Sensitivity Studies Wind tunnel simulations Alan Robins, Paul - - PowerPoint PPT Presentation

Sensitivity Studies

Wind tunnel simulations

Alan Robins, Paul Hayden, Hong Cheng, Matteo Carpentieri, Sando Baldi, Paolo Giambini, Tom Lawton, James Hamilton, Alex Nicolson, Nadia Bahar, James Fabian, Khaled Bashiti, et al Content: The EnFlo Wind tunnel Comparisons with field data Flow and dispersion characteristics Sensitivity studies Conclusions

18 channel air sampling system FID system Flow control system Chilled water supply ~ 10oC Inlet and heater section - 15 layers 400 kW capacity; ambient to ~ 80oC 20 x 3.5 x 1.5 m working section 0 - 4 m/s Mechanical simulation devices Twin fans Heat exchanger Cooled rough wall ~ 10oC Gas supplies Heater control Source Rough wall Computerised control, data collection & data analysis Thermo- couple system Speed control Traverse and turntable control

The EnFlo wind tunnel

Wind speed ratios

Ratio of BT Tower and WCL roof top wind speeds High wind speed limit is about 1:4 essential data for relating wind tunnel and field data

Comparison of concentrations

The scatter is mainly due to the large repeat variability in the field data.

C* = CUH 2 Q

Component along Marylebone Rd, Ut relative to Uref

Comparison of flow fields

Data from WCC House (at 18 m), U(BT) > 3 ms-1 Ut Uref

Component normal to Marylebone Rd, Un relative to Uref

Comparison of flow fields

Data from WCC House (at 18 m), U(BT) > 3 ms-1 Un Uref

Flow structure at an intersection

Mean velocity vectors at height of 5 m

• extract from full data set

Clarification of physical processes in support of field measurements

Wind at 45 degrees to street Δx ~ W at location of measurements Emission released upstream at street centre - dispersion prior to uniform mixing across the street Concentration contours in cross-section; H/W≈1.0 roof level roof level

Near-source concentrations in helical circulation

Short Duration Emissions

Uc ≈ UH/4

Characterisation of advection speed, Uc, and along wind mixing.

ΔT50 ≈ T50/3 Time of flight, T50 - rise

• fall

Rise time, ΔT50

• fall time

7.5(R/H) 2.5(R/H)

Concentration decay

The basic mean concentration decay function, CUH2/Q = 12 (R/H)-2

not fully mixed

R is source- receptor separation Concentration, C* = CUH2/Q Separation, R/H

Fluctuation levels

Separation, R/H Concentration fluctuation, c'/Cmax instrument noise level

Concentration, C* Time, s

Wind tunnel research

EnFlo 20x3.5x1.5 boundary layer wind tunnel 0 to 3.5 m/s; inlet flow heating; surface heating and cooling

Sensitivity studies examples

Basic sensitivity studies

Dispersion repeat run to run variability better than ±10% - so take ± 20% as the target for singificant effects On this basis, results shown to be insensitive to:  modest variations in upwind boundary layer conditions including slightly stable approach flow (LMO ~ Hbl ~ 1 m)  addition of roof roughness  modification to zero-plane level at upwind edge of model  source design and emission rate (speed ratio)  Reynolds, U(H)H/ν, number from about 2,000 to 14,000

Comparison of results from low and higher resolution models; significant differences confined to short ranges.

• 1. Model detail

Model detail

• 2. Street blockages - ineffective

Reference; blockage at A, B or C; blockage at A, B & C. Green, reference; blue, blockage at A; red, at D; yellow at A & D. Wind direction

• 90, degs

Wind direction

• 45, degs

Street blockage

Concentrations along Marylebone Road

• 3. Street blockage - effective

blockage source Wind direction

• 45 degrees

9 7

Wind direction

• 90, degs

Blockage

Run 7, York St a canyon from west of Upper Montagu Street to east of Gloucester Place Run 9, Gloucester Place closed at York Street 9 7 CUref Hbl

2

Q x(mm) basically the same not these

• 4. Scale effects

1:200 scale model 1:500 scale model

Scale effects

CU(H)H2/Q along Marylebone Road, source York Street 1:200 scale 1:500 scale Distance from intersection, m CU(H)H2/Q

• 5. Slightly unstable approach flow

Unstable Hbl/LMO ~ -2.7, u*/Uref = 0.10 Neutral u*/Uref = 0.06

w'2 Uref

2

w'2 Uref

2

Marlylebone Rd

• 5.00E+00

0.00E+00 5.00E+00 1.00E+01 1.50E+01 2.00E+01 2.50E+01 3.00E+01

• 600
• 400
• 200

200 400 600 800 1000 X mm CUA/Q

Unstable Neutral

Slightly unstable approach flow

CUrefHbl

2/Q along Marylebone Road, source York Street

neutral unstable Distance from intersection, mm CUrefHbl

2/Q

Hbl/LMO = -2.7

 C* = CUHH2/Q independent of Re (U) and Q over range examined  C* independent of model scale 1:200 to 1:500  C* repeatability (multi-user) better than ±10%  Effect of shallow boundary layer on C*(x,y,0) ~ -25% (not shown)  Effect of model detail on C* ≤ +30%  Effect of upwind detail on C* < ±25%  Effect of slightly unstable approach flow on C* ~ - 30%  … but C* ~ 12(R/H)-2 robust

Sensitivity studies