rama_upload 4_ cn3135_toxic releases & dispersion models_nus lecture 09.09.2014.pptx

8/10/2019 Rama_Upload 4_ CN3135_Toxic Releases & Dispersion Models_NUS Lecture 09.09.2014.pptx

http://slidepdf.com/reader/full/ramaupload-4-cn3135toxic-releases-dispersion-modelsnus-lecture-09092014pptx 1/65

24/11/2014 NUS CN3135_Consequence Modelling & Explosion Lecture 4

Process Safety, Health

and EnvironmentCN3135

Ramasami Sundaresan(Rama)




Lecture # 4NUS CN3135: Consequence Modelli

Toxic Releases & Dispersion Mode




Learning Outcomes

1. Describe and identify fire and explosion hazards in

processes

2. Analyse and assess the potential impacts of these hon safety and health of People, The Environment anProperty.

3. Develop methods for controlling process hazards tominimize fire and explosion risks.

4. Design measures to safely mitigate the consequenprocess incidents



5-0 Toxic Releases & Dispersion Models

In a process incident, equipment can release toxic materials quickly a

enough quantities impacting:

On-Site Impact: Occupational Health impact Off-Site Impact: Environmental (community) impact

Incident can be in various forms of ruptures:

Process vessel rupture – excessive pressure from runaway rea

Pipeline rupture –

overpressure and/or corrosion failure Storage tank failure – content above its atmospheric boiling po

Bulk rail/road tanker rupture – collision and mechanical failure




Importance of Toxic Release Models

Serious historical accidents (e.g. Bhopal) emphasize the importance orelease models for the prevention of release situations and mitigatingconsequences

Safe plant design to minimize

Probability of occurrence

Magnitude of release consequences

Planning for emergencies aimed at reducing the impact of a releasoccurs

Identifying the release incident (Sections 4-9 and 4-10)

Developing a source model to describe how materials are released and the(Chapter 4),

Estimating the downwind concentrations of the toxic material using a dispe

Pre-Incident Planning




Dispersion Models

Dispersion models describe the airborne transport of toxic material

accident site and into the plant and community.

After a release the airborne toxic material is carried away by the wicharacteristic pattern as

A Plume, or

A Puff

The maximum concentration of toxic material occurs at the release

At ground level, or

At Height

Concentrations diminish downwind due to turbulent mixing and dis

toxic substance with air.




A Characteristic Plume

24/11/2014 NUS CN3135_Consequence Modelling & Explosion Lecture 4Figure 5-1 Characteristic plume formed by a continuous release of mate



A Puff




5-1 Parameters Affecting Dispersion

Parameters that affect atmospheric dispersion of toxic materials:

Wind speed,

Atmospheric stability,

Ground conditions (buildings, water, trees),

Height of the release above ground level,

Momentum and buoyancy of the initial material released.

Wind speed

As the wind speed increases, the plume becomes longer and narro

The substance is carried downwind faster, and

Diluted faster by a larger quantity of air.




Classes of Atmospheric stability

Three stability classes - Unstable, Neutral, and Stable.

Unstable atmospheric conditions

Sun heats the ground faster than the heat can be removed so that the air te

ground is higher than the air temperature at higher elevations, as in the ear Results in unstable stability because air of lower density is below air of grea

Effects of buoyancy enhances atmospheric mechanical turbulence.

Neutral stability

Air above the ground warms and the wind speed increases, reducing the ef

input, or insolation.

The air temperature difference does not influence atmospheric mechanical

Stable atmospheric conditions

Sun cannot heat the ground as fast as the ground cools;

Temperature near the ground is lower than the air temperature at higher ele

Condition is stable because the air of higher density is below air of lower de

The influence of buoyancy suppresses mechanical turbulence.




Ground conditions

Ground conditions affect the

mechanical mixing at the

surface and the wind profilewith height.

Trees and buildings increase

mixing, whereas lakes and

open areas decrease it.

Figure shows the change in

wind speed versus height for

a variety of surface

conditions.


Figure 5-4 Effect of ground conditions on



Height of Release Release height significantly affects ground-level concentrations.

As the release height increases, ground-level concentrations are reduced

disperse a greater distance vertically.


Figure 5-5 Increased release height decreases the ground concentratio



Buoyancy and Momentum

Buoyancy and momentum of the material released change the effective h

release.

The momentum of a high-velocity jet will carry the gas higher than the po

resulting in a much higher effective release height.

Gas density less than air, the released gas will initially be positively buoy

upward.

Density greater than air, then the released gas will initially be negatively b

slump toward the ground.

Temperature and molecular weight of the released gas determine gas de

that of air (with a molecular weight of 28.97).

For all gases, as the gas travels downwind and is mixed with fresh air, a

eventually be reached where the gas will be diluted adequately to be con

buoyant.

At this point the dispersion is dominated by ambient turbulence.




Buoyancy and Momentum


Figure 5-6 The initial acceleration and buoyancy of the released material affects the plum



As the plume is swept downwind, the concentration profile

spreads out and decreases



5-2 Neutrally Buoyant Dispersion Models

Neutrally buoyant dispersion models are used to estimate the conc

downwind of a release in which the gas is mixed with fresh air to th

resulting mixture is neutrally buoyant.

These models apply to gases at low concentrations, typically in the

million range.

Two types of neutrally buoyant vapor cloud dispersion models are c

The plume model - describes the steady-state concentration of ma

from a continuous source.

A typical example is the continuous release of gases from a smokestack.

A steady-state plume is formed downwind from the smokestack.




Neutrally Buoyant Dispersion Models

Given the instantaneous release of a fixed mass of material, Q*, into

expanse of air, the coordinate system is fixed at the source.

Assuming no reaction or molecular diffusion, the concentration C of mfrom this release is given by the advection# equation


#Advection = a transport mechanism of a substance by a fluid due to the fluid's bulk motion.

+

= 0

where

U j is the velocity of the air

Subscript j is the summation over all coordinate directions x, y, and z.



Neutrally Buoyant Dispersion – Velocity Determ


If the velocity u, in Equation 5-1 is set equal to the average wind velocity and

solved, material dispersal will be much faster than predicted. This is due to tu

velocity field.

However, if wind velocity is specified exactly with time and position, includresulting from turbulence, Equation 5-1 would predict the correct concentration.

Unfortunately, no models are currently available to adequately describe turbulen

Hence, an approximation is used.

Let the velocity be represented by an average (or mean) and stochastic (ra

quantity

u j = (u j ) + u' j (5-2)

where

(u j ) is the average velocity and

u' j is the stochastic fluctuation resulting from turbulence.

N t ll B t Di i C t ti



It follows that the concentration C will also fluctuate as a result of the velocity field; so

C = (C) + C

where

(C) is the mean concentration and C' is the stochastic fluctuation.

Because the fluctuations in both C and u t are around the average or mean values, it fol

(u

j

) = 0

(C ) = 0 .

Substituting Equations 5-2 and 5-3 into Equation 5-1 and averaging the result over time

The terms ′and ′ are zero when averaged ′ = ′ = 0 , but the turbu

term ′′ is not necessarily zero and remains in the equation.


(5-3)

(5-4)

(5-5

+

+

′ ′ =

Neutrally Buoyant Dispersion – Concentrati

Neutrally Buoyant Dispersion – Concentratio




Neutrally Buoyant Dispersion Concentratio An additional equation is required to describe the turbulent flux.

The usual approach is to define an eddy diffusivity K j (with units of area/time) such t

Substituting Equation 5-6 into Equation 5-5 yields

If the atmosphere is assumed to be incompressible, then

And Equation 5-7 becomes

Equation 5-9 together with appropriate boundary and initial conditions forms the fundame

dispersion modeling. This equation will be solved for a variety of cases.

′′ =

+

=

=

+

=

Dispersion Coordinates - Plume




Dispersion Coordinates Plume

The coordinate system used for the dispersion models is shown in Figures

The x axis is the centerline directly downwind from the release point and

different wind directions. The y axis is the distance off the centerline, and

the elevation

Figure 5-7 Steady-state continuous point source release with wind.

z

y

x (Downwind)

(Off-wind)

(Vertical)

Wind Direction with

Wind Speed, u

Continuous Release of Material

at Rate Q m occurs here

Puff Movement with Wind




Puff Movement with Wind

The coordinate system used for the dispersion models is shown in Figures

The x axis is the centerline directly downwind from the release point and

different wind directions. The y axis is the distance off the centerline, and

the elevation

Figure 5-8 Puff with wind. After the initial instantaneous release, the puff moves with t

Wind Direction with

Wind Speed, u

Instantaneous Release of

Material at Rate Q*m

occurs here at t=0

Initial Puff at t=0 .

Puff moves downwind

at velocity u

Puff at t=t 1

Ground Influence on Plume




Ground Influence on Plume

The ground represents an impervious boundary. As a result, the concentrati

twice the concentration in plume without influence of ground.

Figure 5-9 Steady-state plume with source at ground level. The

concentration is twice the concentration of a plume without the ground.

z

y

x

Wind Direction with

Wind Speed, u

Continuous Release of Material

at Rate Q m occurs here



Pasquill-Gifford Model




Pasquill Gifford Model

The dispersion coefficients are a function of

Atmospheric conditions

Classified according to six different stability classes that are wind speed and quantity of sunlight.

The distance downwind from the release.

During the day, increased wind speed results in greater atmosp

whereas at night the reverse is true. This is due to a chan

temperature profiles from day to night.



P-G Model – Plume Dispersion Coefficients




G ode u e spe s o Coe c e tsThe dispersion coefficients , and for a continuous source are given in Figures 5-10 a

below. The corresponding correlations given in Table 5-2 on following slide.

Figure 5-10 Dispersion coefficients for Pasquill-Gifford plume model for r

P-G Model – Plume Dispersion Coefficients




pThe dispersion coefficients , and for a continuous source are given in Figures 5-10 a

below. The corresponding correlations given in Table 5-2 on following slide.

Figure 5-10 Dispersion coefficients for Pasquill-Gifford plume model for u

P-G Plume Dispersion Coefficient Correlatio




pTable 5-2 Recommended Equations for Pasquill-Gifford Dispersion Coefficients for Plume Dis

downwind distance x has units of meters)

Pasquill-Gifford stability class σy (m) σ z (m)

Rural conditions

A 0.22 x (1 + 0.0001 x )-1/2

0.20 x B 0.16 x (1 + 0.0001 x )

-1/20.12 x

C 0.11 x (1 + 0.0001 x )-1/2

0.08 x (1 + 0.0002x)-1/2

D 0.08 x (1 + 0.0001 x )-1/2

0.06 x( 1 + 0.0015x)-1/2

E 0.06 x (1 + 0.0001 x )-1/2

0.03 x (1 + 0.0003x)-1

F 0.04 x( 1 + 0.0001 x )-1/2

0.016 x (1 + 0.0003x)-1

Urban conditions

A-B 0.32 x (1 + 0.0004 x )-1/2

0.24 x (1 + 0.001x)+1/2

C 0.22 x (1 + 0.0004 x )-1/2

0.20 x

D 0.16 x (1 + 0.0004 x )-1/2

0.14 x (1 + 0.0003x)-1/2

E-F 0.11 x (1 + 0.0004 x )-1/2

0.08 x (1 + 0.0015x)-1/2

Values for are not provided because it is reasonable to assume that =

P-G Model – Puff Dispersion Coefficients




pThe dispersion coefficients and for a puff release are given in Figure 5-12 and the equ

in Table 5-3. The puff dispersion coefficients are based on limited data (shown in Table 5-2

considered precise.

Figure 5-12 Dispersion coefficients for Pasquill-Gifford puff model.

P-G Puff Dispersion Coefficient Correlations




pTable 5-3 Recommended Equations for Pasquill-Gifford Dispersion Coefficients for Puff Dispe

downwind distance x has units of meters)

Pasquill Rederived Equations




q qThe equations for cases 1 through 10 (pgs. 190 -196) were rederived by Pasquill6 using exp

of Equation 5-37. These equations along with the correlations for the dispersion coefficient

the Pasquill-Gijford model.

Case 11: Puff with Instantaneous Point Source at Ground Level, Coordinates Fixed at R

Constant Wind Only in x Direction with Constant Velocity u

This case is identical to case 7. The solution has a form similar to Equation 5-33:

The ground-level concentration is given at z = 0:

Pasquill Rederived Equations - Case 11

http://c/Users/user/Documents/NUS/F&E%20CM/Text%20Book%20&%20External%20References/Chapter%205%20_%20Toxic%20Release%20&%20Dispersion%20Models.html

http://c/Users/user/Documents/NUS/F&E%20CM/Text%20Book%20&%20External%20References/Chapter%205%20_%20Toxic%20Release%20&%20Dispersion%20Models.html




q qCase 11: Puff with Instantaneous Point Source at Ground Level, Coordinates Fixed at


The center of the cloud is found at coordinates (ut, 0,0). The concentration at the center o

cloud is given by

The ground-level concentration along the x axis is given at y = z = 0:

The total integrated dose Dtld received by an individual standing at fixed coordinates (x, y,

integral of the concentration:




Case 11: Puff with Instantaneous Point Source at Ground Level, Coordinates Fixed at


Frequently the cloud boundary defined by a fixed concentration is required.

The line connecting points of equal concentration around the cloud boundary is called a

For a specified concentration (C)* the isopleths at ground level are determined by:

Dividing the equation for the centerline concentration (Equation 5-40)by the equation for the general ground-level concentration (Equation 5-39

This equation is solved directly for y:

The procedure is

1. Specify ∗, u, and t .

2. Determine the concentrations ( x , 0,0, t) along the x axis using Equation 5-40. Define the bound

the x axis.

3. Set ( x, 0,0, t) = ∗ in Equation 5-45, and determine the values of y at each center-line point d

The procedure is repeated for each value of t required.





Case 12: Plume with Continuous Steady-State Source at Ground Level and Wind Movi

Constant Velocity u


The ground-level concentration is given at z = 0:

The concentration along the centerline of the plume directly downwind is given at y = z = 0:

The isopleths are found using a procedure identical to the isopleth procedure used for case

For continuous ground-level releases the maximum concentration occurs at the release





Case 13: Plume with Continuous Steady-State Source at Height H r above Ground Level and Wind M

Constant Velocity u


The ground-level concentration is found by setting z = 0:

The ground-level centerline concentrations are found by setting y = z = 0:





Case 13: Plume with Continuous Steady-State Source at Height H r above Ground Level and Wind M

Constant Velocity u

The maximum ground-level concentration along the x axis (C)max is found using

The distance downwind at which the maximum ground-level concentration occurs is found from

The procedure for finding the maximum concentration and the downwind distance is to use Equation

the distance, followed by using Equation 5-52 to determine the maximum concentration.





Case 14: Puff with Instantaneous Point Source at Height H r above Ground Level and a Coordinate Sy

That Moves with the Puff

For this case the center of the puff is found at x = ut. The average concentration is given by

The time dependence is achieved through the dispersion coefficients, because their values

puff moves downwind from the release point. If wind is absent (u = 0), Equation 5-54 does correct result.

At ground level, z = 0, and the concentration is computed using






That Moves with the Puff

The concentration along the ground at the centerline is given at y = z = 0:

The total integrated dose at ground level is found by applying Equation 5-42 to Equation 5-

Pasquill Rederived Equations - Case 15ff h h b l





Ground at the Release Point

For this case the result is obtained using a transformation of coordinates similar to the

transformation used for case 7. The result is

where t is the time since the release of the puff.

Worst-Case Conditions For a plume the highest concentration is always found at the release poin




For a plume the highest concentration is always found at the release poin

If the release occurs above ground level, then the highest concentration

ground is found at a point downwind from the release.

For a puff the maximum concentration is always found at the puff center.

For a release above ground level the puff center will move parallel to the ground, and

the maximum concentration on the ground will occur directly below th

center.

For a puff isopleth, the isopleth is close to circular as it moves downwind

The diameter of the isopleth increases initially as the puff travels downwi

reaches a maximum, and

then decreases in diameter.

If weather conditions are not known or are not specified,

then certain assumptions can be made to result in a worst-case resul

that is, the highest concentration is estimated.

Worst-Case Conditions (Cont…)

The weather conditions in the Pasquill Gifford dispersion equations




The weather conditions in the Pasquill-Gifford dispersion equations

included by means of the dispersion coefficients and the wind spee

By examining the Pasguill-Gifford dispersion equations for estimati

concentrations, it is readily evident that the dispersion coefficients a

speed are in the denominator. Thus the maximum concentration is estimated by selecting the wea

conditions and wind speed that result in the smallest values of the

coefficients and the wind speed.

By inspecting Figures 5-10 through 5-12, we can see that the smal

dispersion coefficients occur with F stability.

Clearly, the wind speed cannot be zero, so a finite value must be se

The EPA7 suggests that F stability can exist with wind speeds as lo

m/s.

Some risk analysts use a wind speed of 2 m/s. The assumptions us

calculation must be clearly stated.

Limitations of Pasquill-Gifford Modeling




Pasquill-Gifford or Gaussian dispersion applies only to neutrally bu

dispersion of gases in which the turbulent mixing is the dominant fe

dispersion.

It is typically valid only for a distance of 0.1-10 km from the release

The concentrations predicted by the Gaussian models are time ave

Thus it is possible for instantaneous local concentrations to exceed

values predicted — this might be important for emergency respons

The models presented assume a 10-minute time average. Actual instantaneous concentrations may vary by as much as a fac

the concentrations computed using Gaussian models.

5-3 Dense Gas Dispersion




A dense gas is defined as any gas whose density is greater than the density of t

through which it is being dispersed.

This result can be due to a gas with a molecular weight greater than that of a

a low temperature resulting from autorefrigeration during release or other pro Following a typical puff release, a cloud having similar vertical and horizontal dim

the source) may form.

The dense cloud slumps toward the ground under the influence of gravity, in

diameter and reducing its height.

Considerable initial dilution occurs because of the gravity-driven intrusion of

the ambient air.

Subsequently the cloud height increases because of further entrainment of a

the vertical and the horizontal interfaces.

After sufficient dilution occurs, normal atmospheric turbulence predominates

gravitational forces and typical Gaussian dispersion characteristics are exhib

Dense Gas Dispersion (Cont…)




The Britter and McQuaid model was developed by performing a dimensional ana

correlating existing data on dense cloud dispersion.

The model is best suited for instantaneous or continuous ground-level releas

gases.

The release is assumed to occur at ambient temperature and without aeroso

droplet formation.

Atmospheric stability was found to have little effect on the results and is not

model.

Most of the data came from dispersion tests in remote rural areas on mostly

the results are not applicable to areas where terrain effects are significant.

The model requires a specification of the initial cloud volume,

the initial plume volume flux, the duration of release, and the initial gas dens

Also required is the wind speed at a height of 10 m, the distance downwind,

gas density.





The first step is to determine whether the dense gas model is appli

The initial cloud buoyancy is defined as

= / (5-5

where

g0 is the initial buoyancy factor (length/time2),

g is the acceleration due to gravity (length/time2

),p0 is the initial density of released material (mass/volume), and

pa is the density of ambient air (mass/volume).





A characteristic source dimension, dependent on the type of releas

be defined. For continuous releases

=

/

(5-60)

where

Dc is the characteristic source dimension for continuous releases of dens

(length),

q0 is the initial plume volume flux for dense gas dispersion (volume/time

u is the wind speed at 10 m elevation (length/time).





For instantaneous releases the characteristic source dimension is def

=

/

(5-61)

where

Di is the characteristic source dimension for instantaneous releases of d

(length) and

V0 is the initial volume of released dense gas material (length).



Britter-McQuaid Dimensional Correlation - P




Figure 5-13 Britte

dimensional corr

dispersion of den

Britter-McQuaid Dimensional Correlation - P




Figure 5-14 Britte

dimensional corr

dispersion of den

Equations Used to Approximate the Curves in tMcQuaid Correlations for Plumes




Table 5-4 Equations Used to Approximate the Curves in the Britte

McQuaid Correlations Provided in Figure 5-13 for Plumes

Equations Used to Approximate the Curves in tMcQuaid Correlations for Puffs




Table 5-5 Equations Used to Approximate Curves in the Britter-McQuaid Correlations Provided in

Equations Used to Approximate the Curves in tMcQuaid Correlations for Puffs (Cont…)



24/11/2014 NUS CN3135 Consequence Modelling & Explosion Lecture 4

Table 5-5 Equations Used to Approximate Curves in the Britter-McQuaid Correlations Provided in

5-4 Dense Gas Transition to Neutrally Buoya




In the dense region, thedownstream

concentration is

computed per Dense Gas

(Section 5-3).

After transition, thecompositions are

computed using neutrally

buoyant gas equations for

ground level releases.

As a dense gas moves downstream the concentration decreases.

At the transition x t it becomes a neutrally buoyant gas



Emergency Response Planning Guidelines




ERPGs provides 3 concentration ranges as a consequence of exposu

substance:

ERPG-1 is the maximum airborne concentration below which it is be

all individuals could be exposed for up to 1 hr without experiencing

than mild transient adverse health effects or perceiving a clearly def

objectionable odour.



irreversible or other serious health effects or symptoms that could imabilities to take protective action.



life-threatening health effects (similar to EEGLs).

Sample Emergency Response Planning Gui




Chemical ERPG-1 ERPG-2 ERPG-3

Ammonia 25 200 1000

Benzene 50 150 1000

Chlorine 1 3 20Dimethylamine 1 100 500

Ethylene oxide NA 50 500

Formaldehyde 1 10 25

Hydrogen sulfide 0.1 30 100

Methanol 200 1,000 5,000

Methyl mercaptan 0.005 25 100

Propylene oxide 50 250 750

Styrene 50 250 1,000

Toluene 50 300 1000

Vinyl acetate 5 75 500

5-7 Release Mitigation




The purpose of the toxic release model is to provide a tool for perfo

mitigation.

Release mitigation is defined as "lessening the risk of a release inc

on the source either

in a preventive way by reducing the likelihood of an event that c

a hazardous vapor cloud or

in a protective way by reducing the magnitude of the release an

exposure of local persons or property.

The release mitigation procedure forms part of the consequence m

procedure

After selection of a release incident, a source model is used to dete

release rate or the total quantity released.

Release Mitigation



24/11/2014NUS CN3135_Consequence Modelling & Explosion Lecture 4

Finally, an effect model is used to estimate the impact of the release

a measure of the consequence.

Risk is composed of both consequence and probability. Thus an es

the consequences of a release provides only half the total risk asse

It is possible that a particular release incident might have high cons

leading to extensive plant mitigation efforts to reduce the conseque

However, if the probability is low, the effort might not be required.

Both the consequence and the probability must be included to asse

Release Mitigation



24/11/2014NUS CN3135_Consequence Modelling & Explosion Lecture 4

Inherent Safety

Inventory reduction

Chemical substitution Process attenuation

Engineering Design

Physical integrity of

seals and construction

Process integrity

Emergency control

Spill containment

Management

Policies and proce

Training for vapor

Audits & inspectio

Equipment testing

Routine maintena

Management of ch

Security