CN109641587B - Engine torque smoothing - Google Patents

Abstract

Methods, devices, estimators, controllers, and algorithms are described for estimating a torque curve of an engine and/or for controlling torque applied to a powertrain by one or more devices other than the engine itself to manage net torque applied by the engine and other device(s) in a manner that reduces undesirable NVH. The described methods are particularly well suited for use in hybrid vehicles where the engine is operated with skip fire or other dynamic firing level modulation-however, they may be used in a variety of other situations. In some embodiments, the hybrid vehicle includes a motor/generator that applies a smoothing torque.

Description

Engine torque smoothing

Cross Reference to Related Applications

This application claims priority from U.S. provisional patent application No. 62/379,357(TULAP063P), which is incorporated herein by reference.

Technical Field

The present invention generally relates to a hybrid vehicle powered by an internal combustion engine operating under skip-fire control, the hybrid vehicle having another power source in addition to the internal combustion engine. The torque curve of a skip-fire controlled engine is estimated and an additional power source is used to smooth the torque curve.

Background

The fuel efficiency of an internal combustion engine can be greatly improved by changing the displacement of the engine. This allows maximum torque to be available when required, and also significantly reduces pumping losses and improves thermal efficiency by using smaller displacements when maximum torque is not required. The most common method of implementing variable displacement engines today is to deactivate a group of cylinders substantially simultaneously. In this approach, the intake and exhaust valves associated with the deactivated cylinders remain closed and no fuel is injected when it is desired to skip a combustion event. For example, an 8-cylinder variable displacement engine may deactivate half of the cylinders (i.e., 4 cylinders) so that it operates using only the remaining 4 cylinders. Commercially available variable displacement engines available today typically support only two or at most three displacements.

Another engine control method of varying the effective displacement of the engine is known 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, then may be skipped during the next engine cycle, and then selectively skipped or fired during the next engine cycle. In this way, finer control of the effective engine displacement is possible. For example, firing every second cylinder in a 4-cylinder engine will provide an effective displacement of 1/3 of maximum engine displacement, which is a fractional displacement that cannot be obtained by simply deactivating a group of cylinders.

U.S. patent No. 8,131,445, which is incorporated herein by reference, teaches a skip-fire operating method that allows for the use of a single cylinder deactivation to average out any cylinder fraction to be fired. In other skip-fire approaches, a particular firing sequence or firing density may be selected from a set of available firing sequences or fractions. In skip-fire operating modes, the amount of torque delivered is generally heavily dependent on the firing density or fraction of combustion events that are not skipped. Dynamic Skip Fire (DSF) control refers to skip-fire operation as follows: where the firing/skipping decisions are made in a dynamic manner, for example, at each firing opportunity, each engine cycle, or at some other interval.

In some applications, known as multi-stage skip fire, individual duty cycles that are fired may be purposefully operated at different cylinder output levels, i.e., purposefully using different intake air amounts and corresponding fuel supply levels. For example, U.S. patent No. 9,399,964 (which is incorporated herein by reference) describes some such methods. The single cylinder control concept used in dynamic skip fire may also be applied to dynamic multi-charge level engine operation where all cylinders are fired, but individual duty cycles are purposefully run at different cylinder output levels. Dynamic skip fire and dynamic multi-charge level engine operation can be collectively considered as different types of dynamic firing level modulation engine operation, where the output of each duty cycle (e.g., skip/fire, high/low, skip/high/low, etc.) is dynamically determined during operation of the engine, typically on a single cylinder by duty cycle basis (on a firing opportunity by firing opportunity basis). It should be appreciated that dynamic ignition level engine operation is different from conventional variable displacement, where a defined group of cylinders operate in substantially the same manner as the engine enters a reduced displacement operating state until the engine transitions to a different operating state.

Combustion processes and cylinder firing using skip fire or other firing level modulation techniques may introduce undesirable noise, vibration, and harshness (NVH). For example, the engine may transmit vibrations to the vehicle body where the vibrations may be perceived by the vehicle occupants. Sound may also be transmitted through the chassis to the vehicle cabin. Under certain operating conditions, ignition of the cylinder produces undesirable acoustic effects through the exhaust system and tailpipe. Thus, vehicle occupants may experience undesirable NVH due to structure-borne vibration or airborne sound.

The challenge of skip fire engine control is to achieve acceptable NVH performance. While existing approaches work well, there is a continuing effort to develop new and improved methods for managing NVH during ignition level modulation operation of an engine.

Disclosure of Invention

Methods, devices, estimators, controllers, and algorithms are described for estimating a torque curve of an engine and/or for controlling torque applied to a powertrain by one or more devices other than the engine itself to manage net torque applied by the engine and other device(s) in a manner that reduces undesirable NVH. The described methods are particularly well suited for use in hybrid vehicles where the engine is operated with skip fire or other dynamic firing level modulation-however, they may be used in a variety of other situations. In some embodiments, the hybrid vehicle includes a motor/generator that applies a smoothing torque.

In some embodiments, a period of time is identified in which the instantaneous torque or instantaneous acceleration expected to be produced by the engine exceeds a specified threshold. Then, a reaction torque is applied to the powertrain by the energy source or sink in a controlled manner during the identified period such that an expected powertrain net torque does not exceed a specified threshold. In some embodiments, the specified threshold may vary depending on engine speed and/or transmission gear. In some embodiments, the reaction (smoothing) torque is applied in short pulses timed to react to torque spikes generated during skip fire or dynamic firing level modulation operation of the engine.

In some hybrid vehicle embodiments, the hybrid vehicle is operated using only the output of the internal combustion engine when the estimated engine torque curve is determined to provide acceptable NVH. However, when the estimated engine torque curve is determined to provide unacceptable NVH, both an internal combustion engine and an auxiliary power source/sink are utilized, wherein the auxiliary power source/sink is arranged to provide a smooth torque to reduce the NVH to an acceptable level.

In some embodiments, the total engine torque curve and the determination of the reaction smoothing torque are updated for each firing opportunity such that the demand for reaction smoothing torque and its magnitude are updated for each firing opportunity.

In some skip fire or other dynamic firing level modulation embodiments, the torque curve estimate is used to select the (effective) operating firing fraction. In such embodiments, the fuel efficiency of the various candidate firing fractions may be compared to meet desired drivability criteria after taking into account the fuel efficiency implications of any smoothing torque that may be required when operating at the respective firing fractions.

In some embodiments, the torque curve of the engine may be determined by summing the contributions of each working chamber (e.g., cylinder). In some embodiments, the torque curve for a particular cylinder may be achieved by selecting or determining a normalized torque curve for the operating state of that cylinder, and then scaling the normalized torque curve based on current engine operating parameters. During skip fire engine operation, the normalized torque curve utilized will vary based on the skip/fire firing decision for that particular cylinder. In some embodiments, the normalized torque curve will be based at least in part on intake manifold pressure. In some embodiments, the normalized torque curve may be scaled based on one or more current operating parameters, such as engine speed, spark timing, valve timing/lift, engine firing history, cylinder firing history, and the like.

In some embodiments, the engine torque profile is filtered to identify selected harmonic components of the torque profile. The reaction smoothing torque for application to the powertrain may then be based on the filtering results. In some such embodiments, the filtering results may be amplified based on one or more current engine parameters. The filtered signal may be delayed to align with the predicted torque produced by the engine. The amplified filtered signal may be inverted and used to control the electric motor/generator to source/sink torque based on the inverted torque signal.

In some embodiments, the smoothing torque may be applied as one or more oscillating (e.g., sinusoidal) signals, while in other embodiments, the smoothing torque may be applied as pulses intended to cancel portions of the expected torque spikes.

In various embodiments, the smoothing torque may be effectively applied by devices that draw energy from the powertrain by appropriately increasing or decreasing their respective loads. Similarly, the torque applied by the devices that add torque to the powertrain may be increased or decreased to effectively provide the desired smoothing torque. When a device that can both add and subtract torque is used (such as a motor/generator), either of these methods may be used, or the device may be changed between torque contribution and torque draw states to provide the desired smooth torque.

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 diagrammatic illustration of a representative hybrid powertrain according to an embodiment of the present invention.

Fig. 2 is a diagrammatic illustration of a representative control architecture of a hybrid powertrain, in accordance with an embodiment of the present invention.

Fig. 3A and 3B show cylinder torque curves versus crank angle for multiple firings at different MAP values.

Fig. 4A and 4B show normalized torque curves versus crank angle for combustion strokes for different MAP values according to an embodiment of the present invention.

Fig. 5A and 5B show normalized torque curves versus crank angle for compression strokes for different MAP values according to an embodiment of the present invention.

FIG. 6 illustrates an exemplary table showing values of normalized torque curves for different MAP values, according to an embodiment of the present invention.

FIG. 7 illustrates an exemplary table showing torque scaling factors for different MAP values and engine speeds, according to an embodiment of the present invention.

FIG. 8 illustrates an exemplary torque curve with respect to crank angle at an average engine speed of 1500rpm and a firing fraction of 3/4 for a 4-cylinder engine according to an embodiment of the present invention.

FIG. 9 illustrates the torque curve of FIG. 8 converted to the time domain in accordance with an embodiment of the present invention.

Fig. 10 shows the amount of torque added to (positive) and removed from (negative) the powertrain by the second power source of the hybrid engine according to an embodiment of the present invention.

FIG. 11 illustrates a comparison of total driveline torque between operation of an internal combustion engine only and operation of the engine in conjunction with a second power source, according to an embodiment of the present invention.

FIG. 12 is an exemplary schematic flow chart diagram of a method for selecting a most fuel efficient firing sequence in accordance with an embodiment of the present invention.

FIG. 13 is an exemplary schematic flow diagram of a harmonic cancellation method according to an embodiment of the invention.

FIG. 14 shows a timeline illustrating the timing of a smoothing torque determination for a particular duty cycle relative to an associated duty cycle, in accordance with an embodiment of the present invention.

Fig. 15 shows an exemplary filter characteristic according to an embodiment of the present invention.

FIG. 16 illustrates a representative engine torque curve and the resulting filtered signal suitable for driving an additional power source/sink in accordance with an embodiment of the present invention.

FIG. 17 illustrates rejection of first and second order frequencies according to an embodiment of the invention.

Fig. 18A-18D illustrate examples of cross-fades during firing fraction transitions, in accordance with embodiments of the present invention.

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

Detailed Description

Methods and systems are described for reducing NVH and improving fuel efficiency in a hybrid engine using a skip fire or a firing level modulation controlled internal combustion engine as a power source. The auxiliary power source/sink is capable of adding and/or removing torque from the powertrain in a controlled manner, which helps reduce NVH produced by the engine.

Skip fire operation most commonly includes cylinder deactivation in which the intake and exhaust valves remain closed during the nominal gas exchange phase of the 4-stroke engine cycle. Performing cylinder deactivation requires the engine controller to control the power driver output that activates the cylinder deactivation element. For cam operated valves, cylinder deactivation may be achieved by actuating a solenoid that operates a hydraulic oil control valve that allows the valve lifters to remain rigid (fired cylinders) or collapse (skipped cylinders). Such systems may be referred to as "idle" deactivation systems. Other mechanisms of cam operated valves may be used to achieve cylinder deactivation. Alternatively, electromechanical actuators may be used to control the intake and/or exhaust valves. Regardless of the cylinder deactivation method, there is a time lag between making the firing/misfire determination and the firing cylinder's intake valve opening.

Skip fire control of varying and sometimes irregular firing patterns in an internal combustion engine can result in some firing patterns having unacceptable NVH. One way to deal with such problems is to not use a particular firing fraction or firing sequence that is known to produce unacceptable levels of NVH. Conversely, other firing fractions or sequences are used, and cylinder output is adjusted accordingly (e.g., by adjusting manifold absolute pressure, spark advance, etc.) such that the desired engine output is delivered. These allowed firing fractions are selected based on their desired NVH properties, i.e., the resulting NVH when operating at these firing fractions is acceptable. Various methods of this type are described in commonly assigned U.S. patent application nos. 13/654,244 and 14/638,908, which are hereby incorporated in their entireties for all purposes. Commonly assigned U.S. patent application No. 14/992,779, which is incorporated herein in its entirety for all purposes, describes some systems and methods for integrating additional power sources/sinks with a dynamic skip fire controlled engine. Forcing skip fire engines to operate only at a limited number of firing fractions reduces the fuel efficiency gains that can be achieved with skip fire control because torque control must use other actuators, such as spark timing, MAP, and cams. Using these other actuators to control torque output is generally less fuel efficient than control based exclusively on firing fraction.

This application describes various control methods in which the second power source/sink is operated in a manner that produces a smooth torque applied to the vehicle driveline in addition to the internal combustion engine. A smoothing torque is any torque that is applied to help eliminate or reduce torque variations produced by the internal combustion engine. The smoothing torque may be generated by any suitable energy storage/capture/release device. One example would be an electric motor/generator with a battery and/or capacitor to store and release energy. Alternatively, any system or device that stores and captures/releases energy mechanically, pneumatically, or hydraulically may be used. For example, a flywheel with a variable mechanical coupling, or a high pressure fluid reservoir with a valve controlling fluid flow to and from a turbine or similar device, may be used to capture/release energy from the powertrain. The smoothing torque is applied in a manner such that noise and vibration generated by the skip fire firing sequence is at least partially reduced or eliminated.

FIG. 1 schematically illustrates an exemplary hybrid electric vehicle powertrain and associated components that may be used in conjunction with the present invention. These figures illustrate a parallel hybrid electric powertrain configuration, however it should be understood that the same concepts may be applied to other hybrid powertrains, including series hybrid electric configurations, power split electric configurations, and hydraulic hybrid configurations, although the greatest improvements in fuel efficiency are expected for parallel and series electric hybrid configurations.

Fig. 1 shows a skip fire controlled engine 10 that applies torque to a powertrain driveshaft that is connected to a transmission 12 that in turn drives selected wheels 20 of a vehicle. The motor/generator 14 is also coupled to the powertrain and is capable of either simultaneously generating electrical power (thereby effectively subtracting torque from the drive shaft) or supplementing engine torque, depending on whether the engine is producing excess or insufficient torque relative to the desired powertrain torque output. When the engine produces excess torque, the excess torque causes the motor/generator 14 to generate electricity, which is stored in an energy storage device 24, which may be a battery and/or a capacitor, after being conditioned by power electronics 26. The power electronics 26 may include circuitry for converting the output voltage on the energy storage device 24 to a voltage suitable for delivering/receiving power from the motor/generator 14. When the torque produced by the engine is insufficient, the engine torque is supplemented with torque generated by the motor/generator 14 using energy previously stored in the energy storage device 24. The use of a capacitor as the energy storage device 24 may result in a greater improvement in the overall fuel economy of the vehicle, as it largely avoids the energy losses associated with charging and discharging a conventional battery, which is particularly advantageous when relatively frequent storage and retrieval cycles are contemplated as in the present invention.

FIG. 2 illustrates a hybrid vehicle control system suitable for controlling the hybrid vehicle powertrain shown in FIG. 1, in accordance with certain embodiments. Vehicle control system 100 includes an Engine Control Unit (ECU)130, an internal combustion engine 150, a powertrain 142, and an additional power source/sink 140. The additional power source/sink may include power electronics, a motor/generator, and an energy storage device. ECU 130 receives input signal 114 indicative of a desired engine output. The input signal 114 may be processed as a request for a desired engine output or torque. The signal 114 may be received or derived from an accelerator pedal position sensor (APP)163 or other suitable source, such as cruise control, torque calculator, or the like. The optional preprocessor 105 may modify the accelerator pedal signal before it is delivered to the engine controller 130. However, it should be appreciated that in other embodiments, the accelerator pedal position sensor may be in direct communication with the engine controller 130.

The ECU 130 may include a spark sequencer 202, a torque model module 204, a powertrain parameters module 206, a spark control unit 210, and an NVH reduction module 208. These units and modules communicate with each other and work in conjunction to control the vehicle. The firing sequencer 202 determines a sequence of skips and firings for the cylinders of the engine 150. The firing sequence may be generated based on the firing fraction and the output of the delta-sigma converter, or may be generated in any suitable manner, such as described in U.S. patents 8,099,224, 9,086,020, and 9,200,587, which are incorporated herein by reference in their entirety. In operation, the firing sequence generator may investigate the fuel efficiency associated with various firing sequences and select the firing sequence that provides the best fuel economy while meeting the torque request. In some cases, powertrain torque may be supplemented or reduced by power source/sink 140. The output of the firing sequencer is a drive pulse signal 113 that may be composed of a bit stream, where each 0 indicates skip and each 1 indicates firing for the associated cylinder firing opportunity, thereby defining a firing sequence. Firing decisions associated with any firing opportunity are generated prior to the firing opportunity to provide sufficient time for the firing control unit 210 to properly configure the engine 150, e.g., deactivate the cylinder intake valves at skipped firing opportunities. The torque model module 204 determines an estimated torque based on the firing sequence and the powertrain parameters determined by the powertrain parameters module 206. These powertrain parameters may include, but are not limited to, intake Manifold Absolute Pressure (MAP), cam phase angle, spark timing, exhaust gas recirculation level, and engine speed. The powertrain parameters module 206 may direct the ignition control unit 210 to appropriately set the selected powertrain parameters to ensure that the actual powertrain output is substantially equal to the requested output. The ignition control unit 210 may also activate cylinder ignition. The NVH reduction module 208 may use the output of the torque model module 204 to determine the NVH associated with any particular firing sequence and powertrain parameter set. In some cases, the NVH reduction module 208 may direct the additional power source/sink 140 to add or subtract torque to the powertrain 142. It should be appreciated that the various modules depicted in FIG. 2 may be combined or configured in different ways without affecting the overall functionality of the vehicle control system 100.

Torque curve

In order to determine whether it is necessary to supply a smoothing torque and what the smoothing torque should be, it is advantageous to estimate the total torque curve of the internal combustion engine. This estimation must be done in an accurate, computationally efficient manner so that the engine torque curve can be predicted in real time. The predicted torque curve may then be used to determine what, if any, smoothing torque is required.

In various methods, the above smoothing torque may be selectively applied. That is, many firing fractions and firing sequences deliver an engine torque curve with acceptable NVH levels, and therefore do not require the application of a smoothing torque under those conditions. In other cases, the firing fraction or firing sequence may produce undesirable levels of NVH. Under these conditions, a smoothing torque may be applied to reduce NVH to acceptable levels. In other cases, different firing fractions or firing sequences with acceptable NVH characteristics may be used. Smoothing torque may optionally be used with the firing fraction or sequence. In various embodiments, the smoothing torque system is arranged to analyze the energy costs of the available options and select the most fuel efficient method that also achieves an acceptable level of NVH.

The total torque curve of a skip fire controlled internal combustion engine may be modeled using a single cylinder, normalized torque curve. The normalized profiles for the fired and skipped cylinders may be recorded in a look-up table. Tables may be generated for various levels of intake Manifold Absolute Pressure (MAP), such as MAP increments of 10 kPa. Intermediate values may be determined by interpolating values from these tables. The estimated torque profile for each cylinder may then be determined based on scaling and shifting the normalized torque by several factors, such as spark and cam phase angle, to control the opening and closing times of the intake and/or exhaust valves. Different normalization curves may be used for the fired and skipped cylinders. When different ignition levels are used, the different ignition levels may be modeled differently by starting with different normalized curves for each different ignition level and/or by scaling and shifting differently based on the different spark and cam settings used. The estimated torque curves for all of the engine cylinders may be summed with appropriate phasing to obtain a total engine torque curve. The methods described herein may be used to determine engine torque with a crank angle resolution of 0.5 °, although, as described below, a coarser resolution may often be used to reduce computation time without significantly affecting model accuracy.

Fig. 3A and 3B show torque curves associated with two different MAP values for an engine operating over a range of speeds. FIG. 3A is the average MAP for 70kPa, and FIG. 3B is the average MAP for 40 kPa. In both cases, the vertical scale is torque and the horizontal scale is engine crank angle. Both graphs are for the cylinders that are fired. The torque curves are shown in these figures in 0.5 crank angle increments. The various individual cycle profiles shown represent a range of engine speeds and cam angles. In all cases, spark timing may be adjusted to achieve optimal fuel efficiency.

Fig. 3A and 3B depict cylinder torque curves for a 4-stroke engine. Such engines complete an engine cycle in 720 crank rotations. The engine cycle may be divided into four phases or strokes, intake, compression, combustion (power), and exhaust. Each stroke extends over a crank angle rotation range of 180 °. The stroke transitions correspond to successive Top Dead Center (TDC) and Bottom Dead Center (BDC) piston positions. The torque here is zero because the lever arm on the crankshaft is zero at TDC and BDC.

Examination of fig. 3A and 3B shows that the maximum torque produced during the combustion stroke is significantly higher at 70kPa compared to 30kPa, as more air and fuel is induced into the cylinder at the higher MAP valve. Also, at a smaller MAP value, the pumping loss represented by the negative torque region in the intake stroke is larger. Skip fire engine operation tends to operate at higher MAP values to minimize these pumping losses and thereby improve fuel economy.

The torque curves at each MAP and cam angle may be normalized. Fig. 4A and 4B show such normalized torque curves for the combustion stroke of an engine cycle with a cam angle of 30 ° and average MAP of 70kPa and 40kPa, respectively. In these figures, the vertical axis is the normalized torque and the horizontal axis is the crank angle. An important, unexpected observation is that by normalizing the torque curve to the highest instantaneous torque, all normalized torque curves associated with each firing are substantially the same for all engine speeds. Fig. 5A and 5B show such normalized torque curves for the compression stroke of an engine cycle with a cam angle of 30 ° and an average MAP of 70kPa and 40 kPa. In this figure, the vertical axis is the normalized torque and the horizontal axis is the crank angle. Again, all of the individual torque curves have substantially the same normalized torque curve. Similar normalization curves may be generated for the intake and exhaust strokes of the fired cylinder.

Likewise, similar curves may be generated for skipped cylinders. The skipped cylinders have no power to produce combustion or high temperature exhaust gases. Thus, when a low pressure gas spring is used, the "intake" and "combustion" strokes may have substantially similar curves, as may the "compression" and "exhaust" strokes. The nature of the torque curve during skipped firing opportunities will vary depending on valve motion during skipped opportunities. Skipped cylinders may be deactivated, wherein one or both of the intake and exhaust valves remain closed during the engine cycle such that no air is pumped through the cylinder. If both valves are closed during the cycle, hot exhaust gas may be trapped in the cylinder or may be released before closing the valves. These conditions may be referred to as the formation of a "low pressure" spring (which expels exhaust gas prior to cylinder deactivation) or a "high pressure" spring (which traps exhaust gas by deactivating an exhaust valve prior to venting a previous ignition). These cases will have different torque curves that can be modeled. In some cases, the skipped cylinder may not deactivate the valve but may pump air through the cylinder. Again, this situation can also be modeled. To assist in understanding the invention, the following tables and description will assume that the skipped cylinder is operating in a "low pressure" spring mode, but this is not a requirement.

Fig. 6 shows a table 400 illustrating curves similar to the curves shown in fig. 4A, 4B, 5A and 5B in a table format. In table 400, the rows correspond to crank angles and the columns correspond to different MAP values. Although different types of normalization may be used, the columns are normalized in the table so that the curves associated with each MAP value cover the same area. A separate table may be constructed for each engine stroke (i.e., intake, compression, combustion (power), and exhaust of the activated cylinder). Also, separate tables may be constructed for two different crankshaft rotations (i.e., intake/compression rotation and compression/exhaust rotation) of the skipped engine cycle. Separate tables for each stroke or crankshaft rotation are useful because the scaling factor may be different between different strokes in any given engine cycle depending on engine operating conditions.

Since the normalized torque curve associated with any given firing or skip is known, the estimated torque curve associated with each firing opportunity may be determined by scaling the normalized torque curve by an appropriate scaling factor. FIG. 7 illustrates a portion of an exemplary table 500 for scaling a normalized torque curve. The table entries are proportional to the total torque produced in the stroke for a given MAP (rows in the table) and average engine speed (columns in the table). The average engine speed in a vehicle application is known in real time based on vehicle sensors that monitor engine speed. The table shown in fig. 5 is for a cam phase angle of 30 °. Other similar tables may be constructed for other cam phase angles. In engines using dual cams, different tables may use different combinations of intake and exhaust valve timing.

The effect of spark timing on the torque curve may be handled in different ways. One approach would be to construct tables similar to tables 400 and 500 for different spark timing values. Since spark timing will typically have relatively little effect on other engine strokes, it is likely that only a table of combustion strokes will be needed. An alternative method of dealing with spark timing would be to generate a spark timing multiplier that can be multiplied by the values in table 500 to adjust spark timing. In some embodiments, rather than constructing alternative tables 400 and 500 for different cam phase angles, the effect of varying cam phase angles may be incorporated into the torque model by using simple multipliers.

An alternative approach to include spark timing is to represent the actual torque curve for various spark timings, i.e., a set of tables 400 constructed similar to the table shown in FIG. 6 for several sets of cams and spark timings. Then, all that is required to generate the actual torque curve would be a simple multiplication step between the normalized torque curve of table 400 (FIG. 6) and the scaling factor of table 500 (FIG. 7).

Multiplying the normalized torque curve of table 400 by the appropriate scaling factor of table 500 provides a real-time estimate of the torque curve in degrees of crank angle for any given cylinder. Once the estimated torque curves associated with each cylinder are determined, it is a simple matter to sum the single cylinder torque curves. The cylinder curve will be offset in crank angle and hence time. For a 4-cylinder, 4-stroke engine, cylinder firing will be offset by 180 crank angle. The sum of consecutive firings and skipped firings associated with all cylinders is the engine torque curve. Fig. 8 shows an example of such an engine torque curve for a four cylinder, 4-stroke engine operating at an average engine speed of 1500rpm at a firing fraction of 3/4. The vertical axis is the total net torque from all cylinders and the horizontal axis is the crank angle. In this example, the firing pattern repeats every 720 °. There are three engine torque spikes 813 every 720 ° that are associated with three cylinders firing each engine cycle. The duration of each torque spike is relatively short. The skipped firing opportunity indicates a torque drop (dip). In this example, the cylinder load is about 65% of its maximum value. Often, operation at about 65% of maximum cylinder load corresponds to minimizing Brake Specific Fuel Consumption (BSFC). During the engine cycle, the maximum instantaneous delivered torque is greater than 175N m, which may produce unacceptable NVH performance. Without adding a smoothing torque, a less fuel efficient firing fraction may have to be selected to provide the requested torque.

Scaling multiplier based on firing history

In some embodiments, one or more additional multipliers based on the firing history may be used to further more accurately scale the normalized torque curve model to the delivered torque. These multipliers may be based on the firing history of the particular cylinder and/or the firing history of the immediately preceding engine firing opportunity (firing sequence). During skip fire operation of the engine, the amount of torque provided by any particular firing will vary according to both: (a) the firing history of a particular cylinder; and (b) an ignition history of an immediately preceding engine ignition timing. Typically, when other conditions are equal, the torque produced when a particular cylinder is fired after it was skipped in its previous operating cycle will be greater than the torque produced when the same cylinder was fired after it was fired in its previous operating cycle. This is due in part to the difference between the valve actuation schemes of a fired duty cycle following a skipped duty cycle versus a fired duty cycle following another fired duty cycle. More specifically, when an ignited duty cycle follows another ignited duty cycle, the exhaust valve opening from the previous duty cycle will typically overlap with the intake valve opening in the subsequent duty cycle. This results in a different amount of air being introduced into the cylinder than in the case where the exhaust valve opening does not overlap with the intake valve opening, as typically occurs when a duty cycle fired in the same cylinder follows a skipped duty cycle. Another factor that affects the amount of intake air is the cooling of the cylinder, which allows more air (and correspondingly fuel) to be introduced into the fired cylinder. When the cylinder is skipped in its first two firing occasions, even more cooling may occur and the intake air amount (and therefore the cylinder torque output) may be increased further accordingly. With all other parameters equal, the torque output for different firing occasions for the same cylinder may vary by more than 10% based on the firing history of that particular cylinder. Typically, the skip/fire status of an immediately preceding working cycle of a cylinder has the most significant effect on the torque output of a particular cylinder during a particular working cycle — however, the effect may be viewed based on the skip/fire status of several previous working cycles.

Similarly, the total engine cylinder firing history may also affect the output of any particular cylinder firing. Typically, it does not have an associated intake event when the previous cylinder in the cylinder firing order is skipped. When no intake event occurs, the pressure in the intake manifold will increase slightly-which results in more air being introduced at the intake event of the following cylinder in the cylinder firing sequence. The effect of the intake event (i.e., engine ignition history) associated with several previous cylinders affects the intake air quantity, somewhat like a single cylinder ignition history. Again, based on the current engine firing history, the torque output for different firing opportunities in the engine cycle may vary by more than 10%.

The effect of either or both of the cylinder firing history and the engine firing history may be accounted for by using appropriate firing history-based multipliers derived from a firing history table or other suitable map.

For example, the following two tables illustrate one particular table embodiment that illustrates the effects of an engine firing sequence. The first table illustrates multipliers that are based on the number of firings that have occurred since the last skip. In this example, if the currently fired cylinder is the first firing in the engine firing sequence after skipping, a torque multiplier of 1.05 is used. If the currently fired cylinder is the second consecutive fire in the engine firing sequence after the skip, a torque multiplier of 1.01 is used. If the currently fired cylinder is the third consecutive fire in the engine firing sequence after the skip, a torque multiplier of 0.98 is used. If the currently fired cylinder is the fourth consecutive fire in the engine firing sequence after skip or higher, a torque multiplier of 0.96 is used. It should be appreciated that this table is particularly useful when firing fractions greater than 1/2 are used, where it is contemplated that the resulting firing sequence may include multiple firings in a row.

Number of ignitions

A second table may be used to illustrate the effect of multiple sequence skips in the firing order immediately preceding the currently fired cylinder. In this table, the number of consecutive skips that occur before the current fire is used as an index. In this example, if the currently fired cylinder follows a single skip in the engine firing sequence, a multiplier of 0.98 is used. If the currently fired cylinder follows two consecutive skips in the engine firing sequence, a multiplier of 0.99 is used. If the currently fired cylinder follows three consecutive skips in the engine firing sequence, a multiplier of 1.03 is used. If the currently fired cylinder follows four or more consecutive skips in the engine firing sequence, a multiplier of 1.04 is used. It should be appreciated that this table is particularly useful when firing fractions less than 1/2 are used, where it is contemplated that the resulting firing sequence may include multiple skips in a row.

Number of skips

The particular multiplier used in the engine firing history table described above will vary based on a number of engine related factors, such as intake manifold dynamics, the nature of the engine, and the characteristics of the normalized torque curve.

A separate table may be used to determine the appropriate multipliers to account for the firing history of the cylinders themselves. One such table, illustrated below, that is suitable for use when the fired cylinder was skipped in its previous operating cycle, utilizes intake manifold pressure (MAP) and cam advance angle as its index. In this example, a multiplier of 1.0 is used when the manifold pressure is 50kPA and the cam advance angle is 0 degrees. If the cam advance angle is 10 degrees, a multiplier of 1.02 is used. If the cam advance angle is 30 degrees, a multiplier of 1.07 is used. If the cam advance angle is 60 degrees, a multiplier of 1.10 is used. Suitable values are also provided for other manifold pressures. When the current intake manifold pressure and/or current cam advance angle is between the indexed values in the table, the interpolated value may be used to obtain a more accurate multiplier.

Cam advance angle (degree)

Again, the particular multiplier used will vary based on a variety of engine-related characteristics.

Transforming to the time domain

In some embodiments, it may be desirable to transform the information available in the crank angle domain to the time domain. A rough way to transform the crank angle domain to the time domain is to simply use the average engine speed. We have found that:

Δt_avgΔ (crank angle)/(average engine speed) (equation 1)

For example, if the average engine speed is 1500rpm, then a crank angle of 0.5 ° is equal to about 0.056msec, and the crank angle domain can be easily transformed into the time domain.

Alternatively, a more accurate method of converting the crank angle to time may be used. Most vehicles use an engine speed sensor to monitor engine speed in real time. The sensors typically measure the time between successive marks on the flywheel rotating with the engine passing a fixed sensor to determine engine speed. The mark spacing typically has a crank angle of 6 °. A change in torque supplied to the powertrain will result in a change in engine speed, which can be measured with an engine speed sensor. For example, a torque spike associated with cylinder firing will cause the engine/vehicle to accelerate, while a torque drop associated with skipped firing opportunities will cause the engine/vehicle to decelerate.

The engine controller may compare the most recent change in engine torque determined from the previously described torque model to the most recently measured change in engine speed and establish a correlation between the two. The controller may then extrapolate this relationship for future estimated torque curves to help transform the crank angle domain to the time domain. It will be appreciated that the transformation of the crank angle domain into the time domain is not limited to the previously described method, but any suitable method may be used.

FIG. 9 illustrates the transformation of the torque curve of FIG. 8 to the time domain rather than the crank angle domain. In this figure, the vertical axis is the applied torque and the horizontal axis is time. The change in engine speed with applied torque is included in the time-to-base transformation. The total elapsed time 240msec in the graph corresponds to the same three engine cycles depicted in FIG. 8. In addition to transforming the horizontal axis from the crank angle domain to the time domain, fig. 9 also depicts a more coarse resolution model. In this case, the torque curve is modeled in 6 crank increments instead of the 0.5 increments previously described. The result is a more stepped torque curve. In practice, we have found that 6 ° modeling yields sufficient resolution for engine control and diagnostic purposes. In some cases, an even coarser resolution (such as a resolution of 12 °, 30 °, or even 60 °) may be sufficient. The advantage of using a coarser resolution is the reduction of memory and computational requirements on the engine control unit. It should be noted that whether the crank angle domain is used (fig. 8) or the time domain is used (fig. 9), the overall shape of the torque curve is very similar, with only slight variations resulting from the transformation.

Application of torque curves

Knowledge of the torque curve can be used advantageously in a number of ways. In particular, knowledge of the torque curve associated with the upcoming firing opportunity may be used to control the smoothing torque applied in parallel with the powertrain to eliminate or partially eliminate variations in total powertrain torque. This smoothing torque may be positive (adding torque to the driveline) or negative (subtracting torque from the driveline), or both. The smoothing torque may be supplied by the motor/generator as previously described or some other component.

The engine controller may determine a torque curve for various firing fractions and firing sequences that deliver the requested torque. Some of these curves may require the application of a smoothing torque to provide acceptable NVH characteristics. The engine controller may then select the fraction or sequence from the set of firing fractions or firing sequences that provides the requested torque with minimal fuel consumption. Typically, the selected firing fraction or sequence will provide the desired torque with each cylinder operating at or near its optimum efficiency.

A set of torque limit calibration tables may be constructed for different engine speeds and transmission gears. These tables aggregate the maximum allowed instantaneous torque for different operating conditions. If any point on a torque curve (such as the torque curve shown in FIG. 9) exceeds the torque limit in the calibration table, the firing fraction or firing sequence is not allowed unless a smoothing torque is applied to the powertrain of the vehicle. For example, if the calibrated torque limit 917, corresponding to an engine speed of 1500rpm and when the vehicle is in third gear, is 110N × m, the torque curve depicted in fig. 9 will not be allowed, since the maximum instantaneous value significantly exceeds this value.

Other measures of NVH may be aggregated in addition to or in place of the torque limit calibration table. For example, angular jerk (angular jerk), time derivative of the torque may be determined for different torque curves. If the angular jerk exceeds a certain value within a defined frequency range, the firing sequence may not be allowed or a smoothing torque may be added to reduce the angular jerk. In still other embodiments, the limit may be expressed in terms of a weighted RMS vibration threshold. That is, a weighted RMS average of the instantaneous torque changes may be determined and the value may be compared to a maximum permissible weighted RMS vibration threshold.

Fig. 10 shows a smoothing torque that may be applied to the vehicle's powertrain by an additional power source/sink to reduce the maximum transient torque to a calibrated limit. In this figure, the vertical axis is the applied torque and the horizontal axis is time. A positive applied torque represents torque added to the driveline and a negative torque represents torque removed from the driveline. Examination of fig. 10 shows that there are periods of no applied torque, periods of negative applied torque, and periods of positive applied torque. Three consecutive negative torque periods 1013 in an engine cycle overlap with those portions where three torque spikes 813 corresponding to cylinder firings of the internal combustion engine exceed the torque limit. A positive period of applied torque overlaps with the torque slot (torque ramp) associated with the skipped cylinder. The curve of smoothed torque may be selected to substantially match the shape of the torque curve associated with the firing cylinder. This results in a more repetitive torque curve, which can be perceived as having lower NVH.

It will be appreciated that the duration of those portions of the engine torque spike 813 that are cancelled out by negative torque pulses in the smoothed torque is very short, with each pulse corresponding to less than 180 degrees of crankshaft rotation, and typically less than 90 degrees of crankshaft rotation.

The amount of positive and negative power supplied by the additional power source/sink may be controlled such that they are equal, with less losses associated with the energy capture/storage/release system. Control in this manner will result in the amount of stored energy remaining relatively fixed at some appropriate level. If more stored energy is desired, the amount of power drawn from the powertrain may be increased, and if less stored energy is desired, the amount of power delivered to the powertrain may be increased. In some embodiments, the energy extracted from the powertrain is returned (minus losses) within a cyclic pattern, which in some cases is within the same engine cycle. More specifically, the extracted energy is preferably returned within a period of time as follows: this period is equal to the number of degrees of crank angle associated with each firing opportunity (sometimes referred to herein as the firing opportunity period) multiplied by the denominator of the firing fraction. In an 8-cylinder engine, each ignition timing is associated with a crankshaft rotation of 90 degrees (ignition timing period); in a 6-cylinder engine, each firing opportunity is associated with 120 degrees of crankshaft rotation; and in a 4-cylinder engine, each ignition timing is associated with 180 degrees of crankshaft rotation. Thus, for example, when a firing fraction with a denominator of 5 (e.g., 1/5, 2/5, 3/5, 4/5) is used in an eight-cylinder engine, energy is preferably returned within 450 degrees of crankshaft rotation (90 x 5) -while a 4-cylinder engine operating at the same firing fraction will return its energy within 900 degrees of crankshaft rotation (180 x 5). Of course, the actual period in which energy will be returned will vary depending on both the number of available cylinders and the operating firing fraction.

Fig. 11 shows a comparison of powertrain torque curves between a skip fire controlled engine without smoothing torque and a skip fire controlled engine practicing the present invention. The dashed line depicts only the torque curve of the internal combustion engine (without any compensation). This curve is the same as the one shown in fig. 9. The solid line depicts the torque curve of the engine and motor/generator combination that can both add and remove torque from the powertrain. The solid line depicts the torque curve for the combination of the engine and the motor/generator that can both add and remove torque from the powertrain. It is obtained by adding the smoothed torque of fig. 10 to the internal combustion engine torque curve. Examination of fig. 11 shows that the instantaneous torque curve remains below 110N m (this is the limit in this example) at all times. It will be appreciated that the torque limit varies with engine speed and transmission gear ratio, and may also depend on other variables, such as the tip-in or tip-out rate of the accelerator pedal.

In some embodiments, the predicted torque curve may be determined for a plurality of future firing opportunities assuming different firing fractions or firing sequences. The prediction may extend at least several firings into the future relative to the current firing opportunity. Preferably, the prediction extends far enough into the future so that the engine controller can activate/deactivate the engine valves as appropriate for firing/skipping. This advance period may correspond to 3 to 9 future firing opportunities, depending on engine speed and valve actuation mechanism. In some cases, both longer and shorter prediction periods may be used. In some embodiments, the predicted torque curve may extend over the period between making the firing decision and implementing the firing decision.

The engine controller may determine NVH and fuel consumption associated with a firing fraction or number of firing fractions or firing sequences that deliver the requested torque. For a certain firing fraction or firing sequence, a smoothing torque may be required to provide acceptable NVH. The controller may then choose to run the engine according to the firing fraction or firing sequence and optionally the smoothing torque to provide acceptable NVH while minimizing fuel consumption. The engine controller may also take into account other variables, such as the storage level in an energy storage device associated with an auxiliary power source/sink that provides smooth torque and conversion efficiency to and from the energy storage device, in making the determination of the appropriate firing fraction or firing sequence. The engine controller may use additional knowledge such as whether the energy in the energy storage device is obtained from the internal combustion engine or some other power source (such as the power network in a plug-in hybrid). Use of the present invention will allow operation according to previously unallowable firing fractions, thereby improving fuel efficiency.

FIG. 12 schematically illustrates a method 1200 of determining a most fuel efficient firing sequence in accordance with an embodiment of the invention. In this method, one or more candidate firing sequences may be generated at step 1210 by the firing sequencer 202 (fig. 2) based on the torque request. The candidate firing sequences may be generated by any known method, such as those described in U.S. patents 8,099,224, 9,086,020, 9,200,587, and 9,200,575 and U.S. patent applications 14/638,908 and 14/704,630, which are incorporated herein by reference in their entirety. These sequences are input into the torque model 1220. Also input to the torque model are various engine parameters such as spark timing, cam phase angle, engine speed, MAP, etc. At step 1230, the torque model 1220 determines the torque curves for the candidate firing sequences. Then, it may be evaluated at step 1240 whether the candidate firing sequence requires a smoothing torque to provide an acceptable level of NVH. The vehicle transmission gear setting may be used to make this evaluation. If smooth torque is not required, the flowchart may proceed to step 1260. If a smoothing torque is required, it is evaluated in step 1250 if there is sufficient stored energy to supply the smoothing torque. If insufficient stored energy is available, the candidate firing sequence cannot be used. If the available energy is sufficient, the method proceeds to step 1260 where the fuel efficiencies of the evaluated ignition sequences are compared and the ignition sequence that provides the best fuel efficiency is selected as the operating ignition sequence. The method then proceeds to step 1270, where the engine is operated according to the selected operating firing sequence. Method 1200 may be repeated for each firing opportunity to determine an optimal firing sequence.

Generating a smoothed torque to compensate for internal combustion engine torque variations is one application of the previously described torque model. Variables input to the model during engine operation may include cam angle (control valve timing), MAP, engine speed, spark timing, crank angle, firing sequence, and firing fraction, which are known. The torque model may generate a transient engine torque curve. Knowing the instantaneous torque at a particular crank angle, the engine controller can control the smoothing torque that needs to be removed from the driveline, e.g., by the generator, or added to the driveline, e.g., by the electric motor. The electric motor/generator may be integrated into a single unit that communicates with an electrical energy storage device, such as a battery or capacitor.

In the above description of fig. 9 to 11, the torque curves and the smoothing torque are shown in the time domain. It should be appreciated that in other embodiments, the smoothing torque may be determined and applied in the crank angle domain, rather than being converted to the time domain. This may be advantageous in some applications because the crank angle is always available to the engine controller. In such embodiments, the drawing of torque may involve starting at "x" degrees and ending at "y" degrees, or the addition of torque may begin at "m" degrees and end at "n" degrees. As suggested above, the values of "x, y, m, n" may be arranged as a table and determined according to the current RPM.

Transient conditions

The foregoing description has generally been directed to selecting an optimal combination of engine firing fraction, cylinder load, and smoothing torque during operation at nominal steady-state conditions. While this is important, skip-fire controlled vehicles will often switch between the allowed firing fractions to deliver the required torque. A historical problem with skip fire engines, as well as variable displacement engines, has been the unacceptable NVH produced during transitions between numbers of cylinders fired (i.e., changes in firing fractions).

The smoothing torque may be applied during any transition, such as the transition associated with changing the firing fraction level. As described in co-pending U.S. patent applications 13/654,248, 14/857,371 and U.S. provisional patent application 62/296,451, which are incorporated by reference herein in their entirety, the transition between firing fraction levels may be an unacceptable source of NVH. The use of smoothing torque during those transitions may shorten the required transition time and reduce the use of fuel that consumes (waste) spark retard during the transitions.

One method of dealing with transient conditions may be referred to as harmonic cancellation. In this method, a theoretically predicted engine torque curve is transmitted through a specially designed FIR (finite impulse response) band-pass filter in the crank angle domain to extract DSF frequency components that cause excessive vibration in real time. The engine torque curve may be determined using the previously described method. The filtered signal may be used to generate a smoothed torque via the electric motor/generator to reduce overall powertrain torque variations. The filtering may be implemented using a set of FIR filters that may be run in parallel, each FIR filter extracting a particular frequency band in the crank angle domain. An advantage of the harmonic cancellation method is that the same filter algorithm can be used to quantify the vibrations caused by the DSF in both steady-state and transient conditions.

Harmonic cancellation provides a real-time target torque signal in a digitally efficient manner that may be used in hybrid vehicle applications to reduce vibration. It may be particularly suitable for micro-hybrids where the starter motor acts as a motor/generator and the energy storage capacity is limited. This type of system can handle the relatively small and short duration torque requirements associated with firing fraction transitions, which typically last less than two seconds.

To apply harmonic cancellation, the torque curve may be determined using the previously described method or any other suitable method. For example, once the ECU makes a "firing" or "skip" decision for a cylinder, a torque waveform is generated based on engine parameters in the crank domain (such as engine speed, MAP, cam angle, etc.). The total torque waveform may be assembled by combining the torque waveforms of all cylinders. The total engine torque signal may then be directed through a bank of FIR filters to extract vibrational energy (harmonics) caused by the DSF operation. Since lower frequencies tend to have greater NVH effects, the filter bank may consist of band pass filters of order one DSF and two DSF in the crank angle domain. Filtering in the crank angle domain means that the "frequencies" of the first and second DSF orders may be fixed relative to engine speed, and therefore the filter parameters may not need to be adjusted with engine speed. The FIR filters may have a linear phase shift so that the delays of all filters are similar. This minimizes distortion in the filtered signal. The filtered values of the engine torque curve may be used to help generate a counter or smooth torque in the crank angle domain. When switching between filters to achieve smooth transitions, step-in and step-out functions (sometimes referred to as cross-fades) may be used. Alternatively, the filtered signal may be directed through a quadratic filter to minimize discontinuities during transients.

Fig. 13 shows an embodiment of a harmonic cancellation method. Inputs to the method include various engine parameters such as MAP, cam phase angle, engine speed, and spark timing. Additional inputs to the model are firing fractions or firing sequences, which define patterns of upcoming skips and firings. These values are input into the engine torque model as previously described. The engine speed and the ignition information may be input into a filter coefficient determination module. The module determines filter coefficients for various DSF orders of interest (e.g., first and second orders). In some cases, previously used filter coefficients may be used in the upcoming calculation. The future torque curve and the filter coefficients are input into a filter bank. The filter bank may be a single FIR filter or may consist of an array of FIR filters, one for each frequency band of interest. An advantage of using multiple FIR filters is that it allows different phase compensations to be applied to counteract different phase shifts when physical torque is being produced to the powertrain. The filter bank is configured to calculate an appropriate smoothing torque to eliminate low order torque oscillations in the crank angle domain. The filter coefficients used in this calculation may be sent to a filter coefficient determination module for use in subsequent calculations.

The output of the filter bank relates to a crank angle to time domain conversion module. The module may use the engine speed and the calculated future torque curve to transform an input crank domain signal to an output time domain signal. The transition may simply be based on the average engine speed, or may alternatively include a calculated speed change based on a calculated torque profile. The output of the time domain conversion module may relate to the power electronics unit 26 (see fig. 1) of the motor/generator. The power electronics unit 26 controls a motor/generator that adds or subtracts torque from the powertrain as specified by the time domain conversion module signal. The resulting powertrain torque has been smoothed to remove torque fluctuations that would cause undesirable NVH.

Fig. 14 illustrates some of the timing constraints required to successfully practice the method described in fig. 13. FIG. 14 shows a timeline illustrating decision points, implementation windows, and engine locations associated with implementing some embodiments of the method described in FIG. 13. At point D, a decision is made as to whether to skip or fire a given cylinder. As described in co-pending U.S. patent application 14/812,370, which is incorporated by reference herein in its entirety, the decision is typically made 3 to 9 firing occasions before the decision is made. In general, it is desirable to minimize the lag between making a firing decision and implementing a firing decision to improve engine responsiveness; however, a delay of this magnitude is sufficient for responsive vehicle control. The start of the duty cycle corresponding to the ignition timing associated with the determination of point D is represented on the time line of fig. 14 as point S.

Once a skip or fire decision is made, the cylinder torque profile for that firing opportunity may be determined. In fig. 14, the time for calculating the torque curve is illustrated as window a. The filter bank has a known delay, which is indicated as window B in fig. 14. This represents the time required to process the engine torque signal by the filter bank of fig. 13. Window C in FIG. 14 represents the time required to convert the filtered signal output by the motor/generator to torque on the powertrain. The method described in connection with fig. 13 can be successfully implemented as long as the end point of window C is before point S (the start of the ignition opportunity). Window D in fig. 14 represents additional unallocated time available to complete the process if it becomes necessary.

FIG. 15 shows representative filter responses for various firing fraction denominators for a 4-cylinder, 4-stroke engine. The columns in FIG. 15 correspond to various firing fractions n/2, n/3, n/4, and n/5, where n is an integer greater than zero and less than the denominator, and the numerator and denominator have no common factor. The first row in fig. 15 represents the filter characteristic associated with the first order vibrations of the engine. The horizontal axis on these charts is the normalized frequency expressed in engine order. Here, the engine order corresponds to one cylinder firing per engine revolution. The second row corresponds to the second order engine vibration frequency. The third row corresponds to the composite frequency response of the two filters. Examination of fig. 15 shows that for n/2, the first order frequency is at the engine order of 1, i.e., at the firing fraction of 1/2 in a 4-cylinder, 4-stroke engine, the engine fires once per revolution. For the case of n/3, the first order vibrations are at an engine order of 2/3, for n/4 the first order vibrations are at an engine order of 1/2, and for n/5 the first order vibrations are at an engine order of 2/5. The second order frequency is twice the frequency of the first order. The sum of the two frequency responses is the broader peak curve shown in the bottom row. In fig. 15, the shape of the filter coefficients has been adjusted to provide a substantially constant, linear phase shift for all filters. Although the peak gain is typically close to the harmonic frequencies, the peak gain need not correspond exactly to these frequencies. Instead, the gain at the harmonic frequency may be set at a defined value of 1 in the example shown in fig. 15, and the filter characteristic adjusted to provide a linear phase response.

FIG. 16 shows an exemplary resulting filtered signal for a particular engine operating condition. In this case, the engine was operated at a cam phase angle of 40 °, a speed of 1500rpm, a MAP of 50kPa, and an ignition fraction of 2/3. The resulting engine torque curve under these conditions is shown by curve 1510 in FIG. 16. As expected, curve 1510 shows two torque spikes associated with the firing cylinder followed by a torque drop associated with the skipped cylinder. The red 1520 and purple 1530 curves show the filtered signals for crank angle resolutions of 1 ° and 30 °, respectively. The curves are substantially identical, with a difference of at most 6% in the filtered signal values. The relative insensitivity of the filtered signal to the filter resolution indicates: accurate results can be obtained even with coarse resolution. The use of a coarse resolution significantly reduces the computation time required to perform the computation, for example, the time taken to determine the filtered signal at a resolution of 1 ° is about 130 times longer than the time taken to determine the filtered signal at a resolution of 30 °. This allows calculations to be made in real time in an ECU or some other vehicle control module that has only modest processing power and speed.

Fig. 17 shows the resulting suppression of first and second order vibrations in the powertrain. In this figure, the horizontal axis is the engine order (effectively normalized frequency) and the vertical axis is the amplitude of the powertrain vibrations at that frequency. The grey curve shows the response, without any smoothing torque added. Examination of the figure shows significant vibration at engine orders of 0.5 and 1. The green curve shows the resulting powertrain vibration with the addition of the smoothed torque shown produced by the filtered signal of fig. 16. As is evident in this figure, the first and second order oscillations have almost completely cancelled.

Transient conditions may be handled using a cross-fade technique as shown in fig. 18A-18D. Fig. 18A is the filtered output of one filter bank (denoted as filter a) and fig. 18B is the filtered output of a second filter bank (denoted as filter B). The outputs of the two filter banks are summed according to the switching function illustrated in fig. 18C. Fig. 18D is the sum of the filter outputs of the filter a and the filter B weighted by the switching function shown in fig. 18C. Although the switching function is shown as linear in fig. 18C, this is not a requirement. The use of cross-fading allows the filtered signal to transition seamlessly during the firing fraction transition.

Some advantages of the harmonic cancellation method are that it automatically handles transient conditions. The filtering may be independent of engine speed and cylinder load. It is energy efficient because it suppresses only certain frequency components, which is particularly important in micro-hybrid applications. The step-by-step introduction and step-by-step exit methods can smoothly switch filters. Furthermore, the method has low computational overhead for determining the filtering and gain settings and is digitally efficient both in terms of computation and memory usage.

Retreat from DCCO (deceleration cylinder cut-off)Go out

One particular transient condition that may occur in skip fire controlled engines is DCCO (deceleration cylinder cut-off). Operation of a dynamic skip fire controlled engine during DCCO is described in co-pending U.S. patent application No. 15/009,533, which is incorporated herein by reference in its entirety. The use of the DCCO improves fuel economy because the cylinders are not fueled during deceleration when torque is not requested (e.g., when the accelerator pedal is not depressed). The use of DCCO further improves fuel economy relative to more conventional (DFCO) (deceleration fuel cutoff) because the cylinders have been deactivated during DCCO so that they do not pump air. The pumped air compromises the oxidation/reduction balance required in a 3-way catalytic converter, so its use may be limited and/or additional fuel may be required to restore the catalyst balance.

One problem with DCCO is that the intake manifold is filled with air during a DCCO event. When torque is requested again, a high MAP may result in a high cylinder load, causing a torque ramp up (charge) resulting in unacceptable NVH. Solutions to this problem include reducing engine efficiency by retarding spark timing and/or skipping some cylinders without deactivating valves to help evacuate the intake manifold. Both of these solutions have limitations. Retarding the spark reduces fuel economy. Pumping air through the engine oxidizes the catalytic converter, which may require additional fuel to restore the oxidation/reduction balance, again reducing fuel economy.

During exit from the DCCO event, the MAP will typically drop from atmospheric or near atmospheric pressure to a value suitable for delivering the requested torque, e.g., 70 or 80 kPa. The previously described torque model may be used to determine the engine torque at the exit of the DCCO event. In this case, the MAP will change in successive engine cycles. MAP variation can be modeled using the methods described in co-pending U.S. patent applications 13/794,157, 62/353,218, and 62/362,177, which are incorporated herein by reference in their entirety. Other MAP estimation methods may be used. As MAP decreases, the output of each fired cylinder will generally decrease in a substantially proportional manner.

Torque ramping may be eliminated or reduced by using a smoothing torque. The smoothing torque may be selected such that the powertrain torque gradually increases from zero (the value during DCCO) towards the requested torque level. Unlike some of the previously described cases, the smoothing torque in this case will not necessarily exhibit a regular cyclical behavior, and the smoothing torque will typically remove torque from the powertrain during the transient period associated with the exit DCCO event. The energy associated with the removed torque may be stored in an energy storage device (such as a capacitor or battery) and used to power the vehicle at a future time. Applying a smoothing torque from the additional power source/sink during exit from the DCCO event improves fuel efficiency and does not affect the catalytic converter oxidation/reduction balance.

More generally, the same type of control strategy may be used whenever there is a firing fraction transition from a low firing fraction to a higher firing fraction. These transitions have a tendency to produce engine torque surges that can be mitigated by absorbing some or all of the excess torque in the energy storage device. Similarly, a transition from a high firing fraction to a low firing fraction may cause a torque drop in the engine output. The drop may be partially or completely filled using energy from the energy storage device.

Controlling accessories to help manage torque

In most of the examples given above, the smoothing torque is applied by a bi-directional energy source/sink (such as an electric motor/generator) that is capable of both adding torque to and drawing torque from the powertrain, with excess energy stored in a storage device (such as a capacitor or battery). While electric hybrid vehicles are particularly well suited for applying a smooth torque, in some cases a similar effect may be achieved in non-hybrid vehicles through active control of certain accessories. For example, most non-hybrid vehicle engines include an alternator. When generating electricity, the alternator applies a load to the engine. During normal driving, the alternator is often configured to generate electricity to charge the battery. The output of the alternator may be controlled by controlling the field winding current of the alternator. Thus, in some embodiments, the output of the alternator may be varied to load and unload the powertrain in a manner that effectively applies a smoothing torque to the powertrain.

When more power is required from the engine, the alternator field current can be reduced or removed-this will cause the output of the alternator to drop, thereby reducing the load on the powertrain, which makes more torque available to the driveline. When less power from the engine is required, the alternator may be commanded to produce more power, which provides a higher load on the engine. Thus, the alternator field current can be modulated in a manner that changes its load on the powertrain to counteract vibrations that induce torque surges. The pulse width modulated signal is typically used to drive the alternator field current and can be easily controlled to produce a higher (or lower) alternator output voltage to charge the battery and immediately increase (or decrease) the motoring load applied by the alternator to the powertrain. When the battery charge is already high and it is not desirable to charge more batteries, a device such as a rear window heater or front windshield heater may be turned on to absorb the electrical load. Using the alternator in this manner is particularly effective in handling torque ramping in applications such as transitioning from DCCO operation back to skip fire operation of the engine.

Another accessory that may sometimes be used in a similar manner is an air conditioner in an operating situation where the air conditioner is running. In particular, since the precise output of the air conditioning unit is generally not critical, its output may be modulated to provide some of the described torque smoothing functions.

Other embodiments

The embodiments described above are described primarily in the context of smoothing torque in conjunction with skip fire operation of an engine. However, it should be appreciated that the described techniques are equally applicable to embodiments utilizing multiple charge levels or other types of ignition charge modulated engine operation. Further, many of the described techniques may be used to improve operation during conventional variable displacement operation of the engine (both during transitions between different displacements and during steady state operation at a particular displacement).

Another application of the above described torque model is engine calibration. Using this method, engine calibration is much easier. A calibration table based on engine speed and firing fraction or firing sequence for each gear will indicate what operating conditions provide acceptable NVH. If the engine torque offset (extension) exceeds the allowed torque (i.e., the output of the vibration calibration table), a smoothing torque may be added to bring the total torque curve within acceptable levels.

The invention has been described in connection with specific embodiments, it should be understood that it is not intended to limit the invention to the described embodiments. On the contrary, it is intended to cover alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims. The present invention may be practiced without some or all of these specific details. In other instances, well known features may not have been described in detail to avoid unnecessarily obscuring the invention. For example, there are many forms of hybrid engines, parallel hybrids, series hybrids, micro hybrids, mild hybrids, full hybrids, depending on the relative sizes of the two power sources, the storage capacity of the auxiliary energy source, and the mechanism for storing the auxiliary energy. The invention described herein is applicable to all of these types of hybrid vehicles.

The components, process steps, and/or data structures may be implemented using various types of operating systems, programming languages, computing platforms, computer programs, and/or computing devices in accordance with the present invention. In addition, those of ordinary skill in the art will recognize that devices such as hardwired devices, Field Programmable Gate Arrays (FPGAs), Application Specific Integrated Circuits (ASICs), or the like, may also be used without departing from the scope and spirit of the inventive concepts disclosed herein. The invention may also be tangibly embodied as a set of computer instructions stored on a computer-readable medium, such as a memory device.

Claims (74)

1. A method of controlling a hybrid vehicle having an internal combustion engine and an auxiliary power source/sink, the method comprising:

determining a torque curve of the internal combustion engine by;

determining a normalized torque curve for each stroke of a cylinder in the engine, wherein the normalized torque curve is based on an intake manifold pressure;

scaling the normalized torque curve to determine a cylinder torque; and is

Summing the cylinder torques of all cylinders in the engine to obtain a total engine torque curve;

determining whether the torque curve provides acceptable NVH;

operating the hybrid vehicle only at the output of the internal combustion engine when the torque curve is determined to provide acceptable NVH; and is

Operating the hybrid vehicle with both the internal combustion engine and the auxiliary power source/sink when the torque profile is determined to provide unacceptable NVH, wherein the auxiliary power source/sink provides a smooth torque to reduce NVH to an acceptable level.

2. The method of claim 1, further comprising updating the torque curve accordingly at each firing opportunity of a cylinder in the engine.

3. The method of claim 1, wherein the acceptable NVH limit corresponds to a maximum of instantaneous torque in the torque curve.

4. A method as claimed in claim 3, wherein the maximum value of instantaneous torque varies as a function of engine speed and transmission gear.

5. The method of claim 1, wherein the acceptable NVH limit corresponds to a weighted RMS vibration threshold.

6. The method of claim 1, performed during skip fire operation of the engine.

7. The method of claim 2, wherein the scaling is based on at least one of engine speed, engine ignition history, cylinder ignition history, spark timing, valve timing, and valve lift.

8. A method as in any of claims 2, 6 or 7, wherein the acceptable NVH limit corresponds to a maximum value of instantaneous torque in the torque curve.

9. The method of claim 8, wherein the maximum value of instantaneous torque varies as a function of engine speed and transmission gear.

10. A method as claimed in any one of claims 2, 6 or 7, wherein the acceptable NVH limit corresponds to a weighted RMS vibration threshold.

11. A method of determining an operating firing fraction for delivering a desired engine output during engine operation, the method comprising:

(a) determining an estimated torque curve associated with operating the engine at a candidate firing fraction to deliver the desired engine output at a current engine operating condition, wherein the estimated torque curve is determined by:

determining a normalized torque curve for each stroke of a cylinder in the engine, wherein the normalized torque curve is based on an intake manifold pressure;

scaling the normalized torque curve to determine a cylinder torque; and is

Summing the cylinder torques of all cylinders in the engine to obtain a total engine torque curve;

(b) determining whether a smoothing torque will be required to meet NVH criteria during operation of the engine at the candidate firing fraction under the current engine operating conditions;

(c) determining a fuel efficiency associated with the candidate firing fraction, wherein when a smoothing torque is required;

(d) repeating steps (a) - (c) for each of a plurality of candidate firing fractions; and is

(f) Selecting one of the candidate firing fractions as the operating firing fraction based at least in part on the determined fuel efficiencies of the candidate firing fractions, wherein a fuel efficiency impact of applying a smoothing torque is considered in the determination of the fuel efficiency for each candidate firing fraction requiring a smoothing torque.

12. The method of claim 11, further comprising operating the engine at the selected candidate firing fraction.

13. The method of claim 11, further comprising: determining whether applying a smoothing torque is practical when a particular candidate firing fraction requires the smoothing torque, wherein the candidate firing fraction is rejected from consideration as the operating firing fraction when applying the smoothing torque is not practical.

14. The method of claim 11, wherein each torque curve is based at least in part on intake manifold pressure, engine speed, camshaft phase, and spark timing.

15. The method of claim 11, wherein:

comparing each torque curve to a torque limit associated with operating the engine at the current engine operating conditions to determine whether a smooth torque is required for that torque curve; and is

The smoothing torque is a reaction torque that is expected to prevent the torque delivered during operation at the associated firing candidate fraction from exceeding the torque limit.

16. A method as recited in claim 11 wherein said method is performed during operation of said engine in a skip fire operating mode.

17. The method of claim 11, wherein the method is performed during operation of the engine in a multi-charge level operating mode.

18. The method of claim 11 wherein the candidate firing fraction selected as the operating firing fraction is the firing fraction candidate having the best fuel economy for delivering the desired engine output.

19. The method of claim 15, wherein the smoothing torque is a filtered version of the torque curve.

20. The method of any one of claims 12, or 14 to 19, further comprising: determining whether applying a smoothing torque is practical when a particular candidate firing fraction requires the smoothing torque, wherein the candidate firing fraction is rejected from consideration as the operating firing fraction when applying the smoothing torque is not practical.

21. The method of any of claims 12-13, or 15-19, wherein each torque curve is based at least in part on at least one of intake manifold pressure, engine speed, camshaft phase, and spark timing.

22. The method of any one of claims 12 to 14, or 16 to 19, wherein:

comparing each torque curve to a torque limit associated with operating the engine at the current engine operating conditions to determine whether a smooth torque is required for that torque curve; and is

The smoothing torque is a reaction torque that is expected to prevent the torque delivered during operation at the associated firing candidate fraction from exceeding the torque limit.

23. A method as claimed in any one of claims 12 to 14, or 16 to 18, wherein the smoothing torque is a filtered version of the torque curve.

24. A method of controlling a hybrid vehicle having an internal combustion engine and an auxiliary power source/sink, the method comprising:

operating the internal combustion engine in a dynamic skip fire mode, wherein firing decisions to fire or skip fire one or more cylinders of the internal combustion engine are dynamically made on a firing opportunity by firing opportunity basis for a given reduced effective displacement that is less than a maximum displacement of the internal combustion engine;

determining a torque curve of the internal combustion engine by

Determining a normalized torque curve for each stroke of a cylinder in the engine, wherein the normalized torque curve is based on an intake manifold pressure;

scaling the normalized torque curve to determine a cylinder torque; and is

Summing the cylinder torques of all cylinders in the engine to obtain a total engine torque curve;

determining whether the torque curve provides acceptable NVH;

operating the hybrid vehicle only at the output of the internal combustion engine when the torque curve is determined to provide acceptable NVH; and is

Operating the hybrid vehicle with both the internal combustion engine and the auxiliary power source/sink when the torque profile is determined to provide unacceptable NVH while operating in the dynamic skip fire mode, wherein the auxiliary power source/sink provides a smooth torque to reduce NVH to an acceptable level.

25. The method of claim 24 wherein the torque curve is updated at each firing opportunity for all cylinders of the engine.

26. The method of claim 24, wherein the acceptable NVH limit corresponds to a maximum of instantaneous torque in the torque curve.

27. The method of claim 26, wherein said maximum value of instantaneous torque varies as a function of engine speed and transmission gear.

28. The method of claim 24, wherein the acceptable NVH limit corresponds to a weighted RMS vibration threshold.

29. The method of claim 24, wherein the scaling is based on at least one of engine speed, engine ignition history, cylinder ignition history, spark timing, valve timing, and valve lift.

30. A method of estimating a torque curve of an engine having a plurality of working chambers during operation of the engine, the engine being arranged to operate in a sequence of firing occasions, each firing occasion having a corresponding duty cycle with a corresponding operating state, each operating state having an associated normalised torque curve, the method comprising:

determining or selecting a normalized torque curve corresponding to the operating state of the selected working chamber during the selected working cycle;

determining a torque curve for the selected working chamber based at least in part on scaling the normalized torque curve corresponding to the operating state of the selected working chamber, wherein the scaling varies as a function of one or more current engine operating parameters; and is

Summing the torque curves of all of the working chambers of the engine to obtain an estimated total engine torque curve, the summed torque curves including the torque curve of the selected working chamber.

31. The method of claim 30 wherein torque curve estimation is performed during skip fire operation of the engine and the normalized torque curve is based at least in part on a skip/fire firing decision associated with the selected duty cycle.

32. The method of claim 31, wherein the normalized torque curve is based at least in part on intake manifold pressure.

33. The method of claim 30, wherein the normalized torque curve is based at least in part on intake manifold pressure.

34. The method of claim 30, wherein said one or more current engine operating parameters upon which said scaling is varied comprises engine speed.

35. The method of claim 30, wherein the one or more current engine operating parameters from which scaling changes include spark timing and valve timing.

36. The method of claim 30, wherein the engine includes a crankshaft and the normalized torque curve and total engine torque curve are in a crankshaft angle domain.

37. The method of claim 36, further comprising transforming the total engine torque profile to a time domain.

38. The method of claim 37 wherein the transformation of the total engine torque curve to the time domain is based on the total engine torque curve taking into account the effect of changes in the rotational speed of the engine.

39. The method of claim 30, further comprising:

filtering the total engine torque curve to identify selected harmonic components of the torque curve; and is

A reaction smoothing torque is determined for application to a powertrain including the engine to reduce NVH during operation of the engine.

40. The method of claim 30, further comprising using said estimated total engine torque curve in said selection of a desired operating firing fraction.

41. The method of claim 30, further comprising:

determining whether the predicted engine torque will exceed a torque limit using the total engine torque curve; and is

When it is determined that the predicted engine torque will exceed the torque limit, a reaction smoothing torque is determined that will prevent the predicted engine torque from exceeding the torque limit.

42. The method of claim 41, further comprising applying the reaction smoothing torque during operation of the engine.

43. The method of claim 42, wherein the reaction smoothing torque is applied by an electric motor or an electric motor/generator.

44. The method of claim 41, wherein the torque limit varies as a function of at least one of engine speed and transmission gear.

45. The method of claim 41, wherein the torque limit corresponds to a maximum of instantaneous torque in the torque curve.

46. The method of claim 45, wherein said maximum value of instantaneous torque varies as a function of engine speed and transmission gear.

47. The method of claim 41, further comprising: the determined reaction smoothing torque is used in the determination of the predicted fuel efficiency for operating the engine at the effective firing fraction associated with the total engine torque curve.

48. The method of claim 47, further comprising: the predicted fuel efficiency is used in the selection of the desired operating firing fraction.

49. The method of claim 42 wherein said estimation of said total engine torque curve and said determination of said reaction smoothing torque are updated for each firing opportunity such that the demand for said reaction smoothing torque and its magnitude are updated for each firing opportunity.

50. The method of claim 30, wherein the normalized torque curve is further scaled according to at least one of an engine firing history and a cylinder firing history.

51. A method as in any of claims 31, or 33-50, wherein the normalized torque curve is based at least in part on an intake manifold pressure.

52. A method as in any of claims 31-32, or 34-50, wherein the normalized torque curve is based at least in part on an intake manifold pressure.

53. A method as claimed in any of claims 31 to 33, or 35 to 50, wherein the one or more current engine operating parameters on which the scaling is varied comprise engine speed.

54. A method as claimed in any of claims 31 to 34, or 36 to 50, wherein the one or more current engine operating parameters on which scaling is varied comprise spark timing and valve timing.

55. A method as in any of claims 31-50, wherein the normalized torque curve is further scaled as a function of at least one of an engine firing history or a cylinder firing history.

56. The method of any one of claims 31-38, or 40-50, further comprising:

filtering the total engine torque curve to identify selected harmonic components of the torque curve; and is

A reaction smoothing torque is determined for application to a powertrain including the engine to reduce NVH during operation of the engine.

57. The method of any one of claims 31 to 40, further comprising:

determining whether the predicted engine torque will exceed a torque limit using the total engine torque curve;

determining a reaction smoothing torque that will prevent the predicted engine torque from exceeding the torque limit when it is determined that the predicted engine torque will exceed the torque limit; and is

The reaction smoothing torque is applied during operation of the engine.

58. The method of claim 57, wherein the reaction smoothing torque is applied by an electric motor or an electric motor/generator.

59. The method of claim 57, wherein the torque limit varies as a function of at least one of engine speed and transmission gear.

60. The method of claim 57, wherein the torque limit corresponds to a maximum of instantaneous torque in the torque curve.

61. The method of claim 60, wherein said maximum value of instantaneous torque varies as a function of engine speed and transmission gear.

62. The method of claim 57, further comprising: the determined reaction smoothing torque is used in the determination of the predicted fuel efficiency for operating the engine at the effective firing fraction associated with the total engine torque curve.

63. The method of claim 57 wherein said estimation of said total engine torque curve and said determination of said reaction smoothing torque are updated for each firing opportunity such that the demand for said reaction smoothing torque and its magnitude are updated for each firing opportunity.

64. A method of estimating a torque curve relative to crank angle for a dynamic ignition level modulation controlled internal combustion engine, the method comprising:

determining, for each cylinder of the engine, a normalized torque curve relative to crank angle for each stroke of a piston reciprocating in the cylinder, wherein the normalized torque curve is based on intake manifold pressure;

scaling the normalized torque curve to determine a cylinder torque; and is

Summing the cylinder torques of all cylinders in the engine to obtain a total engine torque curve.

65. The method of claim 64, wherein said scaling is based, at least in part, on engine speed.

66. The method of claim 65, wherein said scaling is further based in part on at least one of spark timing and valve timing.

67. The method of claim 64, comprising transforming the torque profile with respect to crank angle to a torque profile with respect to time.

68. The method of claim 67 wherein crank angle to time conversion includes the effect of a change in rotational speed of the engine according to the torque curve.

69. The method of claim 64, wherein the scaling is based at least in part on at least one of an engine ignition history and a cylinder ignition history.

70. The method of claim 65, wherein said scaling is based, at least in part, on at least one of spark timing and valve timing.

71. A method as claimed in claim 65 or 66, comprising transforming the torque profile with respect to crank angle to a torque profile with respect to time, wherein crank angle to time transformation comprises the effect of a change in rotational speed of the engine according to the torque profile.

72. A method as claimed in any of claims 65 to 68, wherein the scaling is based at least in part on at least one of an engine ignition history and a cylinder ignition history.

73. The method of any of claims 1-7, or 12-19, or 24-50, or 58-69, or 70 performed in a skip fire operating mode or a firing charge level modulation operating mode during operation of the engine.

74. An engine controller arranged to perform the method of any one of claims 1 to 7, or 11 to 19, or 24 to 50, or 64 to 69.

Images