The present invention relates to an intake manifold system for a multi-cylinder four stroke spark-ignition internal combustion engine.

The invention seeks to provide an intake system that improves control over the distribution of fuel and air within the charge in the engine combustion chamber.

According to the present invention, there is provided an intake manifold system for a multi-cylinder four stroke spark-ignition internal combustion engine wherein each cylinder has at least one intake port, each intake port being connected at one end to the combustion chamber by at least one intake valve and at the other end to a first intake duct that is connected to a first plenum common to the first intake ducts of other engine cylinders, wherein each intake port is further supplied with gases containing no fuel by a second intake duct having a fixed outlet cross-sectional area equal to between 25% and 45% of the maximum effective through flow cross-sectional area of the open intake valve, the second intake duct being directed towards the closed end of the intake port whereby, during the periods that the intake valve is closed, gas is drawn in through the second intake duct by manifold vacuum and stored in the first intake duct, and wherein the solid boundary walls of the intake port and of the discharge end of the second intake duct define a U-shaped flow path for guiding the gas entering the intake port from the second intake duct, the geometry of the solid boundaries being such that during operation gas enters from one side of the intake port, performs a single U-turn by flowing around the intake valve stem while adhering to the walls of the intake port and enters the first intake duct from the other side of the intake port, the gas being thereby constrained to scavenge the closed end of the intake port

and to displace substantially all the gases previously present therein into the first intake duct.

The invention is not the first proposal to blow air into the intake port to improve charge preparation. It is known for example to provide an intake manifold system in which an air assisted fuel injector is positioned near the intake valve to introduce air and fuel into the intake port. Small amounts of air, bypassing the butterfly throttle, have also been introduced into the intake port to assist charge preparation by stirring the air near the intake valve when it is closed. In all such systems, the size of the air jet has to be small in order to give the maximum jet velocity for fine atomising of the fuel or deep penetration for good turbulent mixing. Furthermore, the size of such an air jet cannot be made larger because the flow from it must be limited to less than the engine air intake flow during idling; otherwise it would interfere with the idle speed control of the engine. This limits the size of each air jet to less than 1% of the open intake valve area delivering no more than 80% of the throttled idle air flow to the engine.

In the present invention, the effective flow cross-section of the second intake duct is much larger to give significantly larger flows to satisfy at least 60% of the maximum engine air flow requirement and is not limited by the engine idle operating condition. The flow velocity emitted from the duct is kept low to give an organised flow around the end of the intake port with substantially no turbulent mixing.

It is also known to provide an exhaust gas recirculation (EGR) system in a spark ignition engine. In such an engine, the EGR pipe is conventionally connected to the intake plenum where the exhaust gases can mix thoroughly with the intake air before entering the intake ports. This is because it is not an acceptable practice to

introduce the exhaust gases directly into each intake port where there ^' is wet fuel since any fuel entrained within the neat exhaust gases will not find oxygen -for combustion and will result in excessive hydrocarbons emissions in the exhaust gases and unstable combustion.

The present invention proposes taking in a major proportion of the intake charge at much higher flow rates under a wide range of operating conditions through the second intake ducts and storing it in the first intake ducts before it is drawn into the cylinders. This has various advantages that stem from being able to control the mixture strength in different parts of the combustion chamber either to achieve homogeneous mixing or to achieve charge stratification.

A problem encountered with fuel injected engines is that the fuel is not well mixed in the charge under all conditions. If fuel is injected into the intake port when the valve is closed, it tends to forms pools and this wall wetting causes hysteresis problems that affect cold starts and the transient response of the engine.

In the present invention, even when the intake valve is closed, the intake port continues to be scavenged by air drawn in through the second intake duct under the action of the manifold vacuum created by the other engine cylinders connected to the same first plenum. This maintains the intake port dry of fuel and creates in the first intake duct a column containing a stratified mixture of fuel and air.

Depending on the geometry of the intake manifold, in particular the length of the first intake duct, this column can either be stored ready to be introduced back into the same cylinder when the intake valve next opens or the mixture may enter the first plenum to mix thoroughly with the mixture drawn from other cylinders. If the

charged is stored as a stratified column in a long duct, then when the intake valve opens, the column is transferred into the cylinder and determines" the charge stratification within the cylinder. On the other hand if the first ducts are short, not only is better charge preparation achieved in each cylinder but homogeneity between cylinders is improved.

The ability to store a stratified column allows the invention to accentuate the inhomogeneity in the cylinder and thereby permit the engine load to be regulated at least in part by modifying the amount of fuel introduced into the cylinder (rather than the amount of fuel and air mixture) while introducing an excess of air into the engine. This permits losses associated with pumping air in the intake manifold to be reduced.

In an extreme case, it is possible to concentrate all the fuel at one end of cylinder as a homogeneous easily ignitable mixture while filling the lower end of the cylinder with air or exhaust gases into which the flame cannot propagate. Such a divided charge may be used to achieve very lean overall fuel-air mixture or very high overall EGR dilution. If air forms the lower part of the divided charge and a very rich fuel-air mixture forms the upper part, then such division can create incomplete combustion in the upper part of the combustion chamber while still delivering an overall stoichiometric mixture to a catalytic converter or an afterburner. The completion of combustion of such an exhaust gas mixture in the exhaust system can assist in raising the catalytic converter to its light off temperature. If exhaust gases form the lower part of the divided charge and a stoichiometric fuel and air mixture forms the upper part, then the engine can be run with high overall dilution thereby reducing the volumetric efficiency without disturbing the stoichiometry of the exhaust gases.

It should be mentioned that there have been disclosed systems in the prior art in which two intake ducts lead to each intake port and these have generally had one of three aims in mind.

The first aim of the prior art proposals, as exemplified by EP-A-0 098 543, is to achieve manifold tuning by matching the length of the intake ducts to the engine speed, a longer and narrower duct being used at low engine speeds and a shorter and wider duct at high engine speeds. The effect of flow reversal in the first ducts when the intake valve of a cylinder is closed is not appreciated, nor is the des ⁱ gn of the intake port optimised by using the reverse flow to scavenge the closed end of the intake port and achieve charge mixture stratification control .

The second aim, as exemplified by GB-A-1, 195, 060 and US-A- 4,867,109, is to achieve flow balancing. In these systems, in addition to the conventional intake ducts leading from the intake throttle or throttles, there are interconnecting ducts that balance the flow between the cylinders. The flow through these balancing ducts will reverse directions as occurs in the first intake ducts of the present invention but while the first intake ducts in the present invention occupy the major cross sectional area of the intake ports, the balancing ducts in the prior art have a much smaller cross section. Because of the size and location of these balancing ducts the gases flowing into them play little part in scouring the closed end of the intake port or in transporting any significant volume of gas out of the intake port.

The last aim of the prior art in providing two ducts to each intake port, as exemplified by GB-A-2, 114, 221, GB-A- 2,038,415 and GB-A-1,239,264, is to achieve high intake air velocity to increase turbulence at low engine speeds by using a smaller supply duct at low engine speeds and a larger supply duct at high speed. These proposals differ

ω

00

Here again the content at the end of the intake port is not scavenged.

It is found that there is a narrow range of the width of the air jet which can sustain the required deep U-flow pattern at the end of the intake port for a short period of time, even without the solid boundary specified in the present invention. However this is unstable and the flow pattern reverts to one or the other stable flow patterns during the engine cycle and in an unpredictable manner from one cycle to the next. This geometry is therefore not practical for any serious application because it fails to provide reliable and complete scavenging of the intake port which is the necessary condition in order to satisfy the main aim of the present invention for achieving the desired charge mixture stratification.

The consequence of not completely scavenging the content of the intake port during every engine cycle is that clean combustion cannot be guaranteed with the charge mixture stratification of this invention. In the case where liquid fuel is introduced as a spray into the intake port by means of a fuel injector, some of the liquid will be deposited on the back of the intake valve. Whereas a strong scavenging flow scouring the end walls of the intake port will be very effective in evaporating the fuel and transporting the vapour out of the port, a stagnant or recirculating flow would leave a pocket of liquid fuel or saturated vapour remaining inside the port. In the case where a premixed fuel-air mixture is supplied from the first intake duct and exhaust gas is introduced through the second intake duct as the scavenging gas for clearing the port of any fuel-air mixture left behind from the previous intake stroke, a pocket of fuel-air mixture would once again be retained in the port if the scavenging flow is poorly directed. In both these case, the unscavenged pocket of fuel will be drawn into the combustion chamber first at the beginning of the next intake stroke. Because

cross-sectional area of the first intake duct in order to increase the proportion of the charge drawn in through the second intake duct during the intake period.-

It is desirable to design the geometry of the first and second intake ducts in such a manner that the gases containing the dispersed fuel that are drawn in from the first intake duct remain in the vicinity of the spark plug at the end of the compression stroke.

In a preferred embodiment of the present invention, it is desirable to have a high swirl intake port so that the column of stratified intake charge drawn into the combustion chamber wraps itself as a helix to take on the shape of a coil spring within the cylinder. When axially compressed by the piston during the compression stroke, this gaseous mass retains its axial stratification and the regions with high fuel concentration remain in the vicinity of the spark plug near the top of the cylinder.

As a further possibility, stratification may be achieved across the cylinder bore rather .than, or in addition to, the stratification along the cylinder axis.

In an engine having an intake port designed to promote swirl, by directing the fuel-containing part of the charge towards the centre of the combustion chamber and the air or EGR gases towards the outer circumference of the combustion chamber, a radial stratification will be achieved concentrating the fuel in the vicinity of the spark plug, which is generally placed near to the centre of the cylinder head.

In an engine having two intake valves, supplied by separate or siamesed intake ports designed to generate tumbling motion, that is to say rotation in a plane parallel to the plane of symmetry passing between the two intake valves, one can achieve stratification across the

width of the cylinder bore with the fuel concentration being maximum in the latter plane of symmetry and reducing symmetrically on both sides.

In the latter embodiment, it is preferred to provide two second intake ducts, one for each intake valve, each positioned to the side of the associated valve remote from the plane of symmetry between the two valves.

In an intake manifold system of the invention, fuel may either be introduced directly at the first plenum, or it may initially be introduced into the intake port and be subsequently evaporated and transported into the first intake duct and the first plenum during the period that the intake valve is closed. The advantage in either case is that this gives rise to a substantially dry intake port during the intake period. During the intake stroke, the supply of fuel-air mixture always comes from the first intake duct and the supply of diluting air or EGR gases containing no fuel always come from the second intake duct. By ensuring the flow streams from the two intake ducts are kept parallel with one another, stratification of the fuel and air within the combustion chamber can be controlled accurately and in a unique pattern depending on the intake flow motion induced by the intake port design giving radial stratification in the case of swirl or diametrical stratification in the case of tumble.

Thus, the preferred embodiment of the invention achieves charge mixture stratification by packing air or EGR gases containing no fuel around a previously prepared homogeneous mixture of known fuel-air ratio. This is fundamentally different from a conventional in-cylinder injection stratified charge engine where the bulk air is penetrated locally by a high concentration of fuel to achieve charge stratification. The advantage of achieving charge stratification by packing air around, below or to the side of a pre-mixed easily ignitable mixture, is that

- li ¬ the mixture surrounding the spark plug is readily ignitable while the EGR gases or air that is not ignitable just fills out the regions of the combustion chamber remote from the spark plug. This ensures robust ignition in comparison with an in-cylinder injection stratified charge engine in which the location of the ignitable mixture zone in relation with the spark plug is poorly defined. In such an engine, the probability of having either a too rich or too lean mixture occurring in the vicinity of the spark plug is relatively high and this results in unreliable ignition. It is consequently easier to calibrate an engine of this invention to run satisfactorily over a wide range of speeds and loads using stratified dilution as the main control means.

The preferred embodiment of the invention has particular application to assisting light-off of a catalytic converter during cold starts. By calibrating the mixture stored in the first intake duct at a very rich fuel-air ratio and diluting the intake charge with a lot of air from the second intake duct to create mixture stratification across the cylinder, it is possible to achieve incomplete combustion within the mixture in one part of the combustion chamber while leaving sufficient unused air in another part of the combustion chamber. The exhaust gases from the engine containing unburned fuel and air, once thoroughly mixed in the exhaust system, can be re-ignited and burnt completely in an afterburner in front of a catalytic converter, thus assisting in heating the catalytic converter to its light-off temperature. Once the catalyst has reached its light-off temperature, the exhaust gases containing unburned fuel and air can be allowed to react with one another exothermically within the catalytic converter to heat the catalytic converter further either immediately after the flame in the afterburner has been extinguished or to prevent the catalytic converter from dropping below its light-off temperature during idling and part load operation.

In another mode of operation, fuel and air are supplied at stoichiometry directly to the first plenum of the intake manifold system and EGR gases are supplied exclusively to the second intake ducts. By keeping the stratified dilution of EGR gas under control it is possible to achieve stable combustion even at very high overall EGR flow rates. This reduces the N0 _X emission produced during combustion and allow the use of a three-way catalytic converter for further after treatment of the NO _x in the exhaust system.

It is desirable for the flow control valve to be positioned as near to the intake valves as possible so that the flow momentum emitting from the flow control valve and containing the mixture charge from the first intake duct should be transferred directly into the cylinder and dominate the flow motion in the desired region of the combustion chamber.

It is important to distinguish between the function of the individual flow control valves in each first intake duct and the function of the common throttle in the first plenum. The latter does not affect the internal exchange of gases between the first intake ducts across the first plenum and cannot provide the charge dilution control of the invention.

The individual flow control valves to each first intake duct may be linked to move together and used as the main load control means for the engine connected directly to the driver's input. Engine load is set by the quantity of fuel drawn into each combustion chamber from the first intake duct during the intake period and is confined to a compact homogeneous cloud of substantially constant fuel-air ratio surrounding the spark plug region of the combustion chamber, the size of the cloud determining the engine load.

This method bf load control may completely replace the throttle control in the second intake duct as the sole load control means, the throttle control in the second intake duct being originally used to set the engine load by altering the intake charge density within the intake manifold. In this case the intake manifold system may dispense with the throttle control in the second intake duct although such a throttle would still be needed in many other modes of operation.

The invention will now be described further, by way of example, with reference to the accompanying drawings, in which :

Figure 1 is a schematic representation of a first embodiment of an intake manifold of the invention,

Figure 2 is a detail showing the gas flow within a closed intake port connected to first and second intake ducts,

Figures 3, 4 and 5 are views similar to that shown in Figure 1 of further embodiments of the invention,

Figure 6 is a schematic cross section through an intake port and a cylinder head of a combustion chamber of an alternative embodiment of the invention having a variable flow control valve in the first intake duct, the drawing showing the region with high fuel concentration that occurs in an engine designed to promote swirl in the cylinder,

Figure 7 is a vertical section through the intake port in Figure 6 along the axis 98,

Figure 8 is a section similar to that of Figure 6 for an engine having two intake valves in each cylinder and intake ports for the valves designed to promote tumble in the cylinder, and

Figure 9 is a section similar to that of Figure 7 taken along the axis 98 in Figure 8.

Throughout the drawings, like components have been allocated like reference numerals to avoid the need for repetition and where components are modified but serve an analogous purpose, a prime has been added to the reference numerals.

In Figure 1, an engine cylinder head 12 has an intake system 10. The cylinder head 12 is of a four stroke four cylinder engine each cylinder having two intake valves 14, two exhaust valves 18 and a spark plug 16. The intake port 20 for each cylinder is connected via a first intake duct 22 to a first plenum 24 common to all the cylinders but isolated from the ambient atmosphere. Each intake port is further connected via a second intake duct 32 to a second plenum 34 that is connected to ambient through a common intake duct including a throttle valve 40 and a mass air flow meter 42. Fuel injectors 60 are arranged in the intake ports to direct a spray of fuel towards the intake valves 14.

The detailed design of the intake port is better shown in Figure 2 from which it will be seen that the second intake duct 32 has a significant diameter that is approximately 25% of the area of the intake port 20 and the first intake duct 22 makes up the other 75% of the area of the intake port 20. Furthermore, a partition 30 is provided to guide the gases and constrain them to flow along the path represented by the arrows. The gas discharged from the second intake duct 32 is directed along the back wall of the closed end of the intake port behind the intake valve 14 and after scouring the volume of the intake port 20 reverses its direction and enters the first intake duct 22 on the other side of the partition 30. The effect of the partition 30 is to make the positioning and the geometry of the end of the second intake duct less critical, thereby simplifying construction and assembly. Because the scavenging flow path is better defined by the partition 30, the second intake duct need not penetrate into the immediate vicinity of the intake valve.

In the embodiment of Figure 1 the first plenum 24 is always under manifold vacuum created by whichever of the four engine cylinders is in its intake stroke. Air is drawn into the first plenum 24 through the second intake

ducts 32 and the second plenum 34 past the throttle 40. All the ducts 32 therefore have air travelling through them at all times in the direction of the intake valves 14. This air is directly drawn into the cylinder if the intake valve 14 is open but otherwise flows around the intake port as described previously with reference to Figure 2 and is stored in the first intake duct 22 and the first plenum 24. Thus, unlike a conventional engine in which there is no air movement in the intake port when the intake valve is closed, the present invention permits constant air movement in the intake port 20 all the time including the periods of the four stroke cycle of each cylinder when the respective intake valve is closed and the air required for a cylinder is not just drawn in while the intake valve of that cylinder is open, but is drawn in constantly and stored in the first intake ducts 22. It is because air is drawn in during all four strokes of the cycle that the diameter of the second intake duct need only be 25% of the full port cross-section, as it is required to deliver only one quarter of the full intake charge during each of the four strokes of the engine cycle. The relative dimensions of the two intake ducts do not depend on the number of cylinders in the bank or engine.

As soon as the intake valve 14 opens, the stored air in the first intake duct 22 enters through the port 20 into the cylinder, making up some 75% of the intake charge while the remaining 25% is directly drawn in through the second intake duct 32.

Apart from maintaining a dry intake port by evaporating all the fuel deposited in the intake port by the process of continuous convection described above, the embodiment of Figure 1 achieves a thorough mixing within the first plenum of the charge introduced into any cylinder with the charge supplied to the other cylinders. This mixing of the charge between cylinders results in excellent fuel-air

mixture preparation and also improves homogeneity between cylinders . •

In the embodiment of Figure 1, the first intake ducts 22 are relatively short and it is this length of the first intake ducts 22 that results in mixing between the charges supplied to the different cylinders.

In Figure 3 the first intake ducts 22' are intentionally made significantly longer such that the overall volume of each intake duct 22' should exceed 75% of the volume of the cylinder. As a consequence, each duct can store the entire charge to be drawn into the cylinder when the intake valve 14 next opens without the respective air charges in the first intake ducts 22' of different cylinders mixing with one another.

The embodiment of Figure 3 further differs from that of Figure 1 in two other more minor respects namely that the first plenum 24 ' is heated by a heater 70 and is supplied by a purge flow of air through a small pipe 28 connected downstream of the intake throttle 40. The purpose of the heater 70 and the purge pipe 28 is to avoid accumulation of fuel within the first plenum 24' and to maintain the entire intake system dry.

In the embodiment of Figure 3, a column of charge is stored in the first intake duct 22 ' which is transferred as a column into the cylinder when the intake valve 14 is opened. For this reason the fuel distribution within the column will determine the fuel distribution within the cylinder and it is possible in this way to achieve a stratified charge within the cylinder by stratifying the fuel distribution in the column. Charge stratification within the cylinder is assisted by providing a flow straightener 26 within the first intake duct 22' to maintain parallel flow and avoid mixing across the cross- section of the first intake duct 22' and by designing the

intake port 20 to promote high swirl and ensure that the charge enters as a column into the engine cylinder.

In the embodiment of Figure 3, fuel is injected by means of the fuel injectors 60 shortly before or after the intake valve 14 closes. The injection of fuel shortly before the intake valve 14 closes will place fuel at the top of the charge in the combustion chamber in the vicinity of the spark plug. Fuel injected shortly after the valve closes will be swept into the first intake duct 22 ' by the air flow from the second intake duct 32 and will be stored as a stratified column in the first intake duct 22' with the bulk of the fuel contained at the end of the column remote from the intake port . The other end of the column will contain hardly any fuel as the port will have been dried by the air flow. When this column is transferred to the engine cylinder the region with high fuel concentration will therefore once again reside at the top of the cylinder near the spark plug.

The advantages of running an engine with a stratified charge are in themselves well known. In particular one may achieve overall very lean mixtures because the problems normally associated with igniting weak mixtures are avoided. The load within the engine can also be controlled by varying the fuel quantity injected rather than varying the air mass thereby allowing the engine to run with a lesser degree of throttling and lower pumping losses .

There are other advantages that stem from running with a dry intake port, namely improved transient response, better cold starts and reduced hydrocarbon emissions while the engine is overrunning.

The improved charge preparation achieved by the above described embodiments also makes them especially useful in engines in which load control is effected by variable

plenums 24" and 34 is determined by the setting of the throttles 50 and 48 or 84.

If one considers first operation with the throttle 50 nearly closed and the throttle 48 open and the throttle 84 fully closed, then in this mode the system does not differ from the embodiment of Figure 3 except in that the fuel is held at the end of the column remote from the engine cylinder by being injected there in the first place instead of being carried there by the air drawn in through the second intake duct 32.

With the throttle 48 closed on the other hand and the throttle 50 fully open, the intake system operates in the same manner as a conventional intake system with central fuel injection as the second intake ducts 32 will play no part in the operation.

At intermediate settings, that is to say with both of the valves 48 and 50 partially opened, each first intake duct 22" will contain a divided charge with a fuel and air mixture at the end remote from the intake port drawn in through the first plenum 24" and air at the other end drawn in from the second plenum 34.

In this divided charge, instead of air alone being present at the end of the column nearer the intake port, one can introduce an air and EGR mixture with the proportion of EGR determined by the relative setting of the throttles 48 and 84. In the extreme settings the throttle 48 is fully closed and the throttle 84 is opened to achieve a divided charge in which the first part drawn into the combustion chamber consists exclusively of EGR gases.

The intake system as described with reference to Figure 4 offers all the advantages described with reference to Figure 3 that result from being able to run the engine with a stratified charge. A divided charge however has a

Ti Tl

© © ©

55

H υ

tN

oo

<N 4

O

intake valve, without any of the fuel being deposited on the wall of the intake port.

The embodiment of Figure 4 can also be operated in a mode that assists in heating or managing the temperature of a catalytic converter during cold starts in that it can be used to ensure the presence of a high proportion of both air and unburned fuel in the exhaust gases that can react with one another in the exhaust system to give off heat either in the catalytic converter or in an afterburner used to heat the catalytic converter. In this respect one can burn a mixture which is stoichiometric averaged over the entire volume of the cylinder but with an extremely rich composition in the vicinity of the spark plug. This can achieve exhaust gas ignition (EGI) without the need for an additional air pump for supplying air directly into the exhaust system and assists in simplifying the fuel metering system necessary to implement EGI.

The embodiment of Figure 5 is essentially an amalgamation of all the previously described embodiments of the invention, that enables the intake system to benefit from all the advantages detailed above. The intake throttle 40 and mass air flow meter 42 are now common to both the first and the second plenums 24" and 34 and a diverter valve 90 controls the relative proportions of air supplied directly to the second plenum 34 and through a pipe 92 to the first plenum 24". It will readily be seen without further description that depending on the position of the valve 90, this embodiment will operate in the same way as that of Figure 3 or that of Figure 4. The advantage of this embodiment, apart from versatility, resides in the fact that the mass air flow meter 42 measures the total air flow into the engine. The fuel flow can therefore be calibrated against a single air flow measurement.

In Figures 6 and 7 a combustion chamber in the cylinder head 12 has an intake valve 14, an exhaust valve 18 and a

spark plug 16. An intake port 20 designed to promote swirl receives air from a first intake duct 22 and a second intake duct 32. The relative proportions of the charge drawn in through the first and second ducts can be varied by a butterfly valve 96 mounted immediately upstream of the intake valve 14 for rotation above an axis 98. In the position occupied by the valve 96 in Figure 7 the first duct is partially obstructed and the butterfly 96 can be rotated about the axis 98 to lessen or remove the obstruction. The shaded regions in Figure 6 represents a homogeneous fuel-air mixture that is drawn in from the first intake duct 22. The non-shaded regions represent air or EGR gas drawn in through the second intake duct 32. The geometry of the intake port is selected such that the two streams drawn in from the respective ducts do not mix in the intake port and remain separate as they swirl into the combustion chamber. As a result of this motion, the charge within the combustion chamber is stratified with a central cloud of homogeneous mixture in the vicinity of the spark plug 16 and an annular surrounding region containing little or no fuel. By this method an overall high degree of dilution can be achieved while maintaining a readily ignitable mixture near the spark plug. By varying the position of the butterfly 96 the relative proportions of these two regions can be modified and this can be used to set the engine load at least over part of the operating range of the engine.

Because of mixing that continues within the combustion chamber during the compression stroke, the separation between the shaded and un-shaded regions will become progressively less distinct. Consequently, a small volume of either region will not survive in keeping its identity for very long and will merge uniformly with the bulk of the charge. It is therefore desirable to use the flow control butterfly 96 to vary the relative proportion of

either region to at least 25% of the total volume in order to achieve significant stratification.

Charge stratification can be used in a variety of fuelling strategies. If it is desired to achieve very lean burn then one can meter the fuel in dependence upon the amount of air drawn in through the first intake duct only. This air mass can be determined by measuring the overall air intake mass flow and computing the proportion of the charge drawn in through the first intake duct from the setting of the butterfly 96. In this case though the overall mixture is very lean the ignitability is not affected as the part of the charge near the spark plug can be controlled to be stoichiometric.

Alternatively, one can meter the fuel in dependence upon the overall air mass flow such that the overall fuel-air mixture is stoichiometric. Here the cloud near the spark plug may be excessively rich and may not undergo complete combustion, but in such a case the exhaust gases will remain stoichiometric and will contain unburned fuel as well as unused air coming from other parts of the charge remote from the spark plug. Such exhaust gas mixture can be re-ignited or reacted in the exhaust system to assist in heating a catalytic converter to its light-off temperature.

The embodiment of Figures 8 and 9 differs in two major respects from that previously described. First, the engine in this case has two intake valves per cylinder. Second, because in such an engine the intake ports cannot promote swirl when both valves are in use, the intake port or ports promote tumble in the combustion chamber. It should be specified in the interest of clarity that swirl is taken in the present context to refer to a rotation of the air mass about the central axis of the cylinder whereas tumble refers to rotation of the air mass about an axis normal to the central axis of the cylinder, the

rotation taking place in planes parallel to a plane containing the central axis.

To achieve charge stratification in an engine with such tumbling motion, the embodiment of Figure 8 has two symmetrically disposed second intake ducts 32a and 32b that are directed towards the sides of the intake valves 14a and 14b furthest from one another. The gas entering through the second intake ducts 32a and 32b flows into the non-shaded regions of the combustion chamber while the fuel and air mixture entering through the first intake duct 22 remains concentrated about the plane of symmetry passing between the two intake valves 14a and 14b and containing the spark plug 16. The butterflies 96a and 96b once again allow the relative proportions of the gases drawn in from the respective ducts to be varied. When the first intake duct 22 is obstructed to its minimum cross- section, the region containing the fuel-air mixture is restricted to a narrow shaded band in the cylinder about the plane of symmetry. When the obstruction is removed by rotating the butterflies 96a and 96b about the axis 98 the shaded band in the cylinder widens to occupy an increasing proportion of the combustion chamber and reduce the overall dilution.

Once again the butterflies can be used to set engine load over at least part of the operating range of the engine and different fuel metering strategies may be adopted to achieve different effects that take advantage of the charge stratification.