Fast Pressure Calculations on Buildings to Improve … Pressure Calculations on Buildings to Improve Outdoor-to- Indoor Transport & Dispersion Michael Brown 1, Akshay Gowardhan 1,2,

Click to edit Master title style

Click to edit Master subtitle style

LA-UR-05-8025

Fast Pressure Calculations on Buildings to Improve Outdoor-to-

Indoor Transport & Dispersion Michael Brown1, Akshay Gowardhan1,2, Matt Nelson1,

Mike Williams1, and Eric Pardyjak2

1Los Alamos National Laboratory2University of Utah

2007 CBIS ConferenceAustin, TX

LA-UR-05-8025

Presentation Outline

• Why Pressure Important for T&D Applications- Pressure Distribution on Buildings Influences Air Exchange Rate

• Modeling Tools- QUIC-URB Wind Model- QUIC Pressure Solver

• Model Evaluation

• How the fast wind & pressure models could be used to improve Indoor T&D calculations

LA-UR-05-8025

Motivation

5 minute duration outdoor release

½ hour QUIC Salt Lake City simulation

• Outdoor Releases Infiltrate into Buildings

LA-UR-05-8025

Motivation

5 minute duration outdoor release

½ hour QUIC Salt Lake City simulation

• Outdoor Releases Infiltrate into Buildings

LA-UR-05-8025

Motivation

LANL USA Day-NightIndoor-Outdoor Pop DB

McPherson, T., A. Ivey, and M. Brown, 2004: Determination of the spatial and temporal distribution of population for air toxics exposure assessments, AMS 5th Symp. on Urban Environment, Vancouver, BC.

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

91%

9%

Daytime Residential

88%

12%

Daytime Workers

98%

2%

Nighttime Residential

87%

13%

Nighttime Workers

Indoors

Indoors Indoors

Indoors

Out

Out Out

Out9%

91%

88% 87%

98%

2%

12% 13%

LA-UR-05-8025

Motivation

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Exposure estimates can be much smaller if building “protection” considered

Indoor Prediction

Acute Exposure Guideline Levels

Outdoor PredictionGadgil, 2005GMU T&D Workshop

LA-UR-05-8025

Motivation

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Exposure estimates sensitive to building “protection”

• Air exchange for naturally-ventilated buildings is proportional to wind- induced pressure on building walls

wind+ high pressure low pressure

LA-UR-05-8025

Motivation

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Exposure estimates sensitive to building “protection”

• Air exchange for naturally-ventilated buildings is proportional to wind- induced pressure on building walls

wind+

Air flow into building through any openings

Air flow out of building through any openings

LA-UR-05-8025

Motivation

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Exposure estimates sensitive to building “protection”

• Air exchange for naturally-ventilated buildings is proportional to wind- induced pressure on building walls

wind+

Air flow into building through any openings

Air flow out of building through any openings

Chan et al. (2005) – Most residential buildings in US do not have mechanical ventilation systems

LA-UR-05-8025

Motivation

Pressures on surface used as boundary conditions in CFD and multi-zone models, e.g., COMIS

• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Air exchange for naturally-ventilated buildings is proportional to wind-induced pressure on building walls

Orifice equation

Qf = ELAbldg *(2*ΔPbldg /ρ)1/2

Qf = volumetric airflow rateELA = effective leakage area of bldg

In practice

Qf = k *ΔPn 0.6<n<0.7

LA-UR-05-8025

Motivation

The Urban Dispersion Model (UDM)• Outdoor Releases Infiltrate into Buildings

• Population mostly resides Indoors

• Exposure estimates sensitive to building “protection”

• Air exchange for naturally-ventilated buildings is proportional to wind- induced pressure on building walls

The Air Exchange Rate (AER) is due to a Buoyancy (“stack”) Pressure and a Wind-Induced Pressure.

Ignoring the stack pressure effect (e.g., Tindoor =Toutdoor )

AERbldg = (AERref / ΔPref2/3)*ΔPbldg

2/3

Indoor Concentration

AER3600

=τ⎥⎥⎦

⎤

⎢⎢⎣

⎡+= ∫

−− t

t

tout

sout

t

i

s

dtettet ')'()()( ττ

τχχχ where

LA-UR-05-8025

Wind & Pressure Solvers

QUIC-URB

3D Wind Field

QUIC-Pressure Solver

3D Pressure Field

3D Buildings Inflow Wind ProfileIdea:

Use Fast Solvers To Compute Pressure Field on BuildingsandProvide as Input to Indoor Models

LA-UR-05-8025

QUIC-URB Wind Solver

•Based on dissertation of Röckle (1990)

•3D winds obtained from diagnostic/empirical method

•Initial winds based on building spacing and geometry

•Then mass conservation imposed (Sherman,1978)

Upwind Cavity

Downwind Cavity

Wake

Isolated Buildings Densely Packed Buildings

Street Canyon

Upwind Cavity

RooftopBubble

LA-UR-05-8025

QUIC-URB Wind Solver

•Based on dissertation of Röckle (1990)

•3D winds obtained from diagnostic/empirical method

•Initial winds based on building spacing and geometry

•Then mass conservation imposed (Sherman,1978)

Isolated Buildings Densely Packed Buildings

LA-UR-05-8025

QUIC Pressure Solver (Gowardhan et al., 2006)

Assuming steady state and taking divergence of Eqn. 1

Momentum Equation:

jj

i

j

ji

ij

jii

xxU

xuu

xP

xUU

tU

∂∂∂

+∂

∂−

∂∂

−∂

∂−=

∂∂ 2)''(1)(

νρ

0

I II IIIwhere,

I - Advective terms

II -Reynolds stress terms

III -Diffusive terms

(1)

LA-UR-05-8025

QUIC Pressure Solver

• The pressure Poisson equation is solved by iterative method with

• Reynolds Stresses are neglected due to lack of information

• Coefficient of Pressure is calculated using the following formula:

(2)⎟⎟⎠

⎞⎜⎜⎝

⎛

∂

∂−

∂

∂−

∂∂∂

∂∂

=⎟⎟⎠

⎞⎜⎜⎝

⎛∂∂

∂∂

j

ji

j

ji

jj

i

iii xuu

xUU

xxU

xxP

x)''()(2

νρ

0

( )22

1o

op V

PPC

ρ−

=

0=∂∂ np

LA-UR-05-8025

QUIC Wind & Pressure Solvers

Salt Lake City Downtown (Domain:200 x 200 x 50 cells)

Computation Time: QUIC-URB = 67 s Pressure = 46 s

Cp

Pentium 42.5 GHz

LA-UR-05-8025

Model Evaluation Cases

L-shaped

U-shapedCube (90 deg)

Squat

Cube (45 deg)

LA-UR-05-8025

Model Evaluation Cases

7x1 Wide Building ArrayHigh-Rise

LA-UR-05-8025

QUIC vs. Experimental Data: Cube (90 deg.)

Cube

wind

RooftopFront Face Rear Face

( )22

1o

op V

PPC

ρ−

=

Adapted from Richards & Hoxey (2006)

LA-UR-05-8025

QUIC vs. Experimental Data: L-Shaped Building

L-Shape

Side Face BFront Face A

Adapted from Gomes et al (2005)

LA-UR-05-8025

QUIC vs. Experimental Data: L-Shaped Building

L-Shape

Side Face BFront Face A

Adapted from Gomes et al (2005)

LA-UR-05-8025

QUIC vs. Experimental Data: U-Shaped Building

U-Shape

Side Face DFront Face C

Adapted from Gomes et al (2005)

LA-UR-05-8025

QUIC vs. Experimental Data: U-Shaped Building

U-Shape

Side Face BFront Face A

Adapted from Gomes et al (2005)

LA-UR-05-8025

QUIC vs. Experimental Data

ΔCp = Max Cp Front Face – Min Cp Back Face

-0.5

0

0.5

1

1.5

2

Cube(Normal)

Cube (45 deg)

L shaped U shaped High Rise Squat 7x1 Array (1st bldg)

7x1 Array(2nd bldg)

7x1 Array (7th bldg)

ΔC

p

ModelExperiment

( )22

1o

op V

PPC

ρ−

=

LA-UR-05-8025

QUIC vs. Experimental Data

The Maximum Cp on the Front Face

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

Cube(Normal)

Cube (45 deg)

L shaped U shaped High Rise Squat 7x1 Array (1st bldg)

7x1 Array(2nd bldg)

7x1 Array (7th bldg)

Cp

(max

)

ModelExperiment

( )22

1o

op V

PPC

ρ−

=

LA-UR-05-8025

QUIC vs. Experimental Data

The Minimum Cp on the Back Face

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

Cube(Normal)

Cube (45 deg)

L shaped U shaped High Rise Squat 7x1 Array (1st bldg)

7x1 Array(2nd bldg)

7x1 Array (7th bldg)

Cp

(Min

)

ModelExperiment

( )22

1o

op V

PPC

ρ−

=

LA-UR-05-8025

Where the Combined Wind & Pressure Solvers Could Make a Difference

• Off-angle winds

• Dense Urban Areas - Sheltering effects of surrounding buildings

• Detailed analyses of building of interest (where locations of vents, windows, doors are known)

LA-UR-05-8025

Off-Angle Winds

CFD simulations of Gomes et al. (2005)

LA-UR-05-8025

Detailed Analyses of Buildings of Interest

Specify pressure boundary conditions at inlets and outlets for control volume codes.

e.g., COMIS

LA-UR-05-8025

Dense Urban Areas – Sheltering Effect

In city centers, buildings will have much lower natural ventilation rates due to obstruction of wind by surrounding buildings.

LA-UR-05-8025

Dense Urban Areas – Sheltering Effect

In city centers, buildings will have much lower natural ventilation rates due to obstruction of wind by surrounding buildings.

Bauman et al (1988) “Studies show wind pressure reductions of up to 90% resulting from wind blockage by upwind buildings. However, there is a variability of 80% depending on the configuration of the buildings.”

CFD simulations of Yang et al. (2005)

LA-UR-05-8025

Dense Urban Areas – Sheltering Effect

Indoor models often have sheltering correction factors, e.g.,

UDM reduces the ΔP by a fixed amount if the building plan area density is above a specific threshold.

LA-UR-05-8025

Dense Urban Areas – Sheltering Effect

Pressures computed for Madison Square Garden, NYC.

LA-UR-05-8025

Summary

• Wind-induced pressure information on buildings can be used to improve indoor dosage calculations (for outdoor and indoor releases)

• The QUIC wind and pressure solvers are relatively computationally inexpensive and would fit into a fast-response T&D modeling system

• Preliminary evaluation studies indicate that the QUIC wind and pressure solvers generally provide reasonable agreement with experimental studies

LA-UR-05-8025

Technical Challenges

• Rooftop pressures on flat roofs difficult to match

• How about pitched roofs?

• Lack of experimental data in complex building environments

• Is turbulence important?

LA-UR-05-8025

Acknowledgements

This work funded by the JSTO.

Special thanks to John Pace, John Hannan, and Rick Fry for the opportunity to perform this work.

LA-UR-05-8025

CFD vs. Experimental Data

Adapted from Richards & Hoxey (2006)

Rooftop Pressure