Stat emen t
s part of a study on designing, the site of a new railroad tank car and tank truck ammonia loading facility for a very large refinery/chemical plant complex, hazard analyses indicated that possible consequences of hose breakage or accidental disconnection should be investigated. The hose end couid whip around to discharge liquid ammonia in any direction, or the pipe to which the hose was connected could discharge the stream in any direction.
A
Release Attributes
n Chemical reactions? . . . . Yes
Material: . . . . . . . . . . . . Ammonia Method: . . . . . . . , . . . , . . Hole Fluid state: . . . . . . Flashing liquid
--`,,-`-`,,`,,`,`,,`---
A P I PUBL*<4628 96 = 0732290 0 5 6 0 3 0 6 3 6 3 =
S6-2 Chapter 6
spB 1
= e x p k -
p3 (T+SPC)
where SPA, SPB and SPC are constants and Tis the local temperature. Because the SPC constant is zero here (no curvature, Figure S6-l), taking logarithms of both sides of the preceding equation results in the linear equation:
This was fitted to published data by linear regression with a spreadsheet program. For Equation 2, the slope of the left-hand-side quantity with respect to reciprocal temperature, SPB, can be related to the heat of vaporization by the Clausius-Clapeyron equation (3-2), thus
The value of SPB found by regression was -2795 K. The value of SPA is not input to SLAB because the program calculates it from Equation 1 given sPB and the normal boiling point, T&,.
By Equation 3, the fitted heat of vaporization is 23 MJkgmole (R = 8313 Jkgmole). By subtracting the liquid enthalpy from the vapor enthalpy at 300 K (Fzgures S6-4 and S6-3, respectively), the heat of vaporization is about 20 MJkgmole. This difference is acceptable considering other inaccuracies in the overall modeling; the fitted model was used here (converted to the mass basis).
The liquid density and enthalpy were fitted by regression to quadratic functions of temperature;
these are the continuous lines shown in Figure S6-2 and Figure S6-4. Resultant function- calculated values for 300 K were input to SLAB. Because the modeling programs require constant vapor heat capacities, Cp,r this value was found as the slope of the straight line fitted to the vapor
Vapor Pressure of Ammonia
400 to 200 kelvin
P d n h m I r a i Pert'$ Yrth poy 3-170
0.0 1 2.5C-03 3.OE-03 3.5E-03 4.0E-03 T
4.5L-03 5.OE-03 Reciprocal Temperature. k
0.045
(U 0.04
- !
!? 0.035
a - (U
0.03 0 U
Molal Volume of Liquid Ammonia
Figure S6-1. Figure S6-2.
Copyright American Petroleum Institute
--`,,-`-`,,`,,`,`,,`---
A P I P U B L * 4 6 2 8 96 O732290 05bOL07 2 T 8 W
Scenario 6: Ammonia Hose or Piue Break S6-3
Vapor Enthalpy of Saturated Ammonia
A A
A
A
6 1 . , . , , , , , , , , , , I
200 220 240 260 280 300 320 340 360 380 4 Temperalue. kelvin
Enthalpy of Liquid Ammonia
I
Figure S6-3. Figure S6-4.
enthalpy as shown in Figure S6-3. The mass release rate of 11.67 kg/s NH, (Leung-Epstein correlation for flashing liquid choked flow) and expanded jet diameter of 0.0167 m2 were calculated by the methods of Chapter 3.
A summary of these themod'ynamic parameters used for modeling are shown in Table S6-I for an example SLAB input file. (The program had been modified by the author to carry through the identification information shown in addition to the parameter values and symbols.) Lines starting with an asterisk are comments.
Other Parameters
Horizontal jet releases (1 m above the ground) were simulated because they will result in the highest ground level concentrations in the near and far fields compared with vertical releases.
Because the releases are for an industrial com- plex, zr = 0.35 m is conservatively intermediate between about 1 and O. 1. For simplicity, only one value was used for release rate, jet angle, roughness parameter and atmospheric condi- tions; see RECAP box.
Of the three modeling systems, only SLAB has been designed to cope with horizontal and ver-
MODELING PARAMETER RECAP
Released fluid properties Table 6-1 Orifice diameter, m
initiai jet orientation Initiai jet elevation, m Ammonia release rate, kg/s Roughness length, m Atmospheric stability class Ambient temperature, K
Wind speed, mis Release durations
0.05 horizontal 1 11.7 0.35
D 300 4 Steady state and 10,60,120s Averaging times, s 1 O, 60,600,900
tical jet discharges of pure component vaporhquid aerosols. DEGADIS treats only vertical jets, but the information for VLE with varying air content must be calculated externally. HGSYSTEM'S PLUME for turbulent jets cannot handle aerosols. Therefore, SLAB was used for most of the modeling, followed by vertical release examples for both SLAB and DEGADIS.
Table S6-1 shows the base case (steady state) set of remaining input parameters for SLAB.
Simulation
Figure Só-5 shows the predicted effects of ammonia release rate duration on downwind maximum centerline concentrations for four release durations varying from infinite (steady state) to 10 seconds; the averaging time was held constant at 10 seconds. Note that SLAB reports the
--`,,-`-`,,`,,`,`,,`---
API P U B L a 4 6 2 8 96 W 0732290 05bOLOB I134
S 6-4 Chapter 6
maximum concentration of all passing ‘3eah’’ by using finite duration release cor- rection to steady state con- centrations. At any pariicu- lar downwind distance, the program averages the concen- tration passing by that point with respect to time, noting that dispersion is taking place along the downwind path (ax ). Thus the longer the distance from the source, the more dispersion along the x-axis occurs. For example, if the release duration is 60 s and the wind speed is 4 d s , a
“puff’ Will travel 240 m downwind in the 60 seconds.
Beyond the 240 m point (not shown on the graph), a pulse will be observed at further distances; the pulse broadens with respect to downwind distance as that distance in- creases. Thus, the shorter the initiai, square-wave pulse (du- ration of release), the less material will be available for dispersion, so the muximum concentrations will decrease with release duration as shown in this Figure. Figure S6-6 shows the corresponding travel times for peak maxima vs downwind distance.
Figure S6-7 illustrates the
Table S6-1.
BASE CASE INWT FOR SLAB
SLAB Run: Scen. 6. SS, D, 4 m/s, 300K, Tav= los, Ht.= I m F i Le name: ASSGOlL-SLI
*
* Miscellaneous
(Data c o l m i s width = 12.)
*
2 i d s p l = s p i l l type (HORIZONTAL JET = 2) 1 ncalc = sub-step m u l t i p l i e r
* * Release Gas Properties
*
0.01703 ums = molecular weight of source 239.75 tbp = b o i l i n g p o i n t temperature 0.7426 cmed0 = Liquid mass f r a c t i o n
1413.4 cps = vapor heat capacity, Cp, (J/kg-K) 1.3682E+06 dhe = heat o f vaporization (J/kg)
4310.4 cpsl = Liquid heat capacity (J/kg-K) rhos1 = l i q u i d source d e n s i t y ( k g / d ) 2794.69 spb = s a t u r a t i o n pressure constant (K)
= s a t u r a t i o n pressure constant ( K I 239.75 t s = temperature o f source gas ( K I
600.2 0.0 spc
* S p i l l Characteristics
*
*
11.67 qs = mass source r a t e (kg/s)
0.0167 as source area (m2) - EXPANDED JET 20000.0 t s d = continuous source d u r a t i o n i s )
0.0 q t i s = instantaneous source mass (kg) 1.0 hs = source h e i g h t (m)
*
* F i e l d Parameters
*
10.0 t a v = concentration averaging time (SI 10000.0 xffm = maximum downwind distance (m)
0.0 zp(1) = concentration measurement h e i g h t (m) 1.00 zp(2) = concentration measurement height (m) 0.0 zp(3) = concentration measurement h e i g h t (m) 0.0 zp(4) = concentration measurement h e i g h t (m)
*
* Boundary Layer & Meteorological Conditions
*
0.35 z0
10.0 za = ambient measurement height (m) 4.00 ua
50. r h = r e l a t i v e h u n i d i t y (percent)
= surface roughness height (m)
= ambient w i n d speed (m/s) 300.0 t a = ambient temperature ( K I
4.0 stab = atmospheric s t a b i l i t y class value 0.0 a l a = inverse Monin-Obukhov length ( l / m )
O END OF INPUT PARAMETERS
effect of various averaging times for a constant ammonia release duration of 60 s. Here, as the averaging time increases, an observer at a downwind location will “see” a passing peak, and if the averaging time is long enough, many very low (and/or zero) concentrations will contribute to the averaged concentration, thus effecting lower average values the longer the averaging time.
Finally, SLAB and DEGADIS were run to simulate the ammonia release for a vertical jet at the same release height (1 m) for the time conditions shown in F i p r e S6-8. A detailed comparison of the
SLAB curve with the corresponding horizontal release curve in Figure S6-1 showed that the horizontal case gave higher concentrations over the whole distance shown. The DEGADIS curve
Copyright American Petroleum Institute
--`,,-`-`,,`,,`,`,,`---
A P I P U B L a 4 6 2 8 96 0732290 0560307 070 =
Scenario 6: Ammonia Hose or Pipe Break S6-5
differs from the SLAB curve by a factor of about 2. The relative effects of release durations and averaging times would be the same as for the horizontal jet cases. As noted in Appendix II,
DEGADIS does not provide the facility to calculate averaging time effects with respect to downwind distance; this must be calculated externally by the user. Because the DEGADIS results have not been corrected for averaging times at downwind distances, the separation between the two curves in Figure S6-8 is confounded with averaging times and differences.
(Figures S6-5 through S6-8 follow.)
--`,,-`-`,,`,,`,`,,`---
S6-6 Scenario 6: Ammonia Hose or Pipe
O
. O o
v) 10
t i;
g o o
,a- €
U
z C
<
E
m a
-J o 1/)
L 0 0 0 0 0 0 0 0 0
ù O 2 8 % 8 E Z C : O
>
t
. a - , - . - . -
c
.- O c z
c C al U
O O
E
al
- S
n
n O
O n n O O O
*1
r c c c
wdd 'uo!+o~+ua3uo3 i"
Copyright American Petroleum Institute
--`,,-`-`,,`,,`,`,,`---
A P I P U B L * 4 6 2 8 9 b W 0 7 3 2 2 9 0 O S b O L L L 7 2 9 D