A Community-Driven Framework for Wind Farm Flow Models

The Wakebench international framework for wind farm flow models will use online tools to map the state-of-the-art and guide future model development, experiments and validation activities.

Figure 1: Wakebench model evaluation framework tools.

Last Task 31 meeting, hosted by NREL and the University of Colorado Boulder, was organized jointly with Task 29 on 11–14 March to gather experts from rotor and wind farm wake aerodynamics. The objective of the meeting was to design a framework for wind farm flow models that will guide future model development, experiments and validation activities. A schematic with the main components of the framework is shown in Figure 1.

Wakebench is building a formal model validation directed program with the objective of developing and executing a collaborative experimental and numerical research activity, which will lead to quantifying the predictive capability of state-of-the-art models for specific applications [1]. The planning process is composed of four steps:

• 1. Identify the objectives of the model from the perspective of the intended use (application) in terms of relevant quantities of interest and quality-assurance requirements.
• 2. Identify the phenomena of interest that the model should capture and prioritize the assessment based on the expected impact on the objectives.
• 3. Define a validation hierarchy that will allow to assess model performance for the prioritized phenomena.
• 4. Plan experiments to generate data for the validation hierarchy based on how the limited resources can be used most effectively.

The Wind Energy Model Evaluation Protocol (WEMEP) is at the core of the framework to provide documentation and guidance to modelling communities. It is hosted in Github to allow version control and provide credit to contributors. A generic model-agnostic protocol will be drafted in the next few months to harmonize evaluation procedures and best practices. Then each community will have its own space to describe the intended use of their models and document model evaluation repositories. Task 31 will host the wind conditions and wind farm wake modeling communities. Other communities may join later if they wish to do so.

Expert elicitation will be facilitated with the use of online mind-maps, focused on the intended uses of the models developed by the community. Task 31 gathers experts from two focus areas:

• Wind conditions (AEP, siting parameters and flow-correction factors in complex terrain) for wind resource assessment, site suitability and numerical site calibration (Wind community).
• Array efficiency and loads for wind farm design (Wakes community).

Based on the intended-use, Task 31 experts were asked during the last meeting to map relevant phenomena of interest, model building blocks and input quantities. These maps will allow model developers and end-users to describe the state-of-the-art using community-based terminology and identify knowledge gaps.

Counting with the participation of Task 29 experts allowed us to gain a more detailed analysis on relevant phenomena in the rotor and near-wake region. A fundamental question arises then in considering how detailed this region needs to be for wind farm integrated design (power and loads) purposes or, from Task 29 perspective, how detailed the inflow wind conditions (in free-stream or waked conditions) need to be for rotor aerodynamic purposes.

The mind-map provides inputs to the Phenomena Identification and Ranking Tables (PIRT), which:

• ranks relevant physical and other related phenomena for the intended use;
• characterizes the adequacy of the existing model, experiment and validation data; and
• provides a gap analysis as to what are the issues associated to the modeling of these phenomena and how they can be addressed.

PIRT tables will allow Task 31 decide about what validation activities deserve more attention and what should be prioritized in forthcoming experiments. We will unify previous PIRT analyses done in the frame of Ae2 and NEWA research programs [2] to consolidate a unified planning instrument under the umbrella of the IEA-Wind.

The ultimate goal of this framework is to reach consensus from different modeling communities as to where are current gaps and what activities need to be prioritized to obtain the largest impact on the applications of interest. This roadmap for wind farm flow models will be used to justify research needs, align model development and establish better connections between projects from different countries.

What’s Next?

The information gathered from experts participating in the Boulder meeting will be distilled by the Wakebench management team to provide a first draft of Mind-Maps and PIRT tables that will be reviewed by the community through 2019. To extend the expert elicitation to the wider community we will use online brainstorming and voting tools. The results will be reviewed during next in-person meetings:

• Interim meeting, Uppsala University, Visby, 21 May 2019 (from 1 to 5 pm). Just before the Wake Conference 2019 (22–24 May 2019). We will check progress on active benchmarks, review the mind-maps and PIRT tables distilled by the WP leaders and discuss the contents and contributions to the WEMEP.
• Interim meeting, Amherst, Massachusetts, together with NAWEA and WindTech conference, 14–17 October 2019 (exact date to be defined). Call for abstracts deadline: April 17. Review of V&V farmework, define roamap for wind and wake models based on PIRT tables.

Hence, by the end of 2019 the first release of the Wakebench roadmap will be published. This roadmap will be periodically reviewed and updated online based on ongoing research and inputs from the community. If you want to participate in the Wakebench expert elicitation process please sign up to the mail list below.

Reaching community consensus and avoiding bias in the roadmap is a difficult task. If you have run an expert elicitation process and would like to share your experiences or tools please use the discussion below or contact me.

References

[1] Hills R.G., Maniaci D.C., Naughton J.W. V&V Framework. SANDIA Report SAND2015-7455, September 2015.

[2] Sanz Rodrigo J., et al (2016). Report on hierarchy of benchmarks (Version 2.1). Deliverable 3.4 from the New European Wind Atlas (NEWA) project. Zenodo. doi: 10.5281/zenodo.1194744

This work is performed with the support from IEA-Wind Task 31 Wakebench. To receive periodical newsletters about the Task, please subscribe to the mail list.