Question-and-Answer Resource for the Building Energy Modeling Community
Get started with the Help page
Ask Your Question

Modified MOWITT model

asked 2023-05-24 10:34:51 -0500

Tmabir's avatar

updated 2023-05-24 15:25:18 -0500

Just wanted to double check my understanding on the MOWITT model for convective heat transfer coefficient for Windows. MOWITT creates the regression coeffs for forced convection based on wind speed measured at 10m. These coeffs are okay for low rise building since the test facility window height was 2m. Energy plus can convert the convection coeffs based on variable wind measurement heights (depends on convective surface heights) since they have the actual heat transfer data from the experiment. My questions are:

  1. If I want to make my wind height measurements are different and I am not using Energy plus how can I know the proper coeffs for forced convection on an external glazed surface.
  2. If I use the MoWitt model, up to what height of the window is the available forced convection coeffs valid.

Best Abir, UC Davis

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2023-05-26 11:43:04 -0500

updated 2023-05-26 11:53:04 -0500

Not a definitive answer - a few UMH people have direct, hands-on experience with this (hopefully they may chime in).

The following 3x plots may be helpful for your 2nd question: MoWITT vs TARP-predicted hf (forced convection, in W/m2.K ... Y-axis) for very smooth (windward) surfaces, as a function of local Vz (window height, up to 60m). Assuming default E+ weather station conditions (height = 10m, exponent 0.14, boundary layer thickness 270m), vs local building conditions similar to the original studies, further discussed here (exponent 0.22, boundary layer thickness 370m).

At low wind speeds at the weather station (Vmet 2m/s; Vz 1.4m/s @10m), IMHO it's fine to rely on MoWITT up to 60m: it's more conservative at lower heights (i.e. Yazdanian & Klems' thesis), and merges with TARP with increasing height. This is assuming a surface perimeter/area ratio of 3 for TARP (this has a significant impact on results).

image description

At slightly higher wind speeds at the weather station (Vmet 2.5m/s; Vz 1.8m/s @10m), there's a crossover at ~20m. Hard to say if one should definitely switch to TARP above 20m, but it's giving you an idea of the range. Again, TARP's perimeter/area ratio is something to keep in mind (e.g. if it were 2 instead of 3).

image description

Finally, at higher wind speeds at the weather station (Vmet 3m/s; Vz 2.2m/s @10m), the crossover is around 10m (roughly what Yazdanian & Klems would consider low-rise residential buildings, their target building type). The discrepancy between predictions starts to be somewhat significant at greater window heights (without even considering TARP's perimeter/area ratio). In my neck of the woods, average winter wind speeds (weather station) are around 4m/s. I'd likely avoid MoWITT above 4 or 5 stories. But I'm unsure what the alternative is really, given TARP's sensitivity to the perimeter/area ratio. I may have to take a deeper dive, as I can't get my head around this ratio when it comes to fully-glazed, 10-storey buildings with silicon structural joints.

image description

Long story short, it would seem to depend strongly on expected wind speeds. But low-rise, 3-storey residential buildings definitely seem to be the sweet spot for MoWITT.

I hesitate in answering your 1st question, as it's a bit confusing. Are you taking local wind speed measurements, and attempting to estimate hf for glazed surfaces nearby? If so, I'd rely on the initial local wind speed equation while characterizing your local wind environment. I may be misunderstanding the question ...

edit flag offensive delete link more


Thanks for the detailed explanation regarding my second query. It makes a lot of sense. I apologize for not being enough comprehensive in my first query:

Quoting from Kruis et al. The exterior forced convection coefficient in EnergyPlus should be calculated using adjusted MoWiTT regression coefficients that are appropriate for use with near-surface wind speeds.

This paper recalculates the regression coefficient (a,b) used in the MoWITT model based on converting V_10 (Windspeed at weather station

Tmabir's avatar Tmabir  ( 2023-05-26 18:16:43 -0500 )edit

Yes, the curves above were generated using the same updated (a & b) MoWITT coefficients (ref: E+ docs, same as Table 4 in Kruis et al.). Table 3 in the paper also describes the local wind speed calculation/parameters used in E+ , which I think is what you need for your 1st question (no need to rely on E+ per se).

Denis Bourgeois's avatar Denis Bourgeois  ( 2023-05-27 05:51:12 -0500 )edit

... and it may be worth going through past (solved) E+ issues/fixes with MoWITT.

Denis Bourgeois's avatar Denis Bourgeois  ( 2023-05-27 05:57:22 -0500 )edit

Thanks Denis for taking the time to answer my questions and providing all necessary reference. I have one last question the MoWITT model has been designed for smooth vertical surfaces. Are you aware of any model that is sensitive to surface tilt, say for instance- skylights?

Thanks again!

Tmabir's avatar Tmabir  ( 2023-05-27 09:56:55 -0500 )edit

Abir, I don't have field expertise as to the suitability of MoWITT (vs other models) for flat/domed skylights (when it comes strictly to wind-driven convection). If you consider the original (fairly open) MoWITT test setup/location, I wouldn't conclude it to be unsuitable at all < 10m. IMHO all convection models are general purpose. Introduce a 15-storey tower in a 3-storey residential area (I have seen pedestrians blown off their feet from channeled wind gusts), one can expect the wind speed near the base of the tower to be far greater than predicted by any general model.

Denis Bourgeois's avatar Denis Bourgeois  ( 2023-05-27 14:04:15 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Training Workshops


Question Tools


Asked: 2023-05-24 10:34:02 -0500

Seen: 92 times

Last updated: May 26 '23