WO2000065217A1

WO2000065217A1 - Method and system for starting combustion engines

Info

Publication number: WO2000065217A1
Application number: PCT/NL2000/000275
Authority: WO
Inventors: Jan Marten Berendsen; Adrianus Voogd De; Jan Nytomt; Patrik Rask
Original assignee: Nederlandse Organisatie Voor Toegepast-Natuurwetenschappelijk Onderzoek (Tno)
Priority date: 1999-04-27
Filing date: 2000-04-27
Publication date: 2000-11-02
Also published as: DE60011260T2; JP2002543326A; EP1173666B1; AU4624800A; CN1139723C; EP1173666A1; CA2367739A1; AU764458B2; DE60011260D1; CN1349589A; NL1011907C2

Abstract

The invention relates to a method and system for controlling fuel supply to a combustion engine (1) comprising at least two cylinders (4a...4d) during a start attempt. Injectors (13) for individual cylinders (4a...4d) are controlled such that the air-fuel ratio (A/F) obtained varies for each cylinder (4a...4d) by an amount Δf. The fuel injection to the cylinders (4a...4d) varies between the leanest air-fuel ratio and the richest airfuel ratio practically feasible during starting. If the engine (1) does not start within a predetermined number of cycles an incremental increase (ΔfSTEP) of the fuel supply is initiated for each predetermined number of cycles following thereafter.

Description

METHOD AND SYSTEM FOR STARTING COMBUSTION ENGINES

The present invention relates to a method for starting combustion engines according to the preamble of claim 1, and a system used for the application of the method as specified according the preamble of claim 15.

Such a method is known from US-A-5.605.138, in which a method for starting engines with another type of fuel is described, wherein the amount of fuel supplied is established from a basic value. Depending on the change of engine rotation speed during starting, it is determined whether or not a correction to the basic value of the amount of fuel supplied is necessary. For instance, the fuel amount supplied can be increased by some 20% after each 2^nd revolution when no ignition occurs. This gradual increase is repeated until the engine catches up speed.

In US-A-5.579.737 another method is shown, wherein a possible refuelling event is detected. If no refuelling event is detected, the fuel correction established in the latest driving cycle is used. If refuelling is detected then the fuel correction is reset to a default value.

In US-A-3.982.519 yet another method for fuel enrichment during engine cranking is shown, wherein an under-stoichiometric amount of fuel for combustion is applied initially during starting. An incremental increase of the fuel amount is thereafter applied in parallel with the cranking process.

A method of supplying fuel to an internal combustion engine during start-up with a monotonous increase of fuel during cranking is shown in US-A-4.438.748.

In most combustion engines, which may run on either gasoline, diesel, LPG or Natural Gas, the fuel injection system adapts to the quality and properties of the fuel used. Starting engines with a different kind of fuel than the fuel used previously, could cause starting problems. In most cases the fuel system has been able to adapt to the fuel previously used, and has established some kind of correction or adaptation for the fuel supply.

A different kind of fuel could correspond to either of different blends of natural gas having differing calorific values, or different mixtures of gasoline/methanol for flexible fuel engines, or different blends of gasoline or diesel having differing octane or cetane numbers respectively, or gasoline having differing volatility.

Other solutions implemented in systems capable of being started with a different kind of fuel, have used some intermediate fuel tank. The intermediate fuel tank is isolated from the main fuel tank in a controlled manner. The purpose of such an intermediate fuel tank is to be able to start at the same kind of fuel as the fuel used previously, i.e. before shut-off. Even if the engine was run on pure gasoline before shut-off, and refuelled with methanol in the main fuel tank, the engine could be started successfully. Once the engine is started, the content of the intermediate tank will gradually assume the same quality as the fuel contained in the main fuel tank.

Other solutions comprise a fuel-quality sensor, but these sensors have demonstrated a low degree of reliability, and will also increase the cost for the fuel system considerably.

One problem with the prior art using incremental increase of fuel from a basic "lean" air-fuel mixture, is that a number of misfires occurs in all cylinders before the engine is started successfully. This is an unwanted situation with regard to environmental concerns as unburnt fuel will exit and the emission levels of hydrocarbons will increase. Also, the starting of the engine will take longer than necessary, leading to draining of the battery. A problem with solutions using intermediate fuel tanks is that a number of valves and fuel pumps must be implemented, increasing the cost of the system considerable.

An object of the invention is to provide a method and system for controlling fuel supply during start-up, in which the aforementioned disadvantages of the prior art methods and systems are avoided.

This object is achieved by a method according to the preamble of claim 1, in which the air-fuel ratio in each of said combustion chambers is controlled in such a manner that the air-fuel ratio obtained is different for each of said combustion chambers. With the inventive method, the proper air- fuel ratio for successful start of an engine can be found more quickly, by using different air-fuel ratios in each of the combustion chambers. Because of the quicker start-up of the engine, the possibility of draining the battery during the start attempts will decrease. Also, a quicker start-up of the engine will lead to a reduction of levels of hydrocarbons.

In an embodiment of the method according to the invention, the air- fuel ratio in a first of said combustion chambers is controlled to a first air-fuel ratio, and the air-fuel ratio for each of the other of said combustion chambers is reduced from the first air-fuel ratio by respective steps of a first predetermined amount. This will allow an even distribution of air-fuel ratios over the different combustion chambers, allowing a quicker detection of the proper air-fuel ratio needed for ignition. Preferably, the first predetermined amount is equal to the first air-fuel ratio minus a second air-fuel ratio, divided by the number of combustion chambers of the engine minus one. E.g., the first air-fuel ratio is the highest air-fuel ratio perceivable for the engine, and the second air- fuel ratio is the lowest air-fuel ratio perceivable for the engine. In this manner, the full range of possible air- fuel ratios is covered by all combustion chambers, and a successful ignition is likely to occur in one of the combustion chambers.

In a further embodiment of the method according to the invention, the air-fuel ratio of a second combustion chamber with a lower air- fuel ratio than the air- uel ratio in the first combustion chamber, is controlled to a second predetermined amount above the air-fuel ratio in the first combustion chamber during a predetermined number of cycles. This will realise a forced ventilation of the combustion chambers which initially received a too rich air- fuel mixture.

In a still further embodiment, the air-fuel ratio of a second combustion chamber with a lower air-fuel ratio than the air-fuel ratio in the first combustion chamber, is controlled to a second predetermined amount below the air-fuel ratio in- the first combustion chamber during a predetermined number of cycles, after detection of consecutive combustion and non-combustion in the first combustion chamber, in order to verify whether a slightly lower air-fuel ratio is needed for successful ignition. Detection of alternating combustion and non-combustion may be an indication that the air-fuel ratio in that combustion chamber is somewhat low for proper ignition, the alternating combustions being caused by recirculation of fuel from the previous cycle without combustion. This embodiment, therefore, provides a more reliable start-up of the engine. As an alternative, the second predetermined amount is determined by a recirculation model of residual fuel in said combustion chambers of the engine.

In a further embodiment of the method according to the invention, the air-fuel ratio in all said combustion chambers is reduced by a third predetermined amount when no successful ignition has been detected during a predetermined number of cycles. In some internal combustion engines, specifically those using Natural Gas, the tolerance for the proper air- fuel ratio for ignition may be very narrow. It may, therefore, happen that none of the air- fuel ratios in the combustion chambers will lead to a successful ignition. By equally changing all air-fuel ratios by a third predetermined amount (reduction or increment) during a predetermined number of cycles, an air-fuel ratio leading to combustion in one of the combustion chambers may be found. Preferrably, the predetermined number of cycles is in a range between 2 and 10, for example 4.

In a preferred embodiment of the method according to the present invention, the method is only applied after detection or anticipation of a difficult starting condition, such as a refuelling event, a short operation of the engine, or a long period of non- operation of the engine. In these circumstances, a change of the quality or other properties of the fuel used may have occurred, necessitating the use of the method according to the invention. In other circumstances, the engine will probably start quickly with the present settings of the fuel system. A second aspect of the present invention relates to a system for controlling fuel supply to a combustion engine, in which the combustion engine comprises at least two combustion chambers and at least one individual injector for supplying fuel to each individual combustion chamber, the system comprising memory means for storing a predetermined amount of fuel to be supplied dependent of operating conditions of the combustion engine, and correction values established during operation of the engine and needed in order to obtain optimum efficiency from the combustion engine by orrecting the predetermined amount of fuel to be supplied dependent of operating conditions of the combustion engine and combustion detection means for detecting a combustion in any of said combustion chambers, characterised in that the system further includes processor means which are arranged for execution of the method according to one of the claims 1 through 14.

A third aspect of the present invention relates to a computer readable medium characterised in that it comprises a software program, which, after downloading of the program in a motor management system provided with combustion detection means, provides the motor management system with the functionality of the method according to one of the claims 1 through 14.

In the following the present invention is described in more detail by describing exemplary embodiments of the invention. The description of embodiments is made by reference to the figures specified in the following list of figures. Fig. 1 shows schematically a combustion engine with starting and fuel supply systems;

Fig. 2 shows a flow diagram of a control algorithm for the fuel supply according the inventive method; Fig. 3 shows how air-fuel ratio is changed between cylinders according to the inventive method during cranking,

Fig. 4 shows an alternative change of air-fuel ratio according to the inventive method with incremental change of air- fuel ratio during cranking, and Fig. 5 shows an alternative change of air- fuel ratio according the inventive method with a forced ventilation of combustion chambers with indicated excessively rich air-fuel mixture.

In Fig. 1 a combustion engine 1 is shown with four combustion chambers 4a...4d.

The engine is equipped with an inlet manifold 2 and an exhaust manifold 3. Air for combustion is drawn into the combustion chambers via filter 20 and an air mass meter 22. The air mass meter 22 may be of the "hot-wire" type, by which the total mass drawn into the combustion chambers can be determined. Fuel is supplied to the individual cylinders 4a...4d from a main fuel tank 10, using a fuel pump 11 supplying fuel to the fuel rail 12 via fuel supply line 14. In Fig. 1 also an intermediate volume 17 is shown, which volume could visualise any intermediate volume between the main fuel tank 10 and the fuel rail 12. Such intermediate volumes could be formed by fuel feed -lines, fuel filters and fuel pressure accumulators etc., and could add up to considerable volumes.

In conventional fuel supply systems often a fuel return line 15 is installed between the fuel rail 12 and the fuel tank 10. The purpose of the return line 15 is to assure a certain continuous flow of fuel in the fuel rail 12, avoiding vapour lock problems caused by high engine temperatures. In other systems with no risk for vapour lock problems the fuel pump 11 could instead be controlled by a pressure sensor in the fuel rail 12, such that the pressure level is kept at an appropriate level. The fuel tank 10 is further provided with a fill opening 16 and a fuel quantity sensor 18. The amount of fuel being supplied to each individual cylinder 4a...4d is controlled by fuel injectors 13 (only the right hand injector for cylinder 4d indicated in Fig. 1). The duration of opening of the injectors is controlled by an Electronic Control Module ECM depending on detected engine operation parameters. By controlling the amount of fuel injected in each cylinder 4a...4d, the ECM actually controls the air-fuel ratio in each cylinder 4a...4d, as the air mass is known from the air mass meter 22. The ECM is provided with a memory module 21 for storing data.

The ECM detects the prevailing operational parameters of the engine in a conventional manner by using a coolant temperature sensor 6, an engine speed sensor 5, an inlet manifold pressure sensor 23 or the air mass meter 22.

From a lambda sensor 7 located in the exhaust manifold 3 also feed-back information relating to the air-fuel ratio can be obtained. The lambda sensor 7 is conventionally located in the exhaust stream before a catalytic converter 8. In some systems also a second lambda sensor could be used after the catalytic converter 8, mainly for diagnostic purposes of the converter 8. However, this information cannot be used when starting the engine 1, as the lambda sensor 7 does not function reliably when cold.

With a starter motor 9 engaging cogs at the periphery of a flywheel (not shown), the engine can be started. In Fig. 2 a flow-chart is shown of a software program describing the inventive method during starting of an engine of a type as shown in Fig. 1.

The software is installed in the ECM. When a starting attempt is detected in decision block 40, which occurs simultaneously with or shortly before activation of the starter motor, then the program proceeds to program sequence 41, wherein a test is made whether or not a difficult starting condition is detected or anticipated. Otherwise, the procedure returns to its starting position. A difficult starting condition is any type of condition wherein the engine 1 is intended to be started with a different type of fuel quality than the fuel quality used before the shut-off, the latter fuel quality hereafter designated as old fuel. This can be the case after refuelling the fuel tank 10, after a relatively short drive, or after a long parking period.

Combustion engines operating with natural gas could use fuel qualities with substantial difference in calorific value. Combustion engines operating with different mixtures of methanol and gasoline, i.e. all mixtures from 100% methanol to 100% gasoline, could also use fuel qualities with substantial difference in calorific value/octane numbers.

Also combustion engines being operated at standard fuels such as diesel and gasoline, could use fuel qualities with different octane numbers and cetane values, respectively. Also, in the case of gasoline, the volatility of the fuel may be different.

A different fuel could be supplied during starting if the main fuel tank 10 has been refuelled. A flexible fuel engine 1, operating on any mixtures of methanol and gasoline, could have been operated on pure gasoline before stopping for refuelling. If methanol is refuelled to the main tank, only having a small residual volume of gasoline in the order of some litres, then the overall mixture within the main fuel tank 10 would assume an almost pure methanol content.

In most fuel systems, a remaining quantity of the old fuel is always present in the fuel system, and a starting attempt could in most cases be successful. However, the new fuel is immediately supplied by the pump 11, and the combustion chambers 4a...4d are subjected to a transient change from the old fuel to the new mixture in the main tank 10. If such transient change is made with a running engine 1 , an adaptation to the new fuel quality could be made. Once the lambda sensor 7 reaches its operational temperature, a feedback signal is obtained from the combustion process indicative for the present air-fuel mixture. If the new fuel is of inferior quality a correction factor can be established, which correction factor will bring about an increase of fuel amount supplied, and hence a lower air- fuel ratio.

Detection of a refuelling event could be made by using a fuel quantity sensor 18 in the fuel tank 10, as illustrated in Fig. 1. The signal from the level indicator 18 is sent to the ECM, and by comparison with a previous level value stored in a memory 21 of the ECM, a refuelling event can be distinguished. In the case of Compressed Natural Gas engines, refuelling can be detected by a pressure sensor in the gas fuel- tank 10. Alternatively, a refuelling sensor in the fill opening 16 of the fuel tank 10 can be used.

In an alternative embodiment, used in systems where a refuelling detector is not implemented, a difficult starting condition may instead be assumed or anticipated if the engine has been operated during a predetermined short time. The last time period for an uninterrupted operation of the engine could be stored in the memory 21 of the ECM. If the last time period is close to a time period equivalent to the time period needed to consume the old fuel remaining in the fuel system, then a difficult starting condition could be present and indicated. Another difficult starting condition could correspond to a relatively long parking period. Especially if the fuel supply system is of a kind wherein a rather small volume of fuel is retained in the system between the tank 10 and the injectors 13, and/or if the fuel retained in the system between the tank 10 and the injectors 13 is easily drained back to the tank 10. If either of above conditions indicative for a difficult starting condition is detected or anticipated, the program proceeds to program sequence 42 wherein the correction factor established during the latest engine operation period is reset, e.g to 1.0. I.e. any offset of fuel amount determined to be necessary in order to obtain proper operation of the engine, is reset to a zero value.

In the next program sequence 43 the necessary fuel amount to be injected by each individual fuel injector 13 of cylinders 4a...4d (FCYLI-FCYW) is determined from the map stored in the memory 21 and a correction of the fuel amount for individual cylinders 4a...4d according to the present method. As noted earlier, controlling the amount of fuel to be injected by each individual fuel injector 13 also controls the air- fuel ratio in each cylinder 4a...4d.

The map is empirically determined, and for each combination of at least speed, load and temperature of the engine 1 a fuel amount is given by said map. In this example it is assumed that the map is determined for the best possible fuel perceivable by the engine 1. I.e. for each operation condition a rather small fuel volume is needed for proper combustion, and any correction of that basic fuel volume would cause a change in fuel volume. In practice the map would contain an average opening duration for the injectors 13. Program sequence 43 is explained also by reference to Fig. 3. The fuel volume selected according the inventive method for cylinder 4a, F_CYLI, could be given by the amount indicated by the map, i.e. MAP_FUEL, and the correction factor CF. If the above assumption that the map is obtained for the best available fuel quality is correct, the amount of fuel injected to F_CYLI would correspond to the volume given by the map. The relative air-fuel ratio obtained in cylinder 4a is shown in Fig. 3 by the dotted line starting at A/F] at the vertical A/F axis, e.g. at the leanest possible A/F-ratio perceivable by the engine 1.

The fuel volume selected according the inventive method for cylinder 4b, F_{CY 2}, could be given by the same amount as selected for cylinder 4a with a small additional amount Δf. The additional amount Δf is selected such that the air-fuel ratio obtained in the cylinders 4a...4d by the injected volume would differ by a fraction 1/X, where X is the number of cylinders of the engine, and within the range from the substantially leanest possible to the substantially richest possible air- fuel mixture perceivable by said engine 1. The fuel amount selected for cylinder 4b in a four cylinder engine could thus be expressed as;

F_CYL2 := CF * MAP_FUEL + ( ^lA * MAX_FUEL - MAP_FUEL) The relative air-fuel ratio obtained in cylinder 4b is shown in Fig. 3 by the hashed line starting below A/Fi at the vertical A F axis, i.e. below the dotted line initially corresponding to the leanest possible A/F-ratio perceivable by the engine 1 as injected in cylinder 4a.

In a similar manner the fuel amount according the inventive method is selected for cylinder 4c such that the air-fuel ratio obtained in the cylinder 4c is increased by a further amount Δf; i.e. expressed as;

F_CYL3 := CF * MAP_FUEL + ( 2/4 * MAX_FUEL - MAP_FUEL)

The relative air-fuel ratio obtained in cylinder 4c is shown in Fig. 3 by the dash- dot line.

For the last and "richest" cylinder 4d the air-fuel ratio obtained in the cylinder 4d would be increased by a further amount; i.e. expressed as;

F_CY := CF * MAP_FUEL + ( * MAX_FUEL - MAP_FUEL)

The relative air-fuel ratio obtained in cylinder 4d is shown in Fig. 3 by the dash- double-dot line.

As could be seen in Fig. 3 this is the "richest" condition obtained initially in cylinder 4d at a distance Δf from the richest possible condition perceivable by the engine, i.e. indicated by A/F₂ at the vertical axis A/F.

The leanest and richest possible conditions perceivable by the engine correspond in practice to the shortest and longest opening time for the injector, with the fuel qualities commercially available. The limits A F] and A/F₂ could be adjusted if any extreme type of fuel quality should be obtained.

When the engine 1 is cranked by the starter motor, cylinders 4a...4d are given different air-fuel ratios for each cylinder. The actual air-fuel ratio for each cylinder is maintained in Fig. 3 for the first three cycles. In the figure, it is indicated that a successful ignition is obtained in cylinder 4c for the 2^nd and 3^rd cycle. In this example a successful ignition is considered to be established if two consecutive combustions have been obtained in one individual cylinder 4a...4d. If the successful ignition is detected in program sequence 44 (see Fig. 2), then the program proceeds to sequence 47. If no combustion has been detected the other cylinders 4a, 4b and 4d, during the 2^nd and 3^rd cycle, the fuel amount for at least cylinders 4a and 4b is changed immediately to the same fuel amount used for cylinder 4c wherein a successful ignition has occurred.

However, cylinders wherein the air-fuel mixture is somewhat richer than the air fuel mixture of cylinders experiencing successful combustion, may require an alternative approach. For a cylinder with richer than optimal mixture, such as cylinder 4d in Fig. 3 it could be advantageous to subject the cylinder 4d to some forced ventilation. By a control of the fuel supply for a rather limited number of cycles, such that a fuel supply normally would cause a leaner than optimum air-fuel mixture, a ventilation effect can be obtained.

This somewhat leaner than optimum control is indicated in Fig. 3 for cylinder 4d from cycle 3 to cycle 5, with an amount of ΔfcoRR less fuel than the optimal air-fuel ratio

In Fig. 4 an alternative approach for the inventive method is shown, which method advantageously could be implemented in applications where a relatively narrow air-fuel ratio is requested in order to obtain a successful ignition. Such applications could be found with engines 1 being operated on fuel exhibiting a rather wide range of fuel qualities. Even though the injectors 13 are configured initially during cranking such that the air-fuel ratios obtained are evenly distributed within the perceivable range of the engine 1, a starting problem may occur, i.e. no combustion occurs in any cylinder 4a...4d.

In such circumstances it becomes necessary to implement additional routines to overcome this problem. These routines are indicated in blocks 45 and 46 of the flow diagram of Fig. 2. If no ignition occurs the flow diagram continues from decision block 44 to block 45. If the engine refuses to start for a predetermined number of cycles, e.g. 6 cycles in Fig. 4, which is detected in block 46 of Fig. 2, then the fuel amount to each individual injector is increased by predetermined amount ΔfsTEP, leading to lower air-fuel ratios A/F. This can for example be implemented by increasing the correction factor CF in block 46 of Fig. 2, after which the flow diagram continues with block 43 for determining the air- fuel ratio to be supplied to each cylinder 4a...4d.

The predetermined amount ΔfsT_EP by which the fuel amount is increased is only a fraction of the difference in fuel volume Δf supplied to another combustion chamber 4a...4d obtaining the closest but different air-fuel ratio A/F.

An additional incremental increase is initiated in an individual combustion chamber when a number of compression strokes in that combustion chamber have passed a number within a range of 2-10 compression strokes, preferably 4.

If no successful combustion is obtained, then any further incremental increase is interrupted when the increased air-fuel ratio for one cylinder 4a...4d is approaching the air- uel ratio initially used in the next "richer " cylinder 4a...4d. In this case, the first and second predetermined amounts Δf, ΔfsT_EP may be made smaller, and the method is again applied. In Fig. 4 it is shown that two incremental increase steps ΔfsTE have been initiated before a first successful combustion is detected, followed by a detected missing combustion and thereafter a detected second successful combustion. Such a combustion sequence with alternating combustion and non-combustion, is a typical behaviour if the air fuel mixture is somewhat lean, and the residual amount of fuel from a preceding non- combustion event is sufficient to cause the right air-fuel mixture for successful ignition of the air-fuel mixture.

In this case, all combustion chambers 4a...4d with a leaner air- fuel ratios A/F are changed to the same air- fuel ratio as the combustion chamber 4a...4d in which the first successful combustion was detected. The fuel supplied to the combustion chamber or chambers, here cylinder 4d, with richer air-fuel mixture is thereafter decreased in steps. After yet another third incremental increase step Δfsiεp has been initiated to the fuel supply of cylinders 4a, 4b and 4c, a reliable combustion is detected at each ignition event in cylinders 4a...4c. An advantage with such an approach is that at least three cylinders 4a...4c could be controlled such that a successful combustion is obtained rapidly in each of-the three cylinders 4a...4c, and the fourth remaining combustion chamber is only gradually changed in order to be able to determine if a somewhat richer air-fuel mixture is needed.

In Fig. 5 an alternative approach for the inventive method is shown, which method advantageously could be implemented in applications where wall-wetting problems occur or if large amounts of residual fuel could be contained within a combustion chamber 4a...4d after a misfire. In this embodiment an extended forced ventilation is initiated in those combustion chambers 4a...4d found to have been supplied with an excessively rich mixture. When the "ideal" (indicated at the vertical axis) air-fuel mixture A/F,_deai is found for cylinders 4a-4c, causing a successful combustion in each cylinder, an excessive lean-out control is implemented for the richer than ideal combustion chamber 4d. In this embodiment, the fuel amount to be injected is altered to the same amount as supplied initially to the cylinder 4c having a next leaner to the ideal air-fuel mixture A/F,deai- In this case, the fuel amount for cylinder 4d is altered to the same amount as initially supplied to cylinder 4c.

After this alteration the fuel amount is increased by a small incremental amount, ΔfcoRR, for a predetermined number of cycles at a time, and at least until the same fuel amount is supplied to cylinder 4d as to the remaining cylinders 4a..4c. In the embodiments shown, the combustion chamber numbers, i.e. 4a...4d, may correspond to the order of the cylinders in the engine block, as counted from the front end of the engine. In Fig. 1, cylinders 4a, 4b, 4c and 4d would then correspond to cylinders 4a, 4b, 4c and 4d in Fig. 3-5, respectively. Cylinders 4a to 4d as indicated in Fig. 3-5 could also correspond to the ignition order.

Alternative modes of implementation could be made without departing from the inventive concept. If a different air- fuel ratio A/F is obtained in at least two combustion chambers 4a...4d, and a successful combustion is detected in each 2^nd ignition event in any specific cylinder, it could be assumed that natural recirculation of residual fuel could cause the successful combustion. If the engine has a predictable behaviour model as concerns the amount of recirculation of residual fuel from an ignition event without combustion, this model could be used to increase fuel with a step rate ΔfcoRR as a function of the model.

Even after having controlled the air-fuel ratio A/F such that a successful combustion occurs at each ignition event, a further stepwise change of fuel can be implemented. Such a stepwise change of the fuel amount supplied, could be controlled in a closed-loop manner in a fuel increase as well as fuel decrease direction. Such closed loop information could be obtained in a number of ways. One method is to observe the momentary acceleration of the crankshaft, i.e. by using the rpm-sensor 5. Such an rpm- sensor 5 could be of the 72-x type, with 72 cogs arranged at the periphery of the flywheel and detected by said rpm-sensor. By measuring the time between two cogs passing the sensor, the momentary acceleration can be detected. A successful combustion could be detected when the acceleration value is above a certain threshold. An improved combustion could also be detected by an increase of the momentary acceleration value. Another method for detecting a successful combustion could use sensors mounted in the combustion chamber, detecting the actual combustion. Such sensors could be ionisation sensors, pressure-sensors or light-sensors, which detects ionisation, pressure increase or the intensity of light respectively. The invention is not restricted to the kind of combustion detection means. Although the present invention has been described in relation to particular embodiments thereof, many other variations and modifications and other uses will become apparent to those skilled in the art. It is preferred, therefore, that the present invention be limited not by the specification disclosure herein, but only by the appended claims.

Claims

1. Method for controlling fuel supply during start-up of a combustion engine having at least two combustion chambers, characterised in that the air-fuel ratio (A/F) in each of said combustion chambers (4a...4d) is controlled in such a manner that the air- fuel ratio (A/F) obtained is different for each of said combustion chambers (4a...4d).

2. Method according to claim 1, characterised in that the air- fuel ratio (A/F) in a first of said combustion chambers (4a...4d) is controlled to a first air-fuel ratio (A/Fi), and in that the air- fuel ratio (A/F) for each of the other of said combustion chambers (4a...4d) is reduced from the first air-fuel ratio (A/Fi) by respective steps of a first predetermined amount (Δf).

3. Method according to claim 2, characterised in that the first predetermined amount (Δf) is equal to the first air-fuel ratio (A/Fi) minus a second air- fuel ratio (A/F₂) divided by the number of combustion chambers (4a...4d) of the engine (1) minus one.

4. Method according to claim 2 or 3, characterised in that the first air-fuel ratio (A/Fi) is the highest air- fuel ratio (A/F_max) perceivable for the engine (1), and the second air-fuel ratio (A/F₂) is the lowest air- fuel ratio (A/Fmj_n) perceivable for the engine (1).

5. Method according to one of the preceding claims, characterised in that after detection of a successful ignition in a first of said combustion chambers (4a...4d), the air- fuel ratio (A/F) in at least one of the other of said combustion chambers (4a...4d) is controlled to be substantially equal to the air-fuel ratio of the first combustion chamber

(A/Fideal)-

6. Method according to claim 5, characterised in that a successful ignition is detected when a predetermined number of consecutive combustions is detected in one of the combustion chambers (4a...4d).

7. Method according to claim 5 or 6, characterised in that the air-fuel ratio of a second combustion chamber with a lower air-fuel ratio than the air-fuel ratio in the first combustion chamber, is controlled to a second predetermined amount (ΔfcoRR) above the air- fuel ratio in the first combustion chamber (A/F_ldeai) during a predetermined number of cycles.

8. Method according to claim 5 or 6, characterised in that, after detection of consecutive combustion and non-combustion in the first combustion chamber, the air- fuel ratio of a second combustion chamber with a lower air-fuel ratio than the air-fuel ratio in the first combustion chamber, is controlled to a second predetermined amount (Δfco_RR) below the air-fuel ratio in the first combustion chamber (A/F„._eai) during a predetermined number of cycles, in order to verify whether a slightly lower air- fuel ratio is needed for successful ignition.

9. Method according to claim 8, characterised in that the second predetermined amount (Δfco_RR) is determined by a recirculation model of residual fuel in said combustion chambers (4a...4d) of the engine (1).

10. Method according to claim 7 or 8, characterised in that the second predetermined amount (Δfco_Ri is reduced stepwise until the air-fuel ratio in the second combustion chamber is substantially equal to the air-fuel ratio in the first combustion chamber (A/F^eai)-

11. Method according to one of the preceding claims, characterised in that when no successful ignition has been detected during a predetermined number of cycles, the air- fuel ratio (A/F) in all said combustion chambers (4a...4d) is changed by a third predetermined amount (Δf_step).

12. Method according to claim 11, characterised in that the predetermined number of cycles is in a range between 2 and 10, for example 4.

13. Method according to one of the claims 5 through 12, characterised in that a successful ignition in a combustion chamber (4a...4d) is detected by detection of acceleration of the crankshaft of the engine (1), ionisation in a combustion chamber (4a..4d), pressure increase in a combustion chamber (4a...4d) or a change in light intensity in a combustion chamber (4a...4d).

14. Method according to one of the preceding claims, characterised in that the method is only applied after detection or anticipation of a difficult starting condition, such as a refuelling event, a short operation of the engine (1), or a long period of non- operation of the engine (1).

15. System for controlling fuel supply to a combustion engine, in which the combustion engine comprises at least two combustion chambers and at least one individual injector for supplying fuel to each individual combustion chamber, the system comprising:

- memory means for storing:

-a predetermined amount of fuel to be supplied dependent of operating conditions of the combustion engine, and -correction values established during operation of the engine and needed in order to obtain optimum efficiency from the combustion engine by correcting the predetermined amount of fuel to be supplied dependent of operating conditions of the combustion engine; and

- combustion detection means for detecting a combustion in any of said combustion chambers, characterised in that the system further includes processor means which are arranged for execution of the method according to one of the claims 1 through 14.

16. Computer readable medium characterised in that it comprises a software program, which, after downloading of the program in a motor management system provided with combustion detection means, provides the motor management system with the functionality of the method according to one of the claims 1 through 14.

*******