The work shown here was performed using the CFD package called Fire Dynamics Simulator (FDS). FDS is an open source, free CFD package built and maintained by the National Institute of Standards and Technology (NIST). FDS, available here is similar to the capabilities of commercial CFD codes such as ANSYS Fluent.
FDS has capabilities for modeling fluid flow, heat transfer, pyrolysis, and combusion. It is best suited for simulating physical systems with simple geometries, since the physical model used in the simulations has to conform to an underlying rectilinear grid. For most of the work I did in FDS, I was focusing on transport phenomena and the physics of the underlying problem, rather than complex geometries.
For the design of any sort of guidevanes, blades, or devices such as turbines, ANSYS Fluent would likely be a better match for handling these more complicated geometries.
For more complicated geometries, I would use CAD software such as SolidWorks for building or modifying the solid models, and then import the solid model into a CFD program.
The objective was to model the fluid flow, heat transfer, and combustion reactions inside a combustion chamber, and to look at the effect of injected turbulent air inside a combustion chamber of a household cookstove on the emissions of carbon monoxide and soot.
Included below are the following items:
I used Fire Dynamics Simulator (FDS) as the CFD package. The summary statistics and time-series plots were performed using Python.
Whereas Fluent and other commercial CFD packages often include a graphical user interface (GUI) to make using the program more user-friendly, for FDS the user writes an input text file that is executed via the command line.
ig6.fds
&HEAD CHID='ig6', TITLE='Inside Stove; FDS version 6.0.0' /
&DUMP RENDER_FILE='ig6' /
&MISC DNS=.TRUE. /
&TIME T_END=480. /
# MESH
&MESH IJK=240,1,240, XB=0.09,0.21,0.12,0.1205,0.03,0.15 / grid 0.5 mm
&VENT MB='ZMAX', SURF_ID='OPEN', RGB=135,206,235, OUTLINE=.TRUE. /
# REACTIONS
# C2H6O + 1.885 (O2 + 3.76 N2) -> 1.77 CO + 3 H2O + .23 C + 7.0876 N2
# CO + 1/2 O2 -> CO2, CO consumption (oxidation), exothermic rxn
# CO2 -> CO + 1/2 O2, CO production, endothermic rxn (reverse reaction)
# Reaction rates from Westbrook & Dryer, 1981, Units: cm s mol kcal K
# Ethanol Fuel: k = 3.6e12*exp(-30/(RT))[C2H6O]^0.15[O2]^1.6
# CO Oxidation: k = 4e14*exp(-40/(RT))[CO]^1[H2O(g)]^0.5[O2]^0.25
# CO Production: k = 5e8*exp(-40/(RT))[CO2]^1
&SPEC ID = 'NITROGEN' /
&SPEC ID = 'OXYGEN' /
&SPEC ID = 'WATER VAPOR' /
&SPEC ID = 'ETHANOL' /
&SPEC ID = 'CARBON MONOXIDE' /
&SPEC ID = 'CARBON DIOXIDE' /
&SPEC ID = 'SOOT', AEROSOL=.TRUE. /
&SPEC ID = 'AIR', SPEC_ID ='NITROGEN','OXYGEN', VOLUME_FRACTION = 3.76,1.,
BACKGROUND = .TRUE. /
&REAC ID = 'Ethanol Fuel', FUEL='ETHANOL',
SPEC_ID_NU='ETHANOL','AIR','CARBON MONOXIDE','WATER VAPOR','SOOT','NITROGEN',
NU=-1,-1.885,1.77,3,0.23,7.0876, CRITICAL_FLAME_TEMPERATURE = 425. /
&REAC ID = 'CO Oxidation', FUEL='CARBON MONOXIDE', A=4e14, E=167360,
SPEC_ID_NU='CARBON MONOXIDE','OXYGEN','CARBON DIOXIDE', NU=-1,-0.5,1,
SPEC_ID_N_S='CARBON MONOXIDE','WATER VAPOR','OXYGEN', N_S=1,0.5,0.25 / E in J/mol
&REAC ID = 'CO Production', FUEL='CARBON DIOXIDE', A=5e8, E=167360,
SPEC_ID_NU='CARBON DIOXIDE','CARBON MONOXIDE','OXYGEN', NU=-1,1,0.5,
SPEC_ID_N_S='CARBON DIOXIDE', N_S=1 /
# MATERIAL PROPERTIES
&MATL ID='steel',
CONDUCTIVITY=54.0, DENSITY=7833., EMISSIVITY=0.9, SPECIFIC_HEAT=0.465 /
&SURF ID='fuel', MASS_FLUX=0.012, SPEC_ID='ETHANOL', RGB=255,69,0 / mdot" in kg/m^2/s
&SURF ID='wall', MATL_ID='steel', THICKNESS=0.001, BACKING='EXPOSED' /
&SURF ID='Igniter', COLOR='RED', TMP_FRONT=1500., RAMP_T='T_igniter' /
&RAMP ID='T_igniter', T=0.0, F=0.0 /
&RAMP ID='T_igniter', T=1.0, F=1.0 /
&RAMP ID='T_igniter', T=2.5, F=1.0 /
&RAMP ID='T_igniter', T=3.0, F=0.0 /
# IGNITER AND FUEL
&DEVC XYZ=0.155,0.12,0.09, ID='Ignition timer', QUANTITY='TIME', SETPOINT=3.0,
INITIAL_STATE=.TRUE./ Remove OBST after 3 sec when starting with it ON
&OBST XB=0.1500,0.1520,0.12,0.1205,0.0895,0.0905, SURF_ID='Igniter',
DEVC_ID='Ignition timer', BNDF_OBST=.FALSE. /
&OBST XB=0.14,0.16,0.12,0.1205,0.055,0.0555, SURF_IDS='fuel','INERT','INERT',
BNDF_OBST=.FALSE. / As = 0.0004 m^2, SURF_IDS=top, sides, bottom
# AIR INLET THROUGH BOTTOM
&SURF ID='vent air flow', VEL=-0.15, MASS_FRACTION=1.0, SPEC_ID='AIR' / m/s
negative means into the domain
&OBST XB=0.09,0.21,0.12,0.1205,0.03,0.0305, SURF_ID='wall' / bottom
&VENT XB=0.09,0.21,0.12,0.1205,0.0305,0.0305, SURF_ID='vent air flow',
RGB=135,206,235, OUTLINE=.TRUE. /
# OUTPUT
&SLCF PBY = 0.12, QUANTITY = 'TEMPERATURE' /
&SLCF PBY = 0.12, QUANTITY = 'VELOCITY', VECTOR=.TRUE. /
&SLCF PBY = 0.12, QUANTITY = 'VOLUME FRACTION', SPEC_ID='ETHANOL' /
&SLCF PBY = 0.12, QUANTITY = 'VOLUME FRACTION', SPEC_ID='CARBON MONOXIDE' /
&SLCF PBY = 0.12, QUANTITY = 'AEROSOL VOLUME FRACTION', SPEC_ID='SOOT' /
&DEVC ID='v_top', XB=0.09,0.21,0.12,0.1205,0.145,0.150, QUANTITY='VELOCITY',
STATISTICS='MEAN' /
&DEVC ID='Xco_top', XB=0.09,0.21,0.12,0.1205,0.145,0.150,
QUANTITY='VOLUME FRACTION', SPEC_ID='CARBON MONOXIDE', STATISTICS='MEAN' /
&DEVC ID='Xs_top', XB=0.09,0.21,0.12,0.1205,0.145,0.150,
QUANTITY='AEROSOL VOLUME FRACTION', SPEC_ID='SOOT', STATISTICS='MEAN' /
&TAIL /
In [1]:
%pylab inline
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
In [2]:
# Import CSV output file from the FDS simulation
fds = pd.read_csv('images/ig6_devc.csv', skiprows = 1)
fdsdf = pd.DataFrame(data=fds)
fdsdf.describe() # units in s, degC, m/s, mol/mol, kW, kg/s
hrr = fdsdf.HRR.tail(30).mean()*1000 # convert kW to Watts
burn_rate = fdsdf.BURN_RATE.tail(30).mean()*1000*60 # convert kg to grams, seconds to minutes
print(len(fdsdf.Xco_top))
print('Average exiting velocity: %2.1f m/s' % fdsdf.v_top.tail(30).mean())
print('Average exiting carbon monoxide mole fraction: %2.4f mol/mol' % fdsdf.Xco_top.tail(30).mean())
print('Average heat release rate: %2.2f W' % hrr)
print('Average burn rate: %2.4f g/min' % burn_rate)
# Plot the output
f, ax = plt.subplots()
f.set_figheight(6.2)
f.set_figwidth(10.0)
sz_axes = 16
sz_labels = 20
sz_title = 22
plt.rc('xtick', labelsize = sz_axes); plt.rc('ytick', labelsize = sz_axes)
x = fdsdf.Time
y0 = fdsdf.v_top
y1 = fdsdf.Xco_top
y2 = fdsdf.Xs_top
ax.plot(x, y1, 'r-', label = 'X$_{CO}$', linewidth = 1)
ax.axhline(y = y1.tail(30).mean(), linestyle = '--', color = 'r', label = 'X$_{CO} \, \mathrm{(avg)}$', linewidth = 1)
# ax.plot(x, y2, 'ko-', label = 'X$_{s}$', linewidth = 2)
# ax.plot(x, y0, 'b-', label = 'vel [m/s]', linewidth = 1)
ax.legend(loc = 0, fontsize = sz_axes)
ax.set_xlabel('$t \: \mathrm{(seconds)}$', fontsize = sz_labels)
ax.set_ylabel('$X \: (\mathrm{mol/mol})$', fontsize = sz_labels)
ax.set_title('$\mathrm{Mole \, fraction \, of \, carbon \, monoxide \, exiting \, out \, of \, the \, top}$', fontsize = sz_title)
show()
Air flow enters from the bottom and exits at the top. The fuel is injected at the surface represented by the white line in the bottom quarter of the domain.
The first image is the velocity and the second image is the mole fraction of carbon monoxide.
The CFD code and input file capture turbulence and finite rate chemical kinetics, including the reactions for both the production and oxidation of carbon monoxide. So even though the geometry is simple, the simulation can predict the reduction of CO due to increased mixing.
Included additional code:
ig6_pipe2a.fds
...
&SPEC ID = 'PIPE AIR', SPEC_ID ='NITROGEN','OXYGEN', VOLUME_FRACTION = 3.76,1. /
&MATL ID='steel',
CONDUCTIVITY=54.0, DENSITY=7833., EMISSIVITY=0.9, SPECIFIC_HEAT=0.465 /
&SURF ID='pipe air flow', MASS_FLUX=7.5, SPEC_ID='PIPE AIR', RGB=0,0,128 /
mdot" in kg/m^2/s, for Re = 4000
&SURF ID='pipe', MATL_ID='steel', THICKNESS=0.0010, RGB=128,128,128,
BACKING='EXPOSED' /
&OBST XB=0.20,0.22,0.12,0.1205,0.09,0.10, SURF_ID='pipe' / D = 1.0cm
&VENT XB=0.20,0.20,0.12,0.1205,0.09,0.10, SURF_ID='pipe air flow', COLOR='BLUE' /
...
&TAIL /This simulation is the same as before except air was injected through a pipe into the combustion chamber.
The first animated image series below shows the velocity in the combustion chamber, and the resulting added turbulence. The second shows the mole fraction of CO, and how, due to increased mixing of air with the combustion emissions, there is a reduction of the concentration of CO leaving the combustion chamber.
The third animated image shows the mole fraction of CO superimposed on the velocity profile. Much less emissions escape beyond the injected air flow and exit out of the top of the combustion chamber.
The animated gifs are displaying the series of images at a rate of 2 frames per second.
In [3]:
# Import CSV output file from the FDS simulation
fds = pd.read_csv('images/ig6_pipe2a_devc.csv', skiprows = 1)
fdsdf = pd.DataFrame(data=fds)
fdsdf.describe() # units in s, degC, m/s, mol/mol
hrr = fdsdf.HRR.tail(30).mean()*1000 # convert kW to Watts
burn_rate = fdsdf.BURN_RATE.tail(30).mean()*1000*60 # convert kg to grams, seconds to minutes
print(len(fdsdf.Xco_top))
print('Average exiting velocity: %2.1f m/s' % fdsdf.v_top.tail(30).mean())
print('Average exiting carbon monoxide mole fraction: %2.4f mol/mol' % fdsdf.Xco_top.tail(30).mean())
print('Average heat release rate: %2.2f W' % hrr)
print('Average burn rate: %2.4f g/min' % burn_rate)
# Plot the output
f, ax = plt.subplots()
f.set_figheight(6.2)
f.set_figwidth(10.0)
sz_axes = 16
sz_labels = 20
sz_title = 22
plt.rc('xtick', labelsize = sz_axes); plt.rc('ytick', labelsize = sz_axes)
x = fdsdf.Time
y0 = fdsdf.v_top
y1 = fdsdf.Xco_top
y2 = fdsdf.Xs_top
ax.plot(x, y1, 'r-', label = 'X$_{CO}$', linewidth = 1)
ax.axhline(y = y1.tail(30).mean(), linestyle = '--', color = 'r', label = 'X$_{CO} \, \mathrm{(avg)}$', linewidth = 1)
# ax.plot(x, y2, 'ko-', label = 'X$_{s}$', linewidth = 2)
# ax.plot(x, y0, 'b-', label = 'vel [m/s]', linewidth = 1)
ax.legend(loc = 0, fontsize = sz_axes)
ax.set_xlabel('$t \: \mathrm{(seconds)}$', fontsize = sz_labels)
ax.set_ylabel('$X \: (\mathrm{mol/mol})$', fontsize = sz_labels)
ax.set_title('$\mathrm{Mole \, fraction \, of \, carbon \, monoxide \, exiting \, out \, of \, the \, top}$', fontsize = sz_title)
show()
The added turbulent kinetic energy and air coming from the pipe reduces the exiting CO emissions (mole fraction) by an order of magnitude! Both cases maintain the same heat release rate (1.8 W) and fuel burn rate (0.007 g/min), so the reduction in emissions is due to turbulent mixing.
For the design process of a combustion chamber given specific design constraints, combustion chamber geometry, etc., and design goals, simulations can be performed to determine optimal parameters such as the the diameter of the pipe, the location (height) of the pipe, and the velocity of air exiting the pipe.
This visualization shows velocity vectors for the flow field through an opening in a chamber. The plane shown is a vertical slice through the center of the 3-D chamber at the location of the combustion source. Notice the transient turbulent eddies forming as the buoyant-driven flow of CO, CO$_2$, and water vapor expands toward the chamber ceiling (the chamber outline is not shown in the animated image).
Most of the results for this project were focused on plotting quantitative output from the CFD simulations (simulated sensors predicting quantities at point sources). Quantities such as concentrations of emissions of carbon monoxide were plotted and compared at different spatial locations within a room (or chamber), which varied and were heavily dependent on ventilation conditions.
In [ ]: