User Guide

User Guide#

This page provides an overview of what WOMBAT is currently able to model, broken down by general category. Following this, there are separate pages for how this is done as demonstrated through example notebooks. To fully follow the particulars of each example, it is recommended to see how each model's configuration files are composed.

For thorough explanations of the design and implementation ethos of the model, please see our NREL Technical Report: https://www.osti.gov/biblio/1894867 [3], which was published alongside v0.5.1, so some functionality has been updated.

Feature Overview#

For a complete and detailed description of the functionality provided in WOMBAT, it is recommended to read the API documentation. However, that is quite a task, so below is a shortlist of the core functionality that WOMBAT can provide.

Post Processing and the Simulation API#

  • The Simulation class and its Configuration allow users unfamiliar with Python to run the simulation library with minimal code. This was the primary design choice for the model: to create a configuration-based model to enable users with nearly any level of Python exposure.

  • The Metrics class provides a straightforward interface for computing a wide variety of operational summary statistics from performance to wind power plant financials.

Environmental Considerations#

  • Wind speed (m/s) and wave height (m) for the duration of a simulation (user-provided, hourly profile)

  • Reduced speed periods for animal migrations. This is primarily an offshore-focused feature, but can be defined at the simulation, port, or servicing equipment level using the following three variables: reduced_speed_start, reduced_speed_end, and reduced_speed (km/hr). This translates to a maximum speed of reduced_speed being enforced between the starting and ending dates each year of the simulation. For more details, see the documentation pages for the environment, port, unscheduled servicing equipment, or scheduled servicing equipment.

  • Prohibited operation periods. This dictates periods of time when a site or port may be inaccessible, or when a piece of servicing equipment should not be allowed to be mobilized/dispatched. Similar to the speed reduction periods, non_operational_start and non_operational_end dictate an annual date range for when servicing equipment can't be active, or a port or the whole site is inaccessible. The same documentation pages as above can be referenced for more details.

  • Generalized maintenance start date. This enables the first occurrence for all maintenance events to be set universally, if not specifically defined. If the date is prior to the start of the simulation, it merely sets the cadence. The ability to set a maintenance starting date allows for the staggering of the weather timeseries and maintenance timing. For instance, a weather profile starting in January for the Gulf of Maine, would have annual maintenance occurring in the middle of winter, likely causing many unnecessary weather delays in accessing the site, so the cadence can now start in the late spring or summer to ensure there are ideal weather conditions. This feature is made available through the Configuration.maintenance_start field,

System and Subassembly Modeling#

Each substation, turbine, and cable operate with the same core logic of maintenance and failure events. The following will break down the differences between each of these systems' modeling assumptions as well as the basic operations of the maintenance and failure models.

  • Maintenance

    • frequency: based on an amount of time between events and a given starting date.

    • frequency_basis: allows the input of frequency to be configured in days, months, or years.

    • start_date: enables the first occurrence of a maintenance event to be offset from the weather profile, or even well after the start date of a simulation.

    • operation_reduction: the percentage of degradation the triggering of this event cause

  • Failure

    • Randomly samples from a Weibull distribution to determine the time to the next failure

    • operation_reduction: the percentage of degradation the triggering of this event will cause

    • replacement: if True, then all preceding failures and maintenance tasks are cancelled and the time until the next event is reset because a replacement is required, otherwise, each successive failure is added to the repair queue

  • Commonalities between Substations, Turbines, and Cables

    • Maintenance and Failures compose the actual "model"

    • operating_level provides the real degradation of a model, including if it is turned off for servicing

    • operating_level_wo_servicing provides the operating level as if there were no ongoing operations

  • Substations

    • Any failure or maintenance with an operation_reduction of 1 will turn off every upstream connected (and modeled) cable and turbine

    • For substations that are connected by an export cable this, the upstream connections made from the export cable will not be considered upstream connections and therefore shutdown when a downstream substation has a failure

  • Array cables

    • Similar to substations, any failure or maintenance with an operation_reduction of 1 will turn off every connected upstream cable and turbine on the same string

  • Export Cables

    • Similar to array cables, any failure or maintenance with an operation_reduction of 1 will turn off every connected upstream substation and string of cable(s) and turbine(s)

    • As noted in the substation section, export cables connecting substations act independently and not as a string of connected systems

    • The final export cable connecting a substation to the interconnection point can however shut down the entire wind farm

  • Electrolyzers (new in v0.11)

    • Similar to both the turbine model, except that they are only allowed to be connected to a substation. Electrolyzer downtime does not impact the rest of the farm.

    • Requires at least 1 turbine with a power curve and 1 substation to enable H2 production. The turbine and substation may be modeled without any failures as a method to model only an electrolyzer.

    • NOTE: in a future version (likely v0.12) there will be a simplified way to model only an electrolyzer, and provide an external power profile for a more accurate assessment of the capacity factor and hydrogen production.

Repair Modeling#

  • Repair Management: see here for complete details

    • Generally, repairs and maintenance tasks operate on a first-in, first-out basis with higher severity level Maintenance.level and Failure.level being addressed first.

    • Servicing equipment, however, can specify if they operate on a "severity" or "turbine" basis that prioritizes focusing on either the highest severity level first, or a single system first, respectively.

  • Servicing Equipment: see here for complete details

    • Can either be modeled on a scheduled basis, such as a year-round schedule for onsite equipment or equipment with an annual visit schedule during safe weather, or on an unscheduled basis with a threshold for number of submitted repair requests or farm operating level.

    • Mobilizations can be modeled with a fixed number of days to get to site and a flat cost for both scheduled and unscheduled servicing equipment. For scheduled equipment, the mobilization is scheduled so that the equipment arrives at the site for the start of it's first scheduled day.

    • Mooring disconnections and reconnections operate without considering shift timing to ensure that these (typically) very long processes can find an optimal weather window within the simulation. These typically only occur with tugboats.

    • A wide range of generalized capabilities can be specified for various modeling scenarios. It's important to note that outside of the "TOW" designation, there are no specific implementations of functionality.

      • CTV: crew transfer vessel/vehicle

      • OFS: offsite equipment

      • SCN: small crane (i.e., field support vessel)

      • MCN: medium crane

      • LCN: large crane (i.e., heavy lift vessel)

      • CAB: cabling vessel/vehicle

      • RMT: remote reset (no actual equipment BUT no special implementation)

      • DRN: drone

      • DSV: diving support vessel

      • VSG: vessel support group

      • TOW: tugboat and support vessel (triggers the tow-to-port model)

      • AHV: anchor handling vessel (special case that can be modeled from the port or main simulation configuration)

    • Operating limits can be applied for both transiting and repair logic to ensure site safety is considered in the model.

  • Ports

    Note

    Major improvements as of v0.13

    • Currently only used for tow-to-port repairs or tugboat based repairs, which adds the following additional capabilities:

      • TOW: tugboat or towing equipment

      • AHV: anchor handling vessel (tugboat that doesn't trigger tow-to-port)

    • Port access fees can be modeled in the FixedCosts module, or applied as a monthly fee using either the annual_fee or monthly_fee input.

    • Port usage fees are applied when any turbine is at the port for repairs, and can set using the daily_use_fee parameter.

    • Tugboats are mobilized, and will stay at port for the duration of their charter_days parameter.

    • Tugboats are called out on the first request from the requesting system, regardless of the user encoding.

    • All subsequent repair and maintenance requests following a "TOW" request will not be addressed until the system is at the port. This can create a significant backlog of repairs if there are not enough tugboats, crews, or active turbine slots encoded at the port.

    • See the API docs for more details

Examples and Validation Work#

Binder

Below are a few examples to get started, for users interested in the validation work in the code-to-code comparison presentations, the notebooks generating the most up-to-date results can be found in the main repository, where there is a separate analysis for the Dinwoodie et al. [2] comparison, and for the Smart et al. [3] comparison.

Default Data#

As of version 0.13, a default reference data set has been made available for users. Simply specify a simulation as Simulation.from_config("DEFAULT", "osw_fixed.yaml") or Simulation.from_config("DEFAULT", "osw_floating.yaml") to use either of the fixed-bottom or floating offshore wind data sets, respectively. See the examples/COWER_om_workflow.ipnyb for more details.

The default library provides a validated, ready-to-use data set for fixed and floating offshore wind, and an experimental land-based data set. For all three, users can use the pre-configured base models or use them as a starting point for building custom models. Example results can be seen in examples/default_data_demonstration.ipynb or online

Overview#

Offshore#

The default library provides a consistent set of cost and performance inputs for offshore wind energy systems, standardized to 2024 USD. This library supports reproducible analyses in the Cost of Wind Energy Review (COWER): 2025 Edition (forthcoming), including both fixed-bottom and floating offshore wind cases.

Unlike other datasets in this repository, which largely reflect publicly available sources, this dataset incorporates internal adjustments and harmonizations to align with research-focused scenarios for NREL products that may require outputs to be in 2024 USD, like for example, the Annual Technology Baseline, or the Cost of Wind Energy Review.

Land-Based#

The default library provides an experimental set of land-based data that is hybridized from a series of incomplete sources of offshore and onshore costs, and scaling factors between them. It is highly recommended to merely use this data as a starting point for any analysis work.

Core Assumptions#

Offshore#

Material costs for repairs and replacements, as well as failure data, are sourced from Verma et al. [4], Schwarzkopf et al. [5]. Fixed cost data is primarily derived from Guide to a floating offshore wind farm | An informative resource for floating offshore wind [6], Guide to an offshore wind farm | An informative resource for offshore wind [7], while vessel day rate and mobilization assumptions are compiled from a range of public sources. For detailed reference information, please contact Daniel Mulas Hernando (Daniel.MulasHernando@nrel.gov) to request access to WOMBAT_cost_history.xlsx.

  • Cost Year: All monetary values are standardized to 2024 USD.

  • Data Integration: Inputs were consolidated from multiple public sources and internal records (WOMBAT_cost_history.xlsx), with historical exchange rates and inflation applied to transform costs to 2024 USD.

  • Technology Coverage: Includes representative inputs for offshore (fixed-bottom and floating) technologies.

Land-Based#

Failure rates, repair times, and materials costs not provided in Schwarzkopf et al. [5] are largely taken from Carroll et al. [8] with some custom substitutions or modifications. All failure rates are converted to a Weibull scale factor. Using Carroll et al. [8] we adjust offshoreWeibull scale factors for minor repairs for the generator, blades, pitch system, yaw system, and drive train (wind-affected subassemblies) by a factor of 5.912, for non-wind-affected subassemblies a factor of 1.217 is applied, and for all major repairs and replacements, a factor of 2.432 is applied. These scaling factors are primarily derived from Carroll et al. [8]

All costs are rescaled from the 12 MW baseline to a 3.5 MW baseline using the COWER 2025 turbine CapEx figures of $1,117/kw-yr for onshore and $1,770/kw-yr for offshore, a 0.6311 scaling factor.

Additional subassemblies are derived from Carroll et al. [8] and scaled using the above assumptions. Fixed costs are derived from Wiser et al. [9], with labor coming supplemented by (NREL) [10], and equipment costs come from LandBOSSE and internal source adjustments.

Intended Use#

  • Serve as a baseline input for replicable analyses of the offshore fixed and floating COWER-2025 results.

  • Serve as a baseline input for the development of land-based analyses.

  • Support scenario development and sensitivity analyses exploring the impact of cost evolution, operational performance, and logistics assumptions.

Reproducibility#

The accompanying notebook, examples/COWER_om_workflow.ipynb, demonstrates how to replicate O&M results from the Cost of Wind Energy Review: 2025 Edition. It runs 50 simulations per case and summarizes mean and standard deviation results to identify sources of variability within cost components.

Notes#

  • WOMBAT_cost_history.xlsx is an internal NREL document. For questions, contact Daniel.MulasHernando@nrel.gov.

  • This dataset should be treated as a scenario-based reference, not as purely empirical or historical data.

References#

[1]

Iain Dinwoodie, Ole-Erik V Endrerud, Matthias Hofmann, Rebecca Martin, and Iver Bakken Sperstad. Reference cases for verification of operation and maintenance simulation models for offshore wind farms. Wind Engineering, 39(1):1–14, 2015. URL: https://doi.org/10.1260/0309-524X.39.1.1, doi:10.1260/0309-524X.39.1.1.

[2]

Gavin Smart, Aaron Smith, Ethan Warner, Iver Bakken Sperstad, Bob Prinsen, and Roberto Lacal-Arantegui. Iea wind task 26: offshore wind farm baseline documentation. Technical Report, National Renewable Energy Lab. (NREL), Golden, CO (United States), 06 2016. URL: https://www.osti.gov/biblio/1259255, doi:10.2172/1259255.

[3]

Rob Hammond and Aubryn Cooperman. Windfarm operations and maintenance cost-benefit analysis tool (wombat). Technical Report, National Renewable Energy Laboratory (NREL), Golden, CO (United States), 10 2022. URL: https://www.osti.gov/biblio/1894867, doi:10.2172/1894867.

[4]

Jai Verma, José Ignacio Rapha, José Luis Domínguez, and Victor Ferreira. D6.1 General frame of the analysis and description of the new fow assessment app. Technical Report, COREWIND, November 2020. URL: https://corewind.eu/publications/ (visited on 2023-01-30).

[5] (1,2)

Marie-Antoinette Schwarzkopf, Friedemann Borisade, Jannis Espelage, Eve Johnston, Rubén Durán Vicente, Sara Muñoz, Pål Hylland, Wei He, Joaquín Urbano, Francisco Javier Comas, Andrea Arribas, Miguel Somoano Rodriguez, Sergio Fernandez Ruano, Lucía Meneses Aja, Raul Guanche García, and Álvaro Rodríguez Luis. D4.2 floating wind o&m strategies assessment. Technical Report, COREWIND, August 2021. URL: https://corewind.eu/publications/ (visited on 2023-01-30).

[6]

Guide to a floating offshore wind farm | an informative resource for floating offshore wind. October 2023. URL: https://guidetofloatingoffshorewind.com/ (visited on 2025-12-10).

[7]

Guide to an offshore wind farm | an informative resource for offshore wind. June 2025. URL: https://guidetoanoffshorewindfarm.com/ (visited on 2025-12-10).

[8] (1,2,3,4)

James Carroll, Alasdair McDonald, and David McMillan. Failure rate, repair time and unscheduled o&m cost analysis of offshore wind turbines. Wind energy, 19(6):1107–1119, 2016. URL: https://onlinelibrary.wiley.com/doi/full/10.1002/we.1887, doi:https://doi.org/10.1002/we.1887.

[9]

Ryan Wiser, Mark Bolinger, and Eric Lantz. Assessing wind power operating costs in the united states: results from a survey of wind industry experts. Renewable Energy Focus, 30:46–57, 2019. URL: https://www.sciencedirect.com/science/article/pii/S1755008419300092, doi:https://doi.org/10.1016/j.ref.2019.05.003.

[10]

National Renewable Energy Laboratory (NREL). Clean energy employment impacts. 08 2023. URL: https://www.osti.gov/biblio/1995015.

Contents