Basics of Dispersion Modeling

Slides:



Advertisements
Similar presentations
Phoenics User Conference on CFD May 2004 Vipac Engineers & Scientists Ltd COMPUTATIONAL FLUID DYNAMICS Simulation of Turbulent Flows and Pollutant Dispersion.
Advertisements

Extra Large Telescope Wind Engineering. Wind and Large Optical Telescopes Wind is a key factor in the design of large telescopes: larger wind-induced.
Introduction to Computational Fluid Dynamics
Bridging the Gap Between Statistics and Engineering Statistical calibration of CFD simulations in Urban street canyons with Experimental data Liora Malki-Epshtein.
Dominic Hudson, Simon Lewis, Stephen Turnock
Hongjie Zhang Purge gas flow impact on tritium permeation Integrated simulation on tritium permeation in the solid breeder unit FNST, August 18-20, 2009.
Improving and Trouble Shooting Cleanroom HVAC System Designs By George Ting-Kwo Lei, Ph.D. Fluid Dynamics Solutions, Inc. Clackamas, Oregon.
Introduction to SCREEN3 smokestacks image from Univ. of Waterloo Environmental Sciences Marti Blad NAU College of Engineering and Technology.
Introduction to SCREEN3 smokestacks image from Univ. of Waterloo Environmental Sciences Marti Blad.
OpenFOAM for Air Quality Ernst Meijer and Ivo Kalkman First Dutch OpenFOAM Seminar Delft, 4 november 2010.
Module 9 Atmospheric Stability Photochemistry Dispersion Modeling.
Enclosure Fire Dynamics
LES of Turbulent Flows: Lecture 3 (ME EN )
Derivation of the Gaussian plume model Distribution of pollutant concentration c in the flow field (velocity vector u ≡ u x, u y, u z ) in PBL can be generally.
Suspended Load Above certain critical shear stress conditions, sediment particles are maintained in suspension by the exchange of momentum from the fluid.
8 th Conference on Air Quality Modeling – A&WMA AB3 Comments on Nonstandard Modeling Approaches By Ron Petersen, CPP Inc Blue Spruce Drive Fort Collins.
AIAA SciTech 2015 Objective The objective of the present study is to model the turbulent air flow around a horizontal axis wind turbine using a modified.
Air Quality Modeling.
CFD Modeling of Turbulent Flows
1 Predicting and Understanding the Breakdown of Linear Flow Models P. Stuart, I. Hunter, R. Chevallaz-Perrier, G. Habenicht 19 March 2009.
Session 4, Unit 7 Plume Rise
AMBIENT AIR CONCENTRATION MODELING Types of Pollutant Sources Point Sources e.g., stacks or vents Area Sources e.g., landfills, ponds, storage piles Volume.
BsysE595 Lecture Basic modeling approaches for engineering systems – Summary and Review Shulin Chen January 10, 2013.
Air Dispersion Primer Deposition begins when material reaches the ground Material from the lower stack reaches the ground before that of the taller stack.
ENVIRONMENTAL WIND FLOWS AROUND BUILDINGS OUTLINING FLOW MECHANISMS – FLOW STRUCTURE AROUND ISOLATED BUILDING. Why architects need the knowledge about.
Hiromasa Nakayama*, Klara Jurcakova** and Haruyasu Nagai*
Dispersion Modeling A Brief Introduction A Brief Introduction Image from Univ. of Waterloo Environmental Sciences Marti Blad.
Wind Energy Program School of Aerospace Engineering Georgia Institute of Technology Computational Studies of Horizontal Axis Wind Turbines PRINCIPAL INVESTIGATOR:
Mathematical Equations of CFD
Building Aware Flow and T&D Modeling Sensor Data Fusion NCAR/RAL March
A canopy model of mean winds through urban areas O. COCEAL and S. E. BELCHER University of Reading, UK.
Session 3, Unit 5 Dispersion Modeling. The Box Model Description and assumption Box model For line source with line strength of Q L Example.
Air Quality Effects of Prescribed Fires Simulated with CMAQ Yongqiang Liu, Gary Achtemeier, and Scott Goodrick Forestry Sciences Laboratory, 320 Green.
Introduction to Modeling – Part II
Types of Models Marti Blad Northern Arizona University College of Engineering & Technology.
DIMENSIONAL ANALYSIS SECTION 5.
Intro to Modeling – Terms & concepts Marti Blad, Ph.D., P.E. ITEP
Turbulence Models Validation in a Ventilated Room by a Wall Jet Guangyu Cao Laboratory of Heating, Ventilating and Air-Conditioning,
Lecture Objectives: Define 1) Reynolds stresses and
1 LES of Turbulent Flows: Lecture 7 (ME EN ) Prof. Rob Stoll Department of Mechanical Engineering University of Utah Spring 2011.
Lecture Objectives: Accuracy of the Modeling Software.
7. Air Quality Modeling Laboratory: individual processes Field: system observations Numerical Models: Enable description of complex, interacting, often.
CHAPTER 6 Introduction to convection
PowerPoint Presentation
Objective Introduce Reynolds Navier Stokes Equations (RANS)
The Prediction of Low Frequency Rumble in Combustion Systems
Introduction to the Turbulence Models
TERRAINS Terrain, or land relief, is the vertical and horizontal dimension of land surface. Terrain is used as a general term in physical geography, referring.
Types of Models Marti Blad PhD PE
Predictability of orographic drag for realistic atmospheric profiles
Enhancement of Wind Stress and Hurricane Waves Simulation
PCB 3043L - General Ecology Data Analysis.
Variable Air Volume (VAV) Exhaust System Implementation Challenges
Objective Review Reynolds Navier Stokes Equations (RANS)
Fluid Flow Regularization of Navier-Stokes Equations
The application of an atmospheric boundary layer to evaluate truck aerodynamics in CFD “A solution for a real-world engineering problem” Ir. Niek van.
AIAA OBSERVATIONS ON CFD SIMULATION UNCERTAINITIES
AIAA OBSERVATIONS ON CFD SIMULATION UNCERTAINTIES
Suggested Analyses of WRAP Drilling Rig Databases
Models of atmospheric chemistry
Lecture Objectives Review for exam Discuss midterm project
PURPOSE OF AIR QUALITY MODELING Policy Analysis
AIAA OBSERVATIONS ON CFD SIMULATION UNCERTAINTIES
Introduction to Modeling – Part II
CFD computations of liquid hydrogen releases
Objective Discus Turbulence
Accurate Flow Prediction for Store Separation from Internal Bay M
Wind Velocity One of the effects of wind speed is to dilute continuously released pollutants at the point of emission. Whether a source is at the surface.
14. Computational Fluid Dynamics
Low Order Methods for Simulation of Turbulence in Complex Geometries
Presentation transcript:

Basics of Dispersion Modeling Anke Beyer-Lout & Jared Ritter 2017 Colorado I2SL Chapter Education Day Colorado School of Mines August 29, 2017

Outline Describe the problem of effluent dispersion within a turbulent flow field Review the types of dispersion models that are available Explain the strengths and weaknesses of each simulation technique Define which is most appropriate for a given application This presentation covers the following: Also will include occasional discussion on a case study for the recent construction CoorsTek

Effluent Dispersion Primary Objective: Health/Odor Concerns Prevent exhaust re-entrainment within the building and neighboring locations Additional Considerations Cost Savings Capital Cost – Stack height, entrained flow vs. conventional fan Operational cost - Energy consumption Building aesthetic The main concern is to limit exhaust concentration No hazardous or odorous exhaust is re-entrained or affects neighboring locations Additional benefits: Potential cost savings Day 1 – capital cost Long term energy use Architecturally acceptable

Dispersion Modeling Methods Graphical Methods Analytical Methods Numerical Methods (CFD) Physical Modeling (Wind Tunnel) Sort of ranked in terms of complexity Could argue that CFD and wind tunnel are most complex, depending on how many cfd simuluations are run Benefit of better model - most accurate exhaust concentrations – turndown and energy savings

Airflow Around Buildings For all of the dispersion methods, the main driving factor is the complex interaction of the approaching wind and the building 2 main phenomena Rooftop recirc cavity – wind perpendicular to upwind wall Corner vortecies – wind approaching corner

Graphical Method 2015 Handbook – HVAC Applications Chapter 45 From the ASHRAE Handbook Guideline for stack placement Calculate rooftop wake cavity using building dimensions From that you can get stack placement and prelim stack ht ASHRAE also provides preliminary recommendations stacks on roof, intake on ground to take advantage of stagnation region Increasing separation distance, or sidewall receptors

Graphical Method Pros: Identifies ideal placement of exhaust stacks on the roof Provides quick indication of possible stack height requirements Cons: No actual concentration predictions Can only be used for isolated building Can be extremely conservative for large buildings Good starting point to begin to understand how the wind affects the stack design However, no concentrations are predicted Not recommended for hazardous stacks

Analytical Dispersion Methods Gaussian Diffusion Equation Plume Rise, hplume calculation Horizontal Dispersion Coefficient, σy Vertical Dispersion Coefficient, σz Receptor Height, htop More in depth Exhaust concentration predictions achieved by combining plume rise calculations and gaussian dispersion (horizontal and vertical) Usually conservative when used properly, but may not be if there are complexities (upwind structures, screenwalls) Several models fall in this category – including AERMOD

Case Study – CSM CoorsTek Analytical Dispersion Method Pros: Open environment Surrounding buildings similar in size Case study of nearby recently completed project on CSM Because the campus is realtively open with all buildings similar size, would think ok for analytical method If Alderson Hall were 3x size of CoorsTek the approach from the south would be turbulent, which would not be handled well by analytical dispersion

Case Study – CSM CoorsTek Analytical Dispersion Method Cons: Complex rooftop features Critical locations are expected to be within the rooftop wake cavity Intake flow from the AHUs may affect results This is a rendering of the CoorsTek building Relatively complex roof with stacks and intakes within the screenwall In close proximity (Numerical plume rise could over or under predict) AHU’s intake suction could affect the shape of the plume within the screen

Analytical Dispersion Methods Pros: Relatively easy to use Based on Gaussian plume dispersion theory Takes into account building and stack downwash When properly applied, will generate conservative results Less expensive than CFD or Wind Tunnel testing Cons: Building geometries are generic Doesn’t properly account for corner vortices or complex rooftops May be overly conservative for certain situations Most accurate for isolated buildings So to summarize…. Relatively quick/easy method of obtaining exhaust concentration predictions Can be used in basic-building applications, but becomes less accurate with complex rooftops or surrounds In these cases, better predictions can be obtained with CFD or physical wind tunnel modeling - ANKE

Numerical Dispersion Methods - CFD Computational Fluid Dynamics Models (or CFD) solve the equations of motion on a fixed grid in order to calculate the turbulent motion of the air masses. Reynolds Number Averaged Navier Stokes (RANS) Large Eddy Simulation (LES) Direct Numerical Simulation (DNS) Computational Fluid Dynamics Models (or CFD) solve the equations of motion on a fixed grid in order to calculate the turbulent motion of the air masses. CFD simulations are extremely computationally intensive and usually only feasible for small scales. Even then, not all scales of turbulence can be resolved, because the smallest scales of turbulence in the atmosphere are on the millimeter scale (or fraction of an inch scale) and the largest turbulence elements can be on the km scale (i.e. mile). Therefore some or all turbulence needs to parameterized. If all turbulence is parameterized, you end up with a Reynolds number averaged Navier Stokes (RANS) CFD model. If only the turbulence that is smaller than your grid is parameterized and the larger turbulence scales are resolved, you have yourself a large eddy simulation (or LES). If you have the time and money to resolve all turbulence scales, you end up with direct numerical simulation (DNS).

Numerical Dispersion Methods - CFD Steady State Reynolds-Average Navier Stokes (RANS) Does not resolve turbulent fluctuations within the flow field (important for dispersion modeling)

Numerical Dispersion Methods - CFD Large Eddy Simulation (LES) Resolves most of the turbulent fluctuations within the flow field. …but can be expensive and time-consuming to run.

Numerical Dispersion Methods - CFD Critical Issues: Mesh Resolution Turbulence Modeling Boundary Conditions Numerical diffusion

Numerical Dispersion Methods - CFD Pros: Provides a nice visual picture of flow phenomenon RANS useful for interior and mean flow exterior modeling Can model complex environments Can include atmospheric stability and thermal effects Cons: Steady-State RANS does not sufficiently resolve turbulence. (not accurate for plume dispersion modeling) LES shows promise, but the time dependent models are currently too computationally intensive for most applications Every wind speed and wind direction requires a separate simulation

Physical Dispersion Methods Construct Scale Model of Site Building and Surroundings Colorado School of Mines CoorsTek

Physical Dispersion Methods Place in Atmospheric Boundary Layer Wind Tunnel

Physical Dispersion Methods Concentration Measurements in Atmospheric Boundary Layer Wind Tunnel Tracer gas released from stack Sample withdrawn from intake

Physical Dispersion Methods Develop Concentration Profiles of C/m vs. WS and WD C/m vs. Wind Direction for various Wind Speeds C/m vs. Wind Speeds for various Wind Directions

Physical Dispersion Methods Pros: Accurate simulation of air flow patterns around complex buildings Can evaluate impact of screens, stack shape, and other reasonably small structures Multiple wind speeds and wind directions can be readily modeled Cons: Size of wind tunnel limits the sizes (scales) of objects Discrete sampling rather than entire flow field (flow visualization can be misleading) Atmospheric Stability and thermal impact difficult to model

Conclusions Graphical Method Use only as a first cut for estimating stack height and location Do not use for final design if using hazardous chemicals. So, what have we learned? The graphical method is a starting point. I tried the ASHRAE graphical method on the CoorsTek Building. The building outline is shown in blue (screen walls and a rooftop AHU are included). As a rough estimate, a stack height of 40ft above roof is required so the lower edge of the plume stayed above the recirculation zones. The graphical method does not prodce concentration estimates and is therefore not recommended for…

Conclusions Analytical Modeling Provides plume dispersion estimates and thus can be used for final design Applicable for simple buildings with no taller surrounding buildings/features and/or when stack height is not a critical issue A gaussian plume model provides concentration estimates and thus can be used for final design. … If we look at the CoorsTek Building, we can see rooftop equipment and screen walls that a plume model does not include. To account for the increased plume spread due to the screen wall, ASHRAE recommends screen wall reduction factors. Overall, analytical models are designed to be conservative. If desirable concentrations are computed using a Gaussian Plume model, the exhaust and intake design can be finalized at that point. However, advanced modeling methods may be able to optimize the design.

Conclusions CFD Modeling Typical RANS method is not appropriate for dispersion modeling. Rather, only LES models should be used for dispersion modeling Make sure sufficient WS/WD are modeled Validation of CFD results is still very critical If CFD is desired to be used for exhaust dispersion, sufficient WS/WD need to be modeled. And, sufficient means, at least 16 wind directions and multiple wind speeds for each direction foe each stack. We are talking up to 100 simulations to ensure that absolute worst case concentrations are computed. Considering that each simulation can take days to run, depending on the domain size, that project schedule can get out of hand very quickly. Interior air flows are much more suitable to be simulated with CFD. We did not do any CFD simulations for the CoorsTek Building. If we had, it may have looks similar to these two examples here. The top picture shows an LES simulation of atrium smoke. The second picture shows odor associated with a tissue digester. In this case supply air comes from the vents up top and exhaust is pulled through the vent next to the digester.

Conclusions Wind Tunnel Modeling EF-1 Wind Tunnel Modeling Use for complex geometries and/or when stack height is critical Can be fully validated using EPA methodology Use when employing a VAV exhaust strategy Make sure sufficient WS/WD are modeled AHU-3 AHU-2 AHU-1 Alternate designs, such as different screen walls, stack heights and intake locations are easily evaluated in the wind tunnel. Use when employing a VAV exhaust strategy, because the exhaust parameters can be fine tuned. In a project timeline, wind tunnel simulations can be done during initial design, to optimize stack and intake placement, as well as during final design.

Case Study – CSM CoorsTek Method Maximum Predicted Concentration (C/m) Worst- Case Location Graphical 40ft stack height Analytical 1,655 μg/m3 per g/s (16ft stack height) AHU-2 Intake Wind Tunnel 614 μg/m3 per g/s (16ft stack height) We can compare the concentration predictions for the Coorstek Building, That concludes our presentation and we would love to answer your questions.

Senior Project Engineer Questions? For More Information Anke Beyer-Lout Jared Ritter Senior Lead Scientist Senior Project Engineer abeyerlout@cppwind.com jritter@cppwind.com CPP, Inc. 2400 Midpoint Drive, Suite 190 Fort Collins, CO 80525 (970) 221-3371