CN111022196B - Skip fire transition control - Google Patents

Abstract

Methods and arrangements are described for controlling transitions between firing fractions during skip fire operation of an engine to help smooth the transitions. Generally, the transition in firing fraction is performed gradually, preferably in a manner that tracks the manifold filling dynamics relatively closely. In some embodiments, the commanded firing fraction is changed for each firing opportunity. Another approach contemplates changing the commanded firing fraction by substantially the same amount each firing opportunity for at least a portion of the transition. These approaches work particularly well when the commanded firing fraction is provided to a skip fire controller that includes an accumulator function that tracks a portion of a firing that has been requested but not delivered, or vice versa. In various embodiments, the change in commanded firing fraction is delayed relative to the onset of a change in throttle position to help compensate for the inherent delay associated with a change in manifold air pressure.

Description

Skip fire transition control

This application is a divisional application of a patent application entitled "skip fire transition control" and application number 201580048081.5 filed on 9, 15/2015, which is incorporated herein by reference in its entirety.

Cross Reference to Related Applications

This application claims priority to U.S. provisional application No. 62/053,351, filed on 9/22/2014, which is incorporated herein by reference in its entirety.

Technical Field

The present invention relates generally to methods and arrangements for controlling transitions between firing fractions during skip fire operation of an engine.

Background

The fuel efficiency of many types of internal combustion engines can be substantially improved by varying the displacement of the engine. This allows maximum torque to be available when needed, and also significantly reduces pumping losses and improves thermodynamic efficiency by using smaller displacements when maximum torque is not needed. The most common method of varying displacement today is to deactivate a group of cylinders substantially simultaneously. In this approach, fuel is not delivered to the deactivated cylinders and their associated intake and exhaust valves are kept closed as long as the cylinders remain deactivated. For example, an 8-cylinder variable displacement engine may deactivate half of the cylinders (i.e., 4 cylinders), such that only the remaining four cylinders are used for operation. Commercially available variable displacement engines available today typically support only two or at most three fixed modes of displacement.

Another engine control approach to changing the effective displacement of the engine is referred to as "skip fire" engine control. In general, skip fire engine control contemplates selectively skipping firing of certain cylinders during selected firing opportunities. Thus, a particular cylinder may be fired during one engine cycle and then may be skipped during the next engine cycle, and then selectively skipped or fired during the next engine cycle. In this way, even more refined control of the effective engine displacement is possible. For example, firing every third cylinder in a 4-cylinder engine will provide an effective displacement of 1/3 that is the maximum engine displacement, a component displacement that cannot be obtained by simply deactivating a group of cylinders. Conceptually, essentially any effective displacement can be achieved using skip fire control, but in practice most implementations limit operation to a set of available firing fractions, sequences, or patterns. The applicant has filed a number of patents describing a number of different approaches to skip fire control.

One known feature of skip fire control is that engines operating under skip fire control tend to have less desirable noise, vibration and harshness (NVH) than "normal" all-cylinder operation of the engine. Accordingly, there is a continuing effort to develop techniques and mechanisms that can help reduce NVH issues during skip runs while still maintaining certain benefits thereof. Typically, the available skip fire firing fractions/sequences/patterns are selected based at least in part on their preferred NVH characteristics. While reducing NVH when operating at these firing fractions, NVH issues may arise during transitions between different firing fractions. This application describes techniques that may help manage NVH issues while achieving desired performance during transitions between different firing fractions.

Disclosure of Invention

Methods and arrangements are described for controlling transitions between firing fractions during skip fire operation of an engine to help reduce undesirable NVH consequences and smooth the transitions. Generally, the transition in firing fraction is performed gradually, preferably in a manner that tracks the manifold filling dynamics relatively closely.

In some preferred implementations, the commanded firing fraction is changed for each firing opportunity. Another approach described contemplates changing the commanded firing fraction by substantially the same amount each firing opportunity for at least a portion of the transition. These approaches work particularly well when the commanded firing fraction is provided to a skip fire firing timing determination module that includes an accumulator function that tracks a portion of firings that have been requested but not delivered, or that have been delivered but not requested.

In various embodiments, the change in commanded firing fraction is delayed relative to the onset of a change in throttle position to help compensate for the inherent delay associated with a change in manifold air pressure.

In some implementations, the commanded firing fraction is varied such that the product of the skip fraction and the intake manifold pressure remains substantially constant throughout the transition.

In some implementations, the commanded firing fraction is changed using a linear slew rate per firing opportunity such that the amount by which the commanded firing fraction is changed per firing opportunity is the same throughout the transition. The actual slew rate appropriate for any particular transition may be varied based on a number of factors, such as the magnitude of the desired firing fraction change and a number of different engine operating parameters, such as engine speed, etc.

A wide variety of other engine control techniques may be used, including spark retard, feed-forward throttle control, feed-forward camshaft control, pumping air through skipped cylinders, and other techniques to help further smooth the transition.

Techniques are also described to address situations where the target firing fraction changes in the middle of the transition.

Drawings

The invention, together with its advantages, may best be understood by reference to the following description taken in conjunction with the accompanying drawings in which:

FIG. 1 is a functional block diagram of a skip fire controller according to one embodiment of the present invention.

Fig. 2 schematically shows the influence of mode vibration and torque mismatch type vibration during the transition.

3(a) -3(d) are a set of graphs showing requested and adjusted firing fractions, throttle position, intake manifold pressure, and total engine torque during a representative expected constant torque transition between firing fraction l/3 and firing fraction 2/3 when the transition is smoothed by using a first order bandpass filter.

4(a) -4(d) are a set of graphs illustrating two types of adjusted firing fractions, throttle position, intake manifold pressure, and total engine torque during a representative expected constant torque transition between firing fraction l/3 and firing fraction 2/3 when a delayed linear transition is used, in accordance with an embodiment of the present invention.

Fig. 5 is a graph showing the variation of firing fraction (Y-axis) over time (X-axis) during a representative interrupted transition where a second firing fraction transition is requested while a first transition is being made.

6(a) -6(d) are a set of graphs illustrating the adjusted firing fraction, intake manifold pressure, spark timing, and total engine torque during a representative expected constant torque transition between firing fraction l/3 and firing fraction 2/3 when a retarded linear transition is used in accordance with an embodiment of the present invention.

FIG. 7 is a graph illustrating the adjusted firing fraction and intake manifold pressure during a representative expected constant torque transition between firing fraction l/3 and firing fraction 2/3 when a delayed linear transition is used and air is pumped through the deactivated cylinders, according to an embodiment of the present invention.

FIG. 8 is an exemplary lookup table of firing fraction slew rates for different initial and target firing fractions in accordance with one embodiment of the present invention.

In the drawings, the same reference numerals are sometimes used to designate the same structural elements. It should also be appreciated that the depictions in the figures are diagrammatic and not to scale.

Detailed Description

When a limited set of firing fractions is available, transitioning between different firing fractions typically also involves adjustments to selected engine operating parameters. This is because, at any particular firing density, there will be associated operating parameters (e.g., air intake, spark timing, etc.) that are suitable for effectively delivering the desired engine output. Therefore, when making changes to firing density, it is typically desirable to substantially simultaneously adjust selected engine operating parameters to maintain a desired engine output at the new firing fraction. Without such adjustment, operation at the same engine setting typically results in producing more torque than is desired when firing density is increased, and less torque than is desired when firing density is decreased.

From a control point of view, the firing density can be changed very quickly by simply changing the selection of the particular cylinder to be fired, whereas a corresponding change in air intake tends to be effected more slowly due to the latency inherent to filling or emptying of the intake manifold. This is particularly noticeable when the desired firing fraction changes significantly, for example from 1/2 to 1 or from 1/3 to 2/3 (which requires the air intake/manifold pressure to change correspondingly greatly). In general, any mismatch between the spark density and the target cylinder air intake during the transition will result in a low frequency torque disturbance (unless compensated for), which may be perceived as NVH. If such a mismatch results in torque surge, spark timing can be retarded to maintain the desired torque. However, an undesirable negative effect of retarding spark to reduce engine output is that retarding spark generally reduces fuel efficiency. Moreover, excessive spark retard will result in misfire, further reducing efficiency and potentially adversely affecting the engine.

This torque mismatch problem may also be more or less mitigated by slowing down the transition between firing fractions. Slowing down the transition allows the change in firing density to more closely track the change in intake manifold pressure. However, even if the air/torque is exactly matched to the firing fraction, any change from one firing density to another will cause low frequency vibrations because the intermediate firing fraction has undesirable firing patterns. Slowing down the transition tends to exacerbate these types of disturbances. The perceived total NVH can be considered as the sum of these two effects. Component NVH from transition firing mode_Mode(s)And component NVH from transition torque mismatch_MismatchCan be considered to add to form the total NVH_{General assembly}，NVH_{General assembly}＝NVH_Mode(s)+NVH_Mismatch. FIG. 2 depicts NVH as a function of transition time, schematically shownThis situation is shown. For short transition times, NVH_Mode(s)Curve 170 Low NVH_MismatchCurve 172 is high. For long transition times, the situation is reversed. NVH_{General assembly}The curve 174 shows a minimum value, typically around 200 milliseconds as shown in fig. 2. It is therefore generally desirable to have a transition length of around 200msec, for example from about 150msec to about 300 msec. It should be appreciated that certain transitions may be longer or shorter and in some cases the total NVH may not be strictly the sum of the pattern and mismatch NVH, but such concepts are generally qualitatively accurate.

Applicants have previously described a wide variety of skip fire controllers. A skip fire controller 10 suitable for implementing the present invention is functionally illustrated in fig. 1. Skip fire controller 10 is shown to include a torque calculator 20, a firing fraction determining unit 30, a transition adjusting unit 40, a firing timing determining unit 50, and a powertrain parameter adjusting module 60. For illustrative purposes, skip fire controller 10 is shown separately from Engine Control Unit (ECU) 70. However, it should be appreciated that in many embodiments, the functionality of skip fire controller 10 may be incorporated into ECU 70. Indeed, it is contemplated that incorporating a skip fire controller into the ECU or powertrain control unit is the most common implementation.

The torque calculator 20 is arranged to determine the desired engine torque at any given time based on a number of inputs. The torque calculator outputs the requested torque 21 to the firing fraction determining unit 30. The firing fraction determining unit 30 is arranged to determine a firing fraction suitable for delivering the desired torque based on current operating conditions and to output a firing fraction 33 suitable for delivering the desired torque. The firing timing determination unit 50 is responsible for determining the firing sequence that delivers the desired firing fraction. The firing sequence may be determined using any suitable approach. In some preferred implementations, firing decisions may be made dynamically on a firing opportunity by firing opportunity basis, which allows the desired changes to be implemented very quickly. The present application has previously described a wide variety of spark timing determination units well suited for determining a proper firing sequence based on a potentially time-varying requested firing fraction or engine output. Many such ignition timing determination units are based on sigma delta converters, which are well suited to make ignition decisions on a firing opportunity by firing opportunity basis. In some cases, an initial accumulator value in the sigma delta converter may be set at the beginning of a transition to generate an ignition pattern with low NVH during the transition. In other implementations, a pattern generator or predefined pattern may be used to facilitate delivery of the desired firing fraction.

The torque calculator 20 receives a plurality of inputs that may affect or indicate the desired engine torque at any time. In automotive applications, one of the primary inputs to the torque calculator is an Accelerator Pedal Position (APP) signal 24 indicating the position of the accelerator pedal. In some implementations, the accelerator pedal position signal is received directly from an accelerator pedal position sensor (not shown), while in other implementations optional preprocessor 22 may modify the accelerator pedal signal before it is delivered to skip fire controller 10. Other primary inputs may come from other functional blocks, such as cruise control (CCS command 26), transmission control (AT command 27), traction control unit (TCU command 28), and so forth. There are also a number of factors that may affect the torque calculation, such as engine speed. When such factors are utilized in the torque calculation, the torque calculator also provides or may obtain appropriate inputs, such as engine speed (RPM signal 29), as necessary.

Further, in some embodiments, it may be desirable to account for energy/torque losses in the drive train, and/or the energy/torque required to drive engine accessories such as air conditioners, alternators/generators, power steering pumps, water pumps, vacuum pumps, and/or combinations of these and other components. In such embodiments, the torque calculator may be arranged to calculate such values or receive an indication of the associated losses, so that these may be taken into account in the calculation of the desired torque.

The nature of the torque calculation varies with the operating state of the vehicle. For example, during normal operation, the desired torque may be based primarily on driver input, which may be reflected by the accelerator pedal position signal 24. When operating under cruise control, the desired torque may be based primarily on input from the cruise control. When a transmission shift is imminent, a transmission shift torque calculation may be used to determine a desired torque during the shift operation. When a traction controller or the like indicates a potential loss of a traction event, a traction control algorithm may be used to determine a desired torque suitable for handling the event. In some cases, depressing the brake pedal may cause a particular engine torque control. When other events occur that require measured control of engine output, other control algorithms or logic may be used to determine the desired torque throughout such events. In any of these situations, the determination of the requested torque may be made in any manner deemed appropriate for the particular situation. For example, the determination of the appropriate torque may be made algorithmically, using a look-up table based on current operating parameters, using appropriate logic, using set points, using stored curves, using any combination of the above, and/or using any other appropriate approach. The application-specific torque calculation may be performed by the torque calculator itself, or may be performed by other components (within or outside the ECU) and reported to the torque calculator for implementation.

The firing fraction determination unit 30 receives the requested torque signal 21 from the torque calculator 20, as well as other inputs such as engine speed and various powertrain operating parameters and/or environmental conditions useful for determining an appropriate operating firing fraction 33 for delivering the requested torque under current conditions. The firing fraction indicates the fraction or percentage of firings to be used in order to deliver the desired output. Typically, the firing fraction determining unit is limited to a limited set of available firing fractions, patterns or sequences that are selected based at least in part on their relatively more desirable NVH characteristics (sometimes collectively referred to herein as a set of available firing fractions). There are a number of factors that may affect the set of available firing fractions. These typically include requested torque, cylinder load, engine speed (e.g., RPM), and current transmission gear. They may potentially also include a number of different environmental conditions, such as ambient pressure or temperature and/or other selected powertrain parameters. The firing fraction determining unit 30 is arranged to select a desired operating firing fraction 33 based on such factors and/or any other factors that may be considered important by the skip fire controller designer. For example, as described in co-pending application No. 13/654,244; 13/654,248, 13/963,686, and 14/638,908, which are incorporated herein by reference, describe several suitable firing fraction determining units.

The number of available firing fractions/modes, and the operating conditions in which they may be used, may vary widely based on a variety of different design goals and NVH considerations. In a specific example, the firing fraction determining unit may be arranged to limit the available operating firing fraction to a set of 29 possible operating firing fractions, each of which is a fraction with a denominator of 9 or less-i.e. 0, 1/9, 1/8, 1/7, 1/6, 1/5, 2/9, 1/4, 2/7, 1/3, 3/8, 2/5, 3/7, 4/9, 1/2, 5/9, 4/7, 3/5, 5/8, 2/3, 5/7, 3/4, 7/9, 4/5, 5/6, 6/7, 7/8, 8/9 and 1. However, under certain (most if not all) operating conditions, the set of available firing fractions may be reduced, and sometimes this set is greatly reduced. In general, the set of available firing fractions tends to be smaller at lower gears and lower engine speeds. For example, it may be that in some operating ranges (e.g., near idle and/or in first gear) the set of available firing fractions is limited to only two available (e.g., 1/2 or 1) or only 4 possible firing fractions-e.g., 1/3, 1/2, 2/3, and 1. Of course, in other embodiments, the allowable firing fraction/pattern for different operating conditions may vary widely.

Since the set of available firing fractions is limited, it will typically be necessary to change a number of different powertrain operating parameters, such as air intake Mass (MAC) and/or spark timing, to ensure that the actual engine output matches the desired output. In the illustrated embodiment, a powertrain parameter adjustment module 60 is provided in cooperation with the firing fraction calculator 30. The powertrain parameter adjustment module 60 directs the ECU70 to appropriately set the selected powertrain parameters to ensure that the actual engine output is substantially equal to the requested engine output at the commanded firing fraction. For example, the powertrain parameter adjustment module 60 may be responsible for determining a desired MAC, spark timing, and/or other engine settings that are desired to help ensure that the actual engine output matches the requested engine output. Although illustrated as a separate component, the powertrain parameter adjustment module 60 is typically implemented as part of the engine control unit 70. Of course, in other embodiments, the powertrain parameter adjustment module 60 may be arranged to directly control a plurality of different engine settings.

The spark timing determination module 50 is arranged to issue a sequence of spark commands 52 to cause the engine to deliver a percentage of sparks specified by the commanded spark fraction 48. The spark timing determination module 50 can take a wide variety of different forms. For example, a sigma delta converter works well as the ignition timing determination module 50. The assignee's multiple applications and patent publications describe a variety of different suitable ignition timing determination modules, including a variety of sigma delta-based converters that work well as ignition timing determination modules. See, for example, U.S. patent nos. 7,577,511, 7,849,835, 7,886,715, 7,954,474, 8,099,224, 8,131,445, 8,131,447, 8,839,766 and pending application No. 13/774,134 filed on 22/2/2013. The sequence of firing commands (sometimes referred to as the drive pulse signal 52) output by the spark timing determination module 50 may be communicated to an Engine Control Unit (ECU)70 or another module, such as a combustion controller (not shown in fig. 1) that coordinates the actual firing. A significant advantage of using a sigma delta converter or similar structure is that it inherently includes an accumulator function that tracks the fraction of firings that have been requested but not delivered. Such an arrangement helps smooth the transition by taking into account the effects of the previous ignition/zero fire decision.

As described above, abrupt transitions between firing fractions may result in undesirable vibration and/or torque surgeOr droop, i.e., undesirable NVH as discussed above with respect to FIG. 2_{General assembly}. The torque surge/drop occurs because, typically, the change in torque request is less than the change in firing fraction, at least during the transition. The firing fraction change will therefore cause the engine to exceed/fall below the requested torque level. Thus, in the embodiment illustrated in fig. 1, the transition adjustment unit 40 is arranged to help mitigate vibration and torque surge/droop associated with the step change in the requested firing fraction 33. The transition adjustment unit 40 has the effect of extending the firing fraction change over a short period of time when a step change in the requested firing fraction occurs. Such "spreading" (which may include brief delays) may help smooth transitions between different commanded firing fractions and may help compensate for a variety of different delays associated with manifold filling. These may include mechanical delays in engine parameter changes and/or inertia-type manifold fill/drain delays. When the requested firing fraction is at steady state, the commanded firing fraction 48 is the same as the requested firing fraction 33. However, as the transition occurs, the commanded firing fraction 48 effectively gradually changes from the previous requested firing fraction to the target firing fraction.

By using such an arrangement smoother operation can be obtained if the nature of the transition is such that the delay imposed by the transition adjustment unit is acceptable. If the nature of the transition is such that a faster response is desired (e.g., when the driver steps down on the accelerator pedal, or during a traction control event), it may be desirable to bypass or modify the settings of the transition adjusting unit 40 to provide a faster response. Thus, some implementations combine separate "fast path" and "slow path" approaches to managing firing fraction change requests. In such applications, the transition adjustment unit may be used in a "slow path" change in response to the "fast path" bypassing the transition adjustment unit. More generally, the characteristics of the transition adjustment unit 40 may vary depending on the input governing the desired transition, i.e., the desired firing fraction slew rate may vary with the rate and/or magnitude of change of accelerator pedal position.

In order to take account of the intake manifold filling dynamics described above, the applicant has previously proposed to use a filter at the location of the transition adjusting unit 40, which roughly simulates the air filling dynamics in order to smoothen the transition between firing fractions. Such approaches are described, for example, in U.S. patent application nos. 13/654,244 and 13/654,248, which are incorporated herein by reference in their entirety. In general, the requested firing fraction is passed through one or two filters before reaching the firing timing determination unit such that the step change in the requested firing fraction is more gradually presented to the firing timing determination module. Another approach to transition management is described in co-pending application No. 14/203,444.

The filtering approach described in the incorporated patent works well to help mitigate vibration. The applicant has found that sometimes even better results can be obtained with some of the slew rate based techniques described below.

In one aspect, a specified firing fraction slew rate is used to help smooth transitions between firing fractions. In some implementations, a brief delay is also included before starting the converted transition. The appropriate slew rate for any transition may depend on a number of operating parameters, including current engine speed, intake/exhaust valve timing, torque demand, initial and target firing fractions, air charge mass, and so forth. The slew rate may also depend on vehicle parameters such as manifold size, acoustic and vibration transmission paths between the source of NVH and the cabin occupants, and vehicle styling, i.e., car, sports car, luxury car, etc. For example, a linear slew rate on the order of 1% -5% firing fraction per firing opportunity performs well in many applications. A linear slew rate of 2% will cause the transition from firing fraction 0 to 1 to span the course of 50 firing opportunities from the beginning of the transition, which in an eight cylinder engine spans only 6 engine cycles. A slew rate of 1% will take twice as long to transition, while a slew rate of 4% will cause the transition to take half as long. For example, if a transition is made from firing fraction 1/2 to firing fraction 1, a slew rate of 2% would indicate that the commanded firing fraction would be 52% for the first firing opportunity after any imposed delay, 54% for the second firing opportunity, and so on until the desired firing fraction 1 is achieved. Of course, in other cases the slew rate will vary with the initial and target firing fractions.

The appropriate slew rate may be determined by observing manifold pressure in response to throttle movement during the transition. By measuring the change in intake Manifold Absolute Pressure (MAP) over time, it can be determined how quickly the engine can lower manifold pressure. Engines with small intake manifolds or engines operating at high engine speeds may use higher slew rates than engines with large intake manifolds or operating at low RPM. Overall, the intake manifold will fill faster than it empties. Increasing the firing fraction typically requires emptying the intake manifold, while decreasing the firing fraction typically requires filling the intake manifold to avoid torque fluctuations or dips. Therefore, it is generally desirable to utilize a slower slew rate for an increase in firing fraction (typically corresponding to a lower desired air intake/manifold pressure) than for a decrease in firing fraction (typically corresponding to a higher desired air intake/manifold pressure). As described above, other factors may also affect manifold filling/emptying dynamics, including engine speed, intake/exhaust valve timing, current air intake amount, current and target firing fraction, and thus the transition adjustment unit 40 may be arranged to set an appropriate slew rate based in part on any of these factors, or other appropriate factors.

A variety of techniques may be used to determine the appropriate slew rate. In some cases, a lookup table may be used to determine an appropriate slew rate between the initial and target firing fractions. Fig. 8 illustrates a table 800 in which slew rates are expressed as percent change in firing fraction per firing opportunity. Of course, the slew rate may be expressed in other variables, such as time, crank angle, and the like. In table 800The entries of (a) should be considered exemplary only and in practice may differ from that given in fig. 8. Table 800 lists 29 possible first initial firing fractions and 29 possible second target firing fractions. The center diagonal 880 does not display an entry because it corresponds to the initial and target firing fractions being equal, i.e., no transitions. Entries above the diagonal 880 correspond to increasing firing fractions and entries below the diagonal correspond to decreasing firing fractions. As generally mentioned previously, it fills more quickly than intake manifold emptying, so the slew rate above the diagonal is generally less than the slew rate below the diagonal. Another obvious feature in this table is that the slew rate is increased for transitions where the firing fraction changes significantly. This helps to reduce the transition time and thus NVH_Mode(s)And (4) minimizing. Another feature is that some transitions have a small change in firing fraction so that the slew rate can be set to 100%, i.e., a step functional change in firing fraction. These firing fraction values, derived from the slew rates and initial firing fractions in table 800, may be used as inputs to the sigma delta converter to determine the firing sequence. In other embodiments, a lookup table having a plurality of different firing sequences that may be used to transition between different firing fractions may be used directly to determine the appropriate firing sequence.

The actual slew rates used during any given transition may be modified relative to those given in the look-up table 800 based on engine operating conditions and driver inputs. For example, if the driver quickly depresses or releases the accelerator pedal, the actual slew rate of the transition may be increased to make the vehicle more responsive. In some cases, this table may not be used and the firing fraction may be changed to its target value immediately. This may have significant NVH consequences, but large and rapid accelerator pedal position changes may indicate a safety issue, which overrides NVH concerns. High engine speeds produce more ignition opportunities in a given time window. It is possible to have a transition time around 200ms while having a slower slew rate. Changes in intake/exhaust valve timing may affect the amount of air introduced during each firing event, and thus affect the rate of intake manifold filling/emptying. As such, the actual slew rate may be used to modify the values from table 800 based on valve timing. The engine speed also affects the amount of air drawn by the cylinders, so it can also affect the actual slew rate. It should be appreciated that rather than having a single two-dimensional lookup table similar to that shown in FIG. 8 and modifying the actual slew rate value, a higher dimensional table may be used that incorporates additional variables as indices (i.e., valve timing, engine speed, etc.).

As described above, a brief delay may also be added before the transition begins. The length of the delay may vary based on the nature of the change and the design choices of the particular engine, which may involve a variety of tradeoffs in desired responsiveness, NVH considerations, and design simplicity. For example, delays on the order of 1 to 10 ignition occasions have been found to work well in a number of different implementations. Depending on the engine speed and the number of engine cylinders, this delay may be on the order of a few milliseconds to 100 milliseconds. Alternatively, in some situations, it may be desirable to delay the movement of the throttle and the onset of change in manifold absolute pressure until after the transition in firing fraction has begun. This type of retard may be particularly advantageous in transitions from higher to lower firing fractions where spark timing adjustments may be used to reduce the torque output per cylinder. Also, in some cylinder activation/deactivation methods, there may be a delay between the decision to change the firing fraction and the implementation of that decision. Thus, the manifold pressure may have changed by the time the actual change in the firing fraction begins. Suitable values for such delays can be found in a look-up table similar to that shown in fig. 8, where the entries now correspond to the delays associated with a number of different transitions. The actual delay values used may be modified relative to those listed in the table in a manner similar to that described with respect to fig. 8. Alternatively, the delay values may not be listed in a look-up table, but may be determined based on engine parameters and operating conditions.

In the aspect of controlA very significant challenge is when it is desired to vary the firing fraction while maintaining the generated engine torque constant. This may be considered a limiting case where a small change in torque request causes a change in the requested firing fraction. Fig. 3(a) -3(d) depict in simplified form engine operation under such idealized conditions. Fig. 3(a) -3(d) show the requested and adjusted firing fraction, throttle position, intake Manifold Absolute Pressure (MAP), and total engine torque output during the expected constant torque transition between firing fraction l/3 and firing fraction 2/3. In fig. 3(a), the requested firing fraction 210 is shown as a step function between an initial value 1/3 and a final value 2/3. The step occurring at time t₁The time may be set to zero and define the start of the transition. Adjusted firing fraction 212 is shown to follow the requested firing fraction until time t₁Followed by a trace depicted by a first order low pass filter. At the time t at the end of the transition_trThe requested and adjusted firing fractions are again equal.

It will be appreciated that the transition time t_trMay vary depending on a number of conditions, such as torque requests in initial and final states, engine speed, transmission gear, and cylinder load. The transition time is generally selected to provide acceptable NVH performance as discussed with respect to fig. 2.

FIG. 3(b) shows the response of the throttle blade position 220 versus time. The first initial throttle position 220a remains constant until at time t₁The transition begins. The throttle flap position can be controlled using a feed forward control algorithm as this will reduce the total transition time and thus NVH_Mode(s). Because the target MAP is lower than the initial MAP during this transition, the throttle will be closed during the transition to help reduce MAP. Throttle valve in time period T_thUp to the closed position. T is_thIs defined at least in part by the delay in processing the requested firing fraction signal into a new throttle position and the time necessary to physically move the throttle blade. This time can be quite small, approximately 20msec, at 1 or ifOn the order of dry ignition timing. The throttle flap remains in the closed position during most of the transition. It moves to its final target position 220b near the end of the transition. It remains in a substantially constant position throughout the remainder of the transition. In FIG. 3b, the second target throttle position is more fully open than the first initial throttle position. This may seem counter-intuitive because the second target MAP is lower than the first initial MAP; however, because fig. 3a-d depict idealized constant torque transitions, engine air induction should be similar between initial and target operating conditions. The difference in air induction levels will be due to the difference in engine efficiency. At this second target condition, the pumping losses are greater because the MAP is lower and therefore the engine requires more air to produce the same torque. Other factors that affect engine efficiency include spark timing, intake/exhaust valve timing and lift, and cylinder load. In general, these variables may produce more or less efficient operation at initial or target operating conditions, and thus the target throttle position may be opened more or less than the initial throttle position.

It should be appreciated that the throttle trajectory depicted in FIG. 3(b) is a representative idealized throttle trajectory. In practice, other types of throttle trajectory may be used, such as MAP closed-loop control, closed-loop MAP, and additional feed-forward throttle control. The position of the throttle may be changed during the transition, and the final throttle position may be above, similar to, or below the initial value.

Fig. 3(c) shows the change in MAP and skip fraction over time. The skip fraction is defined as one minus the firing fraction. The requested skip score 236 and the resulting requested MAP 230, as well as the adjusted skip score 238 and the resulting adjusted MAP 231 are shown. The two manifold pressures 230 and 231 and the skip fractions 236 and 238 are substantially constant until at t₁The time transition begins. Requested MAP 230 begins to fall in response to the closing of the throttle plate and the removal of air from the intake manifold by introduction into the cylinder. As described earlier, because of the filling/emptying of the intake manifoldMechanics, the response of the MAP 230 is relatively slow. The requested skip fraction 236 reproduces the response of firing fraction 210 shown in fig. 3(a) and has a step function decrease at the beginning of the transition. The adjusted skip score 238 has a more gradual transition. The more gradual transition associated with the adjusted skip fraction 238 causes the intake manifold to evacuate more slowly, causing the adjusted MAP 231 to transition more slowly than the requested MAP 230. If the other engine parameters are fixed, the product of MAP and firing fraction is substantially proportional to engine torque output. For a constant torque output, this implies that MAP and the skip fraction should follow each other, i.e. their product should be substantially constant during the transition. The divergence of the MAP and skip fractions indicates a torque mismatch. Areas 234 and 235 define the mismatch associated with the adjusted firing fraction and are proportional to the excess torque produced by the adjusted firing fraction 212. Similarly, the sum of areas 232 and 234 is proportional to the excess torque produced by the requested firing fraction 210.

The torque surge associated with the mismatch between firing fraction and MAP is more clearly shown in fig. 3 (d). The torque surge occurring based on the intermediate transition of the firing fraction to the target firing fraction is shown by curve 240. In contrast, the torque surge generated by using the adjusted firing fraction is represented by curve 242. It can be seen that by using a filter to smooth the transition, the total torque surge is significantly reduced, but the duration of the surge may be slightly extended because the intake manifold pressure is not reduced by evacuation as quickly as possible. The net torque surge is the integral of the torque mismatch over the duration of the mismatch. While using a first order filter to smooth transitions between firing fractions can significantly reduce the torque surge/droop associated with the transition, it may be difficult to define a filter (or bank of filters) that works well under a wide range of operating conditions and a full range of possible firing fraction changes. One reason for this is that the change due to the linear filter is proportional to the change in the fraction while lasting the same amount of time. Thus, for example, a change from 1/2 to 1 would be twice as large as a change from 1/2 to 3/4, but would occur over the same period of time. Manifold dynamics, particularly evacuation, tend to be different. Rather than a larger change changing faster, the rate of change is similar, but the duration of the change is longer. Another aspect of the linear filter is that the response to a step change (as seen when the firing fraction changes) has the largest output change at the time of the step and a decreasing amount of change with each successive step. This not only does not match the physical behavior, but also exacerbates the consequences of the onset of firing fraction transitions not matching the physical behavior of the manifold.

In various applications, a linear slew rate transition management strategy can be implemented in place of the prior art filter scheme to help further reduce torque surge/droop and vibration. Some of the potential advantages of this approach will be schematically described with reference to fig. 4. More specifically, fig. 4(a) -4(d) generally show the same type of information as shown in fig. 3(a) -3(d), except that these figures graphically compare an exemplary use of a delayed linear slew rate with an exemplary first order filter. It will be appreciated that these figures are of a diagrammatic nature and are intended to illustrate concepts rather than reflect data from particular experiments, as the nature of these curves will in practice necessarily be very dependent on a number of variables, including engine speed and operating conditions, the nature of the transfer function of the filter, the specified slew rate, the delay employed, and so on.

First, fig. 4(a) compares the filtered firing fraction change 212 with the firing fraction change trace 310 that can be seen when using a delayed linear slew rate. The filtered firing fraction change 212 is the same as shown in fig. 3 (a). In the firing fraction trajectory 310, the firing fraction 310 is commanded from t₁(start of transition) for a prescribed delay period T_DRemaining at the original firing fraction. The time delay T_DMay be greater or less than throttle transition time t_th. Appropriate time delay T_DMay vary based on a number of factors, including manifold fill/drain dynamics, throttle response time, and the like. After this delay, the firing fraction 310 increases linearly until it reachesTo the target firing fraction.

FIG. 4(b) shows the response of throttle position 220 versus time. This figure is the same as fig. 3(b), and the description will not be repeated. Fig. 4(c) shows the change in MAP and skip fraction over time. In fig. 4(c), the filtered skip fraction 238 and the resulting MAP 231 are the same as shown in fig. 3 (c). The torque mismatch associated with the filtered skip fraction is depicted by areas 334 and 332 a. This area is equal to that depicted in fig. 3 (c). The delayed linear transition skip fraction 330 reproduces the variation in firing fraction 310. The torque mismatch associated with the delayed linear transformation is depicted by areas 332a and 332 b. Area 332a is associated with torque surge, while area 332b is associated with torque lag. For clarity, this figure assumes that the MAPs 231 associated with both the filtered linear transformation and the delayed linear transformation are equal, whereas in practice they differ slightly due to different evacuation losses associated with different firing fraction trajectories.

Fig. 4(d) shows the resulting torque between these two cases. Curve 242 illustrates the torque produced using the filtered firing fraction and is the same as shown in fig. 3 (d). Curve 340 illustrates the torque produced using a delayed linear slew rate of the firing fraction. The torque mismatch associated with the delayed linear slew-rate change in firing fraction is significantly less than the torque mismatch associated with the filtered change, demonstrating the advantages of this control method.

When using a linear slew rate, the slope of the linear slew and the length of the delay (if any) will each have a direct effect on the magnitude and meaning of the torque mismatch. A number of different engine characteristics and operating parameters will also affect torque mismatch (e.g., engine speed, manifold characteristics, spark timing, valve lift, air/fuel stoichiometry, etc.). When the linear slew rate is selected such that it closely approximates the manifold fill dynamics, the torque surge or lag associated with the transition can be significantly reduced. In fact, it has been observed that in many transitions, a properly chosen linear slew rate can follow the manifold filling dynamics more closely than the first order filter described. It should be appreciated, however, that the magnitude of surge or lag may vary significantly based on how closely (or differentially) the manifold filling dynamics are approximated by the selected slew rate.

There are times when the transition may be interrupted by a new target firing fraction request. That is, in the middle of the transition from the first firing fraction to the second firing fraction, there is a time at which the second change request is made. In such a case, the transition unit may begin implementing the second change from the current state, rather than waiting for the first transition to end. For example, consider the case where the firing fraction determining unit 30 requests a change to the firing fraction 7/8 when the transition adjusting unit 40 is in the middle of the transition from the firing fractions 1/5 to 3/8. Such a situation is presented graphically in fig. 5, which is a graph showing the variation of firing fraction (Y-axis) over time (X-axis). In the illustrated embodiment, the engine is initially operated with a firing fraction 1/5 indicated by the segment "a". At time t₁The requested firing fraction is increased to 3/8. At a specified delay (T in the illustrated embodiment)_D) Thereafter, at time t₂The transition towards the firing fraction 3/8 begins using a specified slew rate (1% in this case, see fig. 8). At time t₃A change request of 7/8 is received, although the transition to 3/8 has not been completed. The transition toward firing fraction 3/8 continues during the specified delay, but once the specified delay is at t₄Ending, the transition to the firing fraction 7/8 begins at the specified slew rate. At t₃The decision to occur changes the target firing fraction and at t₄Delay T between occurring slew rate changes_D1May be different from the delay T associated with the start of the transition_D. In some cases, delay T_D1It may be zero, but in many cases there is an inherent delay between the ignition decision and the implementation of that decision due to the activation/deactivation mechanism of the cylinder. In the illustrated embodiment, the specified delays are the same for both changes, but this is not a requirement. In this case, the slew rate associated with the transition between 1/5 and 7/8 is 2% (see fig. 2). Time t₄There is a knee point for firing fraction slew rate. A generally larger change between the initial and final firing fractions will result in a faster slew rate to avoid excessive, mode-induced NVH. In this example, the magnitude of the change from the current firing fraction to 7/8 is greater than the original magnitude of the change from 1/5 to 3/8, and thus the slew rate is increased. In some cases, the relative slew rate, i.e., the ratio of the slew rate to the total change in firing fraction, may remain substantially constant during the transition. In other cases, it may be desirable to have the slew rate fixed throughout the transition. For reference, the dashed line labeled "b" shows the completion of the change to the ignition score 2/5, provided that a second transition is not requested. While only a single intermediate change in the transition is presented, it should be appreciated that the same principles can be applied to implement any further changes requested in the transition. These may include increases and decreases in the requested firing fraction, multiple sequential change requests that occur fast enough such that intermediate firing fractions are never actually available, and so on.

Torque management

As described above, when the torque delivered during the transition matches the desired torque, the transition is generally smoother. One of the primary reasons for controlling the firing fraction and air charge in the manner described is to help reduce torque oscillations, which tends to help reduce undesirable oscillations. The engine output may be modulated in other ways when an air intake/firing fraction mismatch occurs. One such approach is to control spark timing in a manner that mitigates such torque oscillations. Once operating at the allowed firing fraction level, the spark timing is set at or near a timing that provides the best fuel efficiency, i.e., the maximum torque for a given MAC, typically denoted as the Maximum Brake Torque (MBT) operating point. When the firing fraction increases and the air intake amount decreases, torque surge naturally occurs when the firing fraction increases more rapidly than the corresponding air intake amount decreases. This surge can be mitigated by appropriately delaying the spark during the transition in a manner that provides a more stable torque output. In general, retarding the spark may reduce the output per firing, as will be readily understood by those familiar with the art. If the spark timing prior to the transition does not correspond to the maximum torque timing, the spark may be advanced by a limited amount to provide a slightly greater torque per spark, but knock, misfire considerations, and the like typically limit the usefulness of using spark advance. Thus, the spark retard approach is particularly useful for avoiding torque surge. Such a condition exists when the firing fraction increases faster than the manifold can empty in a low-to-high firing fraction transition, or if the manifold begins before the firing fraction changes in a high-to-low firing fraction transition. In general, the use of firing fractions or air intake delays may be used to mitigate torque mismatches, allowing slightly higher than requested torque (if uncorrected) to be achieved, which may be reduced by moderate spark retard. A relatively short firing fraction ramp (i.e., high slew rate) may then be used to reduce the mismatch between the firing fraction and the air intake charge. The firing fraction slew rate may be defined at time or based on some parameter of engine speed (e.g., crankshaft angle, firing, or firing timing).

One advantage of using spark timing control to help ensure that the engine provides the desired torque throughout the transition is that the spark is easily controlled and can be adjusted very quickly. As described above, spark retard may be used to reduce torque mismatch in transitions that increase the firing fraction. In some cases, spark retard alone may be sufficient to eliminate this mismatch; in other cases, however, the lag in the air intake may be too great to compensate by spark retard without compromising combustion stability. In all cases, the undesirable negative effect of retarding spark to reduce engine output is that retarding spark will reduce fuel efficiency as a whole. Thus, to the extent possible, it is generally preferred to match the air intake to the skip fraction throughout the transition as described above to avoid or at least reduce the efficiency loss associated with spark retard control.

For comparison purposes, fig. 6(a) -6(d) show the fraction of ignition, manifold pressure, spark advance, and total engine output during a given constant torque transition when spark retard and retarded linear transitions are utilized to further help mitigate torque surge. As in fig. 3(a) - (d) and 4(a) - (d), the transition is from firing fraction l/3 to firing fraction 2/3. The change in throttle position is similar to that depicted in fig. 3(b) and 4(b) and is not depicted in fig. 6(a) -6 (d). In these figures, the initial spark timing is optimized to achieve the maximum brake torque, i.e., the most efficient operating point. Fig. 6(a) shows the change in firing fraction 310 during a firing fraction transition managed using the delayed linear transition approach depicted with respect to fig. 4. Fig. 6(b) shows the corresponding changes in the MAP 231 and the skip score 330 during this transition. Areas 332a and 332b illustrate the mismatch between skip fraction 330 and MAP 231. As shown in fig. 4(d), these areas correspond to regions where the torque does not match if no corrective action is taken. Fig. 6(c) shows the change in spark timing during the transition. If the spark timing is maintained at its maximum efficiency, the result is curve 510. However, deviations from this maximum efficiency may be desirable so that torque surge can be eliminated. This type of adjustment is depicted in curve 512. At the beginning of the transition the spark is retarded in order to eliminate the torque ripple seen in fig. 4 (d). Fig. 6(d) shows the resulting effect on torque. The spark adjusted torque 540 does not show a torque surge at the beginning of the transition because the spark retard reduces the output per cylinder. The torque drop 542 near the end of the transition cannot be removed by using the spark timing since the spark timing has been adjusted to its point of maximum efficiency.

In some cases, torque droop 542 may be undesirable. In such a case, the delay T is reduced_DWill cause the MAP 231 to always be above and to the right of the skip score 330 in this transition. In this case, the engine will always produce excessive torque, which can be eliminated by retarding spark timing. This will reduce fuel efficiency while improving NVH. In other cases, the engine generally produces a slightly different maximum efficiency (typical)Spark timing, which is expressed as a Maximum Brake Torque (MBT) point). In this case, the engine has a torque reserve that allows both increased and decreased engine torque by controlling spark timing. Operating the engine with torque reserve has the undesirable effect of reducing fuel efficiency, and therefore the firing fraction transition control strategy described herein minimizes the need to operate without MBT spark timing.

Some engine controllers have the ability to shut off the delivery of fuel to the cylinders while still opening these valves in a regular manner. This technique causes intake air to be pumped through the cylinders and preferably cuts the output of the engine off to zero. Currently, the most common fuel cut is during deceleration, where fuel is typically cut off from all working chambers, which is commonly referred to as Deceleration Fuel Cut Off (DFCO).

During a skip fire transition that would otherwise result in torque surge (e.g., increasing the firing fraction while decreasing the air intake charge), a more or less similar approach may be taken to help balance the air intake charge with the change in firing fraction. Specifically, during skip fire control, the selected duty cycle is not fired. In general, the cylinders associated with the skipped working cycle are deactivated such that they do not pump air through the cylinders during the skipped working cycle. However, if it is desired to reduce the amount of air in the manifold, the valve associated with the selected skipped working cycle may be activated to pump air through the corresponding cylinder during the skipped working cycle. Since the working cycles are intended to be skipped, no fuel is delivered to the working chambers and no combustion occurs.

The number of skipped working cycles that are suitably used to pump air through the engine block for any particular transition may vary based on the nature of the transition. For example, factors such as intake manifold dynamics, initial and target firing fractions, initial and target air intake amounts, firing fraction slew rates utilized, engine speed, torque surge that would otherwise be expected, etc., may all affect both the number of working cycles that are suitably used to pump air through the cylinders, and their respective timing. Some advantages of using the air pumping approach include that it can save fuel relative to other torque mitigation approaches (e.g., spark retard) and help speed up the transition by helping to more quickly reduce the manifold pressure to the desired level. A potential drawback or limitation of this approach is that the exhaust system (e.g., catalytic converter) must be able to handle the air passing through the engine, and not all exhaust systems have this capability at all times. However, when practiced, using skipped cylinders to pump excess air from the manifold during a firing fraction transition may be a synergistic use of skipped working cycles during skip fire control.

Pumping air through some or all of the skipped cylinders has the advantage of faster reduction of MAP, thus allowing faster transitions and potentially lower levels of NVH. Fig. 7 shows MAP and possible skip fractions in two cases, one case pumping no air and one case pumping air. The transition being at time t₁And starting. The case of no pumping is equivalent to that previously discussed with respect to fig. 6 (b). The skip fraction 330 substantially matches the MAP 231 with relatively small mismatch areas 332a and 332 b. The situation of pumping through the skipped cylinder has a faster drop in MAP 731 and thus the skip fraction 730 can change faster and still substantially match MAP 731. The sizes of the mismatched areas 732a and 732b are similar to those associated with the non-pumping transitions 332a and 332 b. Total transition time t_{Pump and method of operating the same}Transition time t before_trMuch shorter. As can be seen in FIG. 2, if NVH_MismatchCan be kept at an acceptable level, shorter transition times are advantageous because they have lower NVH_Mode(s)And thus potentially lower overall NVH. Under certain conditions, the transition time may even be reduced to zero, i.e. the target fraction is reached in the next work cycle in a step function transition.

An alternative method of adjusting the filling/emptying of the intake manifold is to vary the intake and exhaust valve timing. For a cam operated valve, this is accomplished by adjusting a cam phaser, which controls the relative times at which the valve opens and closes. Greater control is possible for engines having variable valve lift or electronically controlled valves. In all cases, the trim valve can be moved to provide the desired MAC (within system control) for a given MAP. This allows another degree of control during the transition. In some cases, the valve timing used during the transition may also be used at the final firing fraction level.

Feed forward air control

As previously discussed with respect to FIGS. 3b and 4b, the engine controller may use throttle feed forward control to accelerate the desired change in manifold pressure. The illustrated example uses a simple step function change in commanded throttle position to adjust MAP. However, more complex control schemes for the throttle may be used to achieve a faster transition of the MAP. Some control schemes may integrate a feedforward control architecture with a number of different types of feedback control, such as a PID (proportional-integral-derivative) controller or a state-space controller to better control the MAP response. In general, feed forward throttle control pairs consider throttle opening or closing more than would be appropriate for steady state operation during the transition, and then return to the appropriate level for steady state operation. The use of feed forward throttle control during transitions between different firing fractions may help accelerate the transition in a controlled manner that may help further reduce vibration. Since transitions occur faster, higher slew rates may typically be used during such transitions.

Much of the discussion above focuses on using the throttle as the primary mechanism for changing the amount of air intake in each cylinder. As will be appreciated by those familiar with engine operation, there are other ways to vary the amount of air intake, including valve timing control, intake/exhaust valve timing and lift control, boost, and the like. In practice, the air intake amount may be controlled using feed-forward control, in addition to or instead of the feed-forward throttle control described, by using these air intake amount control mechanisms. For example, if electronic valves are provided, the opening and closing timing of each valve can be easily controlled to facilitate a more rapid transition in the desired amount of air intake. When the valve train is controlled by one or more camshafts, feed forward control of the cams or camshafts may be used to facilitate faster transitions in air intake charge. Similarly, when the valve train supports variable valve lift, appropriate control of valve lift (including feed forward control) can be used to better match the air intake to firing density. When the engine includes appropriate hardware, any of these air intake amount control mechanisms may be used in parallel.

The described feed forward air control may be used independently or in conjunction with spark retard, and/or pumping air through deactivated cylinders and/or the described firing fraction transition slew rate control. One desirable feature of feed forward air control is that it can be used in conjunction with increases and decreases in firing fractions.

Although several embodiments of the present invention have been described in detail, it should be understood that the invention may be embodied in many other forms without departing from the spirit or scope of the present invention. For example, the transient slew rate limit is described primarily in the context of using a constant linear slew rate throughout the transient. While such an approach works well, it will be appreciated that more complex slew rates may be used when desired, which may be useful to better follow particular manifold fill and/or drain dynamics and/or other design considerations. For example, in some implementations, it may be desirable to divide the transition into two or more linear sections or to define a more complex transition function.

Some skip fire controllers are arranged such that they inherently require a relatively large number of transitions under a variety of normal driving scenarios in order to maximize fuel efficiency. This is particularly true under driving conditions that support a relatively large set of firing fractions. For example, some driving tests by the applicant for skip fire controllers with up to 29 available firing fractions tend to complete a transition on average every second or every two seconds over the course of a variety of different normal driving profiles. This makes it particularly desirable to utilize some of the transition management approaches described herein for driving comfort.

Several different techniques have been described, including ignition fraction management, air delivery management, and spark timing management. Although each may be used independently, better results are generally obtained when used in combination with the objective of avoiding transient torque surges or dips while facilitating rapid transitions between firing fractions.

In the above description, the term "cylinder" is mentioned several times. The term cylinder is to be understood broadly to encompass any suitable type of working chamber. These figures illustrate a wide variety of devices, designs, and representative cylinder and/or engine data. It should be understood that the figures are intended to be exemplary and explanatory and that features and functions of other embodiments may depart from that shown in the figures.

The invention is described primarily in the context of dynamic skip fire operation, where an accumulator or other mechanism tracks a portion of firings that have been requested but not delivered, or that have been delivered but not requested. However, the techniques described are equally applicable to managing transitions between any of the different skip fire firing fractions, or between skip fire firing fractions (where individual cylinders are sometimes fired and sometimes skipped) and all cylinder operations (or using a fixed set of cylinder operations), as may occur when using a variety of different rocking cylinder deactivation (rolling cylinder deactivation) techniques. Similar techniques may also be used to manage effective displacement transitions in the control of variable stroke engines in which the number of strokes in each working chamber is varied in order to effectively vary the displacement of the engine.

The invention may also be useful for engines that do not use skip fire control. For example, although the present invention is described primarily in the context of transitions between different firing fractions during skip fire control, the techniques described may also be used to facilitate transitions between different variable displacement states in more conventional variable displacement engines that use skip fire transition control. For example, an eight-cylinder variable displacement engine capable of operating in a 4-cylinder mode (i.e., 4 fixed cylinders) requires a transition from a firing fraction of 0.5 to 1, and vice versa, and the firing fraction transition management techniques described herein may be advantageously used. Accordingly, the present embodiments are to be considered as illustrative and not restrictive, and the invention is not to be limited to the details given herein.

Claims (21)

1. A method for controlling an engine to transition from an initial first operating firing fraction to a target second operating firing fraction, wherein there is an initial manifold pressure and a target manifold pressure, the target manifold pressure being lower than the initial manifold pressure and the target second operating firing fraction being higher than the initial first operating firing fraction, the method comprising:

operating the engine in a skip fire manner during the transition; and is

Pumping air from an intake manifold through the engine to an exhaust during selected skipped working cycles occurring during the transition to more rapidly reduce intake manifold pressure to the target manifold pressure; and is

Wherein air is not pumped through the engine during skipped working cycles that occur outside of the firing fraction transitions.

2. A method as recited in claim 1 wherein when a request to transition to a third operating firing fraction is received before the transition to the target second operating firing fraction is complete, the method further comprises:

in response to the request to transition to the third operating firing fraction, changing the commanded throttle position to a throttle position that facilitates operation at the third operating firing fraction;

continuing to transition toward the target second operating firing fraction during a specified delay period that delays a start of a change in commanded firing fraction relative to a start of the change in commanded throttle position by a plurality of firing opportunities; and is

Transitioning from the current firing fraction to the third operating firing fraction after the specified delay period has expired by gradually changing the commanded firing fraction from the current operating firing fraction toward the third operating firing fraction.

3. The method of claim 2, wherein the commanded firing fraction is changed for each firing opportunity once the commanded firing fraction has begun to change.

4. A method as recited in claim 3 wherein the commanded firing fraction is changed using a linear slew rate per firing opportunity such that the amount by which the commanded firing fraction is changed per firing opportunity is the same throughout the transition.

5. A method as recited in claim 4 wherein the linear slew rate is in the range of 1% -5% such that the firing fraction increases in the range of 1% -5% per firing opportunity.

6. A method as recited in claim 4 wherein the magnitude of the linear slew rate is selected based at least in part on the magnitude of change in firing fraction and at least one engine operating parameter selected from the group consisting of current engine speed, intake/exhaust valve timing, torque demand, and air intake mass.

7. A method as recited in claim 4 wherein the magnitude of the linear slew rate is selected based at least in part on the magnitude of change in firing fraction and at least one vehicle parameter selected from the group consisting of manifold size, acoustic and vibrational transfer paths between sources of noise and vibration roughness and cabin occupants, and vehicle styling.

8. The method of claim 1, further comprising:

transitioning from the initial first operating firing fraction to the target second operating firing fraction by gradually changing the commanded firing fraction from the initial first operating firing fraction to the target second operating firing fraction, wherein the commanded firing fraction is changed by substantially the same amount for each firing opportunity.

9. The method of claim 1, wherein the commanded firing fraction is provided to a skip fire firing timing determination module that includes an accumulator function that tracks a portion of firings that have been requested but not delivered, or that have been delivered but not requested.

10. The method of claim 1, wherein the transition period is in the range of 150 to 300 milliseconds.

11. A method as recited in claim 1 wherein the engine includes a plurality of working chambers and an intake manifold supplying air to at least the plurality of working chambers, the intake manifold having a manifold air pressure, the method further comprising:

causing a commanded throttle position to change in conjunction with transitioning between different firing fractions to facilitate operation at the target second operating firing fraction, wherein a start of change in commanded firing fraction is delayed relative to a start of change in commanded throttle position by a plurality of firing opportunities, thereby helping to compensate for inherent delays associated with changing the manifold air pressure.

12. The method of claim 1, further comprising:

determining a target manifold pressure associated with the target second operating firing fraction, the target manifold pressure being different than an initial manifold pressure present when a decision is made to change firing fractions; and is

A feed forward throttle control is utilized in conjunction with the transition to accelerate the transition to the target manifold pressure.

13. A method as recited in claim 1 wherein the engine includes a plurality of working chambers and a plurality of intake valves, each intake valve being associated with an associated one of said plurality of working chambers; a camshaft arranged to open and close the intake valve; and an intake manifold supplying air to the working chambers through the intake valves, the method further comprising:

determining a target air intake associated with the target second operating firing fraction, the target air intake being different from the initial air intake present when it is decided to change the firing fraction; and is

A feed forward camshaft control is utilized in conjunction with the transition to accelerate the transition to the target air intake charge.

14. The method of claim 1, wherein the engine includes a plurality of working chambers, each of said plurality of working chambers having an associated spark source, the method further comprising:

determining a target spark timing associated with the target second operating firing fraction, the target spark timing potentially different from an initial spark timing that existed when a decision was made to change the firing fraction; and is

Spark is retarded relative to both the initial spark timing and the target spark timing for the selected firing working chamber during the transition to mitigate or prevent torque surge.

15. A method as recited in claim 14 wherein at least one of the initial and target spark timings is a spark timing that causes the engine to produce maximum brake torque at an associated engine setting.

16. The method of claim 1, further comprising:

transitioning from the initial first operating firing fraction to the target second operating firing fraction by gradually changing a commanded firing fraction from the initial first operating firing fraction to the target second operating firing fraction, wherein the commanded firing fraction has an associated skip fraction that is a complementary fraction of the commanded firing fraction, and wherein each firing opportunity changes the commanded firing fraction in a manner such that a product of the skip fraction and intake manifold pressure remains substantially constant throughout the transition.

17. The method of claim 1, further comprising: causing at least one commanded engine operating parameter that affects a working chamber air intake air quantity to change in conjunction with transitioning between the different firing fractions to facilitate operation at the target second operating firing fraction, wherein the onset of change in commanded firing fraction is delayed relative to the onset of change in commanded engine operating parameter by a delay period specified by a plurality of firing opportunities, thereby helping to compensate for inherent delays associated with increasing or decreasing air quantity in an intake manifold that provides air to the working chamber.

18. The method of claim 17, wherein the specified delay period is a defined period of time.

19. A method as recited in claim 17 wherein the specified delay period is a defined number of firings or firing opportunities.

20. An engine controller arranged to implement the method of any one of claims 1 to 19.

21. An engine controller having a control algorithm arranged to implement the method of any one of claims 1 to 19, the control algorithm being stored in a computer readable medium.

Images