Note: Descriptions are shown in the official language in which they were submitted.
~Z9283
BACKGROUND AND OBJECTS
The time and space available for fuel air mixing in
an internal combustion engine is limited, and "Homogeneous
charge" engines do not burn really homogeneous mixtures.
The incompleteness of mixing, and the unsteadiness of air
fuel ratio delivery to the cylinders under transient response,
degrade engine performance significantly from what would be
possible with quicker transient response and more homogeneous
mixing. Because the physics of mixing processes is compli-
cated, and because mixing states are difficult to measure
experimentally, the great importance of mixing in engines
is not widely understood. It i5 the purpose of the present
invention to use structured turbulent flows to mix fuel and
air in an organized way involving much higher mixing rates
than have previously been possible. The device is character-
ized by excellent transient response and much more homogeneous
fuel air mixtures than have previously been practically availa-
ble. To understand the technical problems which the current
invention has'solved, a discussion of conventional engine
mixing processes is appropriate.
It is characteristic of present day carburetor and
intake manifold systems that the fuel which passes the car-
buretor throttle is rapidly separated from the air and
deposited downstream of the carburetor at the first turn.
At this first turn is generally located an exhaust-heated
hot spot. However, only a part of the fuel can evaporate
on this hot spot surface. The rest of the fuel is not
, - 2 ~ 9~3
evaporated at the hot spot and ~eposits on manifold walls
downstream. Once this fuel is deposited, it proceeds to
the individual cylinders relatively slowly and somewhat
unevenly.
Intake manifolds are carefully designed with contours
on the manifold floor to try to distribute the liquid fuel
evenly between cylinders. Even so, it is usually impossible
to get very tight cylinder-to-cylinder distribution over all
of the relevant speeds, under steady-state conditions of
engine speed and load. Fue~-air proportioning is worse under
transient~conditions. The air velocity in manifold passages
can be a substantial fraction o-f the speed of sound, but
the fuel liquid film velocity i~s generally less than a tenth
of the air velocity. Consequently, if an element of fuel
and an element of air both leave the carburetor throttle at
the same time, the fuel takes much longer to reach the cylin-
ders than the air. On accelerations the mixture delivered
to the cylinders therefore tends to shift lean. Acceleration
enrichment arrangements must therefore be employed. The
greater the acceleration enrichment, the greater the emis-
sion penalty involved.
The two-phase flow situation in a conventional intake
manifold is quite involved. It is practically impossible to
get the transient characteristics and cylinder-to-cylinder
distribution characteristics that are desirable, even with
very laborious development work in each intake manifold
design because of the two phase flow relationships. The
inability of conventional carburetor-intake manifold systems
to perform well with respect to cylinder-to-cylinder and
transient response characteristics has been the main
~ ~ 3 ~ 1~2'~283
motivation for the development of very expensive multi-
cylinder fuel injection sy3~ems.
The need for fast transient response and tight
cylinder-to-cylinder distribution becomes greater as emis-
sion specifications become more stringent and as fuel con-
sumption becomes a more and more important issue. In engines
which employ a three-way catalyst system to control emissions
the "window" of satisfactory operation is of the order of +
or - .1 air/fuel ratio. With the 3-way catalyst system the
penalty for slow transient response and inadequate cylinder-
to-cylinder distribution can be drastic increases in nitric
oxide production. Moreover, with emission control hardware,
fuel and air are no longer the only two fluids to be mixed;
in addition, it may be necessary to secure even distribution
of exhaust gas recirculation from cylinder-to-cylinder.
Another approach to emlsslon control is lean and
dilute ccmbustion. Operation with very lean (or EGR dilute)
mixtures results in very low NOx emissions, and is advan-
tageous from the point of view of the thermodynamic cycle.
If fast and consistent combustionof lean or dilute mixtures
is possible, significant improvements in fuel economy are
achievable simultaneously with excellent NOx control. It
will be shown in the detailed discussion that the level of
enleanment or dilution permissible with good combustion and
good drivability is very sensitive to the details of the
mixing state of the fuel air EGR mixture. As cylinder-to-
cylinder and microscale mixing statistics become tighter,
leaner and more dilute mixtures can be efficiently burned~
Therefore, the excellent mixing of the present invention
~Z9Z83
widens the air fuel ratio limi-ts of satisfactory engine
c~stion and permits significant improvement in emissions
and fuel economy with dilute mixture~. Experi-
mental data with the mixing vortex have been obtained which
indicate that it will be possible to achieve the required
NOx control with much improved fuel consumption with this
lean combustion approach, without any necessity for 3-way
catalysis.
The mixing state inside the cylinders and the cylinder-
to-cylinder variation delivered to the engjne is controllable
by the state of mixing upstream of the manifold runners them-
selves. For some time it has been known that an intake mani-
fold which receives a homogeneous mixture of fuel and air
will distribute a homogeneous air fuel mixture to its indi-
vidual cylinders. This is reasonable, since the flow of
mixed gases through a passage cannot well be expected to unmix
the gases. Condensation rates of fuel from a homogeneous
vaporized mixture are generally quite low, even if the manifold
is below the equilibrium air distillation EAD, temperature.
Also the intake manifold passages quickly warm above the con-
densation temperature of the fuel-air mixture under normal
engine operating conditions. Therefore, if the fuel from
the carburetor or other fuel-air metering system can be homo-
geneously mixed with the air from the carburetor (and with
EGR, Exhaust Gas Recirculation) prior to delivery to the intake
manifold, design of the intake manifold can be very much
simplified.
Designing a manifold for low flow resistance and good
mass-flow distribution from cylinder-to-cylinder is a much
~ 5 ~ 1 1 Z g 28 3
easier problem if the manifold only handles pre-mixed vapors.
The difficulty of desi~ning manifolds presently comes because
they are asked to be at once mixing devices, evaporators, and
flow channels for both liquid fuel on walls and for the much
higher velocity air stream in the flow channels.
It is therefore desirable to design a system where
the fuel is evaporated and homogeneously mixed with the air
and with any EGR prior to introduction of the mixture into the
intake manifold per se. Advances in fluid mechanical know-
ledge based on the fluid mechanical field of "fluidics," and
conceptua~ advances with respect to the interaction between
flow structure, turbulence, and mixing largely made by R.
Showalter, have made design of such a mixing system on the
basis of calculable physical eEfects possible. To understand
how this can be so, it is necessary to describe in a little
detail the physical processes which must occur in order to
evaporate fuel into the air and homogeneously mix the fuel
with the air.
First of all, it is useful to consider the process of
fuel evaporation. Research has shown that, at least so far,
the only practical way to achieve droplet sizes sufficiently
small so as to be in stable aerosol suspension under the
turbulent flow conditions in an intake manifold (below about
1 micron) is to have these droplets form by condensation
from the vapor. Complete mixing requires either complete
evaporation or stable aerosol droplets. Even if the mixture
out of the mixer involves stable liquid droplets,rather than
completely evaporated fuel, the fuel must have been origi-
nally evaporated and then recondensed if it is to be in
stable aerosol sized droplets.
- 6 - ~1292~3
Fuel vapor will evaporate into a surrounding gas
only if the vapor pressure of the fuel at the liquid surface
exceeds the partial pressure of the fuel vapor in the ad-
jacent mixture. It is important to visualize the scale on
which the evaporation process happens. ~t the liquid inter-
face, molecules of fuel are evaporating and condensing con-
tinuously, and the number of molecules is sufficiently large
that the statistical process is governed by exact physical
laws. The volumes of vapor adjacent the liquid surface
determine whether the liquid will evaporate or not. These
vapor volumes are tiny, and are of the order of relatively
few mean-free paths on a side (volumes of the order of cubic
microns). Because vapor pressures in volumes so small,are
determinants of the evaporation process, evaporation is very
closely coupled with mixing processes. Unless the rate at
which fuel vapor leaves the vicinity of the liquid surface
and diffuses into the air-EGR mixture is large, evaporation
rates will be slow, even though the mixture is on gross
volumes well away from saturation conditions. The elements
of gas relevant to the evaporation process can be saturated,
at the same time that the total volume inside the mixer is
far away from saturation.
Evaporation is produced by the interaction of mixing
at all scales, the gross vapor pressure of fuel in the air,
and the vapor pressure of the liquid fuel surfaces. The
vapor pressure of the liquid fuel is clearly related to heat
transfer issues. To evaporate the fuel requires a supply of
the fuel's heat of vaporization. Heat transfer also matters
because liquid vapor pressure is a strongly increasing func-
tion of the liquid temperature. For any given mixing
~ 7 ~ 112~83
situation, evaporation rates will be proportional to the
difference between liquid surface vapor pressure and the
average vapor pressure in the mixing channel. Therefore,
evaporation rates will increase very strongly as liquid
surface temperature increases
However, in an engine, there are strong practical
reasons for wishing to minimize the amount of heat added
to the mixture. Fir~t of all, low mixture temperatures help
control detonation and maximize engine power. Secondly, hot
surfaces tend to form undesirable deposits. The more rapid
the mixing within the mixing section, the less heat or
internal energy need be supplied to evaporate and mix the
fuel (although the heat of vaporization must be suppplied in
any case) and the cooler the surfaces and mixtures can be.
For all these reasons, mixing matters in the evaporation pro-
cess. Fuel evaporation and fuel-air mixing are not the same
thing, but the processes are coupled.~
Mixing is a colnplex process, and is the interaction
in space and time between molecular diffusion, turbulent
diffusion, and the non-random convection of fluid elements
along mean-flow stream lines. Most engineers are taught to
feel that when a flow becomes significantly turbulent, it is
no longer reasonable to attempt (even conceptually) to model
the process in its details. It happens that this view, although
widespread, is also wrong. Flow in many passages of practical
interest can be characterized in terms of mean flow streamlines.
These streamlines define flow patterns which occur because of
basic inertial and pressure equalization physical laws. Tur-
bulence is superimposed on these mean flow streamlines, but
~ 8 ~ 1~29283
the fraction of the kinetic energy in the flow which is in the
form of the mean-flow pattern is often much larger than the
fraction of the flow energy which is in the form of the very
much smaller scale turbulent fluctuations. For certain prac-
tical sections, these mean-flow streamline patterns, or flow
structures, are stable and reasonably predictable over quite
wide ranges of massflow (and identi.cally wide ranges of
Reynolds number). These mean flow patterns can fold mixants
together in an organized way, and greatly increase mixing rates
over what would be possible with turbulence alone.
Mixing is the interaction of molecular diffusion (a
random-walk process), turbulent diffusion (which is modelled
as a random-walk process), and mean flow patterns (which
involve non-random flow processes). It should be mathemati-
cally apparent that a random-walk process is much less effi-
cient for traveling between two specified points than the
properly chosen non-random process. Fundamentally for this
reason, the properly chosen flow patterns can very greatly
increase the rate of mixing in a mixing section, by doing
the large-scale stretching and folding of the fluid elements
in a non-random way. For example, the flow structure can
systematically convect the least saturated air to the liquid
surface for evaporation. The flow pattern can also systema-
tically stretch out interfacial areas within the flow,
increasing mixing area and reducing the mean distance over
which the random-walk processes of turbulent and molecular
diffusion must act. It turns out that properly chosen flow
structures can in this way increase mixing rate~s by very
large factors over mixing rates which would occur if the flow
~ 9 1129~83
relations were purely random. Since there is little time
and space available in an engine for fuel-air mixing, the
more rapid mixing possible with structured turbulent flow
is practically important.
Much background in yroducing desired flow structures
has been worked out by the field of "fluidics,"which uses
flow relations to produce information-handling devices that
operate because there are stable Eluid flow modes in the
flow elements of the fluid circuits. The physical laws
which are most useful for producing predictable flow modes
for flow structures are conservation of linear momentum,
Bernoulli's equation which establishes well-defined rela-
tions between fluid velocities and fluid pressures, and the
wall attachment effect called "Coanda" effect. From con-
servation of linear momentum, conservation of angular momen-
tum can be derived. It should be noted that the fluid
mechanical laws, which fluidics teaches one to manipulate,
continue to be valid when a flow becomes turbulent, therefore
fluidics is very useful in permitting one to produce and
understand flow structures under turbulent conditions.
The present invention uses fluidic fluid mechanics
and mixing and heat transfer theory in the following way.
Spark-fired engines are throttled, and the pressure drop
across the throttle accelerates the flow in a near isentropic
expansion, so that the flow velocity just downstream of the
throttle is very often a significant fraction of sonic
velocity (or sonic velocity itself). Just downstream of
the throttle plate, the flow work across the throttle is
- ' ~ 83
stored in the fluid elements in the form of kinetic energy.
These fluid elements have very siqnificant linear momentum
per unit mass. This flow momentum, properly utilized, is
more than sufficient to produce a very strongly structured
flow pattern downstream o~ the throttle plate. In current
engines, the flow energy and momentum available just down-
stream of the air throttle is dissipated into turbulence and
into unstable vortices. However, correctly designed deflec-
tors can deflect this flow so that a high fraction of the
isentropic velocity past the throttle plate is delivered in
coherent form at high velocity into a channel. In the present
invention, the outlet of this mixing channel is off-center
with respect to the inlet point of the mixing channel by a
distance R. If the flow velocity from the inlet point is
resolved into vector components with respect to radial line R,
including a vector component parallel to R, Vr, and a velocity
vector component perpendicular to the line R, Vt, (or velo-
city tangential), then the fluid introduced into the channel
will have angular momentum with respect to the center of the
outlet. (Angular momentum is defined as ~IVtR). In the present
invention, the channel peripheral walls are roughly concentric
with respect to the mixer outlet The flow from the deflectors
from the throttle plate will progress until it interacts with
the peripheral walls of the channel and it will lose some of
its velocity by drag interactions with respect to this wall.
However, if the passages are properly shaped, much of the
momentum of the fluid will remain. Conservation of angular
momentum is one of the most basic of physical laws. Therefore,
it is relatively easy to design a mixture channel where the
flow velocity from the de1ectors is formed into a stable
vortex flow pattern where the mean flow structure of the
flow pattern is dominated by the physical relations of con-
servation of angular momentum. This flow characteristic will
be discussed in more detail in the Detailed Description.
However, it should be noted here that the vortex flow so
established can be made to be a flow structure which is stable
over a very wide range of Reynolds numbers (wide enough to
cover the entire phase space of engine operation), and that
the flow structure is one with very great advantages with
respect to the mixing and evaporation functions which need
to be served. First of all, the vortex flow pattern will
serve strongly as a separator of the liquid fuel from the
channel air flow, so that the fuel will deposit on the peri-
pheral wall of the vortex chamber in a manner which should
be easy to understand for those who understand cyclone scrubbers:
The centrifugal forces in the vortex at the outside wall will
be in the range of hundreds or thousands of G's. If this out~
side peripheral wall is heated, a very good heat transfer con-
tact with the liquld fuel to be evaporated can be established.
Secondly, the flow pattern is one wherein the air which is
least saturated with fuel is the air which will be thrown to
the outside of the vortex, so that the fuel is constantly
- exposed to the air which has the lowest vapor pressure of fuel
in the chamber. Once the fuel is evaporated, the flow rela-
tions of the vortex (which is a turbulent flow with a pro-
nounced mean flow streamline pattern) efficiently completes
the mixing process. By the time an element of fuel air mix-
ture reaches the central outlet of the vortex chamber it is
_ 1~2 ~ 1~ 83
substantially homogeneous all the way down to mean-free-path
scales .
The operation of the vortex device is somewhat more
advantageous than might at first appear. Fox example, under
cold start conditions, the function of the vortex as a centri-
fugal separator of liquid droplets is most useful. During the
start-up only vapor leaves the vortex. If a rich mixture is
delivered from the carburetor, an equilibrium splash-cloud
of droplets around the outside vortex wall is quickly estab-
lished~ This splash-cloud has a very high surface area of
liquid and very rapidly evaporates the light ends of the fuel
into the air. The vortex at this time functions as an approxi-
mation of an equilibrium air distillation still and makes it
possible to start the engine on a relatively lean mixture
even during choke periods. Since the air fuel ratio coming
out of the cold vortex can be much leaner than the mixture
delivered to it, high CO and ~C emissions are not necessary
during the start-up process. For the same reasonr the fuel
penalties of cold start mixture enrichment need not be experi-
enced since the cylinders need never see an over-rich mixture.
The mixer can be designed to warm up very quickly, and
heating of the outside walls cf the vortex channel is very
advantageous because it permits the liquid surface to be
evaporated to be at a much higher temperature than the mean
temperature of the mixture in the vortex, so that the diffusion
gradients driving the evaporation process can be made very
large. In this way the peripheral walls can be operated so
that evaporation rates are so fast that only a relatively
~ - 13 - 11Z9~83
small fraction of the periyheral wall needs to be wet at
any time. This means that the mass of liquid fuel in the mixer
at any time can be made very small, and therefore the time
lag between fuel transport through the vortex and air transport
through the vortex can be held to a very small value.
Under very low intake manifold vacuum (very high power
demand) the operation of the vortex is beneficial, too. Under
these conditions the liquid fuel evaporates and rapidly re-
condenses in the form of droplets well below one micron, so
that the vortex functions as a smoke generator and mixer and
the volumetric efficiency of the engine is increased because
of the reduction in the mixture temperature delivered.
Experiment has shown that the present invention works
well as an evaporator and mixer~ Certain points are important
with respect to its significance. First of all, it has been
shown that the mixing device will produce cylinder-to-cylinder
air fuel distribution which is so tight that cylinder-to-
cylinder variation cannot be conveniently measured. This is
fuel-air distribution much tighter than that attainable with
fuel injection systems. Secondly, the mixing system has very
rapid transient responses and the rapid transient responses
very much reduce the necessary trade-off between low emissions
and drivability. The microscale homogeneity from the system
widens the lean limits of engine operation, making improvements
in both efficiency and emissions possible. Also, the mixing
device evaporates efficiently enough that it will tolerate
gasolines having end boiling points significantly higher than
those which are tolerable with current engine systems. This
is an important issue, because widening the end point
- 14 ~ ~
specification for gasoline sigrlificantly reduces refining
costs. There are also indications (which will require more
research to establish) that the vortex mixer will permit the
use of napthalenic and other low hydrogen to carbon ratio
hydrocarbons in motor fuel. If this proves to be possible
it will greatly ease the problem of producing motor fuels
from synthetic sources, such as tar sands, oil shale or coal.
Government and industry supported research is now in progress
concerning the vortex to test its ability to burn synthetic
fuels, including alchohols and low hydrogen to carbon ratio
fuels.
The vortex mixer will also permit the design of lighter,
cheaper, and more efficient intake manifolds since these mani-
folds need not handle the complexities of two-phase flow. It
should be mentioned that the mixing relations of the vortex
mixer are such that it can be designed as a relatively low-
drag device, permitting high peak powe-r. The vortex mixer
will also function to mix homogeneously EGR with the fuel and
the air for even EGR distribution cylinder-to-cylinder.
Because of these effects, it has been shown experimentally that
the present invention vortex mixer simultaneously improves
driveability, emissions, and overall fuel economy. The inven-
tors wish to thank 0. A. Uyehara, G. L. Borman, and P. S. Myers
of the University of Wisconsin for many useful discussions
during the development of the device.
- 15 ~ 1~ 3
IN THE DRAWINGS
Figure l is a top llan view of a preferred form of
a vortex mixer, showing particularly fluid mechanical details
in the entrance section which produce the smooth tangential
flow introduction into the vortex mixer passage per se.
Streamline patterns are shown by arrows and parasitic vortices
70 and 72 are surrounded by dotted lines for clarity.
Figure 2 is a side cut-away view along sectional line
C-C showing details of the flow containing passages, particul-
arly with respect to the vortex chamber per se.
Figure 3 is a view along sectional A-A of Figure 1
showing the throttle plate of F:igure l in more detail. The
trapezoidal throttle plate assembly shown can be replaced by
the round-to-trapezoidal entrance section shown in Figs. 4 and
5.
Figure 4 shows a throttle body entrance section which
could be bolted on along the surface of sectional A-A of
Figure 1 as a direct substitution for the assembly shown in
Fig. 3 where the entrance section body involves a transition
from a circular butterfly air throttle to the trapezoidal shape
of the entrance section along sectional A-A.
Figures 4EE 4FF, 4GG, 4HH, 4JJ, and 4KK are half
scale sectional views of the throttle body internal surface
showing the transition between the round throttle plate
area and the trapezoidal entrance area at surface A-A of
_, .. . .
~ 16 -
llZ~283
Figure 1 for the entrance section structure of Figures 4 and
5. These views orrespond to the planes identified by sectional
lines EE, FF, GG, ~H, JJ and KK respectively, shown in Fig.
4.
Figure 5 is a side view of th~ round throttle plate
to trapezoidal entrance passage of Fig. 4, viewing on sectional
5-5 of Figure 4.
Figures 5AA, 5BB, 5CC, and 5DD are sectional views of
the throttle body internal surface, showing the transition
between the round throttle plate area and the trapezoidal en-
trance area at surface A-A of ~F~gure 1 for the entrance
section structure of Figures 4 and 5, and corresponding to
the planes identified by section lines AA, BB CC, DD respec-
ti~ely, shown in Figure 5. Section lines AA to DD are shown
in Figures 4EE, 4FF 4GG, 4HHr 4JJ, and 4KK.
Figure 6 is a projection showing the tangential velocity
in an irrotational flow vortex as a function of radius.
Figure 7, including its exemplary equations, shows the
constant pitch spiral flow streamlines characteristic of an
idealized irrotational flow vortex having neither turbulence
nor molecular diffusion, with the streamline plotted which
would be foll~owed by an ink line tracer from the outside of
the vortex to its central sink.
Figure 8 shows an idealized irrotational flow
vortex analogous to that of Figure 7, but with two stream-
lines plotted to illustrate the nesting of the streamlines.
~ ]-7 ~ 1 ~ Z9 Z8 3
Figure 9 is also analogous to the flow pattern
of Figure 7, and shows the nesting of 10 evenly distributed
streamlines around the idealized vortex flow pattern. Figures
7, 8 and 9 are used to illustrate an important mathematical
point about the mixing patterns within the vortex.
To achieve the flow patterns illustrated in Figures
6, 7, 8, and 9 in a practical mixer requires that boundary
layer flow problems normally encountered in vortex passages
be eliminated.
Figures 10 and 11 illustrate the more complicated flow
pattern which occurs if the boundary layer is not properly
controlled. These figures are taken from Desiqn Theory of
Fluidic Components by J.M. Kirshner and Silas Katz, Academic
Press, 1975, Figure 10 shows a cross section perpendicular to
the axis of rotation of the vortex showing the more compli-
cated vortex flow pattern which occurs because of boundary
layer effects. Figure 11 is a diametral cross section through
FigurelO showing streamlines in the radial and axial direction
which occur in this more complicated flow pattern.
Figure 12 is analogous to Figure 2, showing the fluid
mechanical effects of circumferential weirs, 41, shown in Fig-
ures 1 and 2, showing how these weirs form recirculating vor-
tices at the boundary layer which permit the central flow of
the vortex chamber to be a fair approximation ~f the irrota-
tional flow vortex pattern illustrated in Figures 6, 7, 8, and
9.
Figure 12a is an exploded view of the circled area of
Figure 12, showing the flow between two weirs in more detail.
Figure 13 is a top plan view of a vortex mixer analo-
gous to that of Figure 1, wherein the entrance flow geometry
is arranged for connection with a two venturi down draft
carburetor (carburetor not shown).
Figure 14 is a vertical sectional view of Figure 13
with the outside wall of the finned exhaust heat exchanger
passage removed to show the geometrical arrangement of the
fins.
Figures 15, 16, 17 and 18 show a more primitive vortex
mixer adapted to a downdraft carburetor.
Figure 15 shows a diametral vertical sectional view of
the mixer, with the section bisecting the venturi of the
attached downdraft carburetor.
Figure 16 is a top plan view along sectional 16-16 of
Figure 15.
Figure 17 is a side view along sectional 17-17 of Figure
15 showing the relationship between the carburetor throttle
body and the flow deflectors.
Figure 18 is a sectional view along broken sectional
linel8-~ of Figure 17 further illustrating the shape of t'ne
deflectors.
Figures 19 and 20 illustrate why the homogeneous mix-
ing of the current inventi~n is practically useful, and helps
explain the necessity for tight cylin~er-to-cylinder,microscale
volume, and time sample air-fuel ratio statistics of low N0x
outputs are to be obtained with a fuel efficient lean burn
englne.
-- 19 --
Figure 19 illustrates mixing statistics with the
numerical example of Gaussian air-fuel ratio distributions
A, B. and C. The graph makes clear the advantage of mixing
in extending the lean limit of satisfactory combustion.
Figure 20 plots experimental data which shows the vi-
tal importance of mixing and equivalence ratio on NOX outputs
and engine thermal effiçiency.
1129~:83
DETAILED DISCUSSION
A preferred form of the present invention is shown
in Figures 1, 2, and 3 with an important practical ~odifica-
tion shown in Figures 4, 4EE, 4FF, 4GG, 4HH, 4JJ, 4KK, and
Figures 5, 5AA, 5BB, 5CC, and 5DD. Because the fluid
mechanics involved is far from the training of conventional
automotive engineers, these drawings will have to be treated
several times in this specification from several points of
view.
Figure 1 is a top plan view of the vortex mixer,
with particular attention drawn by the flow lines to the
fluid mechanics in the entrance passage. Figure 2 is a
side sectional view along section line C-C of Figure 1 show-
ing details not visible in Figure 1. Figure 3 is a view
along sectional A-A of Figure 1 showing the rectangular
shaped air throttle passage characteristic of Figure 1.
Referring now to Figure 1 J the airflow to the engine is
delivered to a passage 1 feeding throttle plate 4 which
pivots on throttle plate shaft 2. In this specification the
precise form of the fuel-air metering system is not specified,
and the fuel introduction means in Figure 1 is shown by means
of a fuel spray nozzle 6 spraying into the airflow at 8. The
edges 10a and lOb of the rectangular shaped throttle plate 4
are cut away to produce a smooth, sharp edged orifice effect
for the airfldw flowing between throttle plate 4 and throttle
passage body 12. Throttle passage body 12 connects by fasteners
-20-
' !
llZ9Z83
to vor~ex mixer passage body 14. The details of the fluid mech-
anics within the entrance section conduit 16 and within the
main vortex flow chamber 18 will be discussed in considera-
bly more detail later. The high speed flow past the upper
edge of throttle plate 4 forms high speed wall-attached jet
stream 20, which attaches to the wall of the mixer passage
according to the Coanda effect, which will be discussed
subsequently. The wall attached stream flows smoothly and
is introduced into the generally radially symmetric vortex
mixer chamber 18 in coherent and tangential form. The high
velocity jet 22 formed downstream of the lower edge of
throttle plate 4 also forms a Coallda wall-attached stream
and is deflected by deflection means 24 so that is is de-
flected at cusp 26 into the vortex chamber. The jet 22 flows
past cusp 26 and interacts ballastically with static de-
flector 28 which deflects this jet into a smooth tangential
introduction to the radially symmetric vortex chamber 18.
These tangentially introduced flows from jets 22 and 20 flow
via the Coanda effect in coherent form counterclockwise
around the outside peripheral wall of chamber 18. The flow
around the wall will pass by cusp 26 and merge with the
flow from jet 22, interacting ballastically with deflector
28 and flowing again near the peripheral surface of the
vortex chamber near surface line 30. The flow inside
chamber 18 is therefore a swirling flow. This flow proceeds
in a swirling manner inwardly toward the outlet 32 of the mix-
er chamber 18, which outlet 32 connects the vortex mixer
-21-
t~
~.
llZ~Z83
with the intake manifold of the engine (not shown). The
fluid mechanical relations within vortex chamber 18 form
a flow pattern called an irrotational flow vortex. ~he
details of this flow pattern will be discussed in more
detail later. The swirling flow into outlet 32 is de-
flected by deflection vanes 34 so that the flow in the
outlet section is not the strong irrotational flow vortex
characteristic of the mixing channel 18 between the de-
flection vanes 34 and the outside wall of chamber 18
(going in the radial direction from the center 35 of the
vortex mixing chamber 18). Within the outlet 32
is shown a strictly optional variable geometry deflection
means 36, 38, 40, which can rotate (to a position denoted
by the dotted lines 36a, 38a) to deflect the swirling
flow within the vortex chamber 18 under conditions of
wide-open throttle for engines where peak power is extremely
important. This optional deflector eliminates an unavoidable
flow resistance characteristic of the vortex flow pattern,
and makes it possible for the vortex mixer of Figures 1, 2,
and 3 to have excellent mixing under the lower loaded condi-
tions characteristic of normal engine operation without any
sacrifice of peak power flow capacity. The peripheral wall
surface 42 of the vortex mixing chamber 18 collects the liquid
fuel introduced into the mixing chamber due to centrifugal
forces (analogous to those which operate in cyclone scrubbers).
As was discussed in the Background section, this liquid must
then be evaporated before it may pass through the vortex
-22-
- 23 - 1 1 2 g 2 8 3
mixer. To provide the heat necessary to accomplish this,
exhaust gas i5 passed through chamber 49 formed by peripheral
finned wall 44 of the vortex and outer wall 45. Exhaust heat
transfers to fins46 on wall 42 which heats peripheral vortex
wall inner surface 42. Exhaust gas past these fins 46 is
supplied by the engine, and connection 80 and 82 connect to the
exhaust passages of the engine (not shown). These fins 46, in
connection with the peripheral wall surface 42 form an extremely
effective heat exchanger arrangement which rapidly evaporates
the fuel in the vortex section. In both Figure 1 and Figure 2
circumferential weir protusions 41 are shown on the top and
bottom roughly parallel surfaces of vortex mixing chamber 18.
These circumferential weirs 41 have an important boundary
layer conditioning function which will be discussed subsequently.
Figure 3 is a view along sectional A-A of Figure 1 show-
ing the roughly rectangular throttle plate 4 and showing the
manner in which the throttle passage body 12 may be connected
to the main vortex mixer body 14. A number of fluid mechanical
details with respect to Figure 3 will be discussed later.
However, it is worthwhile pointing out now that the generally
rectangular throttle plate 4 has within it a notch 7 which
forms the flow passage through which most of the idle airflow
passes. Flow from this notch will form part of high-speed
jet 20. The flow from the notch will be attached to both the
side and top walls of the mixer. The spreading of the wall-
attached jet from this notch will be less than the spreading
which would occur at idle flows without the notch, with the
consequence that the vortex flow in main vortex chamber 18
will be faster and more conducive to mixing because of the
provision of notch 7 on throttle plate 4.
~ - 24 - ~1Z9283
A number of practical issues concerning the fastening
and sealin~ of the throttle passage body 12 to vortex body 14
are illustrated in Figure 3. Bolt holes 13 for fastening are
shown. Also shown are pin holes 15 for alignment between
throttle passage body 12 and vortex body 14. A groove 17 for
an O-ring type deformable seal is also shown. O-ring type seals
are useful for sealing pieces of the current invention, since
the O-ring seal does not protrude out between adjoining sur-
faces as conventional gaskets sometimes do. Because of
details of the fluid mechanics in the mixer to be discussed
later, careful alignment of the throttle passage body 12 with
respect to vortex body 14, and smooth transitions between
the airflow contacting surfaces are essential.
Those skilled in the automotive engineering arts
should rapidly recognize an important practical disadvantage
of the rectangular throttle plate in throttle passage body
12 shown in Figure 3. The problem is simply that the seal
between throttle passage body 12 and the upper and lower
sides of throttle plate 4 is likely to produce enough
leakage to be troublesome at idle, and involves difficult
thermal expansion fit problems. Figures 4 and 5 show a
throttle passage body which avoids this seal problem and still
produces flow patternsto the left of the sectional A-A plane
of Figure 1 re~uired for efficient vortex mixing.
For a conventional round butterfly airflow control
valve, the difficulty with side seals does not arise (al-
though tolerances between the throttle plate and throttle
body section must be relatively tight for proper control of
air flow at idle conditions.) Figure 4 and sectional view
figures 4 EE, 4 FF, 4 GG, 4 HH, 4 JJ, and 4 KK combined
1129283
with Figure 5 and sectional views 5 AA, 5 BB, 5 CC and 5 DD
illustrate an entrance section passage which may be bolted on
at sectional AA of Figure 1 exactly analogously to the entrance
throttle passage body 12 of Figure 3,where the air throttle
is a round butterfly valve and where the entrance section
passage surfaces are faired to produce a fluid mechanically
smooth transition between the round throttle plate and the
nearly rectangular opening at the entrance passage plane of -
sectional AA, with the surfaces and deflection arrangements
shaped so that the predominant fraction of the high speed
air flow past the throttle plate is delivered into the corners
of the rectangular opening at gectional AA and along the near
vertical surfaces 50 and 52, with the flow proceeding in a
manner so as to minimize the fraction of the high speed flow
which enters the vortex mixing section 14 adjacent surfaces
54 and 56. Figure 4 is a view from sectional A-A of Figure 1
viewing the round- to-rectangular transition passage which
replaces the rectangular throttle entrance section shown in
Figure 3. Sectional view Figures 4 EE, 4 FF, 4 GG, 4 HH,
4 JJ, and 4 KK are included to illustrate the shape of this
round- to-rectangular entrance passage 58. Figure 5 is a
side view of the entrance passage of Figure 4 through sectional
5-5 which is a view from the other side of sectional GG.
Figures 5 AA, 5 BB, 5 CC, and 5 DD show the internal surface
shapes at sections AA, BB, CC, and DD respectively. In
addition, these sectional figures show in dashed lines de-
flection vane outlines 62 for the top surface 56 and t~e lower
- 26 -
1129Z83
surface 54,~ich vanes serve to deflect the flow so that it
congregates in corners of the entrance passage trapezoidal
entrance and avoids attachment to surfaces 54 and 56. Some
of the fluid mechanics involved in the transition section
58 shown in the series of Figures 4, 4.EE, 4 FF, 4 GG, 4 ~IH,
4 JJ, 4 KK and Figure 5, 5 AA, 5 ~B, 5 CC, and 5 DD must
await further fluid mechanical discussion. However, gross
structure is as follows. The entrance passage assembly 58 consists
of a complexly shaped cast body in which is mounted a general-
ly round throttle plate 60 which pivots on throttle plate
shaft 61. The round flow passage which accommodates throttle
plate 60 is faired to a generally rectangular shape before it
connects at the surface AA forming surface 64. Deflection
vanes 62 serve to push the flow to the outside cor-
ners of the trapezoidal entrance section and minimize high
speed flow adjacent surfaces 54 and 56. Near the mating
surface 64 in Figure 4 in the corner between surface 52 and
surface 56 isan idle flow passage 68 accommodating the majority
~ theidle flow of air to the engine, and arranged so that
this idle flow enters the vortex mixer in the form of a high
speed jet hugging its corner to produce enhanced vortex yelo-
city and mixing in a manner analogous to that discussed in
Figure 3 wlth reference to throttle edge notch 7. This idle
air flow is available to serve the purpose of an atomizing air
flow if fuel is introduced upstream of the critical flow
controlling restriction in this nozzle passage opening 68. Air
flow and fuel flow past noz~le 68 is delivered in a form
conducive.to a very efficient jet at~achment to the corner
- llZ9Z83
wall, so that the high speed flow proceeds efficiently into
the vortex chamber 18, and so that the fuel introduction into
the vortex chamber is extremely rapid. The flow direction from
nozzle 6~ is specified with reference to the flow arrows shown
in Figures ~ KK and 4 EE.
At this point, the gross structure of the mixing
vortex shown in Figures l, 2, 3, 4, 4 EE, 4 FF, 4 GG, 4 HH,
4 JJ, 4 KK, and Figures 5, 5 AA, 5 ~B, 5 CC, and 5 DD ahould
be relatively clear.
High velocit~ air past the air control throttle to
the engi~e flows tangentially into a vortex section and spi-
rals inwardly into the outlet ~2 The vortex flow within
chamber 18 serves to separate t~e fuel and deposit it on
peripheral wall 42 of the vortex chamber per se from whence
it is evaporated and mixed into the inwardly spiraling air-
flow which flows to the outlet 32. However, it happens that
the efficiency of function of the vortex mixer depends to a
very important extent on details of the flow channel design
which are generally not obvious to conventional automotive
engineers.
In the course of discussin~ this invention with skilled auto-
motive engineers, it has been the inventors' consistent experience that,
even after careful instruction on the fluid mechanical de-
tails important to the function of the device, the conven-
tional engineers make fluid mechanical mistakes which signi-
ficantly deqrade the function of the vortex mixer unless the
details of the design are very carefully supervised by some-
one with experience in the field of fluidics. It has been
our experience that mechanical engineers of conventional
. .
1129283
training, even at the Ph.D. level~ have difficulty understanding
and cannot visualize the fluid mechanical details relevant to the
functioning of a fluidic device such as the vortex mixer of the
present invention. Conventional mechanical engineers apparently
find the concept of detailed structure in a turbulent flow
difficult to deal with. Apparently ability to confidently mani-
pulate this sort of fluid mechanics only comes after much study
and some experience in actually building fluidic components.
A careful effort will be made in this specification
to completely describe all the fluid mechanics necessary to
build an efficient vortex mixer. However, the information
involved is complicated enough that it is apparently diffi-
cult to remember and use it all at once. Therefore, it
would be practically very useful for any firm wishing to
make and use the present invention vortex mixer to employ
someone skilled in the arts of fluidics to evaluate the de-
sign and to assist in any experimental development of the
design which may be necessary. It is the inventors' judgement
that this consulting expenditure will be inexpensive insurance
for any serious production development of a vortex mixer.
The mixing and evaporating efficiency of the vortex
mixer depend on the strength and flow pattern of the flow
within vortex mixer chamber 18. What is desired is the
strongest possible irrotational vortex flow rotating about
the center 35. The stronger this flow, the more efficient
the evaporation and the more efficient the mixing of the
device. The strength of this vortex flow witin vortex mixer
-28-
; ~
~ - 29 -
1~2~Z83
chamber 18 depends on the mass mean flow velocity of the flow
entering the mixing chamber tangentially from jets 20 and
~2. The greater the fraction of the isentropic expansion
velocity past throttle plate 4 which can be delivered in
coherent form tangentially past point 26 and point 30, the
stronger the vortex will be and the better the mixer will
function. Relevant here are the mean flow velocity of the
fluid entering the vortex chamber past point 26 and point 30,
the flow patterns past these flow areas, and the turbulence
intensity of the flow past these areas. The velocities,
patterns, and turbulence levels are coupled. In general,
it is desirable to have the flow delivered into the vortex as
coherentlyas possible and with the maximum fraction of the
energy in the flow in the form of mean flow streamlines
(with minimum practical turbulence intensity). In order
to secure the proper flow conditions at the tangential en-
trance of the vortex (points 26 and 30), the detailed geo-
metry between the opening of the throttle plate and these
entrance points must be carefully controlled. With proper
control, it is possible under normal conditions to have
the tangential introduction velocities past point 26 and
point 30 in excess of 80 percent of the isentropic expansion
velocity past the throttle plate. However, unless flow
channel characteristics in the entrance section are properly
handled, the flow velocity into the vortex channel may well
be only a very few percent of the isentropic expansion
velocity. The importance of the fluid mechanics in the
entrance section cannot be overemphasized. Design
- 30 -
~Z9Z83
errors in the entrance section which do not appear to be
significant to a conventionally trained automotive engineer
can have order-of-ma~nitude effects on the mixin~ rate in
the vortex mixer. The flow shapings which are required to
produce fluidic efficiency and high mixing efficiency in
the vortex are not difficult to manufacture, but the geomet-
rical issues involved cannot be i~nored. For this reason,
a fairly detailed description of the fluid mechanical
relations in the entrance section prior to the flow intro-
duction past point 26 and point 30 is required.
It is useful first to have a sense of the magnitude
of the flow velocities to be expected past the air throttl-
ing valve under normal engine operating conditions. The
following chart shows the relationship between intake mani-
fold pressure drop from atmospheric pressure and the isen-
tropic expansion velocity to be expected directly downstream
of the air throttling restriction in the vena contracta.
The chart plots mass velocities and spatial velocities
vs. the pressure drop across the throttle. Throttle pressure
drop is expressed two ways, first as the ratio of manifold
pressure to atmospheric pressure. For each pressure ratio
the corresponclin~ intake manifold vacuum in inches of mercury
is set down. For each pressure drop there corresponds a
ratio of mass flow per unit area past the vena contracta, M
divided by the mass flow per unit area which would occur under
critical flow (sonic) conditions,M*. Also plotted against pres-
sure drop is the velocity of the flow in the vena contractain meters per second.
~ 1V 1~292~3 Mvelocity
Patm. _ m. - Zg g~ Hg ) M _ ~leters/Sec.
.98 .6" .3 53
.94 1.2" .49 g2
.90 3" .62 121
.80 6" .82 181
.70 g" .935 236
.60 12" .985 291
.528 14.1" Hg 1~00 335
~ Even at the lowest intake manifold vacuums
normally encountered in engine operation, the velocities
past the air throttle are quite substantial~ The flow
streams have very high momentum and kinetie energy
because of these high velocities. Under normal condi-
tions the isentropic expansion velocity past the air
throttle will be a substantial fraetion of sonic velocity.
This is enough flow velocity, if the flow velocity of the
jet can be maintained and delivered coherently,to dr.ive
the vortex to produce an extreme]y strong and efficiently
mixing vortex in vortex chamber 18. Elowever, unless de-
sign care is taken, the high isentropic velocity directly
downstream of the air throttle will be rapidly dissipated
into turbulence, with the result that only a very small
fraction of this velocity will be available to drive the
vortex in chamber 1~. To understand how the high velocity
flow directly downstream of the air throttle restrictions
ean be maintained to drive a strong vortex, it is necessary
to understand the Coanda effeet and the efEeet of stable
parasitie vortiees sueh as those whieh are shown in Figure 1.
l~Z9283
High speed flow streams close to walls attach to
these walls according to the Coanda effect which is one
of the bases of fluidics (for a detai]ed explanation of
the Coanda effect, see pages 131 to 139 of Fluidics, Com-
pon _ s and Circuits by N. Foster and G. A. Parker,
Wiley Interscience, 1970: this reference is useful for
many issues in fluidics).
T~e principle of wall attachment is important
enough that it must be described in some detail here.
Basically, a jet entrains flow on both sides. If the jet
is near a wall, fluid entrainment generates a reduced
pressure on the wall side with respect to the outside of
the jet flow. Because of the pressure difference between
the wall side and the outside of the jet, the jet flow
path curves towards the wall (the jet is sucked towards
the wall). As the jet bends towards the wall, the wall
pressure becomes smaller, the suction stronger, and in
consequence the jet attaches to the wall. This attachment
effect is utilized in a number of important digital fluidic
logic circuits, for example, those invented by Raymond Warren,
et al, at Harry Diamond Laboratories. The Coanda, or wall
attachment effect, is of great importance to the design of
the vortex mixer because a wall attached stream (particularly
one adjacent to a properly set up parasitic vortex flow~
spreads much more slowly than a non-wall-attached stream
would. ~ince the jet velocity and the jet area are inversely
11Zg'Z83
reLated for a set mass flow rate, minimizing the jet spread-
ing allg1e minim;zes the jet flow area and maximizes the jet
velocity, which is desired to strongly drive the vortex in
chamber 18.
Wall attached jet fiows attach particularly well in
corners of passages, such as occur at the outlet of throttle
notch 7 in Figure 3, and at the corners of the flow passage
shown in Figures 4 and 5. Flows attached to such corner
passages have smaller losses than jets attached to extended
flat surfaces. Notch 7 shown in Figure 3~ and the deflectors
62 shown in Figure 5 are designed to put wall attached jets in
passage corners for delivery to t:he vortex mixer passage 18.
Idle flow passage 68 shown in Fig~ure 4 serves the same purpose.
Referring specifically to the device shown in Figures
1, 2, 3, flow past throttle 4 flows between edges lOa and lOb
and the channel wall of throttle body 12. Flow past edge lOa
attaches to the channel wall of 12 and flows to the junction
between throttle body 12 and vortex chamber body 14 at the
surface along which sectional A-A is taken. The high speed
2Q jet 22 which is formed past throttle edge lOb flows towards
this surface of sectional A-A in an exactly analogous way.
The surfaces 24 and 25 of the vortex chamber body 14 are set
back so that the high speed jet flows 22 and 20 flowing over
the plane of sectional A-A flow to a surface which is set back
a distance Y from the surfaces of throttle body 12. This set-
back is not explicitly notated in Figure 1, but an exactly ana-
logous setback Y is shown in Figure 5. In practice the surfaces
-33-
s~
1129283
downstream of the throttle body 12 must be slightly set back,
but the setback desirable is smaller than can conveniently be
shown in a patent drawing. Nonetheless, the importance of
the concept of setback must be discussed in detail, since it
is vital to the function of the current invention.
The stability of a wall attached stream is not simply
a function of the intrinsic instabilities which come from
high local Reynolds numbers. Small variations in flow
geometry which might at first appear to be insignificant
can, by producing large disturbances in the jet, cause a
magnification of turbulence which causes the jet to "break
up." Breakup drastically reduces the velocity and kinetic
energy of the jet as it travels downstream from the dis-
turbance, and therefore must be avoided if the vortex is
to function efficiently. The setback (exemplified by
setback distance Y) is important to the successful and
reliable fimction of the vortex mixer shown in Figures 1 to 5.
Streams which flow past a setback will, unless the setback
is too great, reattach cleanly and wi.th relatively small
loss to a surface such as the downstream surfaces of chan-
nel 14. However, if setback distance becomes negative, so
that the deflector protrudes into the wall attached stream,
(this protrusion is denoted as a "step-up"), a high velocity
element of fluid will collide with the step-up as with a
brick wall. This flow element will deflect off the step-up
in a manner which will strongly perturb the flow, break-
ing up and destroying the coherence of the jet and so ruining
-3~-
. ~
~3Z83
the efficiency with whicll flow velocity is delivered to drive
the vortex in the mixer chamber 18. When the jet flow break-
up occurs, the kinetic energy of the jet is quickly lost be-
cause the jet spreads at such an angle that the momentum of
the jet is dissipated in a large mass of turbulent fluid.
In mass production, exact shaping is not possible,
and parts must be made within a range of dimensions called
tolerances. Because step-up between the throttle body 12
and the surfaces of the vortex chamber body 14 is so detrimen-
tal, it is important that the parts be specified with setback in
mind, so that within the actual production range of dimensional
variation between throttle body 12 and vortex channel 14,
step-up never occurs. Since setbacks as large as .020 inch
can be jumped by a wall-attached jet with smooth and low-loss
reattachment on the downstream surface, and since the mixer
is most efficiently manufactured by the intrinsically high
tolerance manufacturing process of die casting there is no
excuse for step-up ever occuring in production. However,
if step-up is permitted to occur adjacent the high-speed
jets 20 and 22 (or the analogous jets which occur from the
entrance section shown in Eigures 4, 4EE, 4FF, 4GG, 4HH,
4JJ, 4KK, 5, 5AA, 5BB, 5CC, 5DD), the fluidic efficiency
and mixing efficiency of the vortex mixer will be
significantly degraded from that which would occur without
the step-up.
Although the geometry of the flow channels with re-
spect to issues such as setback is vital, it is not required
-35-
. ~
1129Z83
that the vortex mixer have particularly smooth surfaces.
Surfaces which come from even a badly controlled die cast
process should be sufficient for the vortex mixer function.
Boundary layer thicknesses are generally great enough to
obscure the effect of most colmmonly encountered surface
roughnesses. However, large surface irregularities, such
as grains of sand protruding in the wall-attached jet,
can significantly degrade the fluidic efficiency (defined
as the ratio of tangential introduction velocity to isen-
tropic velocity past the throttle restriction) of the
device. Generally, it is easy to maintain quality control
well enough to avoid these obstructions.
Referring to the flow patterns shown in the entrance
section in Fig. 1, the fluidic function of the parasitic re-
circulating vortices 70 and 72 should be emphasized. As the
high speed jets 20 and 22 flow along, they entrain fluid
which flows with them. The fluid mechanical relations are
such that recirculating vortices 70 and 72 form in the entrance
section. The geometry of static deflector 28 assures that
the relative positions of parasitic vortices 70 and 72 are
relatively stable for a set air throttle angle (and that the
parasitic vortices enlarge or shrink smoothly during throttle
actuation). This is important: Downstream of a conventional
carburetor throttle are parasitic vortices analogous to vor-
tices 70 and 72. However, there is no geometry which ar-
ranges these flows in steady form and the two vortices fluc-
tuate in relative size with one becoming predominant, washing
-36-
.. ~,~
- 37 -
llZ~283
out, and then the next growing and washing out. This unstable
flow pattern i~ the ~arasitic vortices is responsible for a
number of mode shifts affecting fuel-air distribution in con-
ventional intake manifolds. The flow patterns of parasitic
vortices 70 and 72 are highly desirable, in that these vortices
form stable flow patterns which minimize the shear between the
velocity of high speed jets 22 and 20 and the surrounding
fluid. For this reason, the parasitic vortices 70 and 72
greatly decrease the spreading angle of these jets 20 and
22 and in consequence significantly improve the fluidic
efficiency of the entrance passage geometry. It is important for
the designer of an entrance section to a vortex mixer to
trv to mini~ize the turbulence and maximize the stability
of the parasitic vortices within the entrance section,
since these parasitic vortices have important effects on
the total fluidic efficiency of the mixer. For the generally
rectangular entrance section of Figures 1, 2, and 3, this
stability is a matter of course and does not require experi-
mental checking. ~owever, for a more complex shape such as
the round-to-trapezoidal transition section shown in
Figures 4, 4EE, 4FF, 4GG, 4HII, 4JJ, 4Kl~, 5, 5~ , 5BB, 5CC,
5DD, it is prudent to investigate the fluidic efficiency
of the entrance section experimentally to make sure that
high fluidic efficiency is in fact being obtained. Detailed
discussion of this experimental function will follow. Re-
ferring again to the entrance section flow patterns in
Figure 1, hole 74 shows a desirable place to introduce ex-
haust gas recirculation flow if EGR is employed in the mixer.
~ - 3~ -
l~Z9Z83
Any ~GR flow past hole 74 would be intermixed by parasitic
vortices 70 and 72 and smoothly introduced into the high
speed jets flowing into the vortex mixing chamber 18.
More will be said later about the fluid mechanics
within the entrance section. ~owever, the purpose
of the present invention is to produce a mixe.r producing
essentially perfect fuel-air homogeneity and rapid transient
response. Looking at the flow patterns within the entrance
section and then considering (as will be done later) the
vortical flow pattern within mixing chamber 18, it should
be apparent that the transient response of the mixing device
depends largely on the detailed manner in which fuel is
introduced into the mixing chan~el. Clearly, fuel intro-
duced into the parasitic vortices 70 and 72, or into the
surfaces adjacent these parasitic vortices will be intro-
duced into the vortex mixer more slowly than if the fuel
were directly introduced into the chamber 18 by some fuel
injection means, or if the fuel was injected directly into
one of the high speed jets 22 or 20, as for instance by
introduction into nozzle 68 of Figure 4.
~ eferring again to Figures 1 and 2, in light of
Figures 3 and 4 and 5, it should be clear that the high
velocity flow pattern delivered past sectional AA in Figure
1 should be much the same whether the entrance section 16
shown in Figure 3 is employed or whether this section 12
is replaced by the round throttle to trapezoidal entrance
section shown in Figures 4 and 5. In either case, smooth
high velocity flows will be introduced tangentially past
point 30 and past point 26 into the vortex mixing chamber 18.
- ~ .
. - 39 -
1129283
The flow introduced from jets 22 and 20 will have high
angular momentum with respect to the vortex chamber center
line 35, shown as point 35 in Figure 1. For reasons which
will be discussed in more detail subsequently, the flow field
in the generally annular volume between peripheral vortex wall
42 and deflection vanes 34 will be in the form of an irrota-
tional vortex flow field, where the angular momentum of the
fluid will be approximately conserved as it spirals inwardly
to the vanes 34 for delivery to the outl~t section 32, from
whence the mixture flows to the intake manifold (not shown).
This irrotational vortex flow field within vortex mixing
chamber volume 18 is an extremely high velocity flow field,
where the mean flow streamlines are dominated by the fluid
mechanical relation of conservation of angular momentum, with
turbulence superimposed on the mean flow streamlines. The
swirling vortex flow within mixing chamber 18 is of such high
velocity under normal engine operating conditions that liquid
fuel delivered into the vortex is rapidly thrown to the out-
side of the vortex and impacted against peripheral wall 42.
The flow field is generally so strong, and the boundary layer
adjacent surface 42 so thin, that this liquid fuel generally
atomizes and rotates with the flow field in the form of a
"splash cloud" rotating very near the peripheral wa]l 42,
impacting upon deflector 28 and then re-impacting and rotating
again around the vortex outside wall 42 until it is fully
evaporated. The effect is that fuel evaporates around the
~tireperipheral wall 42 of the vortex chamber. The boundar~
layer turbulence adjacent surface 42 in the vortex is such
- 40 ~ ~ 1 Z9 2 8 3
that this evaporation process has the effect of introducing
fuel vapor into relatively fuel-poor air in a structured tur-
bulent flow mixlng process characterized by extremely high mix-
ing rates. Some of the details of this fluid mechanical situation
inside vortex chamber 18 will be discussed subsequently.
The flow, however, i9 very near to the irrotational vortex
flow pattern which would be predicted analytically and the
flow spirals into pressure recovery and vortex destroying
vanes 34 so that the flow field within outlet 32 is turbulent
and characterized by much lower vorticity than that characteris-
tic of the flow within the annular volume between vanes 34
and peripheral vortex wall 42. Vanes 34 around the outlet 32
serve a dual purpose in the mixing device. First, they reduce
the pressure drop which would otherwise occur across the mixing
channel. Secondly, by eliminating a strong vortex in the out-
let, and substituting a relatively isotropic finer grained
turbulence they very much simplify the fluid mechanical re-
lations in the intake manifoldlng (not shown) downstream of
the vortex mixer. An additional effect of the vanes is to
create in the center of the vortex mixer a volume which is
fluidically decoupled from the fluid mechanics which govern
the flow in the annular volume between vanes 34 and peripheral
wall 42. ~ecause of this decoupling of the fluid mechanics,
it is possible to place within the volume of outlet 32 a
deflection arrangement 38, 3~, 40, which can, when very low
flow resistence is required for wide open throttle operation,
rotate to the position shown in the dashed lines 36 A and
38 A. When this deflector is in this extended position, it
largely destroys the vortex flow pattern within mixing
~ - 41 -
112~'~83
chamber 18, and in consequence reduces the flow resistence
of the total mixing flow channel for wide open throttle
power conditions. Under wide open throttle conditions very
tight cylinder-to-cylinder distribution is of secondary
importance so that destroying the vortex flow pattern is
tolerable. Under the much more common part load conditions
when the engine is throttled, the deflection means 36,38,
and 40 do not in any way impair the mixing of the device.
In Figure 2, it is shown that the lower wall of mixing
chamber 18, wall 43 and the upper wa]l of chamber 18 wall,
wall 47, are surfaces of revolution curved upward, so that
the axial height clearance between wall 43 and wall 47 is
roughly constant in the radial direction from center line 35.
This curvature can be useful under cold start conditions
where liquid accumulates within the mixing chamber 18, because
it allows a significant accumulation of liquid fuel within
the vortex before any liquid can flow out of tne mixer and
into outlet 32, for delivery to the intake manifold. It
should be noted also that it is not a requirement that the
axial height between the top and bottom walls of the mixer
~ constant with respect to radial distance from the vortex
mixer center. However, this constant axial height relation
does simplify the pattern of the mean flow streamlines through
the vortex mixer, in a way which will be made clear in the
subsequent discussion.
~ 1129Z83
Consideration of Figures 1, 2, 3, 4, and 5 should make it clear
that the flow pattern within the vortex mixer of the present invention will
not be a perfect irrotational flow vortex, both because the introduction
means will perturb the flow, and because the flow will be significantly
turbulent. Also, there will be momentum interactions with the peripheral
wall 42, weir sections 41, top 47, and bottom 43. Nonetheless, the physical
relations of conservation of angular momentum are sufficiently important
that the flow within the chamber 18 shown in Figures 1 and 2 will be approx-
imately an irrotational flow vortex, and it is therefore worthwhile to con-
sider the flow streamline relations which occur in an irrotational flow vor-
tex. See Figure 2, and consider polar co-ordinates centered at the center
35 of the outlet 32, so that flow velocity components would be defined in
terms of a velocity in the radial direction V , and a velocity in the tan-
gential direction, Vt. These velocities would be mean flow stream veloc-
ities: the real flow would clearly incIude a fluctuating component in both
the tangential, the radial, and the axial direction. Flow into the vortex
chamber from throttle passage body 12 or 58 would clearly have angular mom-
entum with respect to the outlet 32 center 35. Conservation of angular mom-
entum, MVtR, dictates the increase in the tangential velocity of the fluid
as it flows towards the center. It is easy to verify that the velocity in
the tangential direction as a function of radius r, Vtr, will be expressible
according to the relation
trO o
t(r) r
- 42 -
- 43 -
where Vtr is the tangential velocity at the outside of the
vortex, rO is the radius at the outside of the vortex, and
r is the radius where the velocity tangential is taken.
Figure 6 illustrates the flow velocities which are produced
in an irrotational flow vortex according to the above equa-
tion. The relation is obviously not valid for Fig. 1 and 2
for radii inside the outlet ~2, because of deflector vanes 34,
but the equations describe the flow field in the annulus between
wall 42 and vanes 34.
Because the flow is proceeding rom the outside of the
vortex to a sink at the center of the vortex, the mass flow
rate in the radial direction through any cylindrical cut of
the vortex section will be the same ~outside of the outlet)
so that the radial velocity will be inversely proportional
to the radius for a mixing chamber 18 of constant axial height~
VrrOrO
r(r) r
where Vr( ) is the radial velocity at radius r, and VrrO is
the radial velocity at the outside radius of the vortex.
Clearly the above two equations are of the same form. It
follows that for a set tangential velocity input (set by a
given intake manifold vacuum) and a set volume throughput
through the vortex (set for a specifice engine rpm) the
ratio of the velocity tangential to the velocity radial will
be constant for all the radii of the vortex. Flow streamlines
in the vortex are therefore constant pitch inwardly flowing
spirals for ~ vortex chamber of constant axial height. If
g;~83
- 44 -
axial height of the mixer varies, the spiral pitch will
vary inversely as axial height as a function of radius.
The mean flow streamline pattern described above is
a good approximation of the real flow in a vortex mixer such
as that shown as mixing chamber 18 if certain fluidic
details are tended to. So long as turbulence levels are
sufficiently low and boundary layer flows adjacent top 47
and bottom 43 are controlled, the physical relations of con-
servation of angular momentum make the mean flow streamlines
in the real flow rather close to the pattern of an irrota-
tional flow vortex. In the real flow pattern the mean
pattern has relatively fine scale turbulent perturbations
superimposed upon it. 'rhe resulting combination of large
scale patterning with fine scale turbulence is useful for
mixing. It should be clear that drag interactions between
successive radial elements will tend to reduce the velocity
increase of the flow as it flows towa~ds the center because
the angular momentum as the flow flows towards the center
will decay because of these drag losses. Too much flow
turbulence can increase these drag interactions to the point
that the irrotational vortex flow pattern is destroyed. For
this reason, the entrance section of a vortex mixer must be
designed with care, so that the flow delivered to the vortex
chamber is not too turbulent. Nonetheless, the irrotational
flow vortex form, as a flow mode, is extremely stable, and
is representative of the flow inside a properly designed
vortex mixer over its full operating range.
The interaction between mean flow streamlines and
turbulence is a most important one if one is to understand
- ~5 ~ ~1Z9~3
mixing. We will be considering here turbulence levels small
enough that they do not destroy the basic irrotational vortex
pattern. ~ consideration of Figures 8 and 9 should clarify
some of the points important with respect to understanding
of the interaction between flow structure and molecular and
turbulent diffusive mixing. It should be emphasized that the
graphical illustration of Figure 7, Figure 8, and Figure 9
are exemplary only. However, the examples are important ones.
Figure 7 shows a streamline 154 of a vortex from an outside
radius 150to a sink152 where the streamline obeys the flow
equations previously discussed. This flow streamline would
occur, for example, in an irrotational flow vortex where the
streamline was well away from entrance condition perturbations
and where turbulence in the vortex was zero, if one were at
point156 to introduce, for example, ink into a water vortex
and watch the ink line as it flows towards the sink. The
... ..
streamline,in other words, shows what"the flow path would be
in the absence of any random mixing, either by turbulent
diffusion or by molecular diffusion. If there were any dif-
fusion, the width of the line would increase as it flowed
inwardly towards the sink, as should be clear to those who
understand mixing. In summary, Figure 7 would show a flow
streamline for an irrotational flow vortex if a line of mixant
was introduced at only one point along the outside of the vortex
and in the absence of either molecular or turbulent diffusion.
Figure 8 shows what would happen if the same flow situa-
tion as that of Figure 7 had an additional line of mixant intro-
duced'180 around from the initiai point of introduction. The
~ 4~ 3
vor~c~ would have an outsicle circle157 and a sink 158. At point
160 alonq the circle 157 a line of mixant would be introduced159.
The numbers 159are shown as the flow swirls in towards the sink
to identify that streamline. 180 from point 160 along circle
157 mixant is introduced at162 and produces flow streamline 161.
Flow streamline161 is identified at several points to make it
clear the manner in which the spirall59 and the spiral 161 nest.
~gain, Figure 8 illustrates what would happen in a mathe-
matically perfect irrotational flow vortex with a sink, in
the absence of either molecular diffusion or turbulent dif-
fusion.
Figure 9 is analogous to Figure 8, except now, rather
than having two nested spiral streamlines, mixant would be
introduced evenly around 10 points around the circumference of
the vortex; and therefore, 10 different spiral lirles would
nest as shown.
With respect to Figures 7, 8 and 9, it should be clear
that the presence of small-scale turbulent perturbations and
moiecular diffusion would tend to thicken out the lines as
they flow from the outside towards the sink of the vortex and
therefore that the mixiny pattern would be more and more
homogeneous as the mixture flowed inwardly towards the sink
of the vortex. For example, with respect to Figure 9 it should
be clear that only a relatively small spreading angle of the
mixant lines (corresponding to a relatively small turbulence
intensity) would so smear out the lines of mixant by the time
the flow had spiraled from the outside of the vortex to the
sink, that the mixture at the sink of the vortex would be very
- 47 -
much homo~3eneo~ls. With respect to the nesting of spiral .~tream-
line patterns shown in Figures 8 and 9, it should be pointed
out that tAe mixing chamber 18 of the present invention will
be evapora~ing mixture around the entire periphery ~2 of the
vortex. Because of the physics of the boundary layer tur-
bulence on this ou$side wall, the flow structure will act to
introduce the mixant (fuel), not just around 10 points around the
periphery, but around an effectively infinite number of points
around the periphery. This means that if the liquid is well
distributed around the circumference of the vortex peripheral
wall (a pqint which will be discussed subsequently), the mean
distance across which diffusion needs to occur in order to
achieve essentially perfect homogeneity at the vortex sink is
very short.
A consideration of the turbulent or molecular diffusion
differential equation should make clear that an n-fold decrease
in the mean distance across which diffusion needs to occur, for
a set interfacial area, will decrease the time required for
equilibrium by a factor of n. But the effect is even stronger
Introduction of mixant from many points around the periphery
of the vortex is tantamount to very vastly increasing the inter-
facial area across which diffusion can occur. Of course, this
effect increases mixing rates too. Again, it must be emphasized
that the flow streamlines shown in Figures 7, 8 and 9 are only
exemplary. However, the geometrical reIations with respect to
mixing illustrated by these figures are extremely important and
do not become less important as the flow streamline structures
become more complex; for any given flow structure, the flow
structure will serve to stretch out the concentration gradients
.
r
- 48 - ~129~3
of speci~s to be mixed and therefore, the flow structure will
dramatically affect the rate at which mixing proceeds. Mathe-
matically, the flow structure, or non-random streamline pat-
tern, can be ~lought of as a spatial transform of concentration
fields as a function of time. There are flow transforms which
are very conducive to mixing. The irrotational flow vortex
is such a flow trans~orm. However, it should be clear that
many other flow patterns which are not exactly irrotational
flows can also have flow patterns very much conducive to
mixing. For example, the flow pattern in the vortex of the
present invention will not be a perfect irrotational flow vor-
tex. However, with respect to the spiral streamlines, it will
differ from a conventional irrotational flow vortex only in that
the ratio of tangential to radial velocity will not quite be
constant as a function of radius for the real flow.
The actual flow patterns in the device have been tested
with several Reynolds number models during development of the
invention. Reynolds number modelling of the flow pattern pro-
duced by the present invention using water as the model and ink
as tracer shows that the flow pattern which actually occurs in
the system is much like an irrotational flow vortex and that
the flow pattern is extremely conducive to mixing. In fact,
when a single point mixant introduction (using ink from a
syringe) was used, the mixing was so rapid that the flow looked
effectively homogeneous well before the flow reached the outlet
of the vortex. Consideration of the flow nesting relations in
Figure 7, Figure 8,and Figure 9 should make it clear that the
mixing must have been even better for the multiple mixant
introduction case in the real vortex where fuel is distributed
- 49 ~ 11~ 3
aro~lnd the circumfererlce of the outer wall of the vortex.
While it is recognized that a Reynolds number model operating
on water with ink as a tracer and a plexiglass one-to-one model
is not quite a perfectly analogous modelling ~because the water
is not compressible as the air is~, the analogy is still a
close one, and the mixing observed in the system was very, very
intense, so that even significant decrements in mixing rates
due to compressibility effects (which are not likely) would not
affect the conclusion that the vortex flow pattern actually
produced in the present invention system is extremely conducive
to mixing. Boundary layer turbulence on the outside peri-
pheral wall produces an effecti~e multiple fuel point introduc-
tion characteristic for the evaporation process of the present
invention. The Combination of the fuel evaporation pattern,
and the overall flow pattern produces mixing rates so fast that
the mixture is effectively homogeneous at the mixer outlet.
An early plexiglass Reynolds number model was also used
for modelling the evaporation and mixing process, taking into
account compressibility. The experimental results were impor-
tant ones, and the experiment should be repeated in the engine-
ering development of any vortex mixer. Using a relatively con-
ventional carburetor pressurized by shop air so that the vol-
umetric flow rates and Reynolds numbers through the vortex
channel could be controlled to be in the range of those charac-
teristic of an operating engine, water was introduced to the
vortex. As expected, the water droplets were driven and depo-
sited o~to the outside peripheral wall of the vortex chamber.
The flow pattern meant that water could only leave the vortex chamber
by means of evaporation into the air which passed out the central
1129Z83
- 50 -
outlet of ~he device. With a thermocouple at the peripheral
wall of the vortex, the temperature of the liquid at the wall
was measured. This temperature exactly determined the vapor
pressure of the water at the outside wall. Then the wet bulb and
dry bulb temperature ot the air leaving the vortex chamber was
measured. This determined the vapor pressure of water in
the outlet air. It was found that: (1) for steady-state
conditions, there was no detectable difference in either the
humidity or the temperature of the air exiting the vortex chamber
outlet from radial position to radial position. (2) The vapor pres-
sure of water in the air at the vortex chamber outlet was always
more than 60 percent of the vapor pressure of the water inside
the vortex at the peripheral wall. Under many conditions,
the vapor pressure at the outlet was more than 75 percent of
the liquid water vapor pressure. A number of different flow
conditions were run to collect this data. Consistently, the
smallest fractions of the water vapor pressure occurred under
flow conditions analogous to wide-open throttle type engine
operation.
The water evaporation experimental sequence shows a
great deal about the mixing and evaporation rates in the vor-
tex system. First, mixing in the vortex had to be very
excellent indeed, even down to microscale volumes~ Since the
flow system could not produce supersaturated vapor, the worst
mixing situation would be one where the vortex generated 60
percent plus saturated vapor with the balance dry air on
micronic volume scales. Clearly, the turbulent conditions in
the vortex had to produce a more mixed situation than this.
However, it must be emphasized that even this "worst caæ "
microscale heterogeneity would produce very excellent homo-
5 l 1~Z9~83
geneity ~or volume samples of the order of a cylinder displace-
ment volume. There is reason to believe that the standard
deviation of concentration would be very small on any scale
which could be conveniently measured.
Secondly, the experiment strongly indicates that the
wall temperature in a vortex mixer operating in an engine
burning gasoline would need to be only a little hotter than
the equilibrium air distillation temperature corresponding to
the intake manifold pressure and mixture stoichiometry for
a set engine condition, in ordex tocompletely evaporate , the
fuel under steady-state, steady-flow conditions. If the peri-
pheral wall of the vortex were operated at a temperature sig-
nificantly higher than the equilibrium air distillation temp-
erature corresponding to the mixture conditions, only a frac-
tion of the peripheral wall would need to be wetted by the
fuel in order to produce complete evaporation. This is prac-
tically important. The vapor pressur~ curves for most pure
hydrocarbons are such that an 100F increase in temperature
increases vapor pressure by roughly a factor of 6 (5 to 7).
Suppose the peripheral wall was heated to a temperature 100F
above that of the equilibrium air distillation temperature of
the mixture (which is often below 140F). This would mean that
if the entire peripheral wall of the vortex were wetted with
gasoline, a roughly 6 times stoichiometric mixture would be
produced by the vortex (for equilibrium reasons most of this
fuel would recondense as smoke-sized particles, but this fact
is of little interest to the present argument).
The evaporation time of the fuel under these conditions would
have to be very fast. The following equilibrium argument shows why
the equilibrium process is fast. Consider the equilibrium situation
r~ 2 ~ 9~83
of the vort~ as an evaporator receivillg a se-t fuel-air ratio.
If too much of the peripheral wall is wetted by fuel, fuel will
evaporate off the wall faster than fuel is deposited onto the
wall from the carburetor, and the wall will tend to dry out.
For similar reasons, if less of the wall is wet than that
required for steady~state conditions, fuel will be deposited
on the wall faster than fuel is evaporated, so the surface area
of the wall wet by fuel will increase. Because the wall is
continuously and vigorously sprayed with fuel droplets from the
inlet passage and also with condensing droplets which are too
large to flow through the centrifugally separating vortex flow
field, a too dry wall will be wet to equilibrium very fast.
Evaporation and flow relations are such that a too wet wall will
evaporate to equilibrium wetness very fast. The equilibrium fuel
evaporation process for a hot vortex surface and a strong flow
pattern is therefore very fast.
For an appreciation of how rapidly drop]ets above 5
microns must deposit on the vortex wall, see Environ-
mental Protection Agency report under Contract ~o. EPA-70-20
"A Study of the Influence of Fuel Atomization, Vaporization,
and Mixing Processes on Pollutant Emissions from Motor-Vehicle
Power Plants,".Both the phase-one and phase-two reports were
conducted by Battelle I,abs. In a strong vortex flow field
droplets so impacted need not stick to the vortex chamber
peripheral wall, but may splash in such a way as to form a
"splash cloud" of fuel droplets rotating around the outside
of the vortex at velocities near the velocities of the vor-
tex airflow itself, Such a splash cloud distributes fuel
droplets around the vortex peripheral wall quite evenly.
.
1~29283
When the peripheral vortex wall is hot enough to boil the lighter ends
of the fuel this sort of splash cloud forms readily for a properly
designed vortex mixer. If only one-sixth of the vortex peripheral
wall needs to be wetted by the fuel, the equilibrium amount of fuel
in the splash cloud around the peripheral wall of the vortex will be
quite small, and the mean transit lag of evaporation (the time lag
between fuel passage through the vortex and air passage through the
vortex) will be of the order of 50 milliseconds or less.
The inventors have observed a vortex mixer operating on an
engine on conventional gasoline, by looking inside the vortex mixing
chamber through a viewing window such as will be described with
reference to Figure 12. Fuel was introduced just upstream of the
main vortex flow field by means of an injection nozzle. The window
made it possible to see the bottom and peripheral walls of the vortex
mixer. When the peripheral vortex wall analogous to wall 42 was
more than 20C above the equilibrium air distillation temperature
characteristic of the operating mixture, it was not possible with
the naked eye to detect any trace of liquid phase in the vortex
mixer, even though liquid phase was present in a splash cloud
around the vortex peripheral wall. Neither the shininess
characteristic of wetting nor any blurring was visible to the
eye, testifying to the physical reality of the splash cloud.
Thermocouple readings around this vortex peripheral wall made
it clear that evaporative heat transfer was occurring all
around the peripheral wall, again demonstrating the reality
of the splash cloud. The set-up described above had fairly
sensitive instrumen~ation attached to it. With this
-53-
. _ 54 _ 1129283
instrumentation the evaporation lag in the hot vortex and the
cylinder to cylinder A/F variation were both too small to be
detected.When the vortex was near or below EAD temperature,lags were
of course detected. The present invention vortex system very much
reduces the time lag of the fuel with respect to the air. This is
a substantial practical advantage of the current device. It
should be clear that the time lag of the system will be the
less, the greater the surface temperature about the peripheral
wall of the vortex (the greater the excess of the surface tem-
perature over the equilibrium air distillation temperature
required for the particular fuel and stoichiometry on which
the engine operates), and the stronger the vortex flow pattern.
The above argument refers to the vortex flow passage
per se. No matter how short the lag in the vortex chamber itself,
significant fuel lags can exist in the entrance section passage,
and the system lag will depend on how fuel is introduced into the
. .
vortex chamber. With decent understanding of the fluidics of the
flow, very short lags are attainable.
A designer of a vortex system should pay great atten-
tion to Reynolds number modelling. Reynolds number modelling
with air and water is useful Eor predicting evaporation lags,
as was discussed above. (An advantage of water and air is
that they are non-toxic and non-fla~able). In addition,
Reynolds number modelling in a vortex mixer fabricated of
transparent material with water as the working fluid and
with ink injected at various points through syringe needles
as tracer will give an appreciation for the details of the
flow pattern which can be obtained in no other way. A skill2d
fluid mechanics man, working with a Reynolds nun~er water
` ~ 55 llZ9Z~3
model o~ a syecific vortex system will be able to quickly see
any fluid mechanical problems which occur and should be able
to correct them. Reynolds number modelling with water is in-
dispensible because it permits direct visualization of the
(inescapably somewhat complex) flow patterns. A Reynolds number
model useful for water studies can be readily fabricated out of
acrylic plastic or other transparent plastic. Simple pressure
gauges and water flow gauges will permit Reynolds number
analogies to be established. The water model technique can
also be made to yield mixing statistics data by injecting salt
water into the system at known points and then detecting the
electrical resistance of the water somewhere downstream from
the mixant introduction with a imple ionic probe (linked to
a square wave generator to eliminate ionic corrosion drift).
The ionic probe technique gives resolution of mixing statis-
tics down to volumes of cubic millimeters or even smaller
volumes.
It should be relatively clear to skilled automotive
engineers, and certainly to anyone skilled in the art of
fluidics, that the efficiency of the vortex mixer is largely
dependent on the fluidic efficiency of the vortex entrance
section leading in to the vortex flow chamber per se. This
fluidic efficiency is defined as the fraction of the isentropic
velocity past the throttling restriction delivered tangentially
and in coherent fo~m to drive the vortex in the vor-
tex mixing chamber (for example, mixing cha~ber 1~ of ~igs.
1 and 2). The high speed flow from the air controlling throttle
is desired along the vortex walls roughly parallel to the vor-
tex outlet central axis 35, and not along the ~ortex walls
- 56 - 1~2~Z83
perpendicular to the vortex outlet central axis. For fluidic
rea~ons discussed previously, a particularly convenient place
for the wall attached streams driving the vortex is in the
corners of the inlet section, since wall attached streams in
corners of channels decay relatively slowly. So long as step
up is avoided, the fluidics in a rectangular throttle entrance
section such as that of Figure 3 is relatively straight~
forward. The fluidics in a round throttle to trapezoidal
entrance section channel is more complicated, and it may be
useful to set up an experimental arrangement wherein shop air
is delivered to the throttle plate, and the velocities of the
flow field at the plane of A~ in Figure 1 are analyzed by means
of Pitot tubes. A yet better alternative would be an arrangement
for Pitot tube testing of a flow channel such as that shown
in Figure 5. With simple Pitot tubes made from disposable
syringes, a few inclined manometers and proper measurement
of the pressure drop across the throttle plate, useful flow -
maps of a proposed inlet section can be made. Such flow
maps can be extremely useful in the perfecting of the flow
details in an entrance section analogous to that shown in
Figures 4 and 5.
The basic functioning of the present invention device
should by now be clear, and it should be relatively clear that
the function of the device rests on the interaction of evapora-
tion, molecular diffusion, turbulent diffusion, and the gross
effects of the flow structure. However, it should be said
that the flow structure which is most desirable requires a
bit of design care. The passage-where the flow is introduced
having angular momentum about the channel outlet and where the
passage is shaped to minimize mo~entum loss in the system has
. .
' - 57 -1~29~83
been in existence for a long time. For a detailed discussion
of vortex devices, see Chapter 8 of Design Theory of Fluidic
Components by J. M. Kirshner and Silas Katz, Academic Press,
1975. Figuresl0 and 11 are taken from page 281 of this book
and show the flow pattern which can be produced if the circum-
ferential weirs 41, such as are shown in Figure 1 and Figure 2,
are not used. Figure 10is a view of the flow perpendicular to
the axis of rotation of the vortex, and Figure 11 is a diame-
tral section showing streamlines for the flow of FigurelO. What
is called the developed region (or doughnut) is caused because
of a boundary layer effect. The centrifugal forces in the
flow in the vortexare important to determining the flow pattern.
Centrifugal force is proportional to re , and is therefore
proportional to velocity squared. At the top and bottom sur-
faces of the vortex-containing channel, viscous forces slow
down the flow in and near the boundary layer. This means that
the centrifugal force in the vortex near the wall is much less
than it would be in the center, ~nd the result is t,hat the
radial velocity of flow towards the sink is greater near the
walls'of the vortex than it is in the vortex center. The
effect is so large that the recirculating doughnut flow shown
in Figures 10and 11 often occurs. One of the difficulties is
that this doughnut flow diameter will vary with the Reynolds
number at which the device is operating and as the ratio of
radial velocity to tangential velocity varies and so can
produce unfortunate modal characteristics with respect to
the mixing device. Clearly, the simple irrotational flow
vortex flow form is a preferable flow form. It has better
mixing rates, it is simpler and its equations are not modal,
. . . .
- 58 ~ 1~9Z83
so that the irrotational flow vorte~ will be stable above a
certain minimum Reynolds number. It happens that this mini-
mum Reynolds number is lower than any Reynolds number which
would occur in an operating engine. To achieve ~his approxima~
tion of the irrotational vortex flow pattern, it is necessary
to condition the boundary layer flows.
This boundary layer control can be obtained with circum-
ferential weirs such as 41 shown in Figure 1 and Figure 2.
Figure 12 shows the flow pattern which is produced due to
these weirs where Figure 12 is a diametral half section of a
vortex channel such as that shown in Fig. 14 and shows the
velocities with respect to the radial direction tit should be
clear that very significant tangential velocities, which are
not shown in ~i~ure 12, also exist in and out of the pattern).
The effect of the circumferential weirs 31 is to stabilize
small vortical flows between the weir$ in such a manner that
the effective boundary layer flow is well lubricated and
where the great bulk of the flow energy in the vortex is in
the form of a simple irrotational vortex flow. This has an
advantage with respect to the transfer characteristic of the
device because with the flow in an irrotational flow mode the
statistical variation in transit time from fluid ele~ent to
fluid element can be held to a minimum; this is useful for
servo-mechanical design, particularly for systems such as
3-way catalyst systems. It should be emphasized that the
device w;ll produce significant mixing without circumferen-
cial weirs such as 81. However, operation with the weirs is
preferable.
- 59 ~ ~29283
Figure 12 shows an enlarged sectional view of a
vortex mixer such as that shown in Figures 13 and 14, with
the addition of a transparent window 84 on the top of the
vortex outlet. Window 84 makes possible the viewing of the
vortex mixer flow channel below the point roughly correspond-
ing to point 86, during actual operation of the vo~tex with
a running engine. The window, therefore, made possible viewing
of the bottom of the vortex, including the weirs, as well as
a considerable portion of the vortex peripheral wall. In
the setup where the window was installed experimentally,
the window also made it possible to see the volume in space
into which the injection nozzle was directly in~ecting fuel.
Viewing the operating vortex through window 84 was extremely
revealing. When the vortex was cold, during start-up, liquid
fuel would accumulate around the outside peripheral wall of
the vortex and a small amount of liquid fuel was visible -
between the weirs 81. This liquid fuel between weirs 81`
served as a flow tracer, and made it clear that the flow
between the weirs and the operating vortex was very much like
the flow between weirs 81 shown in Figure 12. Liquid
striations showed the general spiraling pattern between weirs.
The liquid accumulated along the outside of the weirs, at a
maximu~ radial distance from the vortex outlet center, as
would be expected for the weir boundary layer conditioning
flows shown in Figures 12 and 12a. This visual evidence, in
combination with analogous evidence obtained with ink in a
water Reynold's number model, is extremely good evidence that
- 60 ~ ~ 1 ~9 2 8 3
the flow field is as shown in Figure lZ. In addition, when
the vortex mixer was warmed so that its peripheral wall
temperatures were more than 20 C hokter than the equilibrium
air distillation temperature of the fuel air mixture
delivered to the engine, it was not possible, viewing through
the observation window, to see any visually identifiable trace
of liquid phase within the view field corresponding to the
zone below point 86 in Figure 12. This visual data, in com-
bination with thermocouple readings demonstrating that evapora-
tion was occurring all around the peripheral wall of the vortex,
constitutes strong evidence for the reality of the splash-
cloud fluid mechanics discussed previously. It is easy enough
to install an observation window analogous to 84, and easy
enough to view the actual vortex mixer chamber through deflec-
tion vanes 82. It is therefore a worthwhile check on vortex
operations to install such an observation window in any --
serious vortex development program. Generally, the tempera-
tures of the vortex are low enough that the observation
window may be readily fabricated of acrylic plastic, and
sealed with 0-ring type seals.
The vortex mixing chamber shown in Figures l and 2,
in combination either with the entrance geometry shown in
Figure 3 or with the entrance geometry shown in Figures 4
and 5, is the preferred form of the vortex mixer invention.
With the mixer shown in Figures l and 2, all of the throttling
pressure drop across the air-flow controlling throttle is
available to drive the flow patterns of the vortex mixer.
~ - 61 ~ 11Z92~3
For an entrance section similar to that shown in Figures 4
and 5, the mixer of Figures 1 and 2 can be adapted to a
side-draft type of carburetor. However, the arrangement
shown in Figures 1 and 2 is no-t adaptable to the more con-
ventional sort of down-draft carburetor.
In the arrangement shown in Figures 13 and 14, the
fluid mechanics and structure is in many ways analogous to
the fluid mechanics and structure already shown and discussed
with ref~rence to Figures 1 and 2. To emphasize the corres-
pondence of the parts, the numbers in Figures 13 and 14 are
two hundred series numbers with the last two digits corres-
ponding to the two digits of the analogous part referred to
in Figures 1 and 2. For example, weirs 41 in Figur~s l and 2
correspond to weirs 241 in Figures 13 and 14.
See Figure 13. The vortex chamber per se 218 of
Figures 13 and 14 corresponds in all details to the vortex
chamber per se 18 in Figures 1 and 2. The vortex mixer
chamber 218 discharges into outlet 232 which is connected to
an intake manifold (not shown) as before. Vortex chamber
peripheral wall 242, weirs 241, outlet 232 and outlet
deflector vanes 234 each function in a manner analogous to
the corresponding parts in Figures 1 and 2. However, the
mixer of Figures 13 and 14 differs from that shown in Figures
1 and 2 with respect to the entrance section geometry,
although significant fluid mechanical analogies between the
two sorts of entrance geometries are readily apparent. In
- 62 - llZ9Z83
Figure 13, the entrance section geometry is adapted to
receive an intake of fuel-air mixture from a conventional
down-draft two barrel carbuertor (not shown) at openings
250. The flow from the carburet(or flows past openings 250
into entrance source chamber 201, which is heated and inside
which a significant amount of fuel evaporation occurs. Mix-
ture from chamber 201 flows past a curved throttle plate 204
pivotting on shaft 205 and linked to the throttle linkage
of the carburetor (not shown). The linkage between throttle
204 and carburetor must be arranged so that the pressure
drop across throttle plate 204 is enough to drive a strong
vortex, but still a low enough pressure drop to be tolerable
with respect to the metering characteristics of the carburetor.
This carburetor metering constraint on the linkage relations
which control throttle 204 generally limits the pressure drop
across throttle 204 to something less than 5 inches of mercury -
pressure drop. Throttle 204 is a curved roughly trapezoidal
throttle plate with parallel top and bottom sealing surfaces.
On the edge labelled 210 of throttle 204, at the bottom (not
shown) is a notch (not shown) large enough to produce a pres-
sure drop of only a few inches of mercury under conditions
wh~en the carburetor supplying entrance passages 250 is in its
idle configuration. Flow past this notch attaches to a corner
of the entrance section and flows into the vortex in a manner
which has already been discussed with reference to Figures 1,
2, 3, 4 and 5. Provision of this notch, which arranges for
the bulk of the idle flow to flow into the vortex chamber as
'
~lZ9Z83
a wall corner attached stream, produces a stronger vortex
in vortex chamber 218 under idle and off idle conditions
than would occur without the notch.
A major difference between the mixer shown in Figs.
13 and 14 and the mixer shown in Figs. 1 and 2 is that the
downdraft carburetor mixer of Figs. 13 and 14
requires heating of the surfaces of entrance
201. To control lags, fuel air mixture in
entrance 201 must be heated as it flows from opening 250
to the gap controlled by throttle plate 204. For this
reason, exhaust heated fins are arranged all the way
around the vortex mixer, and the entrance section is curved
to allow this circumferential heating. To provide uniform
heat exchange from the heat exchanger fins to a portion of
wall 242 and a portion of the chamber wall of entrance sec-
tion 201, heat piped heat exchanger passage 254 is provi-
ded. This passage contains a small volume of water, but
is otherwise evacuated so that the only gas within the
chamber is water vapor. Because of this, the liquid
interface of water inside 254 is always at its boiling
and at its condensation temperature. The equilibration
of temperature within passage 254 is therefore very fast.
If any surface of passage 254 is cooler than the water
interface, water vapor will rapidly condense upon it,
providing heat of vaporization heat transfer to this
cooler surface. Similarly, if any part of the liquid
contracting surface of the passage 254 is hotter than
-63-
- 64 -
l~Z92~3
another surface on 254, water at this hot surface will
boil, providing very rapid heat transfer from the hot
(evapora~ing) to the cool (condensing) surface. The
result of this is that heat transfer within the volume
of passage 254 is extremely rapid. Therefore the tempera-
ture of the surfaces of the passage 254 is quite uniform.
Heat pipe passages analogous to that of 254 are quite
convenient in providing uniform heating to extended
surfacçs, and are useful in the heat exchanger design for
a vortex mixer.
The mixer shown In Figures 13 and 14 has some disad-
vantages in comparison to that shown with reference in Figs.
1, 2, 3, 4, and 5. First, there is a fuel lag in entrance
source chamber 201 and this introduces a perceptible, though
relatively small, lag in the system no matter how fast the
mixing relationships downstream of auxiliary throttle 204
may beO Secondly, the mixer of Figs. 13 and 14 involve an
additional air throttle 204 downstream of the air flow
throttles already provided with the down-draft carburetor,
and the carburetor throttles and the auxiliary throttle
204 must be carefully linked if proper carburetor metering
is to be maintained. Also, the requirement that the en-
trance section be heated complicates heat exchanger arrange-
ments. However, the vortex mixer described in Figures 13
and 14 works very well. Experimental results obtained on
a mixer similar to that shown in Figures 13 and 14 will be
discussed later.
- 65 -
llZ9283
The vortex mixers shown in Figuxes 1, 2, 3, 4, 5, and
13 and 14 stabilize vortices inside their mixing chambers
18 or 218 which are excellently approximated as irrotational
vortex flows. However, vortex mixers involving a less perfect
vortex flow field also function and produce excellent mixing.
The first vortex mixer which the inventors actually operated
on a multi-cylinder engine is illustrated in Figures 15, 16,
17 and 18. Figure 15 i9 a diametrical cross section-of the
mixer, involving a section through the central passage of a
one venturi down-draft type carburetor. Figure 16 is a top
plan view of Figure 15 taken along sectional line 16-16 with
the top of the device removed to show the disposition of the
angular weirs and deflectors. Figure 17 is a sectional of
the carburetor and its corresponding deflectors 342 and 343
along sectional linel7-170f Figure 15. Figure 18 is a view
of deflectors 342 and 343 along sectional line 18-18 of Figure
17.
Many of the fluid mechanical characteristics of the
mixer shown in Figures 15 to 18 should be clear to the reader.
High velocity flow past the throttle valve 344 is de1ected
by deflectors 342 and 343 and introduced tangentially at a
significant radial distance from the center of the mixing
chamber. This tangentially introduced flow drives a generally
vortical flow pattern in the mixer, which flows to the outlet
3Z0 in an inwardly spiralling fashion until it reacts with
deflector vanes 348~. In the outlet passage of the mixer is
- 66 ~ 9 Z8 3
faired portion 346, which somewhat reduces the flow resistance
of the vortex mixer but is a less efficient flow resistance
reducer than the optional variable spoiler assembly 36, 38,
40 shown in Figs. 1 and 2. The fluid mechanics of the mixer
shown in Figs. 15, 16, 17 and 18 is analogous to ~hat previ-
ously discussed with reference to Figures 1, 2, 3, 4, and 5,
in that it is somewhat sensitive to fluid mechanical details.
Specifically, the arrangement of the deflector vanes 342 and
343 (and specifically the setback y of deElector 343 shown in
Pigure 17) is quite important to the function of the device.
The inventors found experimentally that if these deflectors
were carefully designed, the tangential flow driving the
vortex would be sufficiently high in velocity and low in
turbulence to establish a flow field much like an irrotational
vortex in the mixer. When this approximately irrotational
flow vortex was present, mixing was excellent in the vortex
device. However, if details of the design of the deflectors
343 and 342 were not proper, the turbulence of the flow de-
livered into the vortex chamber was sufficiently high that
the swirling flow within the vortex chamber decayed into a
flow pattern approximating a rotational vortex where the
angular velocity of the streamlines was approximately constant
along the radius from the vortex outlet center Mixing rates
with this rotational vortex were drastically less than those
which occurred with the irrotational vortex, for reasons which
should be clear upon reconsideration of Figures 6, 7, 8 and 9.
, .
- 67 ~ 112g
Figures 15 and 16 show a number of additional features.
The heat exc~anger passage 334 includes within it heat resis-
tant insulated material 33~ to assist the warm up character-
istics of the vortex mixer. In addition, electrical resis-
tance air heating wires 310 are shown schemati~ally upstream
of the choke of the carburetor 326. For a mixed boiling point
fuel including light boiling components a cold-starting mixture
can be achieved by choking until a sufficiency of light end
components are available in the vortex chamber to produce an
ignitable mixture for cold start-up. However, for pure com-
pounds, the maximum air fuel ratio is propor,'ional to the
vapor pressure of the pure compound fuel as a function of
temperature. Below a certain temperature the compound
vapor pressure is too low to evaporate an ignitable mixture,
no matter how much liquid phase is in contact with the air
for evaporation. Therefore, for an engine equipped with a
vortex mixer (which eliminates liquid phase flow to the en-
gine), there exis-ts a minimum start-up temperature for any
pure compound fuel unless there is provision for heat addition
at start-up. The electrical air resistance wires 310 do pro-
vide this heat addition for start-up. The load on the
battery required for this start-up means is moderate, and the
vortex chamber rapidly heats up by exhaust gas flow after
engine light-off, so that electrical resistance wires 310 do
constitute a practical arrangement for cold starting of a vor-
tex mixer.
1129283
With a vortex mixer of Figures 15, 16, 17 and 18, the
mixing is good enough that air fuel ratio variations from
cylinder to cylinder was undetectable experimentally under
many conditions. However, the disadvantages of
this arrangement with respect to wide-open throttle flow res-
istance and fuel evaporation transient response motivated the
development of the vorte,x mixers described in Figs, 1-5, 13
and 14.
..
. . .~ 1
. - 69 ~ 1~292~3
~lixin~ Matters for Emissions_and Fuel Economy
A perceptive man skilled in the automotive engineering
arts may have concluded at this point in the disclosure that
the inventors are guilty of overkill, since they have dis-
closed means to mix fuel and air on much finer scales than
the scales which would be required to achieve essentially
perfect cylinder-to-cylinder air fuel ratio distribution.
It is the conventional wisdom in automotive engineering that
mixing on,ly matters insofar as it generates relatively tight
cylinder-to-cylinder mixture variation. However, the inven-
tors have worked for a long time in the area of super homo-
geneous charge very lean engine combustion, and have shown
that significant advantages are available if very homogeneous
mixtures are burned very lean in an internal combustion engine.
The fuel economy and emission control advantages of mixture
homogeneity will be partially explained with reference to
Figures 19 and 20.
Figure 19 is a mathematically derived numerical
example used to make clear a basic statistical argument.
Figure 19 plots three different gaussian distributions,
curve A, curve B, and curve C, each of which has the same
total area under the distribution curve, but with the curves
having different standard deviations. The curves plotted are
for gaussian distributions (reasonably mixed systems are
nearly gaussian) where the sample size of the distribution is
~129~83
so large that the curve is continuous and smooth. For microscale mixing
discussions, the smooth distribution (high population approximation) is well
justified. Figure 19 attempts to show in a visually clear way that the
lean limit of satisfactory engine combustion shifts leaner (approaching the
lean limit for perfect mixing) as the mixture becomes more and more homogen-
eous. An engine with excellent mixing will operate smoothly and efficient-
ly on mixtures very much leaner than would be tolerable for an engine with
inferior mixing. The details of the argument with respect to Figure 19 will
follow later. However, the importance of mixture homogeneity, and the emis-
sion advantages which mixture homogeneity permits, are made more clear withrespect to the data plotted in Figure 20, which shows that the nitrous ox-
ide outputs available with very homogeneous combustion are very low under
conditions where engine efficiency is optimal.
Figure 20 plots data taken by the inventors at the ENGINE RESEARCH
LABORATORY of the University of Wisconsin Department of Mechanical Engineer-
ing under the close supervision of Professors P. S. Myers and O. A. Uyehara.
All data points in Figure 20 are for minimum best torque spark advance. The
work plots nitrous oxide per indicated horsepower hour (as grams NO2 per
indicated horsepower hour) versus the equivalence ratio of
- 70 -
:,
- 71 - ~ 2~ ~8 ~
mixture burned for three different mixing cases. In the
first case, the propane fuel was introduced into the intake
port of the test engine and flowed past a rather conventional
intake port and intake valve arrangement into the engine
combustion chamber,'where it was burned. In the second
case, fuel was introduced at the intake port and flowed past
a variable flap restriction which fluidically controlled the
flow energy and flow pattern in the combustion chamber (in
fact producing an irrotational vortex in the combustion
chamber) so that mixing was quite good at the time the spark
lit. In the third case, the fuel air mixture was fed to
the engine with the variable restriction flap arrangement,
but was in addition mixed prior to introduction past the
flap with'a primitive early vortex mixer design developed by
the inventors. Since the fuel for these tests was propane,
this vortex mixer was unheated. The nitrous oxide output
level plotted in Figure 20 covers a range of more than a
thousand to one, and for this reason the NOX level is plotted
on a logarithmic scale against equivalence ratio.
It may be worthwhile to relate the concept of equiva-
lence ratio to the somewhat more commonly known concept of
air-fuel ratio. Equivalence ratio is defined as the ratio
of the stoichiometric air-fuel ratio to the actual air-fuel
ratio. The test data of Figure 2~ is for propane. However,
the NOX formation with'gasoline is very similar to what occurs
when burning propane. Plotting equivalence ratio versus air-
_ 72 ~
.
fuel ratio for gasoline gives the following correspondences:
AIR-FIJEL RATIO
EQUIVAL~NCE ~ATIO FOR GASOLINE
1.0 14.7
9 16.3
.8 18.4
.7 ! '~ 1 . O
.6 24.5
.
Figure 20 graphs variation of NOX output versus equiva-
lence ration with the NOX output expressed as grams NO2 per
indicated horsepower hour on a l~garithmic scale, since the
NOX output varied by more than a factor of a thousand over
the range of the equivalence ratio plotted. The ordinate of
the graph is equivalence ratio, as defined above. In Figure
20 it is shown that NOX output is an extremely strong function
of equivalence ratio, and that the relationship between equiva-
lence ratio and NOX output is extremely steep in the very
lean range. Figure 20 also shows that with excellent mixing
the optimal fuel consumption air-fuel ratio shifts leaner and
to a lower NO}~ output value. The data show that the better
the mixing, the better the efficiency and the lower the NOX
output corresponding to the optimal equivalence ratio.
All the equivalence ratios plotted in Figure 20 are
leaner than the stoichiometric ratio. However, all the
mixtures which are relatively near the stoichiometric ratio
will be referred to as relatively rich ratios.
In Figure 20 it is shown that at the relatively rich
ratios characteristic of conventional engine operation, there
~, I
- 73 ~ 3
is not much advantage to extreme air-fuel mixing homogeneity.
Between an equivalence ratio of 1.0 and .8 stoichiometric,
the difference in NOx output between the various mixing-state
cases is not particularly large, nor are the differences very
convincing. However, in the range leaner than .8 equivalence
ratio, the situation changes drastically. For the case of
conventional port fuel injection (the no port flap restriction
case) the best fuel economy happens at an equivalence ratio
of .735, so that enleanment beyond this point results in quite
significant fuel economy penalties (which could not be shown
in this graph, which identifies fuel consumption only at
optimal points). From the equivalence ratios at very leanest
limit where engine operation was possible with this low mixing
arrangement (.59 equivalence ratio for the conventional mixing
case) to the stoichiometric air-fuel ratio, there is only a
little more than a factor of 10 change in NOx output in the
conventional mixing-state case which corresponds to the wide
open flap data. Furthermore, most of this NOx range involves
a fuel consumption penalty. The optimal indicated specific
fuel consumption level for the conventionally set-up engine
mixing case involves NOx outputs of almost half of the maximum
NOx outputs for this engine under the conditions shown in the
graph. It is results like these for conventionally set-up
engines which have convinced the automotive engin~ering pro-
fession that NO control via charge dilution is a relatively
unattractive approach incapable of producing the very low NOx
.
. - 74 - 1129283
output levels which are required by the Federal government
beyond the 1981 model year.
In the equivalence ratio range leaner than .8 equiva-
lence ratio, the NOx performance of the engines equipped with
better mixing is drastically better than that of the conven-
tionally mixed engine. Quite clearly, for a set equivalence
ratio leaner than .8 equivalence ratio, the variable restric-
tion en~ine with its superior in cylinder mixing had signifi-
cantly lower NO levels. For example, for the case plotted
called the .300" flap restriction port injected case, the
optimal fuel consumption (minimum indicated specific fuel
consumption point) occurred at an equivalence ratio of .59
equivalence ratio at an NOx ratio of .27 grams per indicated
horsepower hour. This represents a 48 fold reduction in NOx
level from the maximum for the variable restriction engine,
in contrast to only a 2.4 fold reduction in NOx output from
the maximum for the conventional engine set-up when comparing
the fuel consumption optimal NO output to maximum NOx output.
In addition, the fuel consumption (and therefore the efficiency~
with the variable restriction engine was significantly and
reproducibly better than the fuel consumption with optimal
settings for the engine with conventional levels of mixing
so that the drastic reduction in NOx was obtained with a
simultaneous (although relatively small) improvement in the
fuel economy of the engine. The advantages of mixing to very
complete leveIs of homogeneity are shown even more dramatically
for the case where the variable restriction port engine was
supplied with a mixture homo~enized by a vortex mixer. In
- 75 - llZ~3
this case the optimal indicated specific fuel consumption
is even lower than before, and the reduction in the NOX from
the maximum is a very large factor of 420. The relationship
between equivalence ratio and NO characteristic of the engine
described in Figure 20 is extremely close to that predicted
by chemical kinetic calculations. The small range of NOX
reduction characteristic of a conventional mixed engine is
due to incomplete mixing.
Careful consideration of Figure 20 should make several
points clear. First, in the very lean range mixing has a
strong effect on both efficiency and NOX output. Secondly,
with very homogeneous combustion quite low NO outputs are
attainable if mixing is good enough and the air-fuel ratio
delivered to the engine is properly programmed. The reasons
for the NOX versus equivalence ratio relationship shown in
Figure 20 involve the interaction of chemical kinetics and
mixing statistics. Chemical kinetics do not need to be
discussed in this case. However, if a designer is to take
advantage of the emissions and efficiency control improvements
made possible with the vortex mixerl he must understand certain
statistical issues. Figure 19 is an attempt to illustrate these
statistical issues.
It should be pointed out that the data of Figure 20
relate to an engine where the structure of the flow within the
combustion chamber itseIf was controlledj and that the mixing-
state within the combustion chamber was more homogeneous than
could be achieved with a vortex alone. A vortex, no matter
how well it mixes fuel and air, cannot affect the details of
.
- 76 ~9Z83
flow structure within the combustion chamber itsel and
cannot mix residual gases within the combustion chamber with
the fresh charge air-fuel mixture. The inventors believe that
it would be possible using the mixing vortex and the variable
restriction port in combination, to build engines having
optimal efficiency with NOx levels comfortably below the .4
grams NOx per mile emission level. However, the vortex mixer
used alone with proper fuel air ratio programming appears to
have the potential of passing the 1.0 gram per mile NOx
standard with lean combustion and in a trim involving signifi-
cant fuel economy advantages in comparison to conventional
pre-emission control engines and also advantages with respect
to three-way catalyst equipped engines. The vortex makes these
good results possible by producing extremely homogeneous fuel
air mixing.
Figure 19 gives a graphical explanation of how improved
mixing within the combustion chamber of an engine can widen
the equivalence ratio of dilution limits tEGR limits) which
permit stable and efficient combustion. Figure 19 illustrates
microscale mixture sample variations in a hypothetical engine
where the gross air-Euel-residual ratio from cycle-to-cycle or
cylinder-to-cylinder is fixed but where the mixing inside any
given cylinder is imperfect. Curve A, B, and C of Figure 19
plot distributions for microscale mixing sample volumes, where
the number of sample volumes in each distribution is very large
so that the smooth gaussian distribution shown is produced.
It should be emphasized that the curves of Figure 19 illustrate
~ 77 ~ llZ9~83
a numerical example of a statistical argument, and do not
constitute the results of measurement. Looking at Figure 19,
suppose that if the mixture within the spark plug gap in the
engine is leaner than .55 equivalence ratio at the time of ~'~
sparking, then misfire will infallibly occur (this is a '
worthwhile over simplification for present purpose). Since
there are about 105 spark plug gap volumes in one cylinder volume,
the continuous gaussian distributions shown are reasonable.
The areas under curves A, B, and C are equal. Curve
A has a mean equivalence ratio of .75, but has a standard
deviation of .1 equivalence ratio for its distribution. Under
the assumption that the fraction of the population of curve A
leaner than .55 equivalence ratio will represent the misfire
frequency, this mixture A will misfire in the engine about
2.25% of the time; by usual standards distribution A can be
said to be at its lean misfire limit at a ratio somewhat richer
that .75 stoichiometric. The standard deviation of the mixture
plotted on curve B is half the standard deviation of curve A,
or .05 equivalence ratio. A 2.5% misfire rate for distribution
quality B occurs at an overall equivalence ratio of .65
stoichiometric, and so an engine with mixing such as that shown
for curve B will have a misfire limit richer than .65 equiva-
lence ratio.
Curve C is shown with a standard deviation of .01 equiva-
lence ratio and with the overall equivalence ratio of the mix-
ture at .585. The mixture of curve C is much leaner than that
of curve A or curve B, but because of curve C's tight mixing
- 78 -
llZ9Z83
statistics, a mixture leaner than .55 equivalence ratio will
occur in the spark gap less than 1/lOth of 1% of the time.
Mixture distribution C, with its mean at .585 stoichiometric,
will have a misfire rate 25 times less than the misfire rate
distribution of curve A even though curve A has a mean equiva-
lence ratio of .75 equivalence ratio, and distribution C will
also have a misfire rate only l/25th as great as that of
distribution B with its mean ratio of .65 stoichiometric.
Better mixing (tighter statistics) than that shown in curve C
would permit the misfire limit to be approached even more
closely with satisfactory engine smoothness. Tightening mix-
ture distributions in the cylinder permits a much closer ap-
proach to the ultimate physical misfire limits than can be
achieved with less complete mixing. Satisfactory operation
with very lean mixtures requires extremely tight microscale
statistics. 6
The data of Figure 20 should become more clear in light
of the previous statistical explanation. Tight mixing makes
operation with significantly leaner mixtures possible. More-
over, because the slope of the NOx production-curve as a
function of equivalence ratio is so steep in the lean range,
NOx output of a lean overall air-fuel ratio mixture is much
lower if the microscale mixing distribution is tight than if
the mixing is loose. The reason that tight microscale mixing
is advantageous in the lean regime is that the NOx outputs
for small volumes which are richer than the mean for an over-
all lean mixture are very much higher than the NOx outputs
corresponding to the mean ratio, so that there is an extremely
~ 7g ~ ~ 1 Z9 28 3
non-linear averaging resulting in heavy penalties for in-
complete homogeneity~
The microscale mixing argument just given assumes
essentially perfect cylinder-to-cylinder mixture variation,
and also assumes a steady fuel-air input so that fuel-air
ratio does not vary much under steady-state conditions from
one cycle to another. ~owever, in ! practical engines
neithex of the assumptions may be justified. Lo~king at
the extreme sensitivity of NOx output to equivalence ratio
in the lean regime, it should be easy to appreciate the
importance of stable steady-state fuel air ratios from cycle-
to-cycle, and the importance of tight air-fuel ratio delivery
cylinder-to-cylinder, if minimum nitrous oxide outputs are
to be practicably achievable. With a well-designed vortex
air-fuel mixer, cylinder-to-cylinder air-fuel ratio varia-
tions for a steady fuel-air input are generally too small to
measure and insignificant.
A properly designed vortex mixer has essentially per-
fect cylinder-to-cylinder fuel-air ratio delivery, and can
be designed to deliver microscale fuel air mixing statistics
much tighter than those shown in curve C of Figure 19. These
attributes give the vortex mixer the potential for very low
emlssions in the lean burn regime for a properly set up
engine.
However, this very low NOx potential is contingent on
a rather smooth fuel-air ratio delivery into the vortex as
.. . .
-- ~o --
1~2~283
a function of time. The vortex mixer is an air-fuel mixing
device with small transient lags for both the fuel and the
air. It is not a fuel-air metering device. If, under
steady-state conditions,air-fuel ratio delivered into the vortex
varies as a function of time, the air-fuel ratio of the
mixture from the outlet of the vortex will vary in nearly
the same way as a function of time. For this reason, a
smoothly continuous fuel-air metering into the vortex is
desirable. This smooth fuel input can be achieved in a
number of ways. Conventional carburetors, because of large
air bubbles from the air bleed, produce unstable fuel air
metering over short time periods. However, a small bubble
air bleed system such as that of Toyota (See SAE Paper No.
760757, "The Development of the Toyota Lean Burn Engine")
will produce a smooth fuel-air metering as a function of time.
In addition, there are a number of other continuous flow fuel
air metering systems which do the required accurate and smooth
fuel-air metering.
Consideration of Figure 20 shows that very low N0x
outputs are possible with lean combustion. However, the range
of air-fuel ratios where the N0 is low is relatively narrow,
and is relatively close to the engine misfire levels. In
consequence, the fuel-air metering system required fcr low
N0x operation in the lean reyime requires relatively accurate
fuel-air metering. The air-fuel ratio ranges involved consti-
tute a "small target" for the fuel-air metering device. How-
,
- 81 ~ 1~Z9~3
eVer~ the percentage metering errors tolerable in the lean
regime are considerably wider than those required for adequate
function with a three-way catalyst system.
At present, it should be clear to one skilled in the
automotive engineering arts how to design a vortex air-fuel
mixer. The present invention mixer makes possible signifi-
cant improvements in engine performance with respect to fuel
economy, emissions, and drivability.
Vehicle E perience
The research and development eEfort on the vortex mixer
has been extensive, and has produced too much data to be fully
recounted here. However, the inventors have accumulated a
considerable amount of experience with a vortex similar to
that shown in Figures 13 and 14 mounted on a Buick 1978
Skyhawk vehicle equipped with a 3.8 liter V-6 engine, with
the vortex mated to a high performance after-market intake
manifold. This vehicle has been tested extensively at
Southwest Research Institute in San Antonio, Texas and the
vehicle has produced important and promising results.
With this vehicle set up to operate on a lean air-fuel
ratio (frequently leaner than 21 to l) fueI economy and
emisslon results have both been extremely promising. Perhaps
the most important result has been that the vortex equipped
vehicle seems to be producing a consistent fuel economy
advanta~e of more than 20% when compared to the base-line
stock vehicle. Testing the vehicle at 3,500 pounds inertial
weight over the EPA CVS hot cycle, the vehicle has frequently
.
- ~2 ~ 1~9~3
produced mileages in excess of 23 miles per gallon. This
compares with mileage on the same cycle generally under
18.5 miles per gallon for the conventionally tested engine
at 3,500 pounds. The test vehicle had less than 18.5 miles
per gallon in its stock configuration prior to modification.
The mileage advantage found with the lean burn vortex
equipped engine is due to a combination of factors. With
the leaner mixtures, pumping ~osses are reduced. Lean mix-
tures are thermdynamically more efficient given adequate
combustion. In addition, there is reason to believe that
combustion efficiency at part loads is significantly improved.
Complete mixing completes the combustion process of ~0 burn-
out earlier in the expansion cycle, yielding higher thermal
efficiency. Whatever the details of the explanation of the
vehicle's mileage improvement, the improvement itself has
been verified experimentally many times and at several labora-
tories.
With the vehicle in lean configuration the drivability
of the vehicle has been rated as excellent by experts at
Southwest Research. In addition, since the engine is set up
to have a near stoichiometric mixture under wide open throttle
conditions, the peak power and hence the acceleration of the
vehicle with the vortex is superior to that of the stock
vehicle. It should be emphasized that the vortex mixer per
se will not much increase engine power. However, the vortex
mixer eliminates the necessity to handle two-phase flows
in intake manifold passages, and therefore the vortex mixer
112~3
permits more open intake passa~es than would otherwise be
commercially acceptable. These more open intake passages
ncrease engine power.
~ s of the time of filing of this case, the NOX output
of the vehicle has been somewhat above the 1.0 gram per mile
NOX standard. On the basis of steady-state results, we
believe that this is due to a pronounced irregularity of
the fuel flow from the,carburetor feeding the vortex, and
irregularity due to intermittent fuel flow in the main venturi
due to air bleeds. This fuel flow intermittency due to air
bleed bubbles has been observed by many other workers. A
special sintered metal air-bleed arrangement producing
small bubbles has been developed by Toyota to eliminate this
fuel flow~irregularity. The inventors have reason to believe
that when they eliminate the unsteadiness of the fuel input
of the carburetor feeding the vortex (or replace the carburetor
with a smoo*h fuel input continuous flow air-fuel metering sys-
tem) the vehicle will be comfortably below the 1.0 gram per
mile NO standard with excellent drivability and with fuel
economy which is not worse than that obtained already. The
importance of the smooth fuel air input into the vortex for
low NOX operation in the lean range was emphasized with refer-
ence to the discussion of Fi~ures 19 and 20.
One very interesting experiment with the vortex equipped
vehicle was made possible by the cold weather of Madison,
Wisconsin. The inventors had repeated occasion to test the
cold start capability of the vortex equipped vehicle, after
- 84 _ 11 2 9 28 3
cold soaks at very low temperatures. With a simple choke
arrangement which was very rapidly turned off (within
about one minute) we were able to get extremely excellent
cold start and drive-away performance at temperatures
down to minus 32 degrees centigrade. After the initial
shake-out of the choke programming, the engine invariably
started and drove away smoothly with only a few seconds of
cranking. The inventors do not believe that they have
driven another vehicle with the cold start and cold drive-
away smoothness and ease which they experienced with the
vortex equipped vehicle herein described.
Vortex Designed for Alternative Fuels
The vortex mixer, since it is a very efficient mixer
and fuel evaporator, is well adapted for efficient evaporation
and mixing of any engine fuel. Because of well-known problems
of fuel supply economics, fuels different from conventional
gasoline are of commercial interest. Most of this interest
centers around alcohols and fuels derived from hydrogenation
of coaI, particularly the hydrogen poor high octane fuels
called napthalenics.
Mixing of alcohol in the vortex is straightforward,
and the problems are easy to understand. The fundamental
difficulty with evaporating and mixing an alcohol based fuel
is the significantly higher heat of vaporization of alcohols
in comparison with conventional gasolines. Although the
equilibrium air distillation temperatures required to evaporate
alcohols may be quite low, ~uch heat is required to accomplish
- ~5 - 1~29~83
evaporation. To efficiently burn alcohols in an engine
equipped with a vortex it is necessary that the vortex be
arranged with sufficient exhaust flow capacity and enough fin
to transfer the necessary heat to the vortex peripheral wall
surface to evaporate the alcohol fuel. Since the heat avail-
able in the engine exhaust is very much greater than the heat
required to evaporate an alcohol fuel this is a relatively
straightforward engineering problem involviny heat exchanger
passage sizing. Exhaust flow (and hence heat addition) past
the vortex heat exchanger fins is conveniently controlled by
means of a temperature controlled ported vacuum switch. Such
a ported vacuum switch will automatically control the vortex
exh~aust flow to maintain the desired range of vortex tempera-
ture. Since the evaporation temperatures required for evapora-
ting the alcohol are relatively low, such a ported vacuum
switch controlled exhaust flow control will automatically
adapt itself to varying alcohol percentages and the fuel
supply to a vehicle.
J. P. hongwell and others have pointed out that
napthalenic fuels (very low hydrogen to carbon ratio fuels
derived from coal~ can be derived from lignite relatively
much more cheaply than more hydrogenated fuels. These coal
liquid fuels have significant advantages in a spark-fired
engine. Specifically, the napthalenic and highly aromatic
coal liquid fuels have very high octane numbers. However,
operation with these fuels has heretofore been impractical
becaus~ they form soot when burned even slightly rich.
- ~6 - l~Z9~
There is reason to believe that these napthalenic
hydro-carbon fuels can be efficiently mixed in a vortex and
burned lean, in a way where carbonaceous deposits and soot
would not be formed. If this is in fact practical, the
utilization of coal as a motor fuel source will be signif-
icantly more economic than otherwise. Research attempting to
show that this engine operation is possible with a vortex
mixer is underway at the University of Wisconsin under the
supervision of Professors P.S. Myers and O.A. Uyehara, and
the work is being significantly assisted by the inventors.
There appeared to be two sorts of problems related to
burning the coal liquids in a spark-fired engine. First,
the mixture delivered to the combustion chambers must always
be lean, even at microscale sample volume scales,and it is
desirable that there be no liquid phase delivered to the en-
gine, since ~-vaporation of boundary layers involvesrich
zones which form soot. The vortex mixer is well adapted to
produce this homogeneous mixture supply. The other
problem is that carbonadeous or gummy deposits must be
avoided in the vortex mixing chamber itself. It appears that
it will be possible to avoid these deposits with a vortex
designed so that the fuel in the mixing chamber is in the
form of a splash-cloud (as before described) under the con-
ditions where the vortex peripheral wall is heated enough
to react the fuel. To avoid deposits with any fuel, includ-
- ~7 -
83
ing a napthalenic fuel, it is desirable to have the vortex
peripheral walls as low in temperature as possible, and as
uniform in temperature as possible. It appears to be pos-
sible to burn the napthalenic fuels in a vortex equipped
engine without either combustion chamber carbon deposits or
vortex chamber deposits. It should be mentioned that much
research on fuel deposit formation rates has been done by
William Taylor of Exxon Research and Development Corporation
and that some of this data will be relevant to design if deposit
formation becomes a problem with the napthalenics or with
any other fuels.
Detail Design Issues
A number of detailed design issues are relevant to the
present invention vortex mixer. First, one skilled in the
production arts will clearly see that the vortex mixer is
well adapted to die-casting from aluminum. To minimize
production costs, it is desirable to minimize the amount
of metal in the vortex mixer. This weight minimization is
also highly desirable in terms of rapid warm-up; the light-
er the vortex mixer structure is, the more rapidly it can
be heated up to steady-state temperature after a cold start.
There are some fairly straightforward details relev-
ant to the design of the vortex heat exchanger passages. The
heat exchanger relations in the vortex mixer are such that
the heat exchanger fins can readily be made of aluminum,
~lZ~Z83
- 88 -
and maximum ~in temperatures can readily be held below 175
degrees C. It is desirable that the fin areas be arranged
so that the temperature around the outside peripheral wall
of the vortex mixer is relatively uniform. With a splash-
cloud of fuel droplets, thîs temperature condition is rela-
tively easy to arrange, but a very smooth temperature dis-
tribution around the outlet will involve some degree of em-
perical trimming of fin areas based on data generated by
thermocouples located around the vortex peripheral wa11.
Other issues involving the vortex heat exchanger
sections come into play when water condenses within the heat
exchanger sections during engine shut-down or early in the cold
start sequence. When the vortex is cold, water vapor from
the exhaust will condense and accumulate within the vortex
heat exchanger sections. Because the exhaust gas will also
contain large volumes of carbon dioxide (as well as traces
of ash and sulfates) the PH of the condensate will be acidic,
and somewhat corrosive. Metallurgically, it appears that
very high silicon aluminums (which are easy to die-cast and
relatively inexpensive) have the requisite corrosion resis-
tance for durability in the presence of this relatively acidic
condensate. However, corrosion resistant coatings on the
heat exchanger flns can readily be employed. Another issue
involving condensate in the heat exchanger area around the
vortex fins involves freezing. The passages need to be
designed so that ice does not block the exhaust flow, or the
vortex can not warm up fast enough during start up.
- 89 - ~129283
Summary
A partial summary of a vortex mixer i5 useful here.
The vortex mixer functions as follows:
Part of the isentropic expansion flow velocity past
the engine air-flow control throttle valve is delivered
tangentially and relatively smoothly into a nearly radially
symmetric vortex mixing chamber. This tangentially intro-
duced inlet flow provides annular momentum which, in inter-
action with the geometry of the vortex chamber, drives the
flow within the vortex chamber into an inwardly spira]ling
pattern. If boundary layer geometriesare controlled properly,
for example with the weirs disclosed in this application, the
flow pattern within the vortex mixing chamber is an excellent
approximation to an irrotational flow vortex with turbulence
superimposed on the mean flow streamlines. The air flow from
the outside of the vortex chamber spirals inwardly to the
central outlet in a flow pattern which is very predictable
and extremely conducive to mixing.
The vortex air-flow pattern serves as an inertial
separator with centrifugal forces generally in excess of a
thousand "G's". Fuel droplets are flung to the outside
peripheral wall of the vortex mixing chamber by these cen-
trifugal forces. The combination of the centrifugal force
and the rapidly swirling flow forms an equilibrium splash
cloud around the vortex chamber outside wall. This spl~ash
cloud continuously sprays and wets the outside wall surface,
- 90 ~.129Z83
which is heated for extremely rapid evaporation. The evap-
oration all around the vortex chamber outside wall combines
with the vortex flow pattern of the air to produce exception-
ally rapid and complete fuel air mixing within the vortex
mixing chamber. The mixing chamber functions nicely under
all engine operating conditions. At cold start, the vortex
rapidly evaporates the light ends of the fuel, while the
heavy ends of the fuel accumulate around the vortex pe-
ripheral wall. Excellent cold start characteristics are
therefore possible without ever delivering a rich (pollution
causiny and fuel wasting) mixture to the engine cylinders.
The vortex warms up rapidly. Under normal driving condi-
tions the warm vortex feeds a homogeneous mixture of
vaporized fuel and air to the intake manifold of the engine.
Under very low manifold vacuum wide open throttle conditions -
the vortex flow slows down. Fuel is still evaporated from
the mixing chamber walls, but heat transfer rates at the
vortex wall interfaces are low enough that the air surrounding
the fuel evaporating boundary layers is cold. Much of the
evaporated fuel therefore recondenses, so that the vortex
functions as a smoke generator. In this smoke generating
mode the vortex mixer can function with very low flow re-
sistance and will feed a relatively cold and dense charge to
the engine cylinders for peak power.
- 91 -
1~29~83
~ short discussion of the drawings will assist in
summarizing the subject matter in the present case. Figures
1 and 2 show views of a fluidically efficient vortex mixer
In considering Figure 1, the fluid mechanics within the
vortex chamber per se and the details of the flow streamlines
to the left of section AA should be essentially invarient
regardless of whether the vortex of Fig. 1 is fed with an
entrance section such as that shown in Figure 3 and Figure
1, or whether the substitute round to trapezoidal transition
section shown in Figures 4 and 5 is substituted. Either
throttle body containing entrance section assembly will
deliver high velocity flow to the main body of the vortex
mixer in a form which will efficiently drive an irrotational
vortex within the vortex mixer chamber. A large number of
views of the round throttle plate to trapezoidal entrance
assembly were necessary to clarify its shape.
For the entrance section of Figure 3 a corner notch
opening assured that a significant fraction of the idle air
flow formed a jet which attached to the adjacent walls of a
corner of the passage. For the round throttle to trapezoidal
entrance section shown in Figures 4 and 5 a nozzle 68 is
arranged for the same purpose. In both cases, the high speed
jet so formed attaches to the corner between two adjacent
walls and flows smoothly into the main vortex chamber. Cor-
ner wall attached jets spread relatively slowly. Putting a
- 92 ~ 1 ~ 2 9 Z8 3
large fraction of the idle flow into such a corner wall
attached jet reduces losses in the vortex entrance section
and drives a stronger vortical flow within the vortex mixer
than would otherwise be possible under idle and off-idle
engine operating conditions.
Figures 6, 7, 8 and 9 explain the flow structure of
an irrotational flow vortex, and show how this flow pattern
is useful for mixing. 'Figure 12 explains in light of Figures
10 and 11 the function of vortex mixing chamber weirs, which
permit the irrotational flow pattern to flow and be stable
within the vortex mixing chamber.
Figures 13 and 14 show a vortex mixer having vortex
mixing chamber fluid mechanics nearly identical to the fluid
r mechanics shown with respect to Figure 1, but where the en-
trance section is specifically adapted for a down-draft car-
buretor. Figures 15, 16, 17 and 18 show a more primitive
vortex mixer.
Figures 19 and 2n are included to explain the value
of the vortex mixer operated as the lean burn system, and a
number of issues which must be tended to for satisfactory
lean operation with minimum NOx emissions were discussed.
