Hydrogen Supply Chain Planning with Flexible Transmission and Storage Scheduling (2008.07611v1)

Abstract: Hydrogen is becoming an increasingly appealing energy carrier, as the costs of renewable energy generation and water electrolysis continue to decline. Developing modelling and decision tools for the H${2}$ supply chain that fully capture the flexibility of various resources is essential to understanding the overall cost-competitiveness of H${2}$ use. To address this need, we have developed a H${2}$ supply chain planning model that determines the least-cost mix of H${2}$ generation, storage, transmission, and compression facilities to meet H${2}$ demands and is coupled with power systems through electricity prices. We incorporate flexible scheduling for H${2}$ trucks and pipeline, allowing them to serve as both H${2}$ transmission and storage resources to shift H${2}$ demand/production across space and time. The case study results in the U.S. Northeast indicate that the proposed framework for flexible scheduling of H${2}$ transmission and storage resources is critical not only to cost minimization but also to the choice of H${2}$ production pathways between electrolyzer and centralized natural-gas-based production facilities. Trucks as mobile storage could make electrolyzer more competitive by providing extra spatiotemporal flexibility to respond to the electricity price variability while meeting H$_{2}$ demands. The proposed model also provides a reasonable trade-off between modeling accuracy and solution times.

• The paper presents a MILP model integrating hydrogen production, storage, transmission, and flexible scheduling to enhance cost-efficiency.
• It models trucks as mobile storage and pipelines as linepack, enabling dynamic management of hydrogen production linked to variable electricity prices.
• The approach guides infrastructure planning and operational strategies, highlighting trade-offs between centralized SMR and distributed electrolysis solutions.

This paper, "Hydrogen Supply Chain Planning with Flexible Transmission and Storage Scheduling" (2008.07611), presents a modeling framework to optimize the design and operation of a hydrogen (H₂) supply chain. The primary goal is to determine the least-cost combination of H₂ generation, storage, transmission, and compression facilities required to meet specific H₂ demands. A key innovation is the explicit modeling of flexible scheduling for H₂ trucks and pipelines, allowing them to function not only as transmission methods but also as short-term storage, thereby enabling better spatiotemporal management of H₂ supply and demand. The model is also designed to interface with the power sector by considering electricity prices, which is particularly relevant for electrolytic hydrogen production.

Core Model and Implementation

The model is formulated as a mixed-integer linear program (MILP), a common approach for energy system optimization problems. The objective function aims to minimize the total annualized cost of the H₂ supply chain, which includes:

• Capital Costs (CAPEX): Investment in new facilities like electrolyzers, steam methane reformers (SMR), stationary storage units (e.g., tanks, salt caverns), pipelines, and truck fleets.
• Operational Costs (OPEX): Costs associated with running the facilities, including:
  • Electricity costs for electrolyzers.
  • Natural gas feedstock costs for SMR.
  • Maintenance costs for all equipment.
  • Transportation costs for trucks (fuel, labor).
  • Compression energy costs.

Key Decision Variables in the Optimization:

• Investment Decisions:
  • Capacities of H₂ production facilities (electrolyzers, SMR).
  • Capacities of H₂ storage facilities (stationary storage, pipeline linepack capacity).
  • Capacities of H₂ transmission infrastructure (number of trucks, pipeline diameter/length).
  • Capacities of compression units.
• Operational Decisions (scheduled typically over hourly or sub-hourly time steps for a representative period):
  • H₂ production rate from each facility.
  • H₂ flow rates through pipelines and trucks.
  • Amount of H₂ stored in or withdrawn from storage units (including mobile storage via trucks and linepack in pipelines).
  • Electricity consumption by electrolyzers.
  • Timing and routing of H₂ truck deliveries.

Constraints:

The optimization is subject to various constraints to ensure a feasible and realistic supply chain operation:

• Demand Satisfaction: H₂ supplied must meet the demand at various locations and times.
• Capacity Limits: Operation of each facility cannot exceed its installed capacity.
• Storage Balance: Mass balance for H₂ in storage units (inventory at the end of a period = inventory at the start + inflow - outflow).
• Transmission Limits: Flow rates in pipelines and trucks are limited by their capacities.
• Ramp Rates: Constraints on how quickly production facilities (especially electrolyzers) can change their output.
• Truck Scheduling: Constraints related to truck travel times, loading/unloading times, and availability.
• Pipeline Dynamics: Simplified representation of pipeline pressure and flow relationships, including linepack storage.
• Resource Availability: Limits on natural gas supply or electricity available for electrolysis (though often modeled via price).

Flexible Transmission and Storage Scheduling

A significant contribution of this paper is the detailed modeling of flexibility in transmission assets:

• Trucks as Mobile Storage: Trucks are not just for point-to-point delivery. The model can decide to dispatch trucks to pick up H₂ when it's cheap to produce (e.g., low electricity prices for electrolysis) and hold it (effectively storing it) until it's needed or until prices for alternative supply are higher. This requires tracking the state (location, H₂ inventory) of individual trucks or truck fleets over time.
  • Implementation: This involves time-indexed variables for truck location, H₂ load, and travel status. Constraints would manage travel times between nodes (production, storage, demand centers) and loading/unloading durations.
• Pipeline Linepack: The inherent capability of a gas pipeline to store gas by varying its pressure is modeled. This allows the pipeline itself to buffer short-term fluctuations between H₂ production and consumption.
  • Implementation: Linepack is often modeled as a storage unit associated with the pipeline, with its capacity related to the pipeline's physical characteristics (length, diameter) and allowable pressure range. Constraints link flow rates to changes in linepack inventory.

Pseudocode for Model Logic (Conceptual):

Objective: Minimize Total_Annualized_Cost (CAPEX_Facilities + OPEX_Operations) Subject to: // Investment Variables For each technology (electrolyzer, SMR, storage, pipeline, truck_fleet): Invest_Capacity[tech] >= 0 // Operational Variables (for each time step t, location i, technology tech) Production[tech, i, t] Storage_Level[storage_tech, i, t] Flow_Pipeline[pipeline, t] Flow_Truck[truck, origin, destination, t] Electricity_Consumption[electrolyzer, i, t] // Constraints For each time step t, demand_location j: Sum_Supply_to_Demand[j, t] == Demand[j, t] For each technology tech, location i, time step t: Production[tech, i, t] <= Invest_Capacity[tech, i] // ... and other operational limits (ramp rates etc.) For each storage unit s, location i, time step t: Storage_Level[s, i, t] == Storage_Level[s, i, t-1] + Inflow[s, i, t] - Outflow[s, i, t] Storage_Level[s, i, t] <= Invest_Capacity[s, i] For each pipeline p, time step t: Flow_Pipeline[p, t] <= Pipeline_Capacity[p] // Linepack balance equations For each truck k, time step t: // Truck movement, loading, unloading constraints // Truck inventory balance Electricity_Cost = Sum(Electricity_Consumption[electrolyzer, i, t] * Electricity_Price[i, t]) // ... other cost components (natural gas, O&M) Total_Annualized_Cost = Annualize(Sum(CAPEX_Facilities)) + Sum(OPEX_Operations over representative days/year)

Case Study and Key Findings (U.S. Northeast)

The model was applied to a case paper in the U.S. Northeast. The results demonstrated:

1. Value of Flexibility: Flexible scheduling of H₂ transmission (trucks) and pipeline linepack significantly reduces overall system costs. This flexibility allows the system to absorb variability in renewable electricity prices for electrolysis and to defer or reduce investments in dedicated stationary storage.
2. Electrolyzer Competitiveness: Trucks acting as mobile storage enhance the economic viability of electrolyzers. Electrolyzers can produce H₂ when electricity is cheap, and the H₂ can be stored in trucks temporarily and delivered later, decoupling immediate production from immediate demand. This helps electrolyzers compete against centralized SMR, even if SMR has lower constant production costs.
3. Technology Choice Impact: The model showed that incorporating this flexibility influences the optimal mix of H₂ production technologies. Without flexible mobile storage, centralized SMR might be favored. With it, distributed electrolysis becomes more attractive due to its ability to leverage fluctuating electricity prices.
4. Model Performance: The proposed model provides a reasonable balance between modeling accuracy (capturing operational details like truck scheduling) and computational tractability, making it usable for practical planning exercises.

Practical Applications and Implementation Considerations

• Infrastructure Planning: This model can be used by energy planners, policymakers, and private investors to:
  • Identify optimal locations and capacities for H₂ production, storage, and transmission assets.
  • Assess the economic feasibility of different H₂ supply pathways (e.g., green H₂ via electrolysis vs. blue H₂ via SMR with CCS, though CCS is not explicitly detailed in the abstract, SMR is).
  • Understand the value of investing in flexible assets like truck fleets or pipelines with significant linepack capability.
• Operational Strategy: The model can inform operational strategies, such as optimal dispatch of electrolyzers based on electricity price forecasts and coordinated scheduling of H₂ deliveries.
• Policy Analysis: It can be used to evaluate the impact of policies like carbon pricing or subsidies for renewable H₂ on the supply chain configuration and cost.

Computational Requirements and Limitations:

• Data Intensity: The model requires significant input data:
  • Spatially resolved H₂ demand profiles (hourly or sub-hourly).
  • Time-series of electricity prices (and potentially natural gas prices).
  • Detailed cost and performance parameters for all technologies (CAPEX, OPEX, efficiency, ramp rates, lifetimes).
  • Geographic data for distances and potential infrastructure corridors.
• Computational Complexity: MILPs can be computationally intensive, especially with many time steps, locations, and discrete investment decisions. The paper suggests a good trade-off, likely achieved through:
  • Using representative days or weeks to model a full year.
  • Aggregating certain spatial details or simplifying some operational constraints.
• Assumptions:
  • Perfect foresight of demands and prices is a common assumption in such planning models.
  • The level of detail in pipeline flow dynamics might be simplified.
  • The model focuses on cost minimization and may not explicitly include other objectives like resilience or social acceptance unless formulated as constraints or multi-objective functions.

Deployment Strategy Insights:

The framework can help in deciding deployment strategies by comparing scenarios:

• Scenario 1: Centralized Production Focus: Emphasizing large-scale SMR with extensive pipeline networks.
• Scenario 2: Distributed Electrolysis Focus: Relying on smaller electrolyzers closer to demand centers, potentially leveraging truck-based delivery and storage.
• Scenario 3: Hybrid Approach: An optimized mix determined by the model.

By comparing the costs and infrastructure build-outs under these scenarios, decision-makers can identify robust and cost-effective development paths. The model's ability to quantify the value of flexibility (e.g., trucks as mobile storage) is crucial for making informed choices, especially when considering technologies like electrolyzers whose economics are heavily tied to variable electricity prices.

In conclusion, the paper provides a valuable optimization tool for designing hydrogen supply chains. Its main practical implication is highlighting how flexible operation of transmission assets, particularly trucks acting as mobile storage, can significantly improve the economics of hydrogen supply, especially by making electrolytic hydrogen more competitive in regions with volatile electricity prices. This allows for more efficient utilization of resources and better integration of hydrogen systems with renewable energy sources.