INTERNAL COMBUSTION ENGINE AND A SUPPLY SYSTEM IMPLEMENTING SAID METHOD

Cross-reference to related applications

This application claims priority from Italian Patent Application No. 102018000003377 filed on 08/03/2018, the disclosure of which is incorporated by reference.

Technical field of the invention

The invention relates to the field of methods and systems for managing a supply of a spark ignition engine.

State of the art

In spark ignition engines, also known as Otto engines, the respect of the stoichiometric ratio is crucial in order to observe the limits set by laws on polluting emissions.

This is also important when trying to ensure uniform performances in terms of torque delivery, regardless of unpredictable changes in the quality of the fuel used.

This problem is particularly urgent in the field of engines supplied with gas, for example natural gas, commonly known as methane, which can be stored in the tanks in a gaseous form (CNG) or in a liquid form (LNG) .

Gaseous or liquid fuels have a strong composition variability, which often makes it hard or even impossible to obtain, especially in a short time, a correct stoichiometry of the combustion.

Drivers simply need to change filling station in order to find fuels that are very different from one another and, hence, require a different metering during the relative injection into the cylinders of the internal combustion engine .

The quantity of fuel to be injected depends on the mass of fresh air introduced into the engine, which is measures by means of a dedicated sensor arranged on the intake manifold.

The fuel supply of the internal combustion engine can be operated by means of a closed loop control, which uses a linear lambda sensor as a feedback signal in order to obtain a correct air/fuel ratio (A/F) .

Linear lambda sensors are precise and fast, allow the supply of the internal combustion engine to be properly corrected and ensure the stoichiometry of the combustion. However, linear lambda sensors are technically complicated and, therefore, expensive. Furthermore, they require a dedicated software for the conditioning of the relative generated signal.

Binary lambda sensors, commonly known as ON/OFF sensors, are also known. They do not need a dedicated sensor and their implementation definitely is cheaper.

These ON/OFF sensors also require a closed loop control scheme, forcing the air/fuel ratio (A/F) to oscillate around the stoichiometric value, thus obtaining - as feedback - the switching of the signal generated by the binary sensor. This A/F forcing operation modulates the supply of fuel, increasing or decreasing the quantity of fuel injected, so as to alternatively cause the combustion i taking place in the cylinder to lack oxygen or have excess oxygen.

This forcing operation can be a multiplying non-dimensional correction of the reference signal controlling the mass of fuel to be injected. If its value is 1, there basically is no correction, whereas, if its value is more than 1 or less than 1, there is a correction of the mass of fuel to be injected relative to a reference value, in order to reach a reference air/fuel ratio.

Therefore, the three-way catalyst is forced to work on a stoichiometric condition, in a bi-stable situation, namely continuously oscillating between an oxidizing operation and a reducing operation.

Examples of this concept can be found in US2008319634 and W02010097267.

The closed loop control is carried out by means of a PI (proportional and integral) controller, which corrects the supply of the internal combustion engine based on the switchings performed by the binary lambda sensor upstream of the three-way catalyst.

US4655188 shows a feedback control, in which alpha is a feedback correction control in the formula

3Ϊ- T xCmFXa^ Ts an Tp^Kx Q/N wherein Tp is a "base" fuel injection signal and Ts is a correction of the electric signal due to other factors. Alpha is the median of a plurality of alpha_zero correction values close to the switchings of the binary sensor.

Summary of the invention

The object of the invention is to provide a method for managing a fuel supply of a spark ignition engine, which is capable of better adjusting to the variability of the fuel composition using a binary lambda sensor.

The idea on which the invention is based is that of sampling the air/fuel ratio measured in at least two consecutive switchings of the binary sensor and of calculating an average value. The closed loop control is forced to cause the air/ fuel ratio to converge towards said average value.

Therefore, according to the invention, the air/fuel ratios are sampled in two consecutive switchings, which corresponds to sampling the control signals of the injectors, which has nothing in common with the sampling of the feedback variable.

In other words, whereas in the invention the control is based on the air/fuel ratios, in US4655188 the control is based on the feedback variable. In addition, the invention is independent of the implementation of a feedback control on of the possible features thereof.

Indeed, no variable measured on the exhaust line is acquired, but only the switching instants of the binary sensor are acquired. As a consequence, the control according to the invention, despite being capable of being combined with a traditional feedback control, can actually be compared with an open loop control.

If the air mass flowing into the engine is decided by the driver, then the control manages the supply of fuel based on the air flowing into the engine. If, on the other hand, the throttle valve is motor-driven, then the control can also manage the opening of the throttle valves, besides the injection of fuel.

When dealing with a control carried out by means of a processing unit and of electric injectors, the quantity of fuel clearly corresponds to an electric control signal of the injectors.

The (electric) control signal of the injectors injecting fuel into the cylinders depends on the stoichiometry value towards which the control converges.

Therefore, this average value, hereinafter referred to as

"learned stoichiometry", alters the average quantity of fuel to be injected in a predetermined operating condition of the engine.

The more the composition of the fuel changes relative to an ideal composition, the higher the aforesaid alteration of the mass of fuel to be injected is relative to an ideal condition .

In other words, the reference quantity of fuel to be injected is pre-calculated in ideal conditions and, therefore, is referred to as "nominal stoichiometry".

When the closed loop control involves the use of a multiplying coefficient, referred to as A/F forcing factor, to correct the air/fuel ratio, the method can alternatively comprise the sampling of said multiplying coefficient during said two switchings of the binary sensor and the subsequent calculation of the relative average value of the multiplying coefficient.

The fact of sampling the air/fuel ratios during the switchings of the binary sensor is independent of the type of open loop control adopted.

When the engine point remains stationary for a predetermined amount of time, said "learned stoichiometry" is preferably calculated as the mobile average of a plurality of consecutive switchings, preferably an even number of consecutive switchings. An even number of switchings leads to an advantageous effect, which allows the learning strategy to be simplified. Indeed, in this way, possible delays between the calculation of the A/F forcing factor and the moment of the rich/lean switching of the sensor itself are cancelled out. These delays affect both the rich state, thus causing an overestimation of the air/fuel ratio, and the lean state, thus causing an underestimation of the air/fuel ratio. An even number of samples causes the two delays, leading to an overestimation and to an underestimation of the air/fuel ratio, to compensate one another.

According to a first aspect of the invention, said "learned stoichiometry" is calculated in an operating point of the engine and reflects on the entire operating interval of the internal combustion engine.

For example, a change in the air/fuel ratio can be calculated relative to a reference value of the calculation engine point and the change can be applied to the entire operating interval of the engine or, if the system is working in terms of A/D forcing factor, the entire operating interval of the engine can be subjected to an intervention in terms of A/F forcing factor.

This advantageously leads to a quick adjustment of the injection times of the engine to the changes in the chemical composition of the fuel, so that the internal combustion engine can perfectly work astride a stoichiometric combustion.

The adjustment is preferably operated as soon as a filling of the tank with gaseous or liquid CNG or LNG fuel is detected .

According to a preferred variant of the invention, a 2D map is stored in a memory allocation of the processing unit controlling the fuel supply of the internal combustion engine. This 2D map expresses a reference air/fuel ratio or a correction thereof as a function of the "engine point", namely of the revolutions/minute of the drive shaft and of the delivered torque.

Therefore, the 2D map is provided with a plurality of cells. Similarly, said 2D map can store a central value of said A/F forcing factor, around which it oscillates, in order to determine the oscillation of the air/fuel ratio. Advantageously, when a refuelling procedure is detected, the fuel is renewed with a homogeneous chemical composition and the procedure described above is carried out, thus changing the air/fuel ratio or the central value of the A/F forcing factor in the entire operating field of the internal combustion engine, hence in all the cells making up said 2D map.

According to a preferred variant of the invention, a second correction process is carried out in each engine point, for example in each cell of the aforesaid 2D map, when the engine operates in stationary conditions for a predetermined amount of time. This correction is not applied to all the cells of the 2D map in an indiscriminate manner, but only to the cell corresponding to the engine point in which the process to calculate the aforesaid average value was carried out, and, if necessary, is extended to the cells adjacent to said cell, by means of decreasing weighing factor.

Therefore, whereas the first process preferably takes lace after a refuelling operation and affects the entire map, the second process takes place when the first process has ended and affects the sole cell corresponding to the relative engine point or, at the most, the cells nearby.

The claims describe preferred embodiments of the invention, thus forming an integral part of the description.

Brief description of the figures

Further objects and advantages of the invention will be best understood upon perusal of the following detailed description of an embodiment thereof (and of relative variants) with reference to the accompanying drawings merely showing non-limiting examples, wherein:

figure 1 schematically shows a spark ignition internal combustion engine provided with relative supply means and with an exhaust line, on which there are arranged a three- way catalyst and a binary lambda sensor according to the invention;

figure 2 shows a time diagram of the switchings due to the implementation of a binary lambda sensor and of the method according to the invention;

figure 3 shows a 2D map where reference values are stored, which relate to injection times or to the mass of fuel to be injected in each combustion cycle of a cylinder; figure 4 shows the 2D map of figure 3 displaying a calculation cell and relative adjacent cells;

figures 5, 5a and 5b respectively show a flowchart explaining the control method according to the invention, declined in the two variants of figures 5a and 5b, wherein the variant of figure 5a is preferably carried out when a refuelling procedure is detected, whereas the variant of figure 5b is carried out when the variant of figure 5a is deactivated .

In the figures, the same numbers and the same reference letters indicate the same elements or components.

The blocks of figure 2 represented with a broken line are to be considered as optional.

The broken-line connections of figure 1 represent analogue or digital electric lines.

For the purposes of the invention, the term "second" component does not imply the presence of a "first" component. As a matter of fact, these terms are only used for greater clarity and should not be interpreted in a limiting manner.

Detailed description of embodiments

Figure 1 schematically shows a spark ignition internal combustion engine E provided with supply means J, for example fuel injectors.

The engine E is provided with an exhaust line, on which which a three-way catalyst 3WC is housed. Upstream of the catalyst there is, along the exhaust line, a binary lambda sensor B, which is adapted to switch between two logical conditions indicating a lean and a rich combustion, respectively .

A processing unit CPU monitors the control of the internal combustion engine E and, in particular, controls the injection of fuel into the cylinders of the engine by means of relative injectors J.

Said processing unit implements a closed loop control over said binary lambda sensor B, causing the supply of the internal combustion engine to continuously oscillate around the stoichiometric air/fuel ratio (A/F) .

The processing unit, when a switching of the level of signal generated by the lambda sensor is detected, reverses the sign of the correction by means of the aforesaid "A/F forcing factor", until it detects a new switching of the level of signal generated by the sensor. The forcing factor, if expressed as a multiplying coefficient, oscillates relative to a central value, which, in ideal conditions, is 1. Therefore, 1.01 and 0.99 determine a rich condition or a lean condition of the air/fuel ratio.

In the processing unit there is stored a reference value of the fuel to be injected into the internal combustion engine .

The supply managing method according to the invention comprises a proportional/integral closed loop control over a binary signal (ON/OFF) generated by the binary lambda sensor, the method comprising a procedure for continuously changing an air/ fuel ratio between a lean combustion condition and a rich combustion condition and vice versa, when a switching of said binary signal is detected, the method comprising a procedure carried out in stationary operating conditions of the internal combustion engine and consisting of:

- a first step of registering at least two values of said air/ fuel ratio lΐ and l2 for as many consecutive switchings ,

- a second step of calculating an average value Am of said two registered values,

- a third step of setting said calculated average value as a reference value Ref of the air/fuel ratio.

This setting can evidently be gradual, thus determining a convergence, when the method is carried out repeatedly for some processing cycles.

In other words, the result is a zero value or, in case of a correction carried out by means of a multiplying coefficient, to a neutral value, i.e. equal to 1.

The fact of setting said average value as reference value corresponds to calculating a variation D_Ref = m - Ref. Figure 2 shows a diagram of a supply signal of the internal combustion engine over time. The closed loop control, which is based on the binary signal generated by the binary lambda sensor, determines oscillations around a reference value Ref, which was previously stored in the processing unit .

In the case of figure 2, the reference value of the supply signal is greater than the stoichiometric value m for the currently implemented fuel composition.

In the example, the switchings of the binary lambda sensor take place in lΐ and l2, which are equally spaced apart from Am. Therefore, there is a difference between the previously stored A/F ratio Ref and the current one, which takes into account the actual composition of the fuel.

When the engine remains in stationary conditions for a long time, these circumstances can be exploited by calculating a mobile average between the air/fuel ratios of the ON/OFF switchings of the sensor. The processing unit preferably implements a closed loop control scheme with a proportional and integrative controller PI on the feedback branch, as suggested by the prior art documents for spark ignition Otto engines. This scheme, though, is changed as described above.

The reference value Ref changes upon variation of the engine point and is an ideal value, since it depends on an ideal fuel.

Figure 3 shows an example of a 2D map.

A sensor LI is capable of detecting a refuelling operation of the tank SI and a sensor L2 is capable of detecting a refuelling operation of the tank S2 with a gaseous fuel, which supplies the internal combustion engine. Said sensor can detect a sudden pressure increase in the tank of the fuel gas or it can detect the increase of the level sensor of the tank containing the liquid phase fuel.

When, after having detected said refuelling operation, said first procedure is carried out monitoring the engine point and when the latter is deemed to be stable, said correction of the A/F ratio of the internal combustion engine is performed .

The correction is preferably applied to the same extent to all the reference values of the operating dominion of the engine and, therefore, to the entire 2D map described above, when said procedure is triggered by a refuelling procedure, thus defining a so-called "first process".

According to a preferred variant of the invention, after having ended the first procedure triggered by the refuelling procedure, a second procedure starts, which is preferably carried out in a constant manner until a subsequent refuelling operation.

According to a preferred variant of the invention, which can be combined with the previous one, a "second process" is provided, which is carried out when said first process has ended and involves a similar correction of the reference values Ref, cell by cell.

In other words, the correction, unlike what happens in the first process, is not applied to the entire map, but is applied to the sole cell C in which the engine operates and only optionally to near cells. Said cell C is dark in figure 3. Figure 4, on the other hand, shows the cells adjacent to the cell C.

This is a "punctual correction", which is opposed to the "global correction" performed on the entire map.

Depending on the adopted solution, namely on the degree of division of the operating field of the engine, a more or less thick map can be obtained, i.e. a map with cells with a smaller size.

In case of a very thick map, the method can comprise making corrections in the immediately surrounding cells, which define a first level, and in further cells immediately adjacent to the cells of the first level. These further cells define a second level, etc..

In this case, again, the correction is preferably weighed, i.e. proportionally decreased, when gradually moving away from the cell corresponding to the current engine point. This is a "zone" correction scheme and is slower than the "global correction".

According to a further preferred variant of the invention, the processing unit controlling the internal combustion engine comprises a second 2D map implemented for managing the supply of the internal combustion engine when the latter is supplied by a second tank S2 filled with a fuel that is different from the one of the previous tank SI. Therefore, when the supply is switched between two different fuel sources, a relative 2D map is implemented so as to to avoid excessively long convergence times of the first process, at first, and, then, of the second process. Hence, the vehicle can implement an LNG liquid methane tank and also a CNG gaseous methane tank (so-called "dual fuel" systems) .

In other words, when the supply is switched from CNG to LNG and vice versa, the processing unit starts working on the corresponding 2D map.

This variant can be extended based on the number of tanks installed in the vehicle. In other words, for each tank being used there can be a 2D learning map.

This invention can be advantageously implemented by means of a computer program comprising encoding means for carrying out one or more steps of the method, when the program is run on a computer. Therefore, the scope of protection is extended to said computer program and, furthermore, to means that can be read by a computer and comprise a recorded message, said means that can be read by a computer comprising program encoding means for carrying out one or more steps of the method, when the program is run on a computer.

Figure 5 shows a step 0, during which the engine is in stationary conditions, which means that the number of revolutions and the delivered torque are constant for a predetermined amount of time.

Subsequently, the above-mentioned steps 1 - 3 are carried out and the process starts again from the beginning.

In case of implementation of the first process, which is shown in figure 5a, a refuelling procedure has just been preliminarily performed by means of the step F.

The method can be carried out only once o can be carried out for a predetermined number of times by means of an inner loop with a counter, which joins the input of step 3-

1 to the input of step 0. Step 3-1 specifically involves changing the reference value of the air/fuel ratio in the entire operating dominion of the engine .

On the contrary, step 3-2 of figure 5b, which represents the second process, only involves changing the reference value of the air/fuel ratio of the cell C where the engine operates in the stationary conditions while the second procedure is carried out.

Figure 5b shows the Yes/No outputs of step F reversed relative to one another, as it is carried out when the first process is not being performed or has ended.

The non-limiting example described above can be subjected to variations, without for this reason going beyond the scope of protection of the invention, comprising all equivalent embodiments for a person skilled in the art.

When reading the description above, a skilled person can carry out the subject-matter of the invention without introducing further manufacturing details. The elements and features contained in the different preferred embodiments, drawings included, can be combined with one another, without for this reason going beyond the scope of protection of this patent application. The information contained in the part concerning the state of art only serves the purpose of better understanding the invention and does not represent a declaration of existence of the items described. Furthermore, if not specifically excluded by the detailed description, the information contained in the part concerning the state of art should be considered as an integral part of the invention.