GB2438706A

GB2438706A - A method for controlling the fuelling of an engine.

Info

Publication number: GB2438706A
Application number: GB0708469A
Authority: GB
Inventors: Philip William Husak; Daniel Lawrence Meyer; Venki Nallaperumal; Michael John Cullen; Paul Roth; Michael Smokovitz
Original assignee: Ford Global Technologies LLC
Current assignee: Ford Global Technologies LLC
Priority date: 2007-05-02
Filing date: 2007-05-02
Publication date: 2007-12-05
Also published as: GB0708469D0

Abstract

A method for controlling the fuelling of an engine 10 is disclosed. The method comprising during an engine cold start and before the engine is warmed to a predetermined level, transitioning the engine from open-loop fuelling to closed-loop fuelling and providing a fuelling adjustment to a subsequent engine start in response to fuelling information obtained from the first start. The fuelling information is obtained over at least a complete cycle of closed-loop fuelling following the transition from open-loop fuelling to closed-loop fuelling.

Description

A Method for Controlling the Fuelling of an Engine.

This invention relates to the fuelling of an engine and in particular the fuelling of an engine during a cold start.

Engine starting during cold operating conditions, referred to as a "cold start", can present numerous challenges in maintaining repeatability/reliability and meeting emission requirements. Specifically, providing appropriate engine air-fuel ratio during engine starting conditions can be difficult due to numerous factors, especially given that exhaust gas oxygen sensors used for feedback air-fuel control are typically unavailable during the initial operation of a cold start. As such, the initial fuelling may be referred to as open-loop air-fuel control.

One phenomenon that can degrade cold start air-fuel ratio control is when a portion of injected fuel may not be available for combustion due to fuel vaporisation. This phenomenon may be referred to as "lost fuel" and can be significantly influenced by intake port surface temperature at start-up and fuel volatility (vapour pressure and distillation properties) . Further, lost fuel can significantly impact open-loop fuelling precision and accuracy and cause the observed open- loop air-fuel ratio to deviate from the desired target value.

One approach to provide improved air-fuel ratio control is provided in U.S. 6,266,957. In this example, upon identifying activation of an air-fuel ratio sensor and when an absolute value of the deviation between a target air-fuel ratio and an actual air-fuel ratio is equal to or greater than a predetermined value, a correction value is calculated at that moment and used to update an existing value within the backup RAM.

However, the inventors herein have recognized a disadvantage with such an approach. In particular, the amount of correction at the exact moment of sensor activation may not accurately reflect the open- loop fuelling error caused by lost fuel effects. Further, depending on the type of exhaust gas oxygen sensor provided, it may not be possible to identify how much error is present at the exact moment of sensor activation.

It is an object of the invention to provide an improved method for controlling the fuelling of an engine that at least partially overcomes the problems associated with the

prior art.

A method for controlling the fuelling of an engine wherein the method comprises, during an engine cold start and before the engine is warmed to a predetermined level, transitioning from open-loop fuelling to closed-loop fuelling during which feedback from an exhaust gas oxygen sensor is utilized and the closed-loop fuelling generates a cycling of delivered fuel to maintain an exhaust air-fuel ratio at a desired level and providing a fuelling adjustment to a subsequent engine start in response to fuelling information, the fuelling information being obtained over at least a complete cycle of closed-loop fuelling following the transition from open-loop fuelling.

The subsequent engine start may have conditions similar to that of the engine start.

The conditions may include ambient temperature.

The conditions may include an initial engine coolant temperature before starting commences.

The information may be obtained over at least a first complete cycle of closed-loop fuelling and a fractional part of a second complete cycle of closed-loop fuelling following the transition.

The information may be obtained over at least two or more complete cycles of closed-loop fuelling and a fractional part of a next complete cycle of closed-loop fuelling.

The providing may be enabled in response to whether degraded operation is present.

The complete cycle may be a first complete cycle following the transition.

An engine cold start may be when an engine coolant temperature is below a lower threshold.

The predetermined level may be an upper threshold engine coolant temperature.

The information may include an average fuelling amount during the complete cycle and the complete cycle is a first complete cycle following the transition.

The fuelling adjustment may include adjusting open-loop fuelling of the subsequent engine start.

Providing the fuelling adjustment may be performed during conditions where fuel vapour purging is disabled.

During an engine cold start and before the engine is warmed to a predetermined level may comprise during an engine cold start where engine coolant temperature is below a first threshold and before the engine is warmed where engine coolant temperature is above a second threshold.

Providing a fuelling adjustment may further comprise providing a fuelling adjustment to a subsequent engine start in response to fuelling information, the fuelling information being obtained over at least a complete cycle of closed-loop fuelling following the transition from open-loop fuelling and the complete cycle is a first complete cycle following the transition, the information including an average fuelling amount during the complete cycle and the fuelling adjustment including adjusting open-loop fuelling of the subsequent engine start.

During an engine cold start and before the engine is warmed to a predetermined level may comprise during an engine cold start where engine coolant temperature is below a first threshold and before the engine is warmed where engine coolant temperature is above a second threshold and the complete cycle may be a complete cycle being a first complete cycle following the transition, the information including an average fuelling amount during the complete cycle, the fuelling adjustment including adjusting open-loop fuelling of the subsequent engine start and the providing is performed during conditions where fuel vapour purging is disabled.

The invention will now be described by way of example with reference to the accompanying drawing of which:-Figure 1 shows a schematic engine diagram; Figure 2 shows an example of cold starting operation with accurate open-loop fuelling adjustments; Figure 3 shows an example of cold starting operation with lean errors in open-loop fuelling adjustments; :35 Figure 4 shows an example of cold starting operation with rich errors in open-loop fuelling adjustments; and Figures 5A, 5B and 6 show a control routine in accordance with the invention.

An internal combustion engine 10 comprising a plurality of cylinders, only one cylinder of which is shown in Figure 1, is controlled by electronic engine controller 12. The engine 10 includes a combustion chamber 30 and cylinder walls 32 with a piston 36 positioned therein. The piston 36 is connected to a crankshaft 13.

The combustion chamber 30 communicates with an intake manifold 44 and an exhaust manifold 48 via respective intake and exhaust valves 52 and 54. An exhaust gas oxygen sensor 16 is coupled to exhaust manifold 48 of engine 10 upstream of catalytic converter 20.

The intake manifold 44 communicates with throttle body 64 via throttle plate 66. Throttle plate 66 is controlled by electric motor 67, which receives a signal from ETC driver 69. ETC driver 69 receives control signal (DC) from controller 12. The intake manifold 44 is also shown having fuel injector 68 coupled thereto for delivering fuel in proportion to the pulse width of signal (fpw) from controller 12. The fuel is delivered to fuel injector 68 by a conventional fuel system (not shown) including a fuel tank, fuel pump and fuel rail (not shown) The engine 10 further includes conventional distributorless ignition system 88 to provide ignition sparks to combustion chamber 30 via a spark plug 92 in response to the controller 12. In the embodiment described herein, the controller 12 is a conventional microcomputer including: microprocessor unit 102, input/output ports 104, electronic memory chip 106, which is an electronically programmable memory in this particular example, random access memory 108 and a conventional data bus. The controller may further include a keep alive memory (not shown) for storing adaptive parameters.

Controller 12 receives various signals from sensors coupled to engine 10, in addition to those signals previously discussed, including: measurements of inducted mass air flow (MAF) from mass air flow sensor 110 coupled to throttle body 64; engine coolant temperature (ECT) from temperature sensor 112 coupled to cooling jacket 114; a measurement of throttle position (TP) from throttle position sensor 117 coupled to throttle plate 66; a measurement of turbine speed (Wt) from turbine speed sensor 119, where turbine speed measures the speed of a torque converter output shaft and a profile ignition pickup signal (PIP) from Hall effect sensor 118 coupled to crankshaft 13 indicating an engine speed (N) . Alternatively, turbine speed may be determined from vehicle speed and gear ratio.

Continuing with Figure 1, an accelerator pedal 130 is shown communicating with the driver's foot 132. The accelerator pedal position (PP) is measured by a pedal position sensor 134 and is sent to controller 12.

In an alternative embodiment, where an electronically controlled throttle is not used, an air bypass valve (not shown) can be installed to allow a controlled amount of air to bypass throttle plate 62. In this alternative embodiment, the air bypass valve (not shown) receives a control signal (not shown) from controller 12.

In another alternative embodiment, where a mass air flow sensor is not used, inducted mass air flow may be determined using a variety of computational methods. One example method, "speed-density", computes inducted air mass based on engine speed and throttle position.

As noted herein, during engine starting operation a portion of injected fuel may not be available for combustion due to fuel vaporisation. This phenomenon may be referred to as "lost fuel" and can be significantly influenced by intake port surface temperature at start-up and fuel volatility (vapour pressure and distillation properties) Other factors may influence "lost fuel". These can include, but are not limited to, intake manifold pressure, barometric pressure (altitude effects) and deposits on the intake valves and intake port passages. Further, lost fuel can significantly impact open-loop fuelling precision and accuracy and cause the observed open-loop air-fuel ratio to deviate from the desired target value. Figure 1 further illustrates via arrow 180 an example path where lost fuel is may pass through the engine.

Figure 2 shows an example trajectory of both a desired (or commanded) relative air-fuel ratio (LAMBSE) at 210 and measured relative exhaust gas air-fuel ratio (lambda) at 212 during the first 15 seconds after an engine start. The difference between the open-loop commanded LAMBSE and the measured exhaust gas air-fuel during the first 10 seconds of engine operation after start is primarily a result of lost fuel and labelled as such in Figure 2. As such, the profile of the commanded value 210 is purposefully modified to maintain the desired exhaust air-fuel ratio.

In this example, a closed-loop exhaust gas oxygen feedback signal is provided by a fast light-off HEGO (FLO HEGO) sensor 16. Stoichiometric (lambda = 1.0) is the desired or target open-loop air-fuel ratio during the first seconds of operation. The transition to closed- loop fuelling starts after 10 seconds and is shown as point A in Figure 2. This event occurs upon completion of the HEGO sensor warm-up period. Upon entering closed-loop control, LAMBSE exhibits the classic closed-loop limit-cycle scheduling. Initially, LAMESE integrates in one direction until the HEGO sensor switches, jumps back a specified amount and integrates in the opposite direction, then repeats. The completion of the first complete air-fuel ratio cycle 220, or switching cycle, is denoted at point B of Figure 2. Further, additional cycles are also shown.

In this example, the open-loop fuelling correctly accounts for lost fuel and provides approximately stoichiometric fuelling immediately prior to closed-loop operation. However, variations in lost fuel due to system aging, temperature, altitude and other parameters can cause differences between the open-loop air-fuel ratios, as illustrated in Figure 3.

Specifically, Figure 3 illustrates a lean open-loop fuelling error. In this example, the commanded air-fuel ratio (LAMBSE) trajectory is the same as Figure 2. However, the measured exhaust open-loop air-fuel ratio is leaner than the desired stoichiometric target value (1.1 vs. 1.0). At the transition from open-loop to closed-loop fuelling (point C), the feedback adjustment needs to compensate for approximately a 0.1 relative air-fuel ratio error.

Similarly, Figure 4 illustrates a rich open-loop fuelling error. In this example, the commanded air-fuel ratio (LAMBSE) trajectory is the same as Figure 2. However, the measured exhaust open-loop air-fuel ratio is richer than the desired stoichiometric target value (0.9 vs. 1.0) . At the transition from open-loop to closed-loop fuelling (point E), the feedback adjustment needs to compensate for approximately a 0.1 relative air-fuel ratio error (although in a direction opposite to that of Figure 3) In one example approach, it is possible to learn the above open-loop fuelling errors (e.g., learn variations in lost fuel) by monitoring the first one or more cycles of closed-loop air-fuel control after an engine start to adjust later open-loop cold start fuelling. For example, in the example of Figure 3, a correction of approximately 0.1 relative air-fuel ratio may be stored for those starting conditions (e.g., temperature, barometric pressure, shut-down time, engine speed, fuel type, alcohol content, etc.) so that during a subsequent start under similar conditions, the open-loop fuelling injection amount or timing may be adjusted to better compensate for lost fuel effects. In some cases, this open-loop correction term may be highly temperature dependent and thus may be computed, stored and applied as a function of ambient temperature, air charge temperature (ACT), engine coolant temperature (ECT) and/or cylinder head temperature (CHT) . In this way, conditions of the engine where the error is learned can be used to identify the appropriate correction for subsequent starts with similar conditions.

Similarly, an opposite fuelling adjustment of 0.1 could be used for the conditions of Figure 4. In this way, improved engine air-fuel ratio control may be achieved during engine starting when transitioning from open to closed loop operation.

Referring now to Figures 5A to 6, an example routine to provide fuel injection adjustment and adaptive lost fuel learning is described. Figure 5 provides an example cold-start idle adaptive (CIA) algorithm that begins at 510.

Next, at 512, entry conditions are checked. Example entry conditions requirements include whether the engine is in non-degraded run mode and that the calibratable CIA software selection switch is not set in the by-pass position. If so, an immediate exit from the routine is made. Otherwise, the routine continues to 514 to determine whether open-loop air-fuel control engine idling is present and whether any exception conditions are present.

-10 -Various open-loop exception conditions may be included, such as, for example, the following non-limiting examples: open-loop due to a failure mode (FNEM) condition; open-loop due to the open-loop exception flag being set; open-loop due to drive performance; forced open-loop; open-loop due to exhaust over-temperature; open-loop purge flag set (purge contributes to un-metered fuel) ; purge idle test running; purge monitor rate based idle test running; purge flow monitor test running; purge valve flowing; purge system is not providing expected control response; engine coolant temperature sensor (ECT), cylinder head temperature sensor (CHT), throttle position sensor (TPS), mass air-flow sensor (MAFS), electronic throttle control (ETC), gear selector switch (PRNDL), clutch switch, fuel rail pressure transducer (FRPT) faults or degradation; EGR valve stuck open; EGR intrusive test running; fuel injector and fuel pump faults or failures; deceleration fuel shut-off active; engine on-demand test running; secondary air monitor test running; fuel override enabled; and/or catalyst test running.

If the engine is not idling or if an exception condition is present, the routine exit. Otherwise, open-loop fuelling is scheduled at 516. The open-loop relative desired air-fuel ratio, LAMBSE[], is computed by adding an adjusted open-loop adaptive correction term, CIA OFS[], to the open-loop exhaust lambda, LAMBSEEXH[] . Note that in V-type engine applications, the above parameters and associated error terms may be correlated on a per bank basis -11 -and thus have unique values for each bank, indicated by brackets [1, for example.

Continuing with Figure 5A, in 516, the open-loop adaptive correction term, CIA OFS[], is multiplied by the ratio of KAMRF[] to KAMRFCIALAST[], where KAMRF[] is the closed-loop adaptive air-fuel correction factor stored in keep-alive memory (KAM) at the start and KAMRF CIA LAST[] is the KAMRF[] value stored in memory at the time when CIA OFS[] is computed (see 542) . In this way, it is possible to utilize closed-loop adaptive learning in order to compensate for air-fuel ratio offset errors that are caused by certain events or actions, which may occur subsequent to the completion of the CIA algorithm. For example, consider a refuelling event that occurs while the engine is fully warmed-up. If a significant quantity of fuel (e.g., more than the tank capacity) is replaced with a fuel that has an air-fuel stoichiometry vastly different from the fuel originally in the tank, a HEGO sensor will observe a change in the stoichiometric switching point.

Assuming that sufficient time at closed-loop operation follows this refuelling event, the closed-loop air-fuel adaptation routine will detect and correct the offset error and this will be reflected by a change to KAMRF[] Multiplying CIA OFS[] by KAMRF[] over KAMRF CIA LAST[] will further improve compensation for this air-fuel ratio change on the next cold-start.

The routine then proceeds to 518 where the parameter, CIAOLLAMBSE[] is assigned the value of the most recently scheduled open-loop command LAMESE[]. At 520, the entry conditions for closed-loop fuelling are checked and, if not satisfied, the routine exits. Otherwise, the routine proceeds to 522 where closed-loop fuelling based upon exhaust gas oxygen sensor feedback is invoked using the typical limit-cycle method (e.g., P1 control) . However, -12 -while such closed-loop control is used, the approach described herein may be used with various closed-loop control other than those that use limit-cycle exhaust gas oxygen feedback. For example, closed-loop fuelling can be based on the exhaust air-fuel ratio feedback signal from a proportional-readout sensor, such as, a Universal Exhaust Gas Oxygen (USGO) sensor.

Then, the routine proceeds to 524 to check for closed-loop idle operation and the presence of exception conditions. Excluding those items that are specifically associated with open-loop operation, the exception conditions may be the same as those described in 514, with the addition of certain exhaust gas oxygen (EGO/HEGO) sensor related exception conditions, for example. These may include HEGO sensor degradation or faults and/or upstream EGO monitor high frequency modulation. If the engine is not in closed-loop idle or an exception condition is present, the routine exits.

However, in the example of EGO/HEGO degradation or faults for V-engine applications in which there is a feedback senor in each bank, the routine may still continue to provide adjustment and/or learning for a bank of cylinders with properly functioning sensors only. In still another example, should one bank have a degraded sensor, the routine may continue execution using feedback from the bank that has the functioning sensor to provide control and learning for both banks. Such sensor substitution may be limited to conditions where the difference in air-fuel ratio between engine banks does not exceed a calibratable limit value prior to the EGO/HEGO degradation in the one bank.

Continuing with Figure 5A, if the answer to 524 is Yes, the routine continues to 526 where the limit cycle centre (average), LAMAVE[], for the closed-loop lambse is calculated. For example, the routine may determine the -13 -average value over a first cycle of closed-loop operation following open-loop fuelling during an engine start. The first cycle used may be the first complete cycle of fuelling oscillation during closed-loop control and may begin after an initial correction, as shown in the above Figures. Also, while an average value may be used, various other parameters indicative of an average value or similar value may be used.

Further still, the averaging technique may vary depending on the type of sensor used for feedback control. For example, the averaging techniques for a switching EGO/HEGO type sensor may be different than when a UEGO type sensor is used.

Next, at 528, the routine determines whether a sufficient computational interval for lambda averaging has elapsed. As noted above, the averaging interval may be a first air-fuel limit cycle, or a first number of limit cycles, or may be based on a number of engine combustion cycles of a first one or more air-fuel limit cycles following commencement of closed-loop control, for example.

The size of this interval may further be based on sensor characteristics, statistical significance and other noise factors and thus may be calibratable. If this calibratable interval has not been exceeded, the process returns to 524; otherwise, the routine proceeds to 530.

On Fig.5B at 530, the lambda (fuelling) difference, CIA LAM DIFF[] at the transition point from open-loop to closed-loop fuelling is calculated by subtracting the value of the last open-loop lambda command prior to going closed-loop, CIAOLLAMBSE[], from the averaged closed-loop lambda command, LAMAVE [].

The routine then proceeds to 532 where the value of the open-loop lambda (fuelling) error term at the transition from open-loop to closed-loop fuelling, CIA LAM ERROR[], is computed by subtracting the quantity, (1 -LAMOLDES[]), -14 -from the value of CIA LAN DIFF[] computed in 530.

LAMOLDES[] represents the desired or intended open-loop lambda command value just prior to the open-loop to closed-loop transition. LAMOLDES[] may be both calibration and engine temperature dependent.

The routine next proceeds to 534 where the absolute value of the CIA LAM ERROR[] calculation is compared to the absolute value of a calibratable error hysteresis dead-band term, CIA LAM ERROR HYS. In this way, it is possible to mitigate potential oscillatory behaviour of the control caused by very small error perturbations. If the value of CIA LAM ERROR[] is less than (within) the hysteresis dead-band value, the process proceeds to 536, where the CIA LAM ERROR[] is assigned the stored lambda error value from the last execution of the routine, CIA_LAM_ERROR_LAST H. The process then proceeds to 542. If the value of CIALAMERROR[] is greater than (outside) the hysteresis dead-band value, the process proceeds to 538.

In 538, a proportional, CIAP[], derivative, CIAD[] and integral, CIAI[], controller terms are computed. While this example uses PID control, various other control approaches may be used. Continuing with the PID example, the proportional controller term, CIAP[], is the product of a proportional gain term, CIAGP[] and CIA LAM ERROR[]. The derivative controller term, CIA_D[], is the product of a differential gain term, CIAGD[] and the difference between the current lambda error value, CIA LAM ERROR[] and the stored lambda error value from the last execution of the routine, CIALAMERRORLAST[]. The integral controller term, CIAI[], is the product of an integral gain term, CIAGI[] and the sum of the current lambda error value, CIA LAM ERROR[] and the stored integral controller term value, CIA ILAST[], from the last execution of the routine.

-15 -Note that, as mentioned above, the cold-start "lost fuel" effect, where a large portion of the injected fuel is not available in cylinder for combustion, may be influenced by intake port surface temperature at start-up and fuel volatility (vapour pressure and distillation properties) Therefore, the values for the proportional, differential and integral gain terms may be at least partially dependent upon either engine coolant or cylinder head temperature (ECT or CHT), as well, as upon other conditions These conditions may include a partial dependence on barometric pressure (altitude effects) . Also, the dependencies may be either linear or non-linear.

The routine then proceeds to 540, where an open-loop adaptive offset, CIA OFS[], is computed by combining the proportional, derivative and integral controller terms -CIAP[], CIA DEl and CIAI[].

At 542, CIA LAM ERROR LAST[] is assigned the CIA LAM ERROR[] value from either 532 or 536 and stored in memory. CIA ILAST[] is assigned the CIAI[] value from 538 and stored in memory. CIAOFSLAST[] is assigned the CIA OFS[] value from either 540 or 516 and stored in memory.

KAMRFCIALAST[] is assigned the current value for KAMRF[] and stored in memory. The CIA OFS[] value is further stored in memory. Memory storage may be in the form of a single value, a two-dimensional transfer function (f of x) value; or a multi-dimensional look-up table value. The memory storage locations for the transfer function or look-up table are parameter dependent. These parameters may include, but are not limited to, engine operating temperatures (ECT or CHT) and/or barometric pressures. Parameter dependency may be linear or non-linear. These stored values can then be used upon the next execution of the routine. Finally, the routine exits.

-16 -While Figures 5A and 53 show one example routine, various alternative embodiments may be used. Referring to Figure 6, one example alternative is shown for calculating the cold idle adaptive proportional, integral and derivative controller terms.

Specifically, the routine uses similar acts up through 532, but then continues to 610 where a delta lambda error term, CIA DELTA LAM ERROR, is computed by subtracting a calibratable error hysteresis dead-band term, CIA LAM ERROR HYS from the CIA LAM ERROR[1 term. Next, at 612, the absolute value of the CIA LAM ERROR[] is compared to the absolute value of a calibratable error hysteresis dead-band term, CIALAMERRORHYS. If the calculated value of CIA LAM ERROR[] is within the hysteresis dead-band value, the process proceeds to 614, where the CIA DELTA LAM ERROR[] is assigned the stored delta lambda error value from the last execution of the routine, CIA DELTA LAM ERROR LAST[] The process then proceeds to 620. Otherwise, if the calculated value of CIA LAM ERROR[1 is outside the hysteresis dead-band value, the process proceeds to 616.

At 616, a proportional, CIAP[], derivative, CIA DEl and integral, CIAI[l, controller terms are computed. The proportional controller term, CIAP[], is the product of a proportional gain term, CIAGP[] and CIA DELTA LAM ERROR[] The derivative controller term, CIA DEl, is the product of a differential gain term, CIAGD[] and the difference between the current delta lambda error value, CIA DELTA LAM ERROR[] and the stored delta lambda error value from the last execution of the routine, CIADELTALAMERRORLAST[]. The integral controller term, CIAI[], is the product of an integral gain term, CIAGI[] and the sum of the current delta lambda error value, CIA DELTA LAM ERROR[ I and the stored integral controller term value, CIA ILAST[], from the last execution of the routine. -17 -Again, the values for the proportional, differential and/or

integral gain terms used in 616 may be at least dependent upon either engine coolant or cylinder head temperature (ECT or CHT), as weiLl as, upon other conditions including a partial dependence on barometric pressure (altitude effects) . Also, the dependencies may be either linear or non-linear.

The routine then proceeds to 618, where the open-loop adaptive offset, CIA OFS[], is computed by combining the proportional, derivative and integral controller terms -CIAP[], dADE] and CIAI[] . At 620, CIA_DELTA_LAM_ERROR_LAST [] is assigned the CIA DELTA LAM ERROR[] value from either 610 or 614 and stored in memory. CIA ILAST[1 is assigned the CIAI[] value from 616 and stored in memory. CIAOFSLAST[] is assigned the CIA OFS[] value from either 618 or 516 and stored in memory. KAMRFCIALAST[] is assigned the current value for KAMRF[] and stored in memory. The CIA OFS[] value is stored in memory, as previously described for 542 in Figure 5.

These stored values will then be used upon the next execution of the routine. Finally, the routine exits.

Various advantageous elements are illustrated via the above routines, including the use of adaptive terms having integral and derivative terms, in addition to a proportional term, thereby providing improved learning. Further, updating the adaptive term before adding it to the open-loop lambda term computed from the feedback execution of the open-loop A/F subroutine can provide improved response.

This is accomplished by multiplying the adaptive term by the ratio of the KAMRF[ ] (the closed-loop adaptive air-fuel correction factor stored in keep-alive memory [KAM]) at the start and KAMRFCIALAST[ ] (the KAMRF[ ] value stored in memory at the time when CIA OFS[ I is computed) before it is added to the normally computed open-loop lambda. Also, the routine may suspend computation of the adaptive term while -18 -certain open-loop or closed-loop conditions are present, which can result in the introduction of unmetered air or fuel. These can include, but are not limited to, deceleration fuel shutoff (DFSO), open-loop/closed-loop fuel vapour purge and diagnostic self-tests, for example.

Computation of the adaptive term may also be suspended when certain sensor faults, failures and/or errors are present.

As illustrated by the above example routines, various o operations may be achieved to provide improved results. For example, returning to Figure 2, at the time of transition from open-loop to closed-loop fuelling control, i.e. Point A, the most recent or last value for the open-loop fuelling command is recorded by the routine and stored as the term, CIAOLLAMBSE. In this case, CIAOLLAMBSE would have a value of 1.0. Upon entering closed-loop control, LAMBSE exhibits the classic closed-loop limit-cycle scheduling.

Initially, TJAMBSE integrates in one direction until the HEGO sensor switches, jumps back a specified amount and integrates in the opposite direction, then repeats. The LAMESE value may then be filtered over the first full period of limit-cycle operation in order to obtain an averaged value for LAMBSE. This filtered value, LAMAVE, may be determined at Point B, where 1.0 is the value in this example. While this example uses only the first full cycle, additional cycles may be used under some conditions depending on sensor response characteristics. Further, a second and/or other subsequent cycle or cycles may be utilized in lieu of the first cycle. Once LAMAVE and CIAOLLAMBSE have been determined, a difference term of the two values, CIA LAM DIFF, may be computed. The computed CIA LAN DIFF value is zero for this example, indicating that the initial open-loop fuelling accurately approximated lost fuel and thus no adjustment or adaptation for the present conditions is used.

-19 -Thus, in this example, where stoichiometry is the expected value for the open-loop air-fuel ratio immediately prior to closed-loop operation, the desired CIA LAM 01FF value should be zero. Any deviation from this desired value of zero is considered a system error, CIA_LAM_ERROR. The gain factors can then be applied to the system error and proportional, derivative and/or integral controller terms are generated. As shown in Figures 5-6, these are then combined to produce an open-loop adaptive fuelling correction term, CIA OFS. CIA OFS is stored and subsequently used to offset the open-loop air-fuel commands during the open-loop fuelling period on the next engine start. As noted above, in one example, the various terms used to compute CIA OFS, for example, the proportional, integral and differential gain multipliers, also have temperature and/or barometric pressure dependencies to more accurately account for temperature and/or altitude effects on lost fuel.

Further examples of operation provided by the above routines can be illustrated by returning to Figure 3.

Again, Figure 3 illustrates a lean open-loop fuelling error scenario. In this example, at the transition from open-loop to closed- loop fuelling (Point C), the CIAOLLAMBSE term has a value of 1.0. At Point D, the LAMAVE is determined as described previously, but with a value of 0.9 in this example. After determining CIAOLLAMBSE and LAMAVE, the difference of these two values, CIA LAM DIFF, is then computed. In the example shown, CIA LAM DIFF is assigned a value of -0.1, which is non-zero. Since stoichiometry is the expected value for the open-loop air-fuel ratio immediately prior to closed-loop operation in this example, the desired CIA LAM 01FF value should be zero. Therefore, after comparing the computed and desired CIA LAM DIFF, the system error, CIA LAM ERROR, is equal to the computed CIA LAM 01FF and has a value of -0.1.

-20 -Following the approach outlined herein, the CIA_LAM_ERROR is used to generate the integral, proportional and derivative controller terms. These are combined to produce the open-loop adaptive correction term, CIA OFS, which is stored and used to offset the open-loop LAMBSE commands during the open-loop fuelling period on the next and subsequent cold-starts. The effect will be to reduce the exhaust gas air-fuel ratio error on these subsequent starts. Further, corrective adaptation over subsequent starts will result in an open-loop exhaust air-fuel ratio trajectory that more closely follows the desired or ideal trajectory shown in Figure 2.

Still another example of operation provided by the above routines can be illustrated by returning to Figure 4.

Again, Figure 4 illustrates an air-fuel ratio error similar to Figure 3, but in the opposite direction. CIAOLLAMBSE and LAMAVE are computed at Points E and F, respectively.

Note that the sign of the lambda difference parameter, CIA LAM DIFF, has changed and, when used to generate an adaptive correction term, will shift the exhaust gas air-fuel ratio in the opposite or lean direction on subsequent engine starts. This example also assumes that stoichiometry is the expected value for the open-loop air-fuel ratio immediately prior to closed-loop operation.

Although the examples illustrated herein utilize stoichiometry (lambda = 1.0) as the desired target air-fuel ratio at the end of the open-loop fuelling period, this control methodology can also adaptively correct open-loop fuelling errors for those applications where the desired target air-fuel ratio is either rich or lean of stoichiometry (i.e. lambda < 1.0 or lambda > l.0) Further, for the examples in Figures 2 through 4, a fast light-off HEGO (FLO HEGO) sensor may be used to provide the closed-loop exhaust gas oxygen feedback signal. It -21 -should be noted that this control methodology can utilize the signals from various styles of feedback sensors, including those that can provide a direct reading of the exhaust gas air-fuel ratio, such as, the UEGO (universal exhaust gas oxygen) sensor.

Note that the control routines included herein can be used with various engine configurations, such as those described above. The specific routine described herein may represent one or more of any number of processing strategies such as event-driven, interrupt-driven, multi-tasking, multi-threading and the like. As such, various steps or functions illustrated may be performed in the sequence illustrated, in parallel, or in some cases omitted.

Likewise, the order of processing is not necessarily required to achieve the features and advantages of the example embodiments described herein, but is provided for ease of illustration and description. One or more of the illustrated acts, steps, or functions may be repeatedly performed depending on the particular strategy being used.

Further, the described steps may graphically represent code to be programmed into the computer readable storage medium in controller 12.

Therefore in summary, a method for controlling fuelling of an engine is disclosed in which, during an engine cold start and before the engine is warmed to a predetermined level, the engine is transitioned from an open-loop fuelling to a closed-loop fuelling during which feedback from an exhaust gas oxygen sensor is utilized and where the fuelling generates a cycling of delivered fuel to maintain an exhaust air-fuel ratio at a desired level. The method further providing a fuelling adjustment to a subsequent engine start in response to fuelling information, the fuelling information obtained over at least a complete cycle of closed-loop fuelling following said transition from open-loop fuelling.

-22 -In this way, it is possible to utilize feedback information to obtain a more accurate determination of appropriate fuelling during cold start open-loop conditions, thereby better accounting for variations in lost fuel. For example, as the engine ages, lost fuel can vary, thereby leading to increased emissions if not otherwise corrected.

In one particular aspect, by using cycle average information of first complete fuelling cycle, it is possible to obtain ever more accurate fuelling corrections. In another aspect, the fuelling adjustment is provided only under select conditions to avoid inaccurate readings that may be caused by various conditions.

It will be appreciated that the configurations and routines disclosed herein are exemplary in nature and that these specific embodiments are not to be considered in a limiting sense, because numerous variations are possible.

For example, the above technology can be applied to V-6, I- 4, 1-6, V-8, V-lO, V-l2, opposed 4 and other engine types.

It will be further appreciated by those skilled in the art that although the invention has been described by way of example with reference to one or more embodiments it is not limited to the disclosed embodiments and that one or more modifications to the disclosed embodiments or alternative embodiments could be constructed without departing from the scope of the invention as set forth in the appended claims. :3 ()

Claims

-23 -Claims 1. A method for controlling the fuelling of an engine

wherein the method comprises, during an engine cold start and before the engine is warmed to a predetermined level, transitioning from open-loop fuelling to closed-loop fuelling during which feedback from an exhaust gas oxygen sensor is utilized and the closed-loop fuelling generates a cycling of delivered fuel to maintain an exhaust air-fuel ratio at a desired level and providing a fuelling adjustment to a subsequent engine start in response to fuelling information, the fuelling information being obtained over at least a complete cycle of closed-loop fuelling following the transition from open-loop fuelling.

2. A method as claimed in claim 1 wherein the subsequent engine start has conditions similar to that of the engine start.

3. A method as claimed in claim 2 wherein the conditions include ambient temperature.

4. A method as claimed in claim 2 or in claim 3 wherein the conditions include an initial engine coolant temperature before starting commences.

5. A method as claimed in any of claims 1 to 4 wherein the information is obtained over at least a first complete cycle of closed-loop fuelling and a fractional part of a second complete cycle of closed-loop fuelling following the transition.

6. A method as claimed in any of claims 1 to 5 wherein the information is obtained over at least two or more complete cycles of closed-loop fuelling and a fractional part of a next complete cycle of closed-loop fuelling.

-24 - 7. A method as claimed in any of claims 1 to 6 wherein the providing is enabled in response to whether degraded operation is present.

8. A meLhod as claimed in any of claims 1 to 7 wherein the complete cycle is a first complete cycle following the transition.

9. A method as claimed in any of claims 1 to 8 wherein an engine cold start is when an engine coolant temperature is below a lower threshold.

10. A method as claimed in any of claims 1 to 9 wherein the predetermined level is an upper threshold engine coolant temperature.

11. A method as claimed in any of claims 1 to 10 wherein the information includes an average fuelling amount during the complete cycle and the complete cycle isa first complete cycle following the transition.

12. A method as claimed in any of claims 1 to 11 wherein the fuelling adjustment includes adjusting open-loop fuelling of the subsequent engine start.

13. A method as claimed in any of claims 1 to 12 wherein providing the fuelling adjustment is performed during conditions where fuel vapour purging is disabled.

14. A method as claimed in any of claims 1 to 7 wherein during an engine cold start and before the engine is warmed to a predetermined level comprises during an engine cold start where engine coolant temperature is below a first threshold and before the engine is warmed where engine coolant temperature is above a second threshold.

-25 - 15. A method as claimed in claim 14 wherein providing a fuelling adjustment further comprises providing a fuelling adjustment to a subsequent engine start in response to fuelling information, the fuelling information being obtained over at least a complete cycle of closed-loop fuelling following the transition from open-loop fuelling and the complete cycle is a first complete cycle following the transition, the information including an average fuelling amount during the complete cycle and the fuelling adjustment including adjusting open-loop fuelling of the subsequent engine start.

16. A method as claimed in any of claims 1 to 7 wherein during an engine cold start and before the engine is warmed to a predetermined level comprises during an engine cold start where engine coolant temperature is below a first threshold and before the engine is warmed where engine coolant temperature is above a second threshold and the complete cycle is a complete cycle being a first complete cycle following the transition, the information including an average fuelling amount during the complete cycle, the fuelling adjustment including adjusting open-loop fuelling of the subsequent engine start and the providing is performed during conditions where fuel vapour purging is disabled.

17. A method for controlling the fuelling of an engine substantially as described herein with reference to the accompanying drawing.