Method for correcting a measurement of an air flow passing through a throttle valve of an otto cycle internal combustion engine

Abstract

Method for correcting a measurement of an air flow rate passing through a throttle valve (TV) of an Otto cycle internal combustion engine (E), carried out by means of a mass air flow sensor (MAF), the engine (E) comprising an intake manifold (IM) on which is arranged: the throttle valve (TV), the mass air flow sensor (MAF), upstream of the throttle valve, a first pressure sensor (PU) and a second pressure sensor (PD) arranged respectively upstream and downstream of the throttle valve (TV), a temperature sensor (TU) arranged upstream of the throttle valve (TV), wherein the mass air flow sensor (MAP) is arranged between said temperature sensor (TU) and the throttle valve (TV), the method comprising the implementation, by means of a processing unit (ECU), of a Luemberger observer (OBS) to calculate a signal (xk) representative of the mass air flow rate passing through the throttle valve, wherein the observer is configured to simultaneously acquire at the input a first signal (eyk), a function of a second signal (x̃k) generated by the mass air flow sensor (MAP) and a third signal (ϕthrk), calculated by means of a throttle valve outflow model, representative of said mass air flow passing through the throttle valve.

Description

Field of the invention

[0001] The present invention relates to the field of methods for correcting the measurement of an air flow rate passing through a throttle valve of an Otto cycle internal combustion engine, where the measurement is carried out with a so-called mass air flow meter MAF.State of the art

[0002] Controlling the amount of air trapped in the cylinders of a spark-ignition internal combustion engine, known as "air path control," is essential for controlling the torque generated by the engine. It is of fundamental importance for engine performance because it must contribute to provide the target torque at a predetermined engine rotation speed.

[0003] The torque generated by the engine depends on the amount of air trapped in the cylinders and the amount of fuel injected and the crank angle at which ignition occurs, the so-called ignition advance.

[0004] Since in a spark-ignition engine, the amount of fuel is closely related to the amount of air trapped in the cylinders, it is understood that controlling the amount of air entering the engine is the fundamental factor that affects the engine torque output.

[0005] The implementation of mass air flow sensors, also known by the acronym "MAF" (Mass Air Flow), is known, generally located upstream of the throttle valve and downstream of the air filter. In some applications it is arranged immediately upstream of the throttle valve.

[0006] In turbocharged engines, the MAF may be located between the intake filter and the compressor or between the compressor and the throttle valve.

[0007] However, the mass flow rate of air passing through the throttle valve differs significantly from the mass flow rate of air entering the engine under transient conditions. There are several types of MAFs on the market. The most common and inexpensive are those where a filament is held at a predetermined temperature (HFM) and the mass flow rate is measured as a function of the change in current passing through the filament and where a pressure difference (PFM) is measured to calculate the mass flow rate of air corresponding to that pressure difference.

[0008] It is commonly believed that the MAF is sufficient to determine the mass of air trapped in the cylinders in order to calculate the fuel to be injected based on the desired air / fuel ratio.

[0009] First of all, it must be reiterated that the mass flow rate of air trapped in the cylinders coincides with that measured by the MAF only in stationary conditions.

[0010] Secondly, the measurement is unstable for low flow rates, in particular, significant variations in the signal are detected around the value of the correct measurement, furthermore it is greatly affected by the disturbances caused to the fluid threads by the geometries of the ducts in which it is placed, as well as by the position of the sensor itself with respect to these ducts.

[0011] When the engine is at low load and low rotation speed, the pressure value at the intake manifold downstream of the throttle valve is significantly lower than the value upstream of the throttle valve, causing turbulence in the fluid threads that lead to unstable measurements. This instability therefore does not have an intrinsic electrical nature of the sensor but is related to the fluid dynamics.

[0012] By significantly filtering the electrical signal generated by the MAF sensor, a good reading is obtained in stationary conditions and incorrect in transient conditions, thus also affecting the relative calculation of the estimate of the air trapped in the cylinders.

[0013] The fluid dynamic model based on the well-known equation of isentropic flow through a restriction with a relative efflux coefficient to take into account the differences between ideality and reality is well known.

[0014] By applying this model to the throttle valve, an excellent correspondence is obtained, in stationary conditions, between the calculated mass flow rate and the real and / or measured mass flow rate when the ratio between the values of the pressure downstream and upstream of the throttle valve is low, therefore in conditions of low engine load regardless of the rotation speed.

[0015] When this ratio is high, i.e. close to 1, however, there may be strong discrepancies between the real and / or measured flow rate and that calculated by this model.

[0016] The problem is due to the fact that the quantities used to feed this model are not the real or measured ones but modelled at average value, that is to say ignoring the pumping peaks of the cylinders.

[0017] Unless specifically excluded in the detailed description that follows, what is described in this chapter is to be considered as an integral part of the detailed description.Summary of the invention

[0018] The purpose of the present invention is to present a method for correcting a measurement of an air flow rate passing through a throttle valve of an Otto cycle internal combustion engine, measured by means of an air mass measurement sensor. The basic idea of the present invention is to use a Luemberger observer to correct the measured air mass flow rate, where the observer acquires as input a signal composed of a first contribution proportional to the signal generated by a MAF sensor located immediately upstream of the throttle valve and a second contribution generated by an outflow model of the throttle valve.

[0019] Advantageously, the observer returns a value of the air mass flow rate measured by the sensor and passing through the throttle valve that is stable in any operating condition and fast in following the variations of the air flow rate in transient conditions.

[0020] According to a preferred aspect of the invention, the signal generated by the MAF sensor is low-pass filtered in order to limit the instability described above.

[0021] Preferably, the first contribution is an error calculated between the low-pass filtered signal generated by the MAF sensor and the air flow signal calculated by the observer. The observer can be schematized as a sequence of filters, where a first filter realizes the integral of the error, a second filter realizes a low-pass filter that models the dynamics of the MAF sensor, a third filter realizes a low-pass filter having characteristics similar to the low-pass filter of the signal generated by the MAF sensor.

[0022] Preferably, the first and second contributions are added to each other.

[0023] Preferably, the observer also receives the aforementioned error signal as input.

[0024] Preferably, the low-pass filter applied to the signal calculated by the observer is also a function of the error signal.

[0025] The dependent claims describe preferred variants of the invention, forming an integral part of the present description.Brief description of the figures

[0026] Further objects and advantages of the present invention will be apparent from the following detailed description of an embodiment thereof (and variations thereof) and the accompanying drawings given purely for explanatory and non-limiting purposes, in which: Fig. 1 discloses an example of a calculation scheme temporarily implemented to achieve the present invention; Fig. 2 discloses an example of an internal combustion engine equipped with a calculation system according to the present invention; Fig. 3 discloses a hardware scheme comprising a processing unit in which the method of the present invention is implemented.

[0027] The same reference numbers and letters in the figures identify the same elements or components or functions.

[0028] It should also be noted that the terms "first", "second", "third", "upper", "lower" and the like may be used herein to distinguish various elements. These terms do not imply a spatial, sequential or hierarchical order for the modified elements unless specifically indicated or inferred from the text.

[0029] The elements and features illustrated in the various preferred embodiments, including the drawings, may be combined with each other without departing from the scope of protection of the present application as described below.Detailed description

[0030] Consider the generic continuous-time linear system x ˜ ˙ = A x ˜ + B u y ˜ = C x ˜

[0031] Where x̃ represents the set of state variables of the system, ỹ represents the set of measured quantities of the system, or the set of outputs, and ũ the set of inputs of the system.

[0032] A, B and C represent matrices of coefficients, of which the first is called "dynamic", the second "forcing" and the third "output".

[0033] The linear system can be rewritten in discrete time as 1 T s x ˜ k + 1 − x ˜ k = A x ˜ k + B u k x ˜ k + 1 − x ˜ k = T s A x ˜ k + B u k x ˜ k + 1 = x ˜ k + T s A x ˜ k + B u k y ˜ k = C x ˜ k

[0034] Where T s represents the sampling interval and k represents the generic sampling step.

[0035] A continuous-time first-order filter is given by x ˙ = K u − x

[0036] The same first order filter discretized forward according to Euler, with sampling time T s can be rewritten as x ˙ = x k + 1 − x k T s x k + 1 − x k T s = K u k − x k x k + 1 = x k + T s K u k − x k

[0037] The following system represents the classical Luemberger observer x k + 1 = x k + T s A x k + B u k + L y ˜ k − y k y k = C x k

[0038] Where L represents the gains matrix and the quantity (ỹ k - y k ) = e k represents the estimation error based on the measurement.

[0039] From the latter we obtain: y ˜ k − y k = C x ˜ k − x k = e yk

[0040] Therefore, substituting it in the first of the two, it is obtained x k + 1 = x k + T s A x k + B u k + L e y k

[0041] It is therefore understood that the Luemberger observer differs from the classical linear discrete-time system for the last term Le y k relating to the error.

[0042] A known solution involves excluding the MAF sensor reading at low loads and sometimes even in transient.

[0043] The present solution instead aims at keeping the signal generated by the MAF sensor constantly acquired.

[0044] First, it is necessary to filter this signal with a low-pass filter "Filter_1" to limit the dynamics of the MAF sensor. It is preferably a first-order filter.

[0045] At the same time, the flow ϕ thrk that passes through the throttle valve is calculated using the well-known "Air mass flow Throttle Model".

[0046] However, through bench tests, it was realized that the dynamics of the model is ahead of the MAF sensor readings. The reason is to be found in the fact that the fluid dynamic effects affect the MAF sensor late, as it is located upstream of the throttle valve; the MAF sensor is in itself a low-pass filter with its own slower dynamics than the acquisition of the pressure, temperature and throttle position parameters, which feed the outflow model.

[0047] Therefore, according to the present invention it was necessary to apply a first low-pass filter, named as "MAF Model", which can be of the first or second order, to delay the dynamics of the outflow model.

[0048] The output of this first filter can be compared with the signal generated by the MAF sensor.

[0049] However, keeping in mind that this signal is low-pass filtered by the aforementioned Filter_1 filter, then for consistency a second filter "Filter_2" is applied which has the same gain as Filter_1.

[0050] At this point the two signals generated respectively by the MAF sensor and the outflow model can be compared, generating an error e yk useful for correcting the first "MAF Model" filter.

[0051] Preferably, the error is filtered by a controller "Corrector". From the tests performed it appears that an integrator can be sufficient to perform the controller task. Looking at figure 1, the equations underlying the same scheme can now be constructed: Efflux model: ϕ thr = P us R s T us A α Φ P ds P us C d

[0052] Where P us and T us are respectively the signals representing pressure and temperature measured immediately upstream of the throttle valve; P ds is the signal representing pressure measured immediately downstream of the throttle valve; α represents the throttle valve shutter angle, A(α) represents the throttle valve discharge section; R s is the known gas specific constant, Φ is the so-called pure flow function and C d is the so-called discharge coefficient.

[0053] Considering the first order filter equation mentioned above and applied to the first "MAF Model" filter, we obtain: x k + 1 = x k + T s G d ϕ thr k + δ k − x k

[0054] The first-order filter Filter_2 - similarly to Filter_1, where the variable is ỹ k - can be represented by the following equation: y k + 1 = y k + T s G f x k − y k

[0055] The integral controller can be represented by the following equation: δ k + 1 = δ k + Ge y k

[0056] Where G represents the gain of the integral controller.

[0057] The scheme in figure 1 has been tested on the bench and works with some limitations. We have therefore tried to improve it in various ways.

[0058] Observing equations (1) - (3) we had the idea of adding the error through respective gains L2, L3 also to equations (1) and (2). Therefore, starting from (3) where L1 = G is set, thus we have: δ k + 1 = δ k + L 1 e y k x k + 1 = x s + T s G d ϕ thr k + δ k − x k + L 2 e y k y k + 1 = y k + T s G f x k − y k + L 3 e y k

[0059] Therefore, the scheme of figure 1 is modified according to the scheme of figure 2, in which the error e yk is acquired at the input, not only by the first filter "MAF Model", but also by the second filter "Filter_2".

[0060] With the bench tests of the control scheme of figure 2 it was possible to verify a fast behaviour in transients and, at the same time, precise and stable in stationary and low load conditions.

[0061] Therefore, with reference to figure 2 it can be deduced that the observer performs the following steps in succession: acquisition of the first contribution e yk , integration of the first contribution obtaining an integrated signal δ k of the first contribution, sum of the first contribution and the second contribution ϕ thrk + δ k , filtering of the sum signal ϕ thrk + δ k by means of a second low-pass filter MAF_Dyn so that the low-pass filtered signal has the same dynamics as the signal generated by the air mass flow sensor MAF, obtaining said signal representative of the estimate of an air flow rate x k , filtering of the signal representative of the estimate of an air flow rate x k by means of the third low-pass filter Filter_2 for the calculation of said error e yk .

[0062] Advantageously, the signal measured and corrected according to the present invention, with the flow calculation model through the scheme presented, thus allows to have a stable measurement in stationary and fast in transients.

[0063] The present invention can be advantageously realized through a computer program that includes coding means for the realization of one or more steps of the method, when this program is executed on a computer. Therefore, it is intended that the scope of protection extends to said computer program and also to computer-readable means that include a recorded message, said computer-readable means including program coding means for the realization of one or more steps of the method, when said program is executed on a computer. Variations to the non-limiting example described are possible, without however departing from the scope of protection of the present invention, including all the equivalent realizations for a technician in the field, to the content of the claims.

[0064] From the description reported above, the technician in the field is able to realize the object of the invention without introducing further construction details.

Examples

Embodiment Construction

[0030]Consider the generic continuous-time linear system x ˜ ˙ = A x ˜ + B u y ˜ = C x ˜

[0031]Where x̃ represents the set of state variables of the system, ỹ represents the set of measured quantities of the system, or the set of outputs, and ũ the set of inputs of the system.

[0032]A, B and C represent matrices of coefficients, of which the first is called "dynamic", the second "forcing" and the third "output".

[0033]The linear system can be rewritten in discrete time as 1 T s x ˜ k + 1 − x ˜ k = A x ˜ k + B u k x ˜ k + 1 − x ˜ k = T s A x ˜ k + B u k x ˜ k + 1 = x ...

Field of the invention

[0001] The present invention relates to the field of methods for correcting the measurement of an air flow rate passing through a throttle valve of an Otto cycle internal combustion engine, where the measurement is carried out with a so-called mass air flow meter MAF.State of the art

[0002] Controlling the amount of air trapped in the cylinders of a spark-ignition internal combustion engine, known as "air path control," is essential for controlling the torque generated by the engine. It is of fundamental importance for engine performance because it must contribute to provide the target torque at a predetermined engine rotation speed.

[0003] The torque generated by the engine depends on the amount of air trapped in the cylinders and the amount of fuel injected and the crank angle at which ignition occurs, the so-called ignition advance.

[0004] Since in a spark-ignition engine, the amount of fuel is closely related to the amount of air trapped in the cylinders, it is understood that controlling the amount of air entering the engine is the fundamental factor that affects the engine torque output.

[0005] The implementation of mass air flow sensors, also known by the acronym "MAF" (Mass Air Flow), is known, generally located upstream of the throttle valve and downstream of the air filter. In some applications it is arranged immediately upstream of the throttle valve.

[0006] In turbocharged engines, the MAF may be located between the intake filter and the compressor or between the compressor and the throttle valve.

[0007] However, the mass flow rate of air passing through the throttle valve differs significantly from the mass flow rate of air entering the engine under transient conditions. There are several types of MAFs on the market. The most common and inexpensive are those where a filament is held at a predetermined temperature (HFM) and the mass flow rate is measured as a function of the change in current passing through the filament and where a pressure difference (PFM) is measured to calculate the mass flow rate of air corresponding to that pressure difference.

[0008] It is commonly believed that the MAF is sufficient to determine the mass of air trapped in the cylinders in order to calculate the fuel to be injected based on the desired air / fuel ratio.

[0009] First of all, it must be reiterated that the mass flow rate of air trapped in the cylinders coincides with that measured by the MAF only in stationary conditions.

[0010] Secondly, the measurement is unstable for low flow rates, in particular, significant variations in the signal are detected around the value of the correct measurement, furthermore it is greatly affected by the disturbances caused to the fluid threads by the geometries of the ducts in which it is placed, as well as by the position of the sensor itself with respect to these ducts.

[0011] When the engine is at low load and low rotation speed, the pressure value at the intake manifold downstream of the throttle valve is significantly lower than the value upstream of the throttle valve, causing turbulence in the fluid threads that lead to unstable measurements. This instability therefore does not have an intrinsic electrical nature of the sensor but is related to the fluid dynamics.

[0012] By significantly filtering the electrical signal generated by the MAF sensor, a good reading is obtained in stationary conditions and incorrect in transient conditions, thus also affecting the relative calculation of the estimate of the air trapped in the cylinders.

[0013] The fluid dynamic model based on the well-known equation of isentropic flow through a restriction with a relative efflux coefficient to take into account the differences between ideality and reality is well known.

[0014] By applying this model to the throttle valve, an excellent correspondence is obtained, in stationary conditions, between the calculated mass flow rate and the real and / or measured mass flow rate when the ratio between the values of the pressure downstream and upstream of the throttle valve is low, therefore in conditions of low engine load regardless of the rotation speed.

[0015] When this ratio is high, i.e. close to 1, however, there may be strong discrepancies between the real and / or measured flow rate and that calculated by this model.

[0016] The problem is due to the fact that the quantities used to feed this model are not the real or measured ones but modelled at average value, that is to say ignoring the pumping peaks of the cylinders.

[0017] Unless specifically excluded in the detailed description that follows, what is described in this chapter is to be considered as an integral part of the detailed description.Summary of the invention

[0018] The purpose of the present invention is to present a method for correcting a measurement of an air flow rate passing through a throttle valve of an Otto cycle internal combustion engine, measured by means of an air mass measurement sensor. The basic idea of the present invention is to use a Luemberger observer to correct the measured air mass flow rate, where the observer acquires as input a signal composed of a first contribution proportional to the signal generated by a MAF sensor located immediately upstream of the throttle valve and a second contribution generated by an outflow model of the throttle valve.

[0019] Advantageously, the observer returns a value of the air mass flow rate measured by the sensor and passing through the throttle valve that is stable in any operating condition and fast in following the variations of the air flow rate in transient conditions.

[0020] According to a preferred aspect of the invention, the signal generated by the MAF sensor is low-pass filtered in order to limit the instability described above.

[0021] Preferably, the first contribution is an error calculated between the low-pass filtered signal generated by the MAF sensor and the air flow signal calculated by the observer. The observer can be schematized as a sequence of filters, where a first filter realizes the integral of the error, a second filter realizes a low-pass filter that models the dynamics of the MAF sensor, a third filter realizes a low-pass filter having characteristics similar to the low-pass filter of the signal generated by the MAF sensor.

[0022] Preferably, the first and second contributions are added to each other.

[0023] Preferably, the observer also receives the aforementioned error signal as input.

[0024] Preferably, the low-pass filter applied to the signal calculated by the observer is also a function of the error signal.

[0025] The dependent claims describe preferred variants of the invention, forming an integral part of the present description.Brief description of the figures

[0026] Further objects and advantages of the present invention will be apparent from the following detailed description of an embodiment thereof (and variations thereof) and the accompanying drawings given purely for explanatory and non-limiting purposes, in which: Fig. 1 discloses an example of a calculation scheme temporarily implemented to achieve the present invention; Fig. 2 discloses an example of an internal combustion engine equipped with a calculation system according to the present invention; Fig. 3 discloses a hardware scheme comprising a processing unit in which the method of the present invention is implemented.

[0027] The same reference numbers and letters in the figures identify the same elements or components or functions.

[0028] It should also be noted that the terms "first", "second", "third", "upper", "lower" and the like may be used herein to distinguish various elements. These terms do not imply a spatial, sequential or hierarchical order for the modified elements unless specifically indicated or inferred from the text.

[0029] The elements and features illustrated in the various preferred embodiments, including the drawings, may be combined with each other without departing from the scope of protection of the present application as described below.Detailed description

[0030] Consider the generic continuous-time linear system x ˜ ˙ = A x ˜ + B u y ˜ = C x ˜

[0031] Where x̃ represents the set of state variables of the system, ỹ represents the set of measured quantities of the system, or the set of outputs, and ũ the set of inputs of the system.

[0032] A, B and C represent matrices of coefficients, of which the first is called "dynamic", the second "forcing" and the third "output".

[0033] The linear system can be rewritten in discrete time as 1 T s x ˜ k + 1 − x ˜ k = A x ˜ k + B u k x ˜ k + 1 − x ˜ k = T s A x ˜ k + B u k x ˜ k + 1 = x ˜ k + T s A x ˜ k + B u k y ˜ k = C x ˜ k

[0034] Where T s represents the sampling interval and k represents the generic sampling step.

[0035] A continuous-time first-order filter is given by x ˙ = K u − x

[0036] The same first order filter discretized forward according to Euler, with sampling time T s can be rewritten as x ˙ = x k + 1 − x k T s x k + 1 − x k T s = K u k − x k x k + 1 = x k + T s K u k − x k

[0037] The following system represents the classical Luemberger observer x k + 1 = x k + T s A x k + B u k + L y ˜ k − y k y k = C x k

[0038] Where L represents the gains matrix and the quantity (ỹ k - y k ) = e k represents the estimation error based on the measurement.

[0039] From the latter we obtain: y ˜ k − y k = C x ˜ k − x k = e yk

[0040] Therefore, substituting it in the first of the two, it is obtained x k + 1 = x k + T s A x k + B u k + L e y k

[0041] It is therefore understood that the Luemberger observer differs from the classical linear discrete-time system for the last term Le y k relating to the error.

[0042] A known solution involves excluding the MAF sensor reading at low loads and sometimes even in transient.

[0043] The present solution instead aims at keeping the signal generated by the MAF sensor constantly acquired.

[0044] First, it is necessary to filter this signal with a low-pass filter "Filter_1" to limit the dynamics of the MAF sensor. It is preferably a first-order filter.

[0045] At the same time, the flow ϕ thrk that passes through the throttle valve is calculated using the well-known "Air mass flow Throttle Model".

[0046] However, through bench tests, it was realized that the dynamics of the model is ahead of the MAF sensor readings. The reason is to be found in the fact that the fluid dynamic effects affect the MAF sensor late, as it is located upstream of the throttle valve; the MAF sensor is in itself a low-pass filter with its own slower dynamics than the acquisition of the pressure, temperature and throttle position parameters, which feed the outflow model.

[0047] Therefore, according to the present invention it was necessary to apply a first low-pass filter, named as "MAF Model", which can be of the first or second order, to delay the dynamics of the outflow model.

[0048] The output of this first filter can be compared with the signal generated by the MAF sensor.

[0049] However, keeping in mind that this signal is low-pass filtered by the aforementioned Filter_1 filter, then for consistency a second filter "Filter_2" is applied which has the same gain as Filter_1.

[0050] At this point the two signals generated respectively by the MAF sensor and the outflow model can be compared, generating an error e yk useful for correcting the first "MAF Model" filter.

[0051] Preferably, the error is filtered by a controller "Corrector". From the tests performed it appears that an integrator can be sufficient to perform the controller task. Looking at figure 1, the equations underlying the same scheme can now be constructed: Efflux model: ϕ thr = P us R s T us A α Φ P ds P us C d

[0052] Where P us and T us are respectively the signals representing pressure and temperature measured immediately upstream of the throttle valve; P ds is the signal representing pressure measured immediately downstream of the throttle valve; α represents the throttle valve shutter angle, A(α) represents the throttle valve discharge section; R s is the known gas specific constant, Φ is the so-called pure flow function and C d is the so-called discharge coefficient.

[0053] Considering the first order filter equation mentioned above and applied to the first "MAF Model" filter, we obtain: x k + 1 = x k + T s G d ϕ thr k + δ k − x k

[0054] The first-order filter Filter_2 - similarly to Filter_1, where the variable is ỹ k - can be represented by the following equation: y k + 1 = y k + T s G f x k − y k

[0055] The integral controller can be represented by the following equation: δ k + 1 = δ k + Ge y k

[0056] Where G represents the gain of the integral controller.

[0057] The scheme in figure 1 has been tested on the bench and works with some limitations. We have therefore tried to improve it in various ways.

[0058] Observing equations (1) - (3) we had the idea of adding the error through respective gains L2, L3 also to equations (1) and (2). Therefore, starting from (3) where L1 = G is set, thus we have: δ k + 1 = δ k + L 1 e y k x k + 1 = x s + T s G d ϕ thr k + δ k − x k + L 2 e y k y k + 1 = y k + T s G f x k − y k + L 3 e y k

[0059] Therefore, the scheme of figure 1 is modified according to the scheme of figure 2, in which the error e yk is acquired at the input, not only by the first filter "MAF Model", but also by the second filter "Filter_2".

[0060] With the bench tests of the control scheme of figure 2 it was possible to verify a fast behaviour in transients and, at the same time, precise and stable in stationary and low load conditions.

[0061] Therefore, with reference to figure 2 it can be deduced that the observer performs the following steps in succession: acquisition of the first contribution e yk , integration of the first contribution obtaining an integrated signal δ k of the first contribution, sum of the first contribution and the second contribution ϕ thrk + δ k , filtering of the sum signal ϕ thrk + δ k by means of a second low-pass filter MAF_Dyn so that the low-pass filtered signal has the same dynamics as the signal generated by the air mass flow sensor MAF, obtaining said signal representative of the estimate of an air flow rate x k , filtering of the signal representative of the estimate of an air flow rate x k by means of the third low-pass filter Filter_2 for the calculation of said error e yk .

[0062] Advantageously, the signal measured and corrected according to the present invention, with the flow calculation model through the scheme presented, thus allows to have a stable measurement in stationary and fast in transients.

[0063] The present invention can be advantageously realized through a computer program that includes coding means for the realization of one or more steps of the method, when this program is executed on a computer. Therefore, it is intended that the scope of protection extends to said computer program and also to computer-readable means that include a recorded message, said computer-readable means including program coding means for the realization of one or more steps of the method, when said program is executed on a computer. Variations to the non-limiting example described are possible, without however departing from the scope of protection of the present invention, including all the equivalent realizations for a technician in the field, to the content of the claims.

[0064] From the description reported above, the technician in the field is able to realize the object of the invention without introducing further construction details.

Examples

Embodiment Construction

[0030]Consider the generic continuous-time linear system x ˜ ˙ = A x ˜ + B u y ˜ = C x ˜

[0031]Where x̃ represents the set of state variables of the system, ỹ represents the set of measured quantities of the system, or the set of outputs, and ũ the set of inputs of the system.

[0032]A, B and C represent matrices of coefficients, of which the first is called "dynamic", the second "forcing" and the third "output".

[0033]The linear system can be rewritten in discrete time as 1 T s x ˜ k + 1 − x ˜ k = A x ˜ k + B u k x ˜ k + 1 − x ˜ k = T s A x ˜ k + B u k x ˜ k + 1 = x ...

Claims

1. Method for correcting a measurement of an air flow rate passing through a throttle valve (TV) of an Otto cycle internal combustion engine (E), carried out by means of a mass air flow sensor (MAF), the engine (E) comprising an intake manifold (IM) on which is arranged: - the throttle valve (TV), - the mass air flow sensor (MAF), upstream of the throttle valve, - a first pressure sensor (PU) and a second pressure sensor (PD) arranged respectively upstream and downstream of the throttle valve (TV), - a temperature sensor (TU) arranged upstream of the throttle valve (TV), wherein the mass air flow sensor (MAF) is arranged between said temperature sensor (TU) and the throttle valve (TV), the method comprising the implementation, by means of a processing unit (ECU), of a Luemberger observer (OBS) to calculate a signal (xk) representative of the mass air flow rate passing through the throttle valve, wherein the observer is configured to simultaneously acquire at the input - a first signal (eyk), a function of a second signal (x̃k) generated by the mass air flow sensor (MAF) and - a third signal (ϕthrk), calculated by means of a throttle valve outflow model, representative of said mass air flow passing through the throttle valve.

2. Method according to claim 1, wherein the second signal (x̃k) generated by the mass air flow sensor (MAF) is low-pass filtered, by means of a first low-pass filter (Filter_1).

3. Method according to claim 2, wherein said first signal is an error given by the difference between the second low-pass filtered signal generated by the mass air flow sensor (MAF) (ỹk) and said low-pass filtered signal (xk), representative of the estimate of the mass air flow rate (xk) calculated by the observer (OBS).

4. A method according to any of the preceding claims, wherein the observer performs the following steps in succession: - acquisition of the first signal (eyk), - integration of the first signal obtaining an integrated signal (δk) of the first signal, - sum of the first signal and the third signal (ϕthrk + δk) obtaining a sum signal, - filtering of the sum signal (ϕthrk + δk) by means of a second low-pass filter (MAF_Dyn) so that the low-pass filtered sum signal has the same dynamics as the second signal, obtaining said signal (xk) representative of the mass air flow rate passing through the throttle valve, - filtering of the signal (xk) representative of the mass air flow rate passing through the throttle valve by means of a third low-pass filter (Filter _2) for the calculation of said error.

5. Method according to claim 4, wherein said third low-pass filter (Filter_2) has the same characteristics as said first low-pass filter (Filter_1) so as to achieve a coherent comparison between the second signal, generated by the mass air flow sensor (MAF), and the signal (xk) representing the mass air flow passing through the throttle valve, generated by the observer.

6. Method according to claim 5, wherein said first and third low-pass filters are of the first order and have the same time constant.

7. Method according to claim 6 or 7, wherein said second and third low-pass filtering are operated taking into account said error (eyk).

8. Method according to any of claims 4 - 7, wherein said second low-pass filter is of the first order.

9. Method according to any of the preceding claims 3 - 8, wherein the observer is identified by the following discrete-time equations x k + 1 = x k + T s G d ϕ thr k + δ k − x k + L 2 e y k y k + 1 = y k + T s G f x k − y k + L 3 e y k δ k + 1 = δ k + L 1 e y k Where xk indicates a signal representative of the correct measurement of the air mass flow rate measured by the MAF sensor passing through the throttle valve and k represents the generic sampling step; Ts represents the sampling interval; Gd and Gf indicate gains; ϕthrk indicates the third signal, representative of an air mass flow rate calculated using the throttle valve outflow model; δk indicates a signal representative of the integral of the signal generated by the low-pass filtered air mass flow sensor; L1, L2, L3 indicate multiplicative gains of the error between the estimated air mass flow rate and the measured air mass flow rate.

10. A computer program comprising program coding means for performing all the steps of any one of claims 1 to 9, when said program is run on a processing unit operatively connected to said - first pressure sensor (PU) and second pressure sensor (PD) arranged respectively immediately upstream and downstream of the throttle valve (TV), - temperature sensor (TU) arranged upstream of the throttle valve (TV), - mass air flow sensor (MAF) arranged between said temperature sensor (TU) and the throttle valve (TV).

11. A computer-readable means having stored the program of claim 10.

12. A system for correcting a measurement of an air flow rate passing through a throttle valve of an Otto cycle internal combustion engine (E), the engine (E) comprising an intake manifold (IM) on which is arranged: - the throttle valve (TV), - the mass air flow sensor (MAF), upstream of the throttle valve, - a first sensor pressure sensor (PU) and a second pressure sensor (PD) arranged respectively upstream and downstream of the throttle valve (TV), - a temperature sensor (TU) arranged upstream of the throttle valve (TV), wherein the mass air flow sensor (MAF) is arranged between said temperature sensor (TU) and the throttle valve (TV), - processing means configured to implement a Luemberger observer (OBS) so as to calculate a signal (xk) representative of the mass air flow rate passing through the throttle valve, as a function of - a first signal (eyk), a function of a second signal (x̃k) generated by the mass air flow sensor (MAF) and - a third signal (ϕthrk), calculated by means of a throttle valve outflow model, representative of said mass air flow rate passing through the throttle valve.

Images