Exploring the cost and emissions impacts, feasibility and scalability of battery electric ships

The study addresses how battery electric ships (BESs) can contribute to US and international maritime decarbonization targets and whether large-scale retrofitting of internal combustion engine (ICE) vessels is technically and economically feasible. Despite growing interest in zero-emission shipping, concerns persist about battery size, weight, charging logistics, and scalability. The authors aim to quantify the life-cycle costs and emissions of electrifying US domestic vessels under 1,000 gross tonnage, identify ship and route characteristics suitable for electrification, evaluate infrastructure needs at ports, and assess the impact of serving less than 100% of historical trips on feasibility. The work introduces a capacity-tier framework (serving 100%, 99%, 95%, or 90% of historical trips) to right-size batteries and improve economic viability, and develops a high-resolution model (MariBES) to perform techno-economic analysis (TEA) and life-cycle assessment (LCA) across grid decarbonization and cost scenarios.

Prior research quantified maritime emissions using various methodologies, including operational-mode approaches and region-specific analyses (e.g., Yangtze River, Great Lakes, and US non-oceangoing vessels). Recent deployments of battery electric ferries in Norway, Japan, and Denmark demonstrate feasibility for predictable routes, and the first battery electric tug (eWolf, 6 MWh) emerged in 2021. Comparative studies highlight that grid carbon intensity is a key driver of BES emissions and that BESs can achieve the lowest GHG emissions and costs for certain ship types (e.g., passenger vessels) versus alternative fuels. Lithium-ion batteries show favorable life-cycle environmental performance relative to other chemistries due to higher energy density and longer life. Rapid declines in battery costs bolster prospects for electric interregional container shipping, and system-level models indicate scaling BESs reduces shipping GHG emissions. However, previous work generally focused on emissions estimation or select applications rather than national-scale feasibility, infrastructure requirements, and economic comparisons under varying trip-coverage assumptions. This study fills that gap by evaluating national-scale retrofit pathways, cost-effectiveness, and port electricity needs using high-resolution operational data.

Scope and data: The authors integrate (1) vessel stock and technical specifications from US Coast Guard, US Army Corps of Engineers, and IHS Markit; (2) high-resolution AIS data (location, speed over ground, etc.) for 2021 US domestic operations; and (3) principal port locations (top 150 by tonnage). They define an Integrated Ship Database of 11,687 ships and focus on the 'Domestic Fleet' (6,323 vessels, 50–1,000 GT). An AIS Analysed Subset of 2,722 ships (travel >1,500 km in 2021) underpins detailed modeling.

Model and capacity tiers: The Maritime Battery Electrification Simulator (MariBES) conducts TEA and LCA with high spatial and temporal resolution. BES capacity tiers BESP100, BESp99, BESp95, and BESp90 denote the share of historical ICE trips the BES must cover (100%, 99%, 95%, 90%). Excluding a small fraction of most energy-intensive trips reduces battery sizing needs.

Power and energy modeling: Ship power demand is computed from AIS-derived speeds using an admiralty-style propulsion model and auxiliary demand calibrated by operational mode and adjustment factors (per IMO studies): P_ICED = P_ME + P_AE, with modifiers for weather, fouling, safety factor, and a speed exponent. Trip energy E_tr is aggregated at 5-minute intervals. Battery sizing uses percentile-based reference trip energy E_ref_px at each capacity tier with assumed DOD = 80% and end-of-life SOH = 80%.

Charging scheduling: Charging occurs only when vessels are quasi-stationary (<0.5 knots). An optimization minimizes maximum charging power while maintaining SOC within 10–90% and recharging to 90% when predicted to fall below a 70% boundary before the next trip. The schedule aggregates vessel charging at the nearest ports to estimate port and state electricity and power requirements.

Battery degradation and lifetime: A degradation model (calendar and cycle aging) estimates average lifetimes by tier: ~15.5 years (BESP100), 12.0 (BESp99), 10.4 (BESp95), 10.0 (BESp90).

Emissions (LCA): Emissions for ICE = well-to-tank + combustion. For BES = electricity well-to-tank + grid generation + battery manufacturing − credits for second-life battery (SLB) use. Three grid emissions scenarios from NREL Cambium are applied: BAU, 95% decarbonization by 2050 (DEC50), and 95% by 2035 (DEC35). Analyses consider 2022, 2035, and 2050.

Economic analysis (TEA/LCOT): LCOT ($/km) divides annualized capital and operating costs by annual distance. ICE costs include fuel (ULSD), O&M, and emissions costs (CO2e via social cost of carbon and damages from NOx/SOx). BES costs include battery system CAPEX, charging energy and infrastructure, O&M, emissions (battery + grid), and a net salvage value credit from second-life batteries. Cost scenarios for electricity and battery systems: Optimistic (OPT), Intermediate (INT), and Challenging (CHA). SCC values used: $190/tCO2e (2022), $250 (2035), $310 (2050). Port grid impacts are estimated by aggregating scheduled charging across 150 principal ports.

Physical feasibility (weight): Vessel displacement is estimated from gross tonnage; ICE and BES configurations account for removal/addition of engines, fuel, battery, and electric motor weights. Battery system mass baseline: ~21 kg/kWh; sensitivity includes 25% lighter systems.

• Fleet electrification potential and battery sizing: Among 6,323 domestic vessels (<1,000 GT), excluding a small fraction of long/energy-intensive trips substantially reduces required battery sizes. Omitting just the top 1% of trips (BESp99) cuts required capacity by about two-thirds overall; for passenger ships, battery size can drop by ~85%. Required batteries would be ≤14% of North America’s projected 2030 battery manufacturing capacity.
• Weight feasibility: With 21 kg/kWh batteries, BESP100 raises median vessel weight by ~137% vs ICE, but BESp99 limits the median increase to ~30%; with 25% lighter systems, BESp99 median increase falls to ~21%.
• Emissions impacts: AIS Analysed Subset operational emissions are ~2.1 MMTCO2e (9.5% of 21.9 MMTCO2e from the full domestic fleet in 2021). Under BAU, BES emissions in 2022 are only ~8% lower than ICE, but decline to 1.6 MMTCO2e (2035) and 1.3 MMTCO2e (2050). Under DEC50: 1.4 MMTCO2e (2035) and 0.6 MMTCO2e (2050), corresponding to 42% and 75% reductions below 2022. Under DEC35: ~73% reduction by 2035 and ~75% by 2050. Cumulative 2022–2050 CO2e reductions range from 32% (BAU) to 58% (DEC35). Per-km emissions decline with lower capacity tiers as battery manufacturing’s share shrinks (approx. 53%, 26%, 19%, 15% for BESP100, BESp99, BESp95, BESp90 in 2035 under DEC50).
• Economic feasibility (LCOT): By 2035, most BESp99 vessels are cost effective versus ICE on a ship-by-ship LCOT basis. Under DEC50-INT in 2035, cost-effectiveness ratios are 33% (BESP100), 81% (BESp99), 90% (BESp95), 91% (BESp90). Under DEC35-INT, up to 85% of BESp99 achieve cost parity. Median ICE LCOT lies between BESP100 and BESp99 by 2035. Average ICE LCOT rises ~1%/year while BESp99 falls ~1.5%/year. Passenger and tug (inland-push boats) categories show lower average LCOT than ICE in 2035; many coastal-harbour and ATB tugs also reach parity (76% and 44% respectively). Without emissions costs, BES direct costs remain higher, but carbon pricing materially increases BES cost-effectiveness: at $250/tCO2e (2035), ~80% of BESp99 reach parity; at $157/tCO2e, ~60% do.
• Grid and port infrastructure: AIS subset charging demand is ~3.8 TWh/year; scaling to the Domestic Fleet yields ~7.7 TWh/year. Top 20 of 150 ports account for ~46% of total demand; Port of New York and New Jersey is highest at ~238 GWh/year. Tugboats dominate electricity demand across states and ports (e.g., 99.6% in Louisiana; 99.9% in Texas). In California, passenger ships represent >30% of charging demand. State-level electricity increases are modest (0.1–0.8% in top ten states). About 67% of vessels (and 89% of passenger ships) would require ≤5 MW charging, aligning with emerging megawatt-scale standards (up to ~4.5 MW).
• Policy-relevant insights: Faster grid decarbonization (DEC35 vs DEC50) lowers cumulative shipping emissions by an additional ~15% through 2050 and reduces 2035 BES LCOT by ~10%. Second-life battery values offset ~42–46% of battery system costs, improving economics.

The findings demonstrate that a national-scale retrofit of smaller US domestic vessels to battery electric propulsion can deliver substantial GHG reductions—particularly as the power sector decarbonizes—and can be economically attractive when batteries are right-sized to cover nearly all, but not necessarily every, historical trip. Introducing capacity tiers (e.g., BESp99) addresses the previous assumption that BES must serve 100% of trips, revealing large reductions in battery size, mass, and cost with minimal loss of service coverage. Economic competitiveness improves over time due to falling battery costs, increasing social cost of carbon, and the value recovery from second-life batteries. While BES direct costs often exceed ICE when environmental externalities are excluded, incorporating emissions costs via carbon pricing or regulations shifts many vessels to cost parity or advantage. Grid impacts are manageable at the state level but concentrated infrastructure investments will be needed at a limited number of ports, where tug operations dominate energy demand. Overall, the results offer a practical pathway for rapid maritime decarbonization aligned with US and IMO targets, contingent on continued grid decarbonization, supportive policy frameworks for carbon and air pollution costs, and targeted port electrification planning.

This study presents a high-resolution national assessment of retrofitting US domestic vessels (<1,000 GT) to battery electric propulsion, integrating TEA and LCA within the MariBES framework. Key contributions include: (1) a capacity-tier approach (e.g., BESp99) that right-sizes batteries by excluding a small fraction of long trips, unlocking substantial reductions in battery size, weight, and costs; (2) quantification of GHG reductions up to ~73% by 2035 (DEC35) and strong cumulative reductions through 2050; (3) demonstration that by 2035, the majority of BESp99 vessels achieve LCOT parity with ICEs under realistic decarbonization and cost scenarios, especially when environmental costs are included; and (4) identification of concentrated port charging needs, enabling targeted infrastructure planning. Future research directions include: prioritizing electrification in regions with cleaner grids and favorable tariffs; evaluating multiple smaller BESs as alternatives to single large ICE vessels; optimizing operations (trip scheduling, battery swapping) to mitigate infrastructure peaks; developing renewable microgrids at ports to lower lifecycle emissions; assessing air quality and environmental justice benefits in port communities; and expanding analysis beyond retrofits to compare total cost of ownership for new-build ICE vs BES replacements.

• Activity data represent a single historical year (2021); ship operations vary with economic conditions, potentially affecting energy needs, costs, and feasibility year-to-year.
• Emissions calculations largely use national average grid factors by scenario; real-world charging emissions and costs will vary by region, time of day, and tariff structure.
• LCOT outcomes are sensitive to uncertain future diesel and electricity prices, battery costs, and the social cost of carbon; scenario ranges mitigate but do not eliminate uncertainty.
• The study focuses on retrofitting existing ICE vessels; it does not evaluate new-build BES vs ICE total cost of ownership at retirement.
• Capacity-tier results assume that a small share of energy-intensive trips can be managed operationally (e.g., alternative vessels, scheduling, or other fuels), which may pose logistical challenges.
• Battery lifetime and second-life value assumptions influence economics and emissions; actual degradation, reuse viability, and salvage values may differ.
• Some underlying ship characteristic data are proprietary (IHS Markit), which may limit code reproducibility without licensed access.