WO2020178325A1 - Method to determine the torque of a spark ingnition engine - Google Patents

Abstract

A method of determining the torque output of a vehicle engine using the following equation (I).

Description

METHOD TO DETERMINE THE TORQUE OF A SPARK IGNITION

ENGINE TECHNICAL FIELD

This relates to a method of determining the torque generated in a spark ignition combustion engine. BACKGROUND OF THE INVENTION

There are various prior art methods of monitoring and determining torque in spark ignition vehicles. Patent DE19916725A1 describes a method for torque monitoring in the case of Otto engines in motor vehicles. Patent US7289899B2 describes a method and system for calculating brake torque produced by a turbocharged engine.

US6704639B2 describes a method of calculating engine torque. US8050841B2 describes security for engine torque input air-per-cylinder calculations

The two main problems with the prior art are: that some methods only use one or two sensors to estimate torque which make it inaccurate and proves difficult to detect small torque deviations. Alternatively, prior are systems uses more sensors but does not include fail safe mechanisms for the torque estimator to continue to provide sufficient accuracy in the event of a failure on one of the used sensor is detected.

So in other words prior art using only one or two sensors has got a limited accuracy. That is solved by setting higher limit before detecting unwanted torque production thus compensating for accuracy loss by lowering accuracy requirement. However, this is preventing the system to detect small torque deviation. Prior art is also using desired set points for the actuators to evaluate estimated torque but this is a speculative approach on what will be the engine actual torque and this assumes the actuators deliver the desired set points.

Prior art uses plausibilisation mechanism on intake manifold pressure sensor or air flow sensor through the throttle position but these are the only sensor having fail safe mechanism for the torque estimator to continue to provide sufficient accuracy despite failure of that sensor.

Prior art also captures injected fuel quantity to estimate the air flow then finally estimate the torque based on that. This method is based on the assumption the engine control achieves a stoichiometric combustion which is not the case for example during engine warm-up. As a consequence, that method cannot estimate the torque in all engine conditions.

SUMMARY OF THE INVENTION

In one aspect is provided a method of comtrolling an engine including

determining the torque (Tq) output of a vehicle engine using the following equation:

eff therm where LHV is the fuel lower heat value;

eff_voi is the engine volumetric efficiency;

MAP is the manifold pressure;

MAT is the manifold temperature;

AFR Stoich is the air fuel ratio at stocihiometric conditions.

l opti is the optimum Lambda value that maximizes torque;

Displ is the engine capacity;

r - = 287 J/kg/K;

gas

eff therm is the engine thermal efficiency, and subsequently controlling said engine dependent on said determined torque.

The torque (Tq) output may be determined from the following equation:

eff therm * eff spark where

eff spark is the spark efficiency.

Thermal efficiency (eff therm) may be determined from the parameters of manifold or engine temperature, and engine speed.

A_{o pt}j may be assumed to be stoichiometric.

Spark efficiency may be determined from the parameter of engine speed.

The method may include the step of comparing said determined torque with an expected or reference torque.

Te method may include limiting the torque output of the engine or flagging an error condition dependent on said comparison.

BRIEF DESCRIPTION OF THE DRAWINGS

The present invention is now described by way of example with reference to the accompanying drawings in which:

- Figure 1 shows a simplified block diagram of a method of determining

torque.

DESCRIPTION OF THE PREFERRED EMBODIMENTS The problem of inaccuracy impeding the detection of small torque deviation of an internal combustion engine is solved by multi-sensor torque estimation according to examples below. The problem of providing an accurate torque estimation in case of a malfunctioning sensor is solved by cross-check plausiblisation of sensors and arbitration between sensors and/or recovery sensor models in order to continue to provide an accurate torque estimation. Prior Art

It is known in prior art system to estimate fast indicated torque using the equation 1 below.

„ . .. _, , r. . ^LHV *

Tq indicated fast = -

- -— - 4 _*p * eff— comb * eff therm * eff spark with:

LHV Fuel lower heat value

eff_Voi Engine volumetric efficiency

MAP Manifold pressure

MAT Manifold temperature

A_des Lambda that is injected, part of it will be condensated on walls or will go away if scavenging.

Displ Engine displacement

r 287 J/kg/K

Qcarb_condens Amount of fuel in g that condensate on walls during cold

start and that may not

participate to combustion

Q_carbscaveng Amount of fuel in g that will go out of cylinder in scavenging situation eff comb Correction factor for combustion efficiency when lambda is not at a stochiometric value

eff therm Thermal losses

eff spark Spark efficiency

This equation can provide very accurate estimation of actual torque, up to only few Newton meters error. Achieving that accuracy level demands high complexity of modelling, development and tuning. This complexity adds for malfunctions sources of the final torque estimator because it would be based on a too large number of inputs.

Invention In one aspect is a method of determining torque which has a reduced number of inputs while keeping sufficient accuracy for torque monitoring.

In one aspect is provided torque estimator equation simplified enough to ensure resilience to one sensor malfunction. In one example the troque of an engine is determined according to the following simplified equation 2:

- - MAP *Displ

LHV * (^eTTvo1 * r *MAT

AFR_sto ic * _nr.†; ' , ,

Tq (mdicated slow) fast = - - -— * eff therm

LHV Fuel lower heat value. This may be assumed. In advanced systems there may be on board systems to determine this from the fuel (e.g. detection systems for ethanol mix/cetane or octane numbers) eff_VO[ Engine volumetric efficiency. This may be determined from calibration. For volumetric efficiency e.g. 4 dimension mapping may be used with the inputs of i)engine speed ii) -Pressure ratio of Inlet Manifold Pressure/ Atmospheric pressure; iii) -position Cam exhaust iv) position Cam intake. Of course this is only applicable to engines fitted with cam phasers. Also if they are fitted on an engine then it is mandatory to consider them in the volumetric efficiency maps. Where cam phaser information is not available the engine vol. efficiency may be estimated form just i) and ii)

MAP Manifold pressure

MAT Manifold temperature

^_opti This is the Lambda value that produces the most torque. It is generally to stoichiometric.

Displ Engine capacity / (e.g. 1.0L)

r 287 J/kg/K

eff therm Thermal efficiency. This may be determined from estimated thermal losses and e.g. from calibration e.g. from a calibration curve map with e.g.

manifold/engine temperature and engine speed e.g as inputs.

AFR Stoich Air fuel ratio at stocihiometric conditions.

Further Refinement

In a refined embodient, the torque may be determined from the following equation:

MAP *Displ

I m * r^e> > ^{vo 1} * r *MAT L

v 1 AFR sto _tch * l . t

Tq (indicated fast) = - — : - ^0ptl * eff therm * eff spark

eff spark Spark efficiency This may be determined from calibration e.g. from a calibration curve map.

A) One or more sensor inputs may be used to determine this e.g. 1) Slow

Indicated mean effective pressure; -2) Operating mode 3) position Cam exhaust; 4) -position Cam intake 5) engine speed 6)_-lambda 7)-coolant temperature; 8) -Manifold temperature; 9)-injection pulse data including split injection pulses data; 10)-injection angle

B) The operating mode can be assumed . They can be considered for accuracy reasons without securing them because they will have little influence on torque estimator if they get corrupted. So in one example one can assume normal mode so torque estimator. In one example preferably only the following parameters are used to determine spark efficiency:

i) Slow Indicated mean effective pressure; ii) Operating mode; iii) position Cam exhaust; iv) position Cam intake; v) engine speed; vi) coolant temperature; vii) -Manifold temperature

C) preferably in a further simplified example only the following parameters are used to determine Spark efficiency: a) Slow (IMEP) Indicated mean effective pressure; b) Operating mode; c) position Cam exhaust; d) position Cam intake; e) engine speed;

D) Again c) and d) do not have to be used if not available and operating mode may be assumed thus in a further simplified example only the slow IMEP and the engine speed are used as parameters are used to determine Spark efficiency

In addition optionally the parameter of actual spark angle may be used in the above cases A, B, C, D above.

As far as slow IMEP is concerned this is an internally computed parameter and may be provided directly by the torque estimator itself, so in the above the value of IMEP may be considered not to be an (external) parameter which is required. Thus in sub-examples A, B , C, D, the value of IMEP can be disregarded so in example D) only external input is engine speed in a simple example. Torque and IMEP are equivalent so slow Torque/IMEP is just an intermediary step in the fast torque/IMEP calculation

- - MAP *Displ

LHV * (^{eTT m l} * r *MAT

^ AFR_stoic * l_ht,_† ,

Tq indicated slow = - : - -— * efif therm eff spark = f(Tq_indicated_slow, Engine speed) then

Tq indicated fast = Tq indicated slow * eff spark

MAP *Displ

MV (^e’’^vo1 * r *MAT

AFR_stoich )

4 *Pi * eff therm * eff spark .

In combustion engine, assuming it has stoichiometric ratio, the produced torque mainly (and this is really simplistic) depends on the air trapped inside cylinder and spark angle applied. Changing air inside a cylinder is slow. It takes up to few seconds. Torque achievable by air inside cylinder is called slow torque. If a particular spark angle is the best one for torque (this is called maximum best torque spark angle or MBT) then engine torque = slow torque because spark efficiency =1 because spark angle is best one for torque. Changing spark angle is a faster command. It can be changed at each cylinder revolution. Torque achieved according to actual spark angle is called fast torque. In some conditions if the spark angle used is not the best one for torque so engine torque = slow

torque(achievable by air) * spark efficiency(actually achieved by spark) with spark efficiency<l. In one example a map (2D calibration) map is used to evaluate spark efficiency. Spark efficiency is defined as spark efficiency = f(spark angle at mbt - actual spark angle) . Having a secured torque is then all about determining the spark at mbt and actual spark. Spark mbt can be mapped (2D calibration) in e.g. a safe memory area. These calibrations may depend on a lot of variables as described above. In an example spark mbt = f( engine speed, slow IMEP/torque). The actual spark angle may be read back or computed from hardware.

In one aspect a plausibility test may be provided which comprises comparing spark feedback reading of coil command with demanded spark. In another test it is checked whether the spark angle is inside the expected range i.e between minimum spark for not stalling the engine and maximum brake torque spark.

These tests are important to have a fully secured FAST indicated torque.

To secure that hardware feedback we need a plausibilisation that is as follows:

It is assumed that the engine operates at stoichiometric conditions and so the eff comb has disappeared from equation 1.

In examples, torque estimation determination is based on the following inputs/parameters which may be derived from actual sensors or models.

i) Engine speed;

ii) Atmospheric pressure; This is optional. The atmospheric pressure may be fixed or assumed e.g. at average or a general sea level value. Alternatively in refined embodiments a pressure sensor may be used or the pressure determined from GPS (with optionally GPS plus weather data maps including pressure at the location) . It is preferable to use estimates of pressure at altitudes or use GPS and other data we have no choice than use this. Typical example : sea level : Patmo = 1013 hPa, High altitude (4700m)

: Patmo = 580 hPa.

iii) Intake manifold pressure (MAP);

iv) Intake manifold temperature (MAT) .

v) Intake and exhaust cam position angle (if applicable i.e. this is optional) Cam phaser are not fitted on every engines (even this is becoming quite rare not to have both intake + exhaust phasers). Examples are applicable to all type of engines (no cam phasers or one cam phasers or both cam phasers if fitted==applicable here).

vi) Engine temperature

vii) Slow (IMEP) - this may derived from a model.

Other

In providing the above parameters for such simplified torque estimation; in refined embodiments for improved robustness (e.g. resilient to a sensor failure) there is for one or more sensors a cross-check plausiblisation and arbitration and/or recovery method or model.

The cross-check plausiblisation and arbitration and/or recovery models used in the application example are the following:

The engine speed may be computed through more than three available sources (these could be could be crank sensor, intake and exhaust cam sensors, combination between vehicle speed and transmission ratio). After correlation checks, an arbitration mechanism selects the most plausible and accurate inputs to be used as source of the engine speed information.

Figure 1 illustrates this and shows a simplified block diagram of a method of deteminging torque from e.g. sensor inputs . A main physical sensor is input A to a plausibility check block 1 along with optionally the inputs B from one or more other sensor/or indications of enviromental conditions. Block 1 checks whether the sensor input agrees with the expected sensors input computed or estimated from the one or more other sensors., and thus gives a plausibilty output D which may be an indication that the actual sensor output A is plausible or not. Optionally or alternatively or additionally a model 2 may be provided which models the parameter in respect of the sensor. Again this parameter provided by the model can be compared with the actual sensor input in block 3, and in embodiments depeding on the output D from plausibility block 1 various conclusion or indications are made. In one example, in block 4 a value of the parameter E which is a sensor failure resilient parameter value is determined. This may be the actual value A if the plausiblity check in 1, i.e. result D concludes the actual sensor value is plausible. Or if not plausible the value C from the model is used to provide E. E is then used to compute torque in block 5.

Atmospheric pressure sensor information may be correlated with the boost and exhaust pressures in non-boosted conditions. According to the result, an arbitration mechanism may be used to determine the value to use. The intake manifold pressure may be correlated with a modelization of the pressure based on the throttle position and boost pressure then compared to the sensor value. If the sensor value does not match the model, an environmental check is used to determine which information is correct and use it. If the environmental check fail, the highest pressure value of the two will be used. The intake manifold temperature is correlated with a modelization of the temperature based on boost temperature and then compared to the sensor value. If the sensor value does not match the model, an environmental check is used to determine which information is correct and use it. If the environmental check fail, the lowest temperature information will be used.

The intake and exhaust cam position is correlated with then current engine speed and the reference default mechanical cam position. If the signal is not plausible, the default mechanical position will be used instead as a recovery value.

The engine temperature information is correlated with oil and water temperature sensors. It the signal is not plausible, a warm-up engine temperature model will be used as a back up. The initial value of the model is based on a average value of the engine-related temperature sensors used at start-up.

As an example of application of the torque estimator, we have secured our torque estimation from its inputs acquisition down to its output calculation, we can safely use it to detect undesired torque even in case of one sensor malfunction. Used in the context of a safety monitoring and compared to prior art this torque estimator can detect any unwanted torque production leading to a vehicle unwanted acceleration even in case of one sensor malfunction. The advantage of this invention is that smaller torque deviations are detected. Furthermore, all sensors are covered against malfunctions which could cause the unwanted acceleration of being undetected.

Claims

1. A method of controlling an engine comprising determining the torque (Tq) output of a vehicle engine using the following equation:

eff therm

where LHV is the fuel lower heat value;

eff_voi is the engine volumetric efficiency;

MAP is the manifold pressure;

MAT is the manifold temperature;

AFR Stoich is the air fuel ratio at stoichiometric conditions.

^opti is the optimum Lambda value that maximizes torque;

Displ is the engine capacity;

r 287 J/kg/K;

eff therm is the engine thermal efficiency, and subsequently controlling said engine dependent on said determined torque.

2. A method as claimed in claim 1 wherein the torque (Tq) output is determined from the following equation:

eff therm * eff spark where

eff spark is the spark efficiency.

3. A method as claimed in claims 1 or 2 where thermal efficiency (eff therm) is determined from the parameters of manifold or engine temperature, and engine speed.

4. A method as claimed in claims 1 to 3 where A_opti assumed to be stoichiometric.

5. A method as claimed in claims 1 to 4 where spark efficiency is determined from the parameter of engine speed.

6. A method as claimed in claim 1 to 5 including the step of comparing said determined torque with an expected or reference torque.

7. A method as claimed in claim 6 including limiting the torque output of the engine or flagging an error condition dependent on said comparison.