practical implementation.
• Economic Feasibility: The sensitivity analysis on H2 production cost (Tables II & III) provides practical targets. Achieving $1/kg H2 requires significant reductions in electrolyzer CAPEX (below$100/kW) and low electricity costs ($12-20/MWh from cheap renewables). This sets clear goals for R&D and manufacturing scale-up.

Implementation Considerations and Challenges:

• Computational Requirements: Solving the optimization model (likely an OPF-constrained unit commitment and economic dispatch problem with storage) can be computationally intensive, especially for larger networks or longer planning horizons. Real-time implementation might require approximations, decomposition methods, or advanced solvers.
• Data Dependency: The model relies heavily on accurate load and renewable generation forecasts. Errors in forecasts can lead to suboptimal dispatch and potential reliability issues.
• Model Complexity: Accurately modeling the non-linear dynamics of the distribution network and the complex operational constraints of various DERs (including startup/shutdown costs, degradation, maintenance) adds significant complexity to the optimization problem.
• Asset Sizing: While some cases include optimal sizing, real-world deployment involves complex decisions based on long-term forecasts, site-specific conditions, and grid needs. The optimization approach provides a valuable tool for informing these decisions.
• Interconnection and Grid Code Compliance: Integrating large-scale DERs requires careful planning to meet grid codes, including requirements for fault ride-through, reactive power capability, and grid stability.
• Control and Communication Infrastructure: A robust communication and control system is needed to gather data, send commands to DERs, and execute the optimal schedule in near real-time.

In summary, the paper provides a practical demonstration via simulation of how strategically deploying and operating hydrogen and battery storage alongside PV can enable significant decarbonization in distribution networks. The paper highlights the distinct roles of short-term batteries and long-duration hydrogen storage and provides quantitative insights into their technical and economic impacts on grid operation, including costs, green energy penetration, DG utilization, and voltage profiles. The core implementation challenge lies in developing and deploying the sophisticated optimization and control systems required to manage these diverse assets effectively under variable conditions, alongside achieving the necessary cost reductions for storage technologies.