Improved Engine Management
The invention relates to a system and method providing improved engine management and in particular using real time cylinder pressure data. The aspects discussed herein are an extension of the concepts disclosed in
International patent application no. PCT/GB02/02385 entitled "Improved Engine Management" commonly assigned herewith and incorporated herein by reference.
Known engine management systems (EMS) monitor and control the running of an engine in order to meet certain pre-set or design criteria. Typically these are good driveability coupled with high fuel efficiency and low emissions. One such known system is shown schematically in Fig. 1. An internal combustion engine 10 is controlled by an engine control unit 12 which receives sensor signals from a sensor group designated generally 14 and issues control signals to an actuator group designated generally 16. The engine control unit 12 also receives external inputs from external input block 18 as discussed in more detail below.
Based on the engine performance data derived from the sensor input from the sensor block 14 and any external input from the external input block 18 the engine control unit (ECU) optimises engine performance by varying the relevant performance input variable within the specified criteria.
Typically the sensor block 14 may include sensors including mass air flow sensors, inlet temperature sensors, knock detection sensors, cam sensor, air/fuel ratio (AFR) or lambda (λ) sensors, and engine speed sensors. The external input block 18 typically includes throttle or accelerator sensors, ambient pressure sensors and engine coolant temperature sensors. In a spark- ignition engine the actuator block 16 typically comprises a fuel injector control and spark plug operation control. In a compression ignition engine the actuator
block typically comprises a fuel injector.
As a result, for example, in a spark-ignition engine, under variable load conditions induced by the throttle under driver control, the sensors and actuators enable effective control of the amount of fuel entering the combustion chamber in order to achieve stoichiometric AFR, and of the timing of combustion itself.
Known engine management systems suffer from various problems. EMS technology remains restricted to parameter based systems. These systems incorporate various look-up tables which provide output values based on control parameters such as set-points, boundaries, control gains, and dynamic compensation factors, over a range of ambient and engine operating conditions. For example in spark ignition engines spark timing is conventionally mapped against engine speed and engine load and requires compensation for cold starting. In compression ignition engines fuel injection timing is mapped in a similar manner. As well as introducing a high data storage demand, therefore, known systems require significant initial calibration. This calibration is typically carried out on a test bed where an engine is driven through the full range of conditions mapped into the look-up tables. As a result the systems do not compensate for factors such as variations between engine builds let alone individual cylinders, and in-service wear. Accordingly the look-up tables may be inaccurate ab initio for an individual engine, and will become less accurate still with time.
In one aspect it is known how to control engine performance based on a consideration of conditions in the cylinders and its associated actuators, in particular combustion events, either by monitoring relevant conditions or relying on calibrated or mapped values. This approach suffers from the problems identified above. In addition in many cases extrapolations are being taken from indirect measurements.
Examples of such cylinder and actuator conditions include maximum cylinder pressure (Pmax), fuel injection timing, combustion noise and conditions in the cylinder during cranking (i.e. starting of the engine before fuel is injected).
For example Pmax is generally specified for an engine build and the performance input variables are controlled to ensure that the cylinder pressure does not exceed Pmax which can otherwise cause engine damage. Those variables include, in a compression ignition engine, fuel injection timing. As a result, fixed upper thresholds are provided so that the performance for a given engine and indeed cylinder will in some cases be non-optimal as, in practice, the Pmax achieved using the pre-determined performance input variables may be less than the absolute Pmax allowed for the engine, for example as a result of engine build or individual cylinder differences and engine instantaneous condition. In addition, engineering margins have to be applied based on production tolerances thus compromising on engine performance.
Furthermore, in known systems controlling injection timing a (generally) constant (in time terms) delay is generally present between actuation of the injector and actual combustion. Again, however, this delay can vary significantly both as a result of the state of the injector and of the cylinder, and will also vary as a result of production tolerances from engine to engine, variations between cylinders and wear over time as a result of which, once again, non-optimal control is achieved.
Known systems have also been developed to minimise combustion noise. For example the Multec (TM) DCR 1600 diesel injection system manufactured by Delphi Corp. of Troy, Michigan, USA uses an accelerometer to monitor
combustion noise and, dependent on the detected values, controls performance input variables based on pre-calibrated values.
A further problem with known engines relates to the generation of white smoke and/or unstable conditions during cranking i.e. start up of an engine, for example a compression ignition engine, particularly under cold conditions. Co- pending International patent application no. PCT/GB02/02385 relates to an arrangement in which engine warm up is controlled by varying the combustion timing within combustion stability limits.
In another aspect it is known to control engine sub-systems to improve engine performance but once again this requires additional sensors.
Such sub-systems include valves controlling opening and closing of the various ports to the engine cylinders, exhaust gas recirculation sub-systems, compressors and turbines and after-treatment systems.
In relation to the cylinder valves, for example, current arrangements require a cam sensor or valve sensor to sense camshaft position or the time when a valve opens and closes which information is then used by the EMS. In addition valve lift sensors are provided to sense valve impact and match this to a calibrated value which is subject to production tolerances and differences between individual cylinders and which does not take into account wear.
US4531499 relates to a control device for exhaust gas re-circulation (EGR) in which a pressure signal is differentiated and the values obtained matched against calibrated values to identify an exhaust gas percentage in the cylinder. However this is based on limited data derived from the signal.
In known compressor or turbine systems respectively up and downstream of the engine itself, in order to obtain the relevant pressure ratios various sensors are currently provided in relation to each component and used in conjunction with calibrated values to obtain the existing pressure ratio for control purposes.
Known after-treatment systems such as NOx traps in the engine exhaust stream require regeneration periodically to maintain their trapping efficiency. The regeneration regime is dependent on exhaust temperature and the air to fuel ratio (AFR) in the exhaust stream. In known systems multiple sensors are generally required to monitor these values in order that regeneration can be triggered at an appropriate time.
The invention is set out in the claims.
Embodiments of the invention will now be described, by way of example, with reference to the drawings, of which:
Fig. 1 is a block diagram representing a prior art EMS;
Fig. 2 is a schematic diagram representing an EMS according to the present invention;
Fig. 3 is a schematic view of a single cylinder in cross section according to the present invention;
Fig. 4 is a trace of pressure against crank angle for a cylinder cycle of a four stroke engine; Fig. 5 shows a more detailed trace of cylinder pressure (bar) against crank angle (0ca);
Fig. 6 shows a plot of heat release against crank angle (0ca);
Fig. 7 shows additional detail relating to the compression and expansion strokes of the four stroke cycle shown in Fig. 4;
Fig. 8 shows pressure and temperature variation in a cylinder over multiple cycles;
Fig. 9 shows a portion of the cylinder pressure superimposed on a graph of inlet and exhaust valve profiles Fig. 10 shows schematically a valve and valve seat on an engine cylinder;
Fig. 1 la is a block diagram showing an engine cylinder and upstream component;
Fig. 1 lb is a block diagram showing an engine cylinder and downstream component; Fig. 12 is a block diagram showing a cylinder and downstream after-treatment system;
Fig. 13 is a block diagram showing control modules in an engine according to the present invention;
Fig. 14 is a block diagram showing the components of an EMS according to the present invention;
Fig. 15 is a block diagram showing individual cylinder control in an EMS according to the present invention;
Fig. 16 shows the pressure cycle for the selected cylinder in a six-cylinder engine; Fig. 17 shows a quasi propagatory model of a gas path;
Fig. 18 a solution according to the quasi propagatory model; and
Fig. 19 is a further solution according to the quasi propagatory model.
The following discussion of an embodiment of the invention relates to its implementation in relation to a four stroke combustion ignition engine comprising a diesel engine. However it will be appreciated that the invention can be applied equally to other stroke cycles and types of internal combustion engines including spark-ignition engines, with appropriate changes to the model parameters. Those changes will be apparent to the skilled person and
only the best mode presently contemplated is described in detail below. Like reference numerals refer to like parts throughout the description.
Fig. 2 is a schematic view showing the relevant parts of an engine management system according to the present invention in conjunction with a six cylinder engine. An engine control unit is designated generally 20 and controls an engine designated generally 22. The engine includes six cylinders designated generally 24. Each cylinder includes a pressure sensor 26 which connects to the ECU via a line 28. In addition the ECU provides electronic control to each of the cylinder points, e.g. inlet and exhaust valves, as well as fuel injectors, (not shown) via control lines 30 allowing variable valve actuation (WA). This may be a camless engine or alternatively the port valves can be actuated by cams as is well known. The ECU 20 can also receive additional controls and actuator inputs 32 as discussed in more detail below. The engine management system monitors the pressure in each cylinder through each complete engine cycle, namely 720° rotation of the crankshaft in a four- stroke engine. Based on this data the injection timing for each cylinder 24 is varied by varying the timing of each injector via control lines 30.
In Fig. 3 there is shown schematically a more detailed view of a single cylinder 24 of the engine. The in-cylinder pressure sensor 26 comprises a piezoresistive combustion pressure sensor with a chip made of silicon on insulator (SOI) available from Kistlerlnstrumente AG, Winterthur, Switzerland as transducer Z17619, cable 4767A2/5/10 and amplifier Z18150. It will be appreciated that any appropriate in-cylinder pressure sensor can be used, however. For example the sensor can be of the type described in co-pending application number DE 100 34 390.2. The pressure sensor 26 takes continuous readings through the four strokes of the piston 40. The readings are crank-synchronous and triggered by crank teeth 42a of the crank 42, detected by a crank tooth sensor 44 which sends an appropriate signal via line 46 to the ECU 20. In the preferred embodiment readings are taken every 2° of
crankshaft rotation although any desired resolution can be adopted, the limiting factors being processing power and crank angle sensing resolution. For each cylinder the readings are taken across a cycle window of width 720°. As discussed in more detail below with reference to Fig. 16, the window is selected to run from a point substantially before engine top dead centre (TDC) for each cylinder.
The data obtained from the in-cylinder sensor 26 is processed as discussed in more detail below and a high resolution plot of pressure versus crank angle (which can be simply converted to time if the engine speed is known) is obtained for each cylinder and each cycle. From this information, monitoring and control of engine performance is greatly enhanced.
In overview, in a first aspect the maximum cylinder pressure Pmax can be consistently raised to the engine build threshold, maximising performance.
In a second aspect it is possible to derive the point at which start of combustion commences from which the instantaneous delay for each cylinder and each cycle can be calculated allowing injected timing to be tuned.
At the same time in a third aspect, it is possible to derive information concerning the rate of combustion, a value which has not previously been explicitly taken into account. This can then be controlled to a given target for improved efficiency or stability and in particular the combustion noise and exhaust emissions, which are both related to the rate of combustion, can be controlled. Identification of the start of combustion allows the relevant portions of the pressure data to be processed.
In a fourth aspect, in relation to start control of the engine, the pressure data can be processed to provide information on cylinder conditions during cranking such that supply, timing and quantity of fuel can be optimised for combustion stability and emissions targets.
Naive events such as valve opening and closure can be derived in a fifth aspect from the pressure. This information can furthermore be used as feedback for variable valve timing (WT).
In a sixth aspect, in camless engines whereby valves are actuated electronically, further information can be derived in relation to the valve motion. The cylinder pressure signal enables a force-balance calculation to be made on each valve. This provides a means of controlling valve impact velocities thus minimising mechanical wear.
In a seventh aspect, yet further information concerning the amount of exhaust gas in the cylinder can be derived from the pressure information allowing optimisation of the relevant performance input variables to obtain a desired EGR operation.
In an eighth aspect, compressor or turbine pressure ratio control can be enhanced by identifying the cylinder pressure at relevant valve events. For example the intake manifold pressure can be obtained from the cylinder pressure when the intake valve is open and the exhaust manifold pressure can be derived in a similar manner. From this, together with ambient and turbine exit pressures, pressure ratios can be based on an instantaneous cylinder - by - cylinder pressure value.
Yet further, in a ninth aspect, temperature and AFR data can be derived from the pressure information and other sensors allowing optimisation of these values for after-treatment regeneration systems again allowing improved control.
The pressure data derived is shown in Fig. 4 which shows the cylinder pressure variation against crank angle for one full cycle between -360° and +360°. As is well known the engine cycle is divided into four regions, induction from -360° to -180°, compression from -180° to 0° (TDC), expansion from 0° to +180° and exhaust from +180° to 360°, defining a full 720° cycle. Theoretically, for instantaneous combustion occurring over an infinitely small period of time the optimum point for combustion is at 0° TDC, but in practice injection timing can vary by several degrees from TDC. The cylinder pressure curve C is derived from the filtered pressure signal Pf obtained by passing the raw digital pressure signal values through a zero phase-shift low pass filter to attenuate signal noise.
Medium and long-term sensor drift is then compensated for by applying the polytropic gas principle, applying the law:
(Pf + Po)Vn - constant (1)
where: P0 = offset pressure
Pf = filtered pressure signal V = cylinder volume n = polytropic index
To derive the offset pressure P0 three values of Pg;k;ι are selected together with the corresponding cylinder volumes Yj at each of those points (which can of course be derived directly from the crank angle value). It can be
shown that the equation can be solved neatly if the values are chosen at points for which the respective values of volume satisfy the equation:
The subscripts j, k, 1, represent the respective data points corresponding to CA positions θj; θk and θj. It can be shown that, in this case, equation (1) yields the relationship
As a result P0 can be obtained effectively estimating n at every cycle for each cylinder, compensating for variation of n with engine temperature, exhaust re-circulation and so forth.
The pressure curve obtained is then processed to provide additional engine performance data allowing enhanced control.
In a first aspect the pressure curve can be processed to obtain Pmax as can be understood in more detail with reference to Fig. 4. It will be seen that the value is easily derivable simply by selecting the maximum value of pressure. The relevant point can be identified in any appropriate way, for example by differentiating the curve and identifying the crossover point between positive and negative gradient. Depending on the resolution of the measured data, the maximum can be interpolated between adjacent data points for example by using a polynomial curve fitting technique as will be well known to the skilled reader.
As the absolute Pmax limit is typically known for a given engine build, once the instantaneous Pmax for a given cylinder is known from one cycle it can be assessed whether this is at the threshold. If not the relevant performance input variables can be adjusted to enhance performance whilst maintaining Pmax within the threshold. This can be done, for example, by adjusting any one or more of the fuel injection quantity, fuel injection timing and/or variable compression ratio actuator. The values adopted for the variables can be based on a model or on a mapping, the nature of which will be apparent to the skilled person and which is not described in detail here. As a result a simple feedback control system can be adopted whereby the performance input variables are adjusted to converge the instantaneous measured Pmax on a target Pmax. Accordingly engine performance is maximised within the pre-set limits providing efficient use of the engine structure and compensating dynamically for in-service wear whilst optimising performance.
In a second aspect of the invention the start of combustion is estimated from the pressure data obtained. Referring to Fig. 5 a more detailed representation of the pressure curve is designated 50. In a compression ignition system there is a delay between start of injection 52 and combustion itself. This is formed of two factors, the injection delay caused by the time lag in the injection actuator itself and the ignition delay between injection of the fuel and combustion. The combined delay can give rise to an error of several degrees crank shaft angle. Accordingly by identifying the start of combustion it is possible to extrapolate back to ensure that the actuator is triggered an appropriate time before the desired start of combustion time.
Start of combustion can either be identified by examining the heat release curve derivable from the pressure curve, or by identifying the "kick" in the pressure curve which takes place when combustion starts.
The manner in which heat release is processed can be understood with reference to Fig. 6 in conjunction with Fig. 5, as shown in Rassweiler, G.M. and Withrow, L., "Motion Pictures of Engines Flames Correlated with Pressure Cards", SAE Journal (Trans), Vol.42, pp. 185-204, May 1938 (Reprinted SAE Paper 800131, 1980) and Colin R. Ferguson "Internal Combustion Engines", Wiley the heat release (HR) is related to the difference between the pressure curve 50 and the "motoring curve" 54 shown in dotted lines in Fig. 5. At a given data point i the proportion of heat released, HRj can be approximated by:
HRj = -±L _ i (4)
where Pj is the firing cylinder pressure and Pm is the motoring pressure.
The motoring curve is the pressure curve that would be obtained if combustion did not take place in the cylinder, representing purely the varying pressure resulting from the compression stroke in the cylinder. Equation (4) gives rise to the curve 56 shown in Fig. 6 representative of the heat release against crank angle CA. It will be seen that at the start of injection (point 52) there is no difference which would be expected as combustion does not immediately take place as explained above. After the injection delay 58 there is an initial negative heat release value at region 60 which results from evaporation of the fuel in the cylinder. This value crosses into a positive region 62 when combustion takes place. Typically the crossover point 64 is taken as marking the start of combustion and the interval represented by the negative portion 60 of the heat release curve represents the ignition delay 65. Alternatively the point where the heat release curve gradient changes sign within the region 60
may be taken as the start of combustion. As a result the start of combustion point 64 is easily derived from processing of the pressure curve.
The motoring curve 54 can be derived in various ways known to the skilled person. For example it can be pre-calibrated, or obtained by "skip firing" in which at certain intervals fuel is not injected into the cylinder for one cycle (eg during cranking or overrun) and the resultant pressure curve obtained. Alternatively, the approach set out in International patent application no. PCT/GB02/02385 is adopted whereby a pre-injection portion of the pressure curve is extrapolated to obtain the motoring curve. This can be better understood with reference to Fig. 7 in which the pressure curve is designated by reference letter C. In particular the portion A of the curve C from which the motoring pressure curve E is derived extends between the point θxvc at which the intake valve to the cylinder closes (intake value closure) and the value of θinj at which injection is triggered which in a preferred embodiment can vary from cycle to cycle to allow better estimation of the motoring pressure curve under heavily retarded conditions. Alternatively, however θinj can be limited to the most advanced possible value. Accordingly data values Pj and Nj are taken across the range iivc < i < ilnJ . By taking logarithms on both sides of the polytropic gas equation, a least-squares curve fitting technique is applied over the range of pressures during the compression stroke, to yield the constants n and K. This enables the rest of the motoring pressure Pm curve E to be constructed from the values of each measurement point i:
Pπri = Nfn K (5)
Alternatively the system could use iterative techniques to estimate P0, n and K together but the preferred embodiment is the curve fitting technique described above as it is found to reduce the processing time necessary.
In the alternative approach start of combustion is identified by observing the kick designated generally 66 in the pressure curve of Fig. 5. In the preferred dP approach this is achieved by taking — , and identifying where it starts to dθ increase after its minimum at TDC (ie. Point of inflection). This point represents the start of combustion. Once the start of combustion point 64 has been identified in whatever manner then the full delay can be worked out either in the CA (θ) domain or in the time domain simply by calculating when the injection command was issued to the injector. As a result when a known start of combustion point is identified, it is simply necessary to issue the injection command to the injector the calculated delay period before that. This can be done either by appropriate feedback control of start of combustion to a target start of combustion point, or by replacing a pre-set delay calibrated for each injector with the detected delay. Such an arrangement, for example, can be used in an engine fuel efficiency mode in which start of combustion is identified from a preceding pressure curve and tracked to a target value in an appropriate manner in which case the required injector trigger point can be extrapolated back.
In a third aspect further information is derived from the traces shown, for example in Figs. 5 and 6, to assess the rate of combustion and the related feature of combustion noise. From Fig. 5 it will be seen that the rate of pressure rise, which can be related directly to the rate of combustion, is given dP by — at point 64, i.e. the gradient can be obtained by simple differentiation dθ
once start of combustion has been identified at 64. Alternatively, referring to Fig. 6, an equivalent measure is to obtain the rate of heat release at the start of fJfJD combustion given by . This is obtained simply by differentiating dθ
Equation (4) with respect to crank angle θ giving:
In either case, once the rate of combustion is identified this can be compared to a target value which can, for example, be calibrated on the test bed in the manner that will be apparent to the skilled person. The measured combustion rate can then be tracked to the target value by adjusting the appropriate performance input variables accordingly. Those variables may include injection timing, exhaust gas recirculation proportion or variable compression ratio as appropriate, again relying on mappings between the selected performance input variable values and a desired rate of combustion.
One particular aspect of this approach is control of the combustion noise - a significant environmental factor that may also be indicative of wear on the cylinder. Once start of combustion 64 has been identified then the combustion noise is related to the rate of combustion such that variation of the combustion rate as identified above can be targeted to maintaining the combustion noise below a predetermined level again, for example, relying on appropriate mappings. Yet further information can be obtained by analysing the pressure signal at the start of combustion using band-pass filtering or Fast Fourier Transforms (FFT) which will provide yet further information concerning the combustion noise allowing actuator control to be tailored yet more effectively.
Various advantages are associated with the second and third aspects. The identification, measurements and control of these various combustion events, namely start of combustion, rate of combustion and combustion noise give rise to numerous benefits. Compensation is provided for fuel property variation and manufacturing tolerances, rather than relying on pre-set values based on an assumption that all engines and all cylinders are alike. This in turn can give rise to improvements in emissions and fuel economy and allows the system to compensate for changes arising from in-service wear. The approach can be particularly beneficial in certain engine types for example in the case of advanced combustion systems such as low temperature homogenous charge, compression ignition (HCCI), or Late Compression Ignition (LCI) combustion which can be unpredictable such that greater certainty as to the point of start of combustion and the rate of combustion is required to control emissions, stability and other output variables. Similarly where it is required to regenerate after-treatment systems such as NOx traps the point of combustion can be significant in determining the cylinder temperature /AFR which in turn can affect the regeneration regime such that greater control over the start of combustion is desirable. In relation to the noise control aspect it will be seen that improved noise vibration and harshness (NVH) characteristics can be provided particularly during cold-starts and that there is potential for a switch to a lower combustion noise regime (for example by dropping to a lower predetermined combustion rate by corresponding adjustment of the performance input variables) to allow quiet mode in circumstances such as a vehicle passing through urban areas.
In a fourth aspect illustrated by Fig. 8 the pressure curve 120 is processed to obtain additional control of the engine starting process during which the engine
is cranked without fuel injection. In particular in a compression ignition system such as a diesel engine, it is desirable to obtain an appropriate fuelling strategy, that reflects favourable temperature conditions in the cylinders before fuel injection takes place to avoid unstable starts or white smoke generation (i.e. where unburnt fuel vapour comes out of the engine, which occurs during incomplete or no combustion). The cylinder temperature can be derived from the pressure curve 120 in the cylinder. This is made possible by applying the ideal gas equation, ie:
PV.
T, = -^ (7) mR
where m is the trapped mass of air and inert gas and R is the gas constant of the mixture. This trapped cylinder mass can be obtained based on a mean- alue or a quasi-propagatory approach.
In the mean- value approach, the trapped mass can be obatined from:
Λ int
where ηv is the volumetric efficiency at cranking, Vcyι is the cylinder swept volume, Pjnt is the intake manifold pressure and Tint the intake manifold temperature. The intake manifold pressure can be obtained from the cylinder pressure at intake valve opening as discussed in more detail below. The method for obtaining the value of ηv under applicable conditions, and correcting for intake temperature and coolant temperature is well known and discussed, for example in Taylor, C.F., "The Internal Combustion Engine in Theory and Practice, Volume 1", 2nd Ed., MIT Press. Finally, the gas constant
R, of the air and inert gas can be easily derived from assumed or measured air humidity as discussed in Taylor (ibid).
In the quasi-propagatory approach, the airflow through the intake, cylinders and exhaust is modelled as a complete process. In-cycle events resulting in pressure waves along the various ducts are modelled. The theory behind this is described by R. Cipollone and A. Sciarretta in 'A New Modelling for the Air and Gas Dynamics in ICE Manifolds Oriented to Air-Fuel Ratio Control', ASME 1999, Paper 99-ICE-170.
As a result the correct conditions for combustion can be identified. In particular those conditions may be mean cylinder temperature above a predefined threshold. As can be seen in Fig. 8, in a preferred approach the cylinder temperature 122 at TDC can be extrapolated along cranking cycles 128, 130 to obtain a cylinder temperature curve 132. When this curve crosses the threshold temperature 124, i.e. at 134, it can be assumed that injection can take place in the next firing cycle 136 (it can be seen in this cycle that the cylinder pressure 120, shown in dotted lines in the preceding cycles, departs from the motoring pressure on firing). The above approach enables more reliable start-up conditions thus avoiding the problems of white smoke (ie. Unburnt fuel in the exhaust system). This is particularly relevant to low compression ratio engines.
Once the correct conditions have been identified fuel can be supplied as indicated at 126. Yet further, dependent upon the sensed pressure conditions, once fuel supply has started, the appropriate performance input variables such
as timing and quantity of fuel can be controlled to provide variable compression ratio control input (such as by WA or other appropriate means) to ensure effective combustion during start-up and stability of idling.
It has been proposed in PCT/GB02/02385 to control spark timing (or analogously, in a compression ignition system, injection timing) to obtain combustion stability at start-up by looking at the variance of the work done in the cylinder which can be taken from the IMEP derived from the pressure curve, where IMEP is a value known to the skilled reader representing the work done during the stroke. Similarly the value of Pmax derived as discussed above can be used as a measure of combustion stability such that a consistent level of Pmax is maintained from cycle to cycle by an appropriate control strategy based on the variance of Pmax or some other measure of its cycle-to-cycle variation.
The relevant performance input variables include variable valve timing or other appropriate means of varying the compression ratio, fuel timing, fuel quantity, cylinder charge or exhaust gas ratio which can be varied according to an appropriate model or calibration map to obtain a desired level of combustion stability on start-up.
As a result the arrangement provides improved stability in particular in relation to low compression ratio engines while eliminating cold start white smoke problems. Because it is based on dynamic pressure measurements, the invention also compensates for in-service wear.
In a fifth aspect the pressure curve is processed to obtain information concerning valve events such as valve opening or closing events. Information of this type is useful, for controlling valve opening and closing events on a WT engine.
Opening and closing of the valves can be seen by the valve lift for the exhaust valve 140 and inlet valve 141 of Fig. 9. This will give rise to pressure variations in the portion of the pressure curve 142 that as can be seen in Fig 9. This figure indicates the effect of the exhaust valve closure 146 on the pressure trace. Careful signal filtering is required to distinguish this from system noise. Note that the point 144 indicating 'Exhaust Valve Closed' is defined based on closure at the beginning of the ramp-down. Detection of the events can be in any appropriate manner, for example pattern recognition, identification of predetermined frequencies using an approach such as FFT or bandpass filtering and so forth as will be well known to the skilled. In addition other events such as fuel injection and start of combustion can be filtered out, once they are identified as discussed above in relation to Figs. 5 and 6, so as not to be confused with the relevant valve events.
Once the valve events have been sensed then any delays inherent between triggering the valve actuators and the actual valve event can be identified and factored into subsequent triggering commands such that more accurate control of valve events is obtained. This can be based on a pre-determined control strategy in which fixed delays are replaced by the measured delays or a feedback loop. The strategy itself will be apparent to the skilled person and can comprise a modelled or pre-calibrated strategy dependent on other measured engine parameters as appropriate.
As a result the need for a cam sensoris removed. More accurate control of valve events is provided allowing tracking of valve operation on a cylinder-by- cylinder and cycle-by-cycle event such that differences between individual cylinders, differing cylinder conditions during a drive cycle and in-service wear are all compensated.
According to a sixth aspect of the invention the pressure curve can be processed to provide improved valve lift control in a camless engine. In such engines independent valve actuators are triggered to open and close valves on a cylinder allowing WA and hence enhanced engine control. The valve force on the valve seat must of course be sufficient to close the valve fully, but the greater the force and in particular the greater the impact, the more damage can be caused to the cylinder head and valve seat which can reduce engine lifetime.
Fig. 10 shows such an arrangement in which the cylinder is designated generally 70 and includes an aperture 72 defining a valve seat. The valve itself includes an actuator (not shown) which controls the valve velocity and a valve head 74 which closes against the valve seat 72. In addition a valve lift sensor is shown at 76 which senses valve displacement. In order to balance the valve impact, the force of the valve Fvaιve must equal the force created by the cylinder pressure PCYL across the valve seat area Avaιve, as represented by the following equation
Fvaive - (PcYL - PpθRτ)- Avaιve (9)
where PPO T is the pressure behind the valve obtained from manifold pressure by, for example, an appropriate sensor.
PcYL can be derived by taking the appropriate reading from the pressure curve. This can be done either by identifying the crank angle at which valve closure is triggered and then reading this value against the pressure curve shown in for example Fig. 4, or by identifying the valve event as discussed above with reference to Fig. 9 and taking the valve pressure at that moment. Alternatively still, when valve closure is required, the instantaneous pressure value can be measured directly.
As a result Fvaιve for valve closure can be identified from the specific pressure conditions in the cylinder and equation (5). The corresponding velocity is obtained by integrating the valve head acceleration avaιve which can be derived from Fvaive using F=ma and used to control the actuator. As a cross-check, avaive can be twice integrated to provide the corresponding displacement and the displacement valve compared against the sensed valve lift value. As a result a further mapping can be built up for example in a neural network to refine the model or mapping. The valve lift can be governed by a feedback loop with the position sensor to obtain an optimised balanced F^^ by adjusting the valve lift as will be apparent to the skilled reader.
Because the invention uses an instantaneous pressure value rather than, for example, a mapped value, a more accurate measure of the valve forces is attained as a result of which damaging impact on the cylinder valve seat can be
minimised. In addition a useful cross check is provided to refine the valve lift distance required for a given impact force.
In a seventh aspect EGR distribution is improved relying on data derived from the pressure curve. In particular the inert gas fraction distribution, from cylinder to cylinder can be derived from the pressure trace, if desired in conjunction with a sensed air mass intake or alternatively based on a predetermined model. The inert gas fraction is the mass-fraction of gas in the cylinder that will not burn for example CO, C02, NOx, H20. This can be derived based on the principles described by R. Cipollone and A. Sciarretta (see above reference), or any other appropriate technique, but applied to more than one species in the ducts and cylinders. Alternatively the pressure trace can be used to provide a correlation with modelled values and if necessary to correct the model.
For example an observer based strategy can be applied to estimate the air mass and inert gas fractions within each cylinder. This consists of an in-cycle model of the gas path (air/exhaust path) and a correction algorithm that uses the cylinder pressure variation function as well as other sensor data to correct the corresponding outputs from the model by updating the model.
The in-cycle model of the engine gas path can be based on the Quasi- Propagatory Model (QPM) approach described by R. Cipollone and A. Sciarretta. In-cycle models, such as the Ricardo WAVE engine model available from Ricardo Consulting Engineers Ltd. of Shoreham, United Kingdom, are generally one-dimensional models. These compute the mass flows, pressure and temperatures in the manifolds and connecting pipes, based on the method of characteristics, typically within 3% accuracy. However, due
to the high computational overhead, this cannot be applied to real-time application such as in an EMS.
The QPM approach relies on segmenting the various volumes along the gas path into a series of inter-connecting capacitances and ducts whereby each is considered as a lumped-parameter (ie. zero dimensional model) (see Figure 17). Air mass and inert gas fractions are computed for each capacitance including the cylinders. Furthermore, the method of characteristics is substituted by a simple linear relationship between pressure and gas velocity (see Figures 18 and 19). The speed of execution is governed by how finely the manifolds and pipes are segmented. A trade-off has therefore to be achieved between model accuracy and execution time. Validation tests that shown that the QPM model is able to ran in real-time with sufficiently accurate results to be used in an EMS.
The QPM model described above can be extended to incorporate a relationship between inert gas fraction and the rate of heat release function. The latter can be a Wiebe model as will be well known to the skilled reader and as described in Colin R. Ferguson "Internal Combustion Engines", Wiley, that is adapted for the particular combustion process being considered. This relationship may be an empirical mathematical formula or a look-up table derived from engine test results. Since cylinder pressure can be obtained from the heat-release curve, a direct relationship can be established between the inert gas fraction and the estimated cylinder pressures.
The model correction algorithm utilizes the sensed cylinder pressures in order to correct on the model estimates. In a real engine application there will be errors between the model and the real engine because of model assumptions, engine production build variations, in-service wear, etc. This means that the
gas path model will therefore have to be corrected using sensor data. By comparing the sensed cylinder pressures with the estimated cylinder pressures, the Wiebe model, as well as other aspects of the model, such as heat loss through the cylinder walls, can be updated. This, in turn, enables the inert gas fractions to be updated. Another correction can be applied to the estimated inducted air temperature using the inlet temperature sensor. Inert gas fractions can therefore be estimated for each cylinder and continually updated. Corresponding corrections on the air masses can also be made.
Once the overall inert gas fraction, and inert gas fraction distribution is determined then this can be compared against a pre-determined mapping and performance input variables such as EGR valve setting, fuel injection timing, variable valve actuation, boost pressure actuation and so forth used to optimise the smoke/ AFR on a cylinder-by-cylinder and stroke-by-stroke basis. In the case of fuel quantity, the fuelling can be adjusted within the smoke limit for each cylinder. This ensures that the overall fuelling, and hence performance, is not restricted by the cylinder with most inert gas and therefore lowest fresh air quantity.
The inert gas quantity in each cylinder can be balanced for example by using early intake valve opening or late exhaust valve closing (or any other suitable combination of valve timings).that are independent for each cylinder. However, this requires independent cylinder WA such as can be offered by camless engines.
As a result of this approach there are improvements in NOx emissions/smoke in particular for compression ignition engines such as diesel engines, additionally providing an improved driveability/smoke trade-off. These improvements are particularly relevant for HCCI engines where stability is an issue. Furthermore because measurements are constantly being updated or correlated, in-service wear is compensated.
In an eighth aspect the pressure curve is used to provide additional pressure ratio information in systems where a compressor or turbine is used. Fig. 11 A shows a compressor system and Fig. 11B shows a turbine system. In the compressor system of Fig. 11 A a cylinder 80 includes an inlet manifold 82 fed by a compressor 84. Where the compressor is driven by a turbine forming, for example, a variable geometry turbo charger (VGT), as shown and in Fig. 1 IB a cylinder 80 has an exhaust manifold 86 and, downstream thereof, a turbine 88. In the case of both a compressor 84 and a turbine 88 the pressure ratio (ratio between the pressure at compressor inlet or turbine outlet and the cylinder inlet or exhaust manifold respectively) is used as a feedback control variable to govern operation of the compressor/turbine.
In the case of a compressor, the process of present invention measures the cylinder inlet manifold 82 pressure simply by taking the relevant pressure reading from the cylinder during the period when the inlet valves are open. This can be derived from the pressure curve at a crank angle in the window where the inlet valve is known to be open, or, in WA systems can be based on sensing of the relevant valve event and determining the pressure as discussed in more detail above. The pressure value obtained can then be fed, together with compressor inlet pressure, into the pressure ratio calculation
governing control of the compressor which is then controlled according to any appropriate manner as will be apparent to the skilled reader. In the case where additional components such as an air filter or after-cooler introduce further pressure drops between the air intake and the intake manifold 82 then the contributions of these can be factored in using a mapping or in any other appropriate manner.
In relation to the turbine system shown in Fig. 11B a similar approach is adopted except that the exhaust manifold pressure is derived from the pressure curve at an appropriate instant and compared with the turbine exit pressure (measured by an appropriate sensor, for example) to obtain a pressure ratio. The effect of any intermediate component 86 can again be compensated by appropriate calibrated values.
Once the ratios are obtained then operation of the compressor or turbine can be controlled to achieve an optimum pressure as will be well known to the skilled reader, but based on instantaneous, accurate cylinder pressure measurements rather than calibrated values as a result of which performance of the components can be optimised. As a result improved VGT or compressor control is obtained leading to optimisation of fuel economy. Under every operating condition the operating point of the compressor or turbine within its allowable operating map can be known, since the coordinates of such an operating point on the operating map are mass flow (known by estimating the inducted air mass using the principles described by R. Cipollone and A. Sciarretta (see above reference) or any other appropriate technique), and pressure ratio. On board diagnostic (OBD) readings are similarly improved as a
result of the direct rather than indirect measurement and engine and cylinder variations, as well as in-service wear, are compensated.
In a ninth aspect the pressure curve is processed to improve the after-treatment process. Referring to Fig. 12 a typical after-treatment arrangement is shown including an engine cylinder 80 which exhausts to a regeneratable aftertreatment device such as a lean NOx trap 90. In the embodiments shown the turbine 88 of a VGT is provided in the exhaust path between the cylinder 80 and the trap 90. The exhaust passes through the NOx trap which reduces NOx emissions but requires regeneration at certain intervals. The regeneration takes place under appropriate conditions of air-fuel ratio and temperature of the exhaust.
Accordingly the pressure curve is used to derive the exhaust gas temperature. This can be derived by re-applying Equation (7) at exhaust valve opening (EVO):
where this time, me, is the mass of the trapped exhaust gas, made up of the charge mass and injected fuel mass, and -R-g is the corresponding gas constant.
As indicated here, the point at which the temperature is derived is when the exhaust valve is open which can be identified either by assessing the relevant value when the valve is triggered, or by detecting a valve open event as discussed in more detail below and deriving the temperature at that point. Similarly the AFR can be derived from the pressure trace at an appropriate instant. This is made possible by estimating the
inducted air mass using the principles described by R. Cipollone and A. Sciarretta (see above reference) or any other appropriate technique. Where there are intermediate components between the cylinder 80 and the trap 90 such as the turbine 88, any compensation for the temperature values can be built in for example by taking appropriate values from a map, or by using additional sensors or an appropriate model as will be apparent to the skilled reader.
The data obtained can be used in various ways. In one embodiment the data can be used to identify more accurately when appropriate conditions have been reached for triggering regeneration. Alternatively appropriate performance input variables such as injection timing can be adjusted so as to improve the conditions for regeneration based on mappings or a model as will be apparent to the skilled reader. As individual cylinders can be independently controlled, the system can be yet further enhanced by running certain cylinders to a common exhaust manifold so as to optimise regeneration and others to optimise other aspects of vehicle performance such as fuel efficiency. In the case of a "V" or other engine with multiple aftertreatment devices this can be achieved by running one group of cylinders under optimum conditions for regeneration and the other under optimum conditions for fuel efficiency or other output targets. Yet further, performance input variables to a given cylinder can be controlled so as to obtain a trade-off between regeneration optimisation and fuel efficiency or other output targets.
An example of one of these output targets is related to the smooth running of the engine, in particular, to avoid undesirable torque fluctuations when regenerating. Since indicated torque can be obtained directly from cylinder pressures, in the form of indicated mean effective pressure (IMEP) (as would
be apparent to the skilled person and as described in PCT/GB02/02385), more precise torque control during regeneration can therefore be achieved.
Any appropriate control mechanism and strategy can be adopted to implement the various enhancements discussed above, as will be apparent to the skilled person. One appropriate system is discussed in overview with reference to Fig. 13 and includes a controller 100, one or more actuators 102, cylinder 104, processor 106 and a module 108 supporting a model or map correlating predetermined values. The measured pressure from the cylinder together with the corresponding crank angle θca are fed to the processor 106 which derives a pressure curve and/or pressure value and from those performance output variables such as temperature, heat release, AFR and so forth as discussed above. These parameters are output to a controller 100 together with other necessary sensor inputs from a sensor or sensors 110.
Where necessary the controller takes these inputs and feeds them to the model or mapping module 108 in_ order to obtain the desired adjusted performance input variables. The module 108 can be calibrated during engine prototyping on the test-bed, for example, to provide mappings between performance output values such as Pmax, combustion noise and so forth and desired performance input variables such as fuel injection timing and quantity and WT.
The adjusted performance input variables are then fed to the relevant actuators
102 which control the conditions in cylinder 104. As a result a feedback loop is provided in which the measured pressure value provides a performance output value which is either controlled to track a target performance output
value, or which can be used as a check or correlation against values obtained from the module 108.
It will be appreciated that, where appropriate, instead of closed loop control the pressure value can simply be fed through the processor to obtain a calibrated performance input value at pre-determined intervals or otherwise. It will be further appreciated that the module 108 can be formed at various levels of sophistication, for example providing multiple dimensional mapping tables allowing trade-offs between a plurality of desired performance output values. For example in relation to the embodiment discussed above with reference to Fig. 12, where the exhaust temperature of the cylinder is measured, the performance input variable value may be further controlled based on the detected rate of combustion such that these two factors are traded off in subsequent operation of the cylinder to obtain a compromise between a favourable regeneration regime and combustion noise level.
A platform for an engine management system according to the present invention is described in more detail with reference to Figs. 14 to 16 for a system monitoring the pressures in all six cylinders of an engine and providing information concerning fuel quantity and injection timing which override the corresponding outputs of a production engine control unit 170.
Cylinder pressure sensors 172 are digitised by processing means comprising in the preferred embodiment an EMEK II intelligent data acquisition system 174. The data acquisition system also receives signals from sensors 176 which may include, for example, a mass air flow sensor, inlet temperature sensor, cam sensor, valve lift sensor air/fuel ratio or lambda sensor or any other appropriate sensors of known type. As can be seen from Fig. 15 the data acquisition system 174 yet further receives a crank tooth signal
providing a value of the crank angle (CA).
The digitised signals from the data acquisition system 174 are transmitted to a control and diagnostics unit 178 which may comprise a C40/C167 prototyping unit developed by Hema Elektronik GmbH of Germany. The control and diagnostics unit 178 further receives data including production sensor data from production engine control unit 170 and all input data is received in external input block 180. The control and diagnostics algorithms are configured, in the preferred embodiment, in MatrixX/SystemBuild, a high level simulation and algorithm development tool, and downloaded as compiled code to one or more digital signal processing (DSP) boards generally designated 182. The processed control data is transmitted from an external output block 184 of the control and diagnostics unit 178 to the modified production engine control unit 170 which controls the production actuators including, for example fuel injectors according to their control systems and algorithms discussed above.
It will be seen that the control and diagnostics unit 178 further includes a calibration block 188 which interfaces with an external calibration system 190 connected, for example, to a host PC 192. The calibration system 190 can carry out various calibration steps. For example the performance input variables for obtaining a performance output variable a desired Pmax, a desired rate of combustion, a desired start control strategy, a desired valve force, a desired EGR, a desired compressor or turbine performance or a desired exhaust temperature or AFR can be stored. It will be appreciated that any other appropriate calibration steps can equally be performed, or a model derived equivalently.
The DSP shown generally at block 182 runs separate cylinder pressure based EMS algorithms to implement the control strategies outlined above.
The plot in Fig. 16 is of cylinder pressure against crank angle and it will be seen that, for each cylinder, the cycle window 200 runs over a full 720°
cycle from a crank angle significantly before TDC to a crank angle shortly after TDC. This is followed by a data acquisition period 202 allowing the finite processing time required which runs up to a first "TN interrupt" 204. A second TN interrupt 206 occurs 120° later (ie 720° divided by the number of cylinders). Crank synchronisation timing and fuel quantity commands derived from the data acquired in the previous cycle window are applied at the second interrupt 206 as a result of which signal processing 208 must take place within the interval between the first and second interrupts. It will be noted that as the engine speed increases, although the crank angle interval between the first and second interrupts remains the same, in the time domain the interval decreases accordingly such that the signal processing step 208 must be implemented efficiently so as not to overlap the second TN interrupt. For example referring to the second plot of Fig. 16, in cylinder 4, it will be seen that the signal processing step 208 is carried out at a higher engine speed and hence falls closer to the second TN interrupt.
The ordering of the cylinders in Fig. 16 is 1, 4, 3, 6, 2, 5.
In the preferred embodiment the timing commands generated in control and diagnostics unit 178 are transmitted via the control area network (CAN) bus 194 to the production ECU 170 where they "bypass the normal commands generated by the production control algorithms. As a result the system can be "bolted on" in a preferred embodiment to an existing production ECU 170 with the logic appropriately modified to allow priority to the modified system in controlling production actuators.
It will be appreciated that the various embodiments discussed can be combined or interchanged and components therefrom combined or interchanged in any appropriate manner. In particular multiple control regimes can be combined and traded off against one another so as to achieve a compromise mode of
operation meeting more than one target output performance value. The approach can be applied in engine types of different configurations, stroke cycles and cylinder numbers and to different fuel type or combustion type internal combustion engines including natural gas engines and spark or compression ignition type engines and to different injection processes such as port-injection, direct-injection, a combination of both, multi-injection and multi-injector engines in which case the in-cylinder pressure data can be processed generally as discussed above but modified appropriately to obtain data on the equivalent parameters, which data can then be applied to appropriate actuation points dependent upon the engine type. Although the discussion above is principally applied to taking readings and applying on a cylinder-by-cylinder and cycle-by-cycle basis, averaging techniques can be applied over multiple cylinders or cycles as appropriate.