DESCRIPTION

The present invention refers to an apparatus and to a method for controlling the quantity of fuel injected in an internal combustion engine; in particular, a Diesel cycle engine.

As is known, the operation of an internal combustion engine, in particular of a Diesel cycle engine, produces a considerable quantity of polluting substances which, in addition to polluting the environment, directly cause negative effects on the human body, especially on the respiratory apparatus; indeed, the unburned hydrocarbons in the form of micro-particulate are capable of reaching the pulmonary alveoli and compromising the operation thereof, inducing pathologies such as asthma, chronic obstructive pulmonary disease (COPD) and cardiovascular diseases of various type. In order to reduce this type of pollution, manufacturers of automobiles and/or of engines have created various solutions, such as anti-particulate filters (FAP), traps or other solutions. However, these solutions are effective on PM100 but are not very effective on PM10. There is therefore the need to find alternative solutions that improve the efficiency of the engines by having them produce less unburned hydrocarbons, but which at the same time do not significantly alter the power and drive torque supplied by an internal combustion engine as well as the efficiency of the same.

This problem can be faced in a satisfactory manner by controlling, among other things and with a high level of precision, the quantity of fuel introduced into the cylinders of an engine. In this manner, it is possible to reduce the quantity of unburned hydrocarbons produced by an internal combustion engine, since it is always injected a quantity of fuel that the engine can (almost) completely burn.

In addition, the injection systems according to the state of the art are capable of carrying out more than one injection in the course of every single operation cycle, so as to improve the polluting emissions, the efficiency and reduce the noise of the engine during its operation. Since the multiple injections per operation cycle have a lower duration with respect to a single injection, the capacity to estimate the quantity of fuel injected in the course of every single injection becomes, in current injection systems, even more important .

In order to improve the control of the quantity of injected fuel, there are various approaches (see Jorach RW et al . , 2011, "Common Rail system from Delphi with solenoid valves and single plunger pump", MTZ 03/2011, Vol. 72.) which are based on the development of injectors having particular shapes and/or on the addition of pressure sensors inside an injector (see Matsumoto S. et al . , 2013, "The New Denso Common Rail Diesel Solenoid Injector", MTZ 021/2013, Vol. 74) .

However, these solutions have the problem of only being usable on newly-designed engines or upon substitution of the injectors, while they do not make possible an easy and quick implementation of the control of the quantity of fuel injected on engines that are already existing or even already mounted on vehicles in circulation.

The present invention solves these and other problems by providing an apparatus (also referred as engine control unit) and a method for controlling the quantity of fuel injected in an internal combustion engine as in the enclosed claims.

The idea underlying the present invention is to (directly or indirectly) control the pressure of the fluid in the high- pressure fuel circuit, such that said pressure follows a pre- established trend, wherein said trend is determined on the basis of the quantity of fuel that it is necessary to inject into an internal combustion engine.

In this manner, it is possible to control the quantity of fuel injected in the course of each injection, without having to use sensors dedicated to measuring the flow rate and/or the pressure. As already mentioned, this makes it possible to implement the control of the quantity of fuel injected per single injection on engines that are already existing and/or even already installed on circulating vehicles, reducing the emissions of unburned hydrocarbons and thus making it possible to improve the operating efficiency of said engines.

In addition, the presence of this possibility to control considerably simplifies the step of adjusting an internal combustion engine, i.e. the step in which the technicians make the so-called engine maps (or similar representations of information that defines the control) which are then stored in the electronic control unit that manages the operation of the engine; indeed, by using this invention, technicians can directly specify the quantity and/or flow rate that must be introduced in an internal combustion engine which is situated in a certain operating condition, without having to specify control variables such as the time of activation (energization) of the injector which, as is known, does not allow a direct control on the flow rate and/or the quantity of fuel injected into an engine, thus making the control of the flow rate and/or of the quantity of injected fuel imprecise. Further advantageous characteristics of the present invention are the object of the enclosed claims.

These characteristics and further advantages of the present invention will be clearer from the description of an embodiment thereof shown in the enclosed drawings, provided as a merely non-limiting example, in which:

- fig. 1 illustrates a Common Rail injection system comprising an apparatus according to the invention;

- fig. 2 illustrates the architecture of the engine control unit according to the invention included in the system of fig. 1 ;

- fig. 3 illustrates a mathematical model of part of the injection system of fig. 1; - fig. 4 illustrates a block diagram of a system for controlling the quantity of injected fuel, comprising a first embodiment of the unit of fig. 2 which uses the model of fig. 3 ;

- fig. 5 illustrates a first variant of the injection system of fig. 1 comprising a second embodiment of the unit of fig. 2 which uses a mathematical model different from that of fig. 3 ;

- fig. 6 illustrates a variant of the mathematical model of fig. 3;

- fig. 7 illustrates a block diagram of a control system comprising the second embodiment of the apparatus according to the invention which uses the model of fig. 6;

- fig. 8 illustrates a block diagram of a control system comprising a third embodiment of an apparatus according to the invention;

- fig. 9 illustrates a block diagram of a control system comprising a fourth embodiment of an apparatus according to the invention.

The reference to "one embodiment" in this description indicates that a particular configuration, structure or characteristic is included in at least one embodiment of the invention. Therefore, the terms "in one embodiment" and the like, presents in different parts in this description, are not all necessarily referred to the same embodiment. In addition, the particular configurations, structures or characteristics can be combined in any suitable manner in one or more embodiments. The reference numbers used hereinbelow are only for the sake of ease and do not limit the protective scope of the range of the embodiments.

With reference to fig. 1, an injection system 1 according to the invention comprises the following parts:

- a low-pressure fuel pump 2 positioned inside a tank 5 and adapted to generate a pressure capable of transferring the fuel (diesel, gasoline or other) , which is situated inside the tank, towards a high-pressure pump 11;

a fuel filter 3, which is in fluidic communication with the delivery side of said pump 2 and is adapted to remove the solid impurities that are found suspended in the fuel and/or to precipitate, on the bottom thereof, the water that is situated in solution in said fuel;

an injection pump 11 positioned downstream of the filter 11 and adapted to increase the value of the fuel pressure, preferably to several hundred bar, if the fuel is a gasoline, and to a value even greater than 2000 bar if the fuel is a diesel fuel;

a fuel rail 13 (also termed common rail or more simply rail), which is in fluidic communication with the delivery side of the injection pump 11 by means of a feed duct 131, wherein these ducts 13, 131 have mechanical characteristics (such as the material with which they are made, the shape, or other characteristics) such to allow the passage of a flow of fuel that is situated at the pressure generated by the injection pump 11;

a pressure sensor 12, which is at least partly positioned in the rail 13 and is configured for detecting the pressure of the fuel that flows at its interior;

a valve FMV (fuel metering valve) 17 for controlling the flow rate at the pump inlet and preferably capable of damping the pressure oscillations due to the variation of the flow rate of the fuel flow;

at least one electro-injector 14 (which will be referred to in the course of this description with the term 'injector')^ which is in flow communication with the fuel rail 13 via an injection duct 141 which has mechanical characteristics similar to the ducts 13,131, and wherein said injector 14 is configured for injecting upon command, i.e. only when it is activated, a predetermined quantity of fuel into the engine; - engine control unit 16 (ECU, also referred as apparatus for controlling the quantity of injected fuel) adapted to supervise the operation of the engine, wherein said unit 16 is in signal communication, preferably by means of a network of CAN-BUS type or of another type, with at least the injection pump 11, the pressure sensor 12 and the injector 14, in a manner so as to be able to detect a fuel rail pressure p _ra ii _r ex _P (i.e. the pressure of the fuel rail 13) and preferably the number of revolutions n _pump of the pump 11, so as to process, on the basis of control rules that will be better described hereinbelow, an injector activation signal act± _n j and, optionally, a control signal dutycycle _F1 4v for the valve FMV 17 which ensure that a predefined quantity of fuel is injected into the internal combustion engine.

In fig. 1, it is indicated that both the injection pump 11 and the injector 14 are in fluid communication with the fuel rail 13 (high pressure) as well as with the fuel recovery duct 51 (low pressure) . This because an injection system of a Diesel cycle engine is illustrated in fig. 1, in which the pressures within the fuel rail 13 have an order of magnitude of thousands of bars, and therefore they can cause fuel leakage. In addition, the Common Rail injectors for Diesel cycle engines comprise a pilot solenoid valve driven by a solenoid or by a piezoelectric element, wherein said valve is configured in a manner such that, when the injector actuation system is activated, the pilot valve is opened and allows the transit of a certain quantity of fuel coming from the rail 13 (high pressure) which causes the opening of the injector and, therefore, the injection of the fuel into the Diesel cycle engine. This mechanism (necessary for operating at such high pressure values) produces, for each injection, a flow of fuel that is unloaded at low pressure from the pilot valve and which must be recovered by means of the fuel recovery duct 51. It is also indicated that this invention can be applied to Diesel cycle engines as well as to all other engine types (such as the Otto cycle engines, Atkinson cycle engines or engines of another type) that provide for the use of fuel injectors which also may not include the pilot valve.

With also reference to fig. 2, the engine control unit 16 (also referred as apparatus for controlling the quantity of fuel injected into an internal combustion engine) will now be described; such unit 16 comprises the following components:

- processing and control means 161 (also termed computing means), such as one or more CPU, which govern the operation of the injection system 1, preferably in a programmable manner, by means of the execution of suitable instructions;

- memory means 162, such as a random access memory RAM and/or a Flash memory and/or memory of another type, which are in signal communication with the processing and control means

161, and wherein preferably stored in said volatile memory means 162 are at least the instructions which can be read by the processing and control means 161 when the unit 16 is in an operating condition;

- acquisition means 163, such as an interface for a network of CAN-BUS type or of another type, which are in signal communication with the processing and control means 161 and are configured for detecting at least the pressure p _ra ii _r ex _P of the fuel rail 13 and preferably the number of revolutions n _pump of the pump 11 or even for detecting the current signal actually given to the injector;

- actuation means 164, such as an injector drive circuit, which are in signal communication with the processing and control means 161 and are configured for generating an electric current such to be able to energize the solenoid or the piezoelectric element of the injector 14 or to drive the operation of an injector drive circuit outside said unit 16 and/or such to drive the valve FMV 17;

- input/output means (I/O) 165 which can for example be used for connecting said device 16 to peripheral devices or a programming terminal configured for writing instructions (which the processing and control means 161 will have to execute) in the memory means 162 and/or allowing the diagnostics of the injection system 1 and/or of the engine and/or of the entire vehicle on which said apparatus 16 is mounted; such input/output means 16 can for example comprise a CAN-BUS, USB, Firewire, RS232, IEEE 1284 or other adapter;

- a communication bus 167 which allows the exchange of information between the processing and control means 161, the memory means 162, the acquisition means 163, the actuation means 164 and the input/output means 165.

As an alternative to the communication bus 167, it is possible to connect, with a star-like architecture, the processing and control means 161, the memory means 162, the acquisition means 163, the actuation means 164 and the input/output means 165. The injection pump 11 can comprise means for measuring the rotation speed (such as an encoder, a tone wheel or other means) which allow measuring the rotation speed n _pump of the pump 11.

As an alternative to or in combination with this solution, the rotation speed n _pump can also be determined, if the pump 11 is driven by means of a service belt driven by the engine axle, by detecting the number of revolutions of the engine and multiplying it by a mechanical transmission ratio; meanwhile, if the pump 11 is driven by an electric engine - preferably of synchronous type with permanent magnets, whose speed is controlled by the engine control unit 16, preferably by means of an Electronic Speed Controller (ESC) - the rotation speed n _pump of the pump 11 might not be detected physically (i.e. it is not detected by using speed measurement means) but only by means of reading speed information from the memory means 162 (see Fig. 2), since said speed information is generated by the computing means 161 in order to then be sent to the electronic speed adjuster which provides for adjusting the pump 11 speed. In the memory means 162 of the engine control unit 16, engine control information is stored which allows the unit 16 to control the operation of the engine in the different work conditions, associating injection information with each work condition. More in detail, this information is preferably stored in the form of maps which allow associating one or more inputs (such as the rotation speed of the engine, the temperature of the air, the temperature of the cooling water, the position of the gas pedal, the setting selected by the vehicle driver, the position of the butterfly valve if present or other inputs) with the respective injection information (quantity to be injected, nominal rail pressure, number of injections, etc.) .

It is still possible for a technician to store at least part of the engine control information in non-map form (such as in the form of instructions that allows the computation in real time of the injection information), without however moving away from the teachings of the present invention.

The injection information defines the quantity of fuel that must be overall injected in the course of at least one part of an engine cycle; such quantity of fuel to be injected can be represented by means of a numeric value (integer or with floating point) corresponding to the quantity of fuel that must be injected in the course of a part of the engine cycle and, optionally, by a second numeric value which represents the number of injections that must be completed in the course of said part of the operation cycle. In this manner, the unit 16 is capable of computing the trend (i.e. at least one value) of the flow rate of fuel that must traverse the injector 14 in the time interval of said part of the engine cycle. For example, if it is necessary to complete three injections in said operation cycle, the control unit 16 can be configured to determine the trend of said fuel flow rate in a manner such that in the course of the first and/or third injection, a fuel quantity is injected equal to a tenth of the quantity of fuel injected in the course of the second injection.

As an alternative to this representation, the quantity of fuel to be injected into the engine in the course of at least one part of an engine cycle can also be represented by means of a sequence of fuel flow rate values, wherein said sequence allows defining the trend of the flow rate of fuel that must traverse injector 14 in the course of said part of the engine cycle. In this manner, it is possible to define the trend of the flow rate of the fuel, so as to make possible the specification of a number of injections per (part of) arbitrary cycle, wherein in the course of each of said injections the injector 14 injects a pre-established quantity of fuel into the engine.

As already described above, it is indicated that the injection information defines the quantity of fuel that must be injected in the course of at least part of an engine cycle and not of a complete engine cycle. This signifies that the unit 16 can be configured for varying the trend of the flow rate of the fuel flow to be injected in the course of a same engine cycle. In this manner, it is possible to ensure the correct combustion of the injected fuel, since the unit 16 can monitor preferably by means of combustion sensors (which can for example comprise the piezoelectric elements of the injectors or dedicated piezoelectric sensors flush-mounted on the combustion chamber or other sensors) in signal communication with said unit 16 - that the injected fuel is actually burned/ignited and, in the case of poor combustion of the fuel injected in the course of at least one part of the engine cycle, vary the trend of the flow rate of the fuel flow preferably before the conclusion of said operation cycle, though also starting from the next operation cycle, so as to reduce the unburned fuel emissions and extend the operating lifetime of the lubricant. In other words, the acquisition means 163 are also configured for acquiring ignition information detected by the combustion sensors which are configured for detecting the pressure variations in the combustion chamber, and wherein the processing and control means 161 are configured for determining the flow rate of fuel to be injected q _in j also on the basis of said ignition information which describes the evolution of the combustion in the cylinder.

With reference also to fig. 3, a mathematical model M will now be described which is used by the engine control unit 16 for determining the trend of the pressure p _r aii over time in the fuel rail 13 (wherein p _rail is diffentiated from p _ra ii _r ex _P due to the fact that it is determined by means of the model M, while Praii, exp ίs detected by means of the pressure sensor 12) on the basis of the instantaneous flow rate q _pU m _P for the injection pump 11 and of the reference curve of the flow rate of the injected fuel q± _n j which is preferably determined on the basis of the excitation time (ET) of the injector 14 actuation circuit and of the value of nominal rail pressure p _n0 m (wherein ET and p _n0 m are stored in the maps of the engine control unit); in addition, since the model M is preferably used for Diesel cycle engines, the model M described hereinbelow determines the trend of the pressure p _ra ii over time, also on the basis of the trend over time of the flow rate of the pilot valve q _pv of the injector 14. Therefore, the engine control unit 16 determines the trend over time of the pressure p _ra n by means of the model M on the basis of the instantaneous flow rates qi _njf q _P ump and preferably also of the flow rate q _pv .

The model M is preferably a linear time-invariant model (LTI) with concentrated parameters, where part of the injection system 1 was modeled, from the hydraulic standpoint, as a zero-dimensional chamber network in which the pressure is uniform and can vary only with respect to time. These chambers are connected by means of single-dimensional ducts (these elements have a hydraulic capacity equal to zero in the model) and calibrated orifices. In light of these assumptions, model M is constructed, which will now be described in more detail. As can be observed from fig. 3, some parts of the injection system 1 were modeled in the following manner:

- the injection pump 11 was modeled as an entity having only a hydraulic capacity C _lf wherein a fuel flow enters into said pump 11 through the port i having a flow rate q _pU m _P , and wherein said flow is introduced into the feed duct 131;

- the feed duct 131 is modeled as an entity positioned downstream of the injection pump 11 and having a hydraulic resistance R ₂ and a hydraulic inductance L ₂ in series with each other, wherein a fuel flow transits in said rail 131 having a flow rate q ₂ ;

- the fuel rail 13 is modeled as an entity positioned downstream of the feed duct 131 and only having a hydraulic capacity C ₃ ;

- the injection duct 141 is modeled as an entity positioned downstream of the fuel rail 13 and having a hydraulic resistance R ₄ and a hydraulic inductance L ₄ ;

- the injector 14 is modeled as an entity positioned downstream of the injection duct 141 and having four zero- dimensional chambers connected by three single-dimensional tubes .

More in detail, in the modeling of the injector 14, it is assumed that such injector comprises the following parts:

- an injector inlet modeled with a hydraulic capacity C ₅ ;

- a pilot valve control chamber modeled with a hydraulic capacity C ₇ , wherein a fuel flow enters into said control chamber having a fuel flow rate q ₆ and a fuel flow exits through a port P2 having a flow rate q _pv , wherein said flow with flow rate q ₆ enters into said control chamber by means of a duct which departs from the injector inlet and which is modeled with a hydraulic resistance R ₆ and a hydraulic inductance L ₆ in series with each other;

- an injection chamber modeled with a hydraulic capacity C ₅ , wherein a fuel flow enters into said injection chamber having a fuel flow rate q ₈ and a fuel flow exits having a flow rate q ₁₀ , wherein said flow with flow rate q ₈ enters into said injection chamber by means of a duct which departs from the injector inlet and which is modeled with a hydraulic resistance R ₈ and a hydraulic inductance L ₈ in series with each other, while the flow with flow rate q ₁₀ exits by means of a duct which is modeled with a hydraulic resistance R ₁₀ and a hydraulic inductance L ₁₀ in series with each other;

- a pulverizer modeled with a hydraulic capacity C _{ll r} wherein a fuel flow enters into said pulverizer, by means of the above-described duct, having a fuel flow rate q ₁₀ , and a fuel flow exits through a port P3 having a flow rate q± _n j. It is indicated that, as already described above, the hydraulic inductance L ₆ , the hydraulic resistance R ₆ and the hydraulic capacity C ₇ also might not be present if the injector 14 does not comprise the pilot valve (e.g. if the injector 14 is used in an Otto cycle engine or if the diesel injector is of piezoelectric type with direct actuation) .

A mathematical description of the model M can be carried out by means of differential equations. More in detail, the law of conservation of mass can be applied in the following manner to the parts of the model M that are described by means of a hydraulic capacity C _j with j = {1, 3, 5, Ί, 9, 11} :

where j and q _ne t,j are respectively the pressure and the rate in net mass that enters into the j-th chamber and the derivative of the pressure with respect to time.

The hydraulic capacity C _j can be computed by means of the formula reported below:

wherein V _j and a-,- are respectively the volume of the j-th chamber and the speed of the local sound in the j-th chamber. With regard to the mathematical description of the parts of the model M which are described by means of a resistance R _j and a hydraulic inductance L _j with j = {2, 4, 6, 8, 10} (i.e. the ducts that connect the different chambers), the balance equation of the motion quantity can be applied in the following manner:

wherein q _j is the mass flow rate that flows into the j-th tube, Ap _j is the pressure drop along the j-th tube.

The hydraulic inductance L _j can be computed by means of the formula reported below:

where l _j and A _j are respectively the length and the section of the tube j-th, where the section of a tube with circular form can be computed in this manner A _j = where d ₃ is the diameter of the tube j-th.

The hydraulic resistance R _j is computed in the case that the fuel flow flows into the tubes with a laminar motion. In this manner, it is possible to preserve the linearity of the equations and, therefore, prevent the structure of the control system implemented by the control unit 16 (described hereinbelow) from being excessively complicated; still in this case, the distributed load losses due to friction depend on the average properties (over time) of the fuel and on the geometric characteristics of the j-th tube. Therefore, the processing and control means 161 compute at least one trend over time of a reference pressure p _ra ii by executing a sequence of instructions (which implements the model M) in which the motion of the fuel is considered along at least the injection duct 141 as a laminar motion. In a more complex variant of the algorithm, a modeling of the friction can also be considered that also accounts for the type of flow (laminar or turbulent) .

In addition, the high-pressure hydraulic system represented by means of the model M also accounts for the effect of concentrated losses, which damp the pressure waves, by lengthening the tubes by a suitable quantity with respect to their actual lengths. These concentrated losses correspond to the presence of calibrated orifices or to the resistances encountered by the flow of fuel inside the injectors and in the ports of the fuel rail 13 which connect said rail 13 to the injection duct 141. Therefore, given the heterogeneity of the tubes employed in the feed system 1, the resistance Rj is defined, by means of the Poiseuille formulas, in a different manner for each tube as follows:

where μ and p are respectively the dynamic viscosity and the density of the fuel, so as to make it advantageously possible to model the behavior of the injection system 1 in the presence of fuels of different type and possibly account for thermal effects.

The tubes 2, 4, 6 and 8 preferably have circular sections, while the tube 10 preferably has an annular section which is defined by the inner radius R± _n t and outer radius R _ex t- For the tubes 2, 4, 6 and 8, R _j includes the contribution of the local concentrated resistances, whose effect is taken under consideration by means of the addition of a virtual tube of (reduced) length lj, i _oc and internal diameter dj _rloc .

In addition, since the model M provides that the tubes do not have hydraulic capacity, the volume V _j of each chamber has been increased by a quantity equal to half the (physical) volume of each tube connected to said chamber. In this manner, the reliability of the model M is increased without increasing the complexity thereof, thus allowing the engine control unit 16 to control the quantity of fuel to be injected into the engine in a more precise manner.

Finally, if the injection system 1 in Fig. 1 comprises more than one injector 14, the hydraulic capacity C ₃ can be defined in a manner so as to be a function of the number of injectors of the system 1, so to allow the modeling of multi-injector injection systems advantageously without increasing the number of degrees of freedom (i.e. the size of the state vector) of the model M. From tests carried out by the Applicant, it results that the above-described model M is reliable for providing the trend of the pressure p _ra ii in the fuel rail 13 even in multi-injector systems; indeed, since the injectors do not function simultaneously, the injectors which are not active and the relative feed tubes can be modeled by increasing the value of the hydraulic capacity C ₃ .

When the injection system 1 is in an operating condition, said system 1 receives a flow of fuel through the port Pi (Fig. 3) , while it preferably expels two flows of fuel through the ports P ₂ and P ₃ ; q _pump is the mass flow rate supplied by the injection pump 11, while q _pv and q± _nj are the flows that exit from the injector respectively through the pilot valve and the injection holes of the nozzle. As already mentioned above, the flow rates q _pU m _P , q _pv and q _in j are the forcing terms of the model with concentrated parameters M and, therefore, they are able to vary the internal fluid-dynamic state of the modeled system M, and hence also that of the (physical) injection system 1. The dynamic behavior of the injection system 1, if such system is linear time-invariant and with concentrated parameters, can be summarized by the following system of formulas:

The first vector equation contains the equations present in formulas 1 and 3 expressed in matrix form. The state vector {x} is defined in the following manner:

As can be observed from the above-reported formula, the state vector {x} is constituted by pressures and flow rates and, since the model M comprises six chambers and five tubes, the state vector {x} is a column vector with 11 components

( {x} e ¾ ^llxl ) .

The vector {u} is the vector containing the external forcing terms that operate on the system and it is defined in the following manner:

The matrix A ( [A] e R ^llx11 ) represents the physical system and includes all the hydraulic capacitances C _j , the hydraulic resistances R _j and the hydraulic inductances L _j ,- such matrix is defined in the following manner:

The matrix B ( [B] e 9i ^llx3 ) defines how the forcing terms {u} operate on the physical system, and it is defined in the following manner:

1

0 0

Ci

0 0 0

0 0 0

0 0 0

0 0 0

[ ] 0 0 0 (10)

1

0 0

c ₇

0 0 0

0 0 0

0 0 0

1

0 0 The second vector equation of formula 6 is tied to the strategy of control of the injection system 1 employed by the control unit 16; such equation which comprises the terms {y}, [C] and [D] , where the vector {y} contains the observation variables, i.e. the parameters of the injection system 1 that must be controlled or at least monitored which, in this case, are reduced to the pressure p _ra ii of the fuel rail 13, that is

{y} => y ₁ =p ₃ . Consequently, the matrix [C] ([c]eR ^lx11 ) becomes a row vector which is defined in the following manner:

[c] = [0 0 1 0 0 0 0 0 0 0 0] (11)

The matrix [D] is instead zero, since none of the entering signals relative to the vector {u} has an effect on the observation variable p _r aii (p3) without simultaneously also modifying the state variables contained in the state vector { x } .

On the basis of Formulas 6-10, it is possible to obtain the explicit expression of the transfer function ( [G] e R ^lx3 ) that defines p _ra ii as a function of each component of the vector {u} in the frequency domain. Indeed, if {u}= {u}e ^!ffli is a harmonic function, the transfer function [G(ico)] e C ^lx3 ) is also harmonic, which is defined in the following manner:

[GM] = {[0](,ω [l] - [A] [B]) (12)

The transfer function [G(ico)] as defined above can be reformulated in a manner so as to underline the different components that contribute to the observable variable p _ra ii in a distinct manner.

[G(iw)] := {G _pufflp (i«), G _pv (i«), G _inj (i«) } (13)

Since all the entering signals relative to the vector {u} simultaneously operate on the injection system 1, the relation between the injection flow rate q _inj and the pressure p _ra ii can be similar to that of a Multi-Input Single Output system (MISO) and, therefore, the transfer function (in the frequency domain) G _inj between Q _inj and P _ra ii {Qi _nj and P _ra ii respectively correspond in the frequency domain to g _inj (t) and p _aiI (t)) can be expressed in the following manner:

Qin ^{j la} >)

The function G± _nj is a Single Input Single Output transfer function (SISO) which connects the injected fuel flow rate ζ)± _η3 to the pressure of the P _ra ii of the fuel rail 13 in the frequency domain (by executing the inverse Fourier transform of P _ra ii it is then possible to obtain the trend of p _ra ii as a function of time) , while the second expression accounts for the effects that the other two forcing terms, i.e. Q _pump and Q _pv , have on the injected fuel flow rate Q± _nj .

With reference also to fig. 4, a preferred control system SC will now be described which allows the control unit 16 to control the quantity of fuel introduced by the injector 14 into the engine by means of only controlling the pressure of the fuel rail 13.

The control system SC receives in inlet the desired flow rates q _inj and q _pvf and by means of the transfer function G _inj such flow rates are used by the engine control unit 16 for computing a reference pressure p _ra ii (t) of the fuel rail 13, where t is the time. As already described above, in order to carry out the computation of the p _ra ii, one also uses the flow rate q _pump which is preferably computed by means of a transfer function H2, since the injection pump 11 is preferably of volumetric type and, therefore, allows knowing the flow rate of the fuel that flows at its interior on the basis of its number of revolutions n _pumPr which is detected by the acquisition means 163 on the injection system 1 (represented in fig. 4 by the block S), and (optionally) also on the basis of the nominal rail pressure p _n0 m (this is the value reported in the engine maps) . It is indicated that if the injection pump 11 is controlled by the ECU by means of a fuel metering valve FMV placed at the inlet of the pumping chamber and integrated in the pump, then q _pUmp also depends on the work position of the FMV.

Once the p _r aii is computed, the control unit 16 computes an average deviation £ _ra n (also termed error information) between said pressure p _ra ii and the pressure p _ra ii _r ex _P detected by the pressure sensor 12 in fig. 1 (represented in fig. 4 by the block HI); after this, the average deviation £ _ra n enters into a regulator Rl (implemented by the control unit 16) which, on the basis of said deviation £ _ra n, determines whether or not to correct the state of the injector 14 by generating the opportune injector activation signal act _in j .

The regulator Rl can be of PI (Proportional Integral) or PID (Proportional Integral Derivative) type, but it could also be of MPC type or another type. It is evidenced that the teaching provided by this invention does not regard how to make the controller Rl nor how to control the pressure in the rail 13, but rather it regards how to use said controller Rl to control the pressure in the fuel rail 13 as well as the quantity of fuel injected by the injection system 1 into the engine.

It is evidenced that the flow rate q _pv is tied to the flow rate qinj by a transfer function which can be determined with hydraulic test stand of the injectors on the basis of the technical characteristics provided by the manufacturer of the injector 14; therefore, q _pv could be computed on the basis of the desired flow rate gi _nj and, therefore, the reference pressure p _ra ii can be computed on the basis of only the flow rate qinj, since the flow rate q _pv is a function of the flow rate qi _n j . Alternatively, one could provide for measuring q _pv by installing an instantaneous flow rate meter on the fuel recirculation tube from the pilot stage of the injector to the tank .

When the injection system 1 is in an operating condition, the engine control unit 16 executes a method for controlling the quantity of injected fuel according to the invention comprising the following phases:

- acquisition phase, in which the pressure of the fuel rail 13 is detected by means of a pressure sensor 12;

- an error computation phase, in which, through the processing and control means 161, a piece of error information {£ _rail , £ _in j , S _q t _y _inj) is computed on the basis of at least the following information:

o reference information, which preferably comprises the reference pressure p _ra n or an average value thereof, determined on the basis of a flow rate gi _nj or a quantity qty _iri j of fuel to be injected into the engine;

o feedback information determined on the basis of the fuel rail pressure p _ra ii _r ex _P (e.g. by computing the average pressure p _ra ii _r ex _P in a certain time interval) .

- control phase, in which it is determined, through the computing means 161, whether or not to vary the electrical excitation time (ET) of the injector, which varies the fuel rail pressure p _ra n, ex _P , through the actuation means 164 so as to keep the amplitude of the error information £ _ra n below a certain threshold.

It is evidenced that an increase of the electrical excitation time ET of the injector 14 through the actuation means 164 causes a decrease of the pressure p _ra ii _r ex _P in the fuel rail 13, while the reduction of said excitation time ET causes the rise of said pressure p _ra ii _r ex _P . In this manner, it is possible to control the pressure p _ra n _r ex _P and, therefore, also the flow rate qinj by means of the greater or lower activation of the injector 14.

Of course, numerous possible variants are possible of the embodiment discussed up to now. A first variant is that illustrated in fig. 5 in which the injection system 1' is illustrated; for the sake of brevity in the following description, only the parts which differ this and the subsequent variants with respect to the main embodiment described up to now will be highlighted; for the same reason, the same reference numbers will be used where possible, with one or more primes for indicating elements structurally or functionally equivalent to the new variants. The injection system 1' comprises a control unit 16' similar to the above-described control unit 16 and, as an alternative to or in combination with the valve FMV 17, a pressure control valve 15 installed on the rail (PCV) in signal communication with said control unit 16' and positioned in a manner so as to place the fuel rail 13 in fluid communication with a fuel recovery duct 51 (which has a pressure close to that of the tank 5) when the pressure of the duct reaches a certain pressure threshold.

The valve 15 (PCV) preferably comprises the following parts:

- a seal piston, on which a force is exerted by the pressurized fuel contained in the rail 13;

- a preload spring which operates on said piston, generating a force capable of overcoming the force generated by the fuel in the rail 13 when this fuel is situated at a pressure typically less than 100 bar (for a Diesel cycle engine) ;

- a coil which, when traversed by an electric current, is capable of generating a magnetic field that produces a force on the piston which, combined with that generated by the preload spring, is capable of overcoming the force generated by the fuel when this is situated at a pressure greater than 100 bar and up to over 2000 bar.

In summary, the greater the average current that circulates in the coil, the greater the intensity of the electromagnetic force on the seal piston and the greater the pressure threshold beyond which the piston changes position and places the fuel rail 13 in fluid communication with a fuel recovery duct 51.

This pressure threshold is settable by the control unit 16' and allows varying the level of the nominal pressure p _n0 m reachable by the rail 13 so as to make it possible to control the pressure thereof in the injection system.

It is evidenced that this control unit 16' can be used alternatively or in combination with the preceding described embodiment .

With reference to figures 6 and 7, a control system SC will also be described that is similar to the control systems SC previously described in figs. 3 and 4.

The control unit 16' is configured in order to execute a model M' (fig. 6) which is similar to the previously described model M, but which additionally models the pressure control valve 15. In such case, it is assumed that fuel metering valve FMV is not present and that the pressure level is controlled by the valve 15. In this manner, a term is added that is considered forcing with respect to the model of the injection system 1, along with a port P ₄ through which a fuel flow exits in order to enter into the fuel recovery duct 51, making the trend over time of the pressure p _r aii function not just with Qpumpr Qpv and pinjr but also with a pressure control flow rate q _P cv that represents the flow rate of fuel which flows through the pressure control valve 15 towards the duct 51.

In order to model this new forcing term of the model M', the formulas 1, 8, 10 and 14, which define the model M used in the preceding embodiments of the invention, are modified in the following manner: g net, j

P _j =

Qne t,l g pump - g ₂

Qne t,9 = g ₈ - gio

Qne t,l l = gio ^{■ ~~} Qinj 'Jpump 'Jpv Qinj 'Jpcvr

0

0 0 0 0

0 0 0 0

[ ] 0 0 0 0 (10 '

1

0 0 0

c ₇

0 0 0 0

0 0 0 0

0 0 0 0

G _pump ji^Q _pump ji ) + G _py {i ) Q _py {iω) + G _fcy (io)Q _fcy (i<a) (14')

When the injection system 1' is in a certain operating condition, the control unit 16' is capable, preferably by executing a sequence of instructions which implement the model M', of determining the trend of the reference pressure p _ra i i on the basis of the flow rates q _pump , q _pvf q± _n j and q _PC v -

The unit 16' comprises, in addition to the regulator Rl of the first embodiment, also a second regulator R2 which is configured for generating the valve control signal du tycycl epcv _r which actuates, through the actuation means 164, the pressure control valve 15, preferably specifying the position of said valve, on the basis of the pressure p _ra ii _r ex _P detected by the pressure sensor 12 so as to keep the latter as close as possible to a nominal rail pressure p _n0m - Such nominal rail pressure p _n0m is preferably the ideal work pressure of the injection system 1', i.e. the pressure for which such system 1' was designed and is preferably defined by the injection information; in practice, p _n0m can coincide with the average integral of p _ra n or with its maximum value.

As for the regulator Rl, the regulator R2 can be of PI (Proportional Integral) or PID (Proportional Integral Derivative) type, but it could also be of MPC type or of another type.

The valve control signal du tycycl e _PC v is then read and used, preferably together with the nominal rail pressure p _n0m , by the computing means of the unit 16' in order to determine the flow rate q _PC v by means of a transfer function H3 preferably implemented by a sequence of instructions which are executed by the computing means 161.

It is evidenced that the transfer function H3 could also be substituted by a flow rate meter configured for measuring the average flow rate of fuel that flows through the pressure control valve 15.

The flow rate q _PC v is then used by the computing means for computing a reference pressure p _ra ii by means of the transfer function Gi _nj ' reported above. As can be observed by the formula 14', such transfer function Gi _nj ' takes into account the transfer function G _PC v of the pressure control valve 15 and the trend of the flow rate through said valve 15 in the frequency domain .

The presence of the PCV ensures an improved dynamic response of the injection system during engine transients with respect to the case of the FMV, making possible an accurate control of the fuel quantity injected by the injector 14, with respect to the case of the FMV.

With reference to fig. 8, a third embodiment of a control unit 16'' according to the invention will now be described; such apparatus 16'' is comprised in a control system SC' and differs from the units 16 and 16' of the preceding embodiments with regard to the computation of the error information. This is made not on the basis of the reference pressure (generated by the transfer function G _in j) and the pressure p _ra ii _r ex _P of the fuel rail, but rather on the basis of the desired flow rate qinj (which is included, as with the other embodiments, in the injection information) and of an estimated flow rate q± _n j*.

The flow rate q± _n j* is estimated by means of a transfer function F _ln1 reported below:

( G _pUmp (i» ) Q (ί ω) + G (ico ) Q (ίω)

¾ ( ^iG) ) (15)

G

( ^< °H " rail, exp (ίω)

This transfer function F _in ^ is obtained by solving the formula 14 with respect to Q _in j in a manner such that Q _ln j* = F _inj P _rallfeKp .

As can be observed, F _inj allows estimating the flow rate q± _n j* of fuel injected by the injector 14 on the basis of the forcing terms q _pU m _P , Praii _r ex _P and (optionally) q _pv (in the frequency domain Q _pump , P _ra ii _r ex _P and Q _PV ) ; such terms are determined/detected as in the second embodiment of the invention, i.e. p _ra ii _r ex _P is detected by means of a pressure sensor 12 (represented in fig. 8 by the block HI), q _pump is determined by means of a transfer function H2 which allows computing said flow rate q _pU m _P on the basis of the number of revolutions n _pump of the injection pump 11 and possibly of p while q _pv is for example determined on the basis of a transfer function which can be determined on the basis of the technical characteristics provided by the manufacturer of the injector 14. As an alternative to that just described above, the transfer function F _in - can also be obtained by solving the formula 14'

(and not the formula 14) with respect to Qi _n jr in a manner so as to control the quantity of injected fuel also on the basis of the position of the valve PCV 15, so as to improve the control precision.

Once the flow rate of fuel considered to be injected q _inj is defined, the control unit 16' ' computes an average deviation £inj between said flow rate of fuel to be injected q _inj and the flow rate q± _n j* estimated by means of the function F _inj ; after this, the average deviation £ _inj enters into a regulator R3 (implemented by the control unit 16' ') which, on the basis of said deviation Sinj, determines the activation signal acti _nj that operates on the injector 14.

The regulator R3 can be of PI (Proportional Integral) or PID (Proportional Integral Derivative) type, but it could also be of MPC type or another type. It is evidenced that the teaching provided by this invention does not regard how to make the controller R3, but rather how to use said controller R3 for controlling the quantity of fuel injected by the injection system 1 into the engine.

When the control system SC ' is in an operating condition, the processing and control means 161 of the apparatus 16' ' cyclically execute the following steps:

- determine the flow rate of fuel to be injected qi _nj on the basis of at least the injection information contained in the unit 16' ' ;

- determine the error information S _ln , for example by computing the difference between the flow rate of fuel to be injected q _in j (reference signal) and the estimated fuel

- determine whether to activate (and for how much time) or deactivate the injector 14 (e.g. by increasing or decreasing the energization time of the injector 14 with respect to the nominal value reported in the engine maps) on the basis of the error information 8 _lnj and of the transfer function (control law) implemented by the controller R3;

- detect the pressure by means of the pressure sensor 12;

- detect/determine the instantaneous flow rate q _pU m _P of the injection pump 11;

- updating the estimated fuel flow rate qinj* by executing a set of instructions that implements the transfer function F _inj .

This embodiment allows varying the flow rate of injected fuel in a quicker manner with respect to the preceding embodiments, i.e. it is able to reduce the time between a variation of the desired flow rate gi _nj and the consequent activation or deactivation (or correction of the energization time) of the injector 14, thus improving the control of the quantity of injected fuel; indeed, the unit 16'' carries out the computing of the estimated flow rate q± _n j* downstream of the feedback branch, in a manner so as to reduce the response time to a variation of the desired flow rate qinj, since the reference signal qi _n j does not have to first be transformed into pressure, as instead occurs for the previously described embodiments, but it is the feedback signal that is transformed by means of an opportune transfer function into a value of flow rate qi _n j* comparable with that of the desired flow rate

With reference to fig. 9, a fourth embodiment of a control unit 16''' according to the invention will now be described; such apparatus 16' ' ' is comprised in a control system SC ' ' and differs from the units 16'' of the preceding embodiment just described above due to the possibility of computing a piece of error information £ _{qty inj} - not on the basis of the desired flow rate q _in j and an estimated flow rate qinj*, but rather on the basis of the quantity to be injected qty _inj (integral over time of qinj) r preferably in the course of an operation cycle or of a single injection, and the estimated injected quantity qty _in j* (integral over time of q± _n j*) , preferably in the course of the same operation cycle or of the same single injection.

Once the quantity of fuel to be injected qty _in j is defined, the control unit 16''' computes a deviation £ _gty _i _nj between said quantity of fuel to be injected qty _in j and the estimated quantity qtyin _j * by means of the integration in a time interval of the result of the function q _in j* (which is obtained by anti- trans forming the function Qinj*) , i.e. of the estimated flow rate; after this, the average deviation £ _inj enters into a regulator R4 (implemented by the control unit 16''') which, on the basis of said deviation £ _{qty in} j, determines whether or not to activate the injector (e.g. by correcting the energization time ET of the injector) 14, by generating the injector activation signal act _in j ·

The regulator R4 can be PI (Proportional Integral) or PID (Proportional Integral Derivative) type, but it could also be of MPC type or another type. It is evidenced that the teaching provided by this invention does not regard how to make the controller R4, but rather it regards how to use said controller R4 to control the quantity of fuel injected by the injection system 1 into the engine.

For such purpose, the processing and control means 161 of the apparatus 16''' are configured for determining the quantity to be injected tyinj on the basis of at least the injection information, and for determining the estimated injected quantity qty _in j* by computing the integral over time (e.g. by means of a summation operation) of the estimated flow rates qinj* by means of the transfer function F _lnj .

Since the value of estimated injected quantity qty _in j* is representative of only one operation cycle of the propulsor or of a single injection, such integration operation must preferably start from the beginning of each injection or at least of each operation cycle of the propulsor; for such purpose, the value of the estimated injected quantity qty _in j* _r which is obtained by means of process of integration of the estimated flow rate qinj*, must be zeroed at the start of each injection or at least of each propulsor operation cycle, e.g. by generating an initialization signal RST.

When the control system SC ' ' is in an operating condition, the processing and control means 161 of the apparatus 16''' cyclically execute the following steps:

- determine the quantity of fuel to be injected qty _inj on the basis of at least the injection information contained in the unit 16' ' ' ;

- determine the error information £ _{qty inj} -, for example by computing the difference between the quantity of fuel to be injected qtyinj and the quantity of estimated injected fuel qty _inj *;

- determine whether to activate (and for how much time) or deactivate the injector 14 (for example by increasing or decreasing the energization time of the injector 14 with respect to the nominal value reported in the engine maps) on the basis of the error information £ _{qty inj} and of the transfer function (control law) implemented by the controller R4 ;

- detect the pressure by means of the pressure sensor 12;

- detect/determine the instantaneous flow rate q _pU m _P of the injection pump 11;

- update the estimated fuel flow rate q± _n j* by executing a set of instructions which implements the transfer function F _in - ; - zero the value of the estimated injected quantity qty _in j* if the propulsor is at the start of a new operation cycle and/or if it is at the start of a new injection (for example by raising the logic level of the signal RST) ;

- update the value of the estimated injected quantity qty _in j* by integrating the values of the estimated fuel flow rate qi _nj* over time that were determined in the course of the current operation or injection cycle.

If it is desired to reduce the computation time necessary for updating the value of the estimated injected quantity qty _inj , it is possible to add, to the current value (computed up to that moment) of said estimated quantity gty _inj* , the time integral of the estimated fuel flow rate qi _nj* computed over a sufficiently small time interval At, thus carrying out an incremental computation of the estimated quantity qty±„ _j* .

This embodiment allows defining the quantities of fuel to be injected in a simpler manner with respect to the preceding embodiments, since the control system SC ' ' controls the quantity of fuel to be injected and no longer controls the fuel flow rate. In this manner, the generation of the injection information is simplified, thus simplifying the adjustment of the propulsor.

As an alternative to the models M and M', the Applicant has observed that by using a spectral model, which will now be described, it is possible to more quickly obtain an alternative model with respect to said mathematic models with concentrated parameters Μ,Μ', having a high level of immunity to the disturbances mainly caused by resonance phenomena due to the geometric and structural characteristics of the injection system 1,1'.

More in detail, the spectral model is obtainable starting from a series of measurements obtained by means of a test stand of the injectors, wherein each measurement comprises a duct pressure measurement and a flow rate measurement, and wherein all the pressure and flow rate measurements respectively form the (point) functions in the time domain p(t) (equivalent to

Prail , exp ) and q (t) (equivalent to q± _n j) ·

Once these measurements are obtained, said functions p(t) and q(t) are preferably multiplied by a window function w (t) in the following manner: where w(t) is defined in the following manner: w(t) = ∑w _T (t + k(T - At _ov )) (17) k = 0

where W _T (t) is a window function, preferably the Hanning window function, T is a time interval much smaller than the rotation period of the injection pump 11 (which is preferably equal to double the rotation period of the internal combustion engine fed by the injection system 1,1'), K is the number of window functions that one wishes to use, while At _ov is the superimposition time for two subsequent windows, which is preferably greater than half the time interval T. By using the function w(t) thus defined, the Applicant observed that the frequency content of the functions p(t) and q(t) is minimally altered, thus permitting a spectral estimate with a high level of significance.

After this, it is possible to compute in the frequency domain, on the basis of the signals p„(t) and q _w (t), the power cross- spectral densities S _QP (ω) , S _PQ (ω) and the power auto-spectral densities S _pp (ω) , S _QQ (ω) of said signals p _w (t) and q _w (t):

Ξ _ΡΡ (ω) := \_ ⁺ κ _ρρ (τ) θ- ^ιωτ άτ,

where _R _PP is the autocorrelation function of p _w (t), R _qq is the autocorrelation function of q _w (t) and R _pq and -R _gp are the cross- correlation functions between p _w (t) and q„(t) and T _d is the integration domain.

Finally, it is possible to estimate the transfer function G _ln j{o) on the basis of the power auto-spectral and cross- spectral densities, e.g. by means of the Shin-Hammond estimate which is summarized in the formula reported below: where S _pp , S _QQ and S _PQ are estimates of S _pp (ω) , S _QQ (ω) and S

Such estimates can be evaluated as follows:

Q _k * ((D _f Τ) Q _k (f , Τ)

¾ο ( ^ω ) ~ ¾ _Q (co )

κ

∑-

S _QP (CD) « ¾ _P (co ) where P _k (o) and are the fast Fourier transforms (FFT) of the signals p„(t) and q„(t) in the interval T, whose initial instant depends on the value of k, while P _k ^* (ω) and Q _k ^* (ω) are the conjugated complex functions of P _k (ω) and Q _k (ω) .

In this manner, it is possible to obtain, in an automated manner, the transfer function G _lnj {o) and, therefore, a model which in addition to being an alternative to the models with concentrated parameters Μ,Μ', is also capable of offering a high level of immunity to disturbances, thus simplifying the step of adjustment of an internal combustion engine.

Some of the possible variants have been described above, but it is clear to the man skilled in the art that, in the practical achievement of the invention, other embodiments exist, with different elements that can be substituted by other technically equivalent elements. The present invention is therefore not limited to the above-described embodiments, but is susceptible of various modifications, improvements, substitutions of parts and of equivalent elements without deviating from the underlying inventive idea, as specified in the following claims.