DE102012219353B3 - Calculation and forecasting method of knocking and super-knocking processes and control device for the control of combustion processes in internal combustion engines, in particular in gasoline engines - Google Patents

Abstract

The invention relates to a method for determining a local probability (P 'knock) for the occurrence of super knock processes in a fuel-air mixture introduced into a combustion chamber of an internal combustion engine, in particular a supercharged gasoline engine, comprising the following steps: a) determining the local Probability density (P) of a spatial temperature gradient (∂T / ∂x) of a fuel-air mixture under consideration and of a length (l) assigned to this spatial temperature gradient (∂T / ∂x), over which the temperature gradient (∂T / ∂x) in each case is substantially constant, the probability density (P) depending on at least one local turbulence parameter (k, <∊>, <T'2>, <Z '2>) and at least one local thermodynamic parameter (p, T, λ), b) Determining the local probability (P'Klopf) by integrating the local probability density (P) into a local knock probability.

Description

The present invention relates to a method for determining a local probability for the occurrence of super-knocking processes in a fuel-air mixture introduced into a combustion chamber of an internal combustion engine, in particular in a supercharged gasoline engine. The invention further relates to a control device for carrying out such a method and to a motor vehicle having such a control device.

For a better understanding of combustion processes of fuel-air mixtures in modern internal combustion engines, it is customary to simulate such combustion processes by means of suitable simulation methods. For this purpose, a plurality of numerical simulation methods are known from the prior art, which depends on various parameters such. B. the stoichiometric composition and pressure and temperature of the fuel-air mixture, but also depending on engine-specific parameters such as the geometric structure of the combustion chamber o. Ä., Simulate the combustion process. Also, various operating parameters of the engine such. As the speed can be considered in these procedures. Such numerical methods are of crucial importance in the development of modern internal combustion engines, for example in the design of the geometry of the combustion chamber of the internal combustion engine. The same applies to the optimization of internal combustion engines in terms of their fuel consumption and the drive power provided by it.

Many modern internal combustion engines use fuel-air mixtures to optimize their drive performance with the lowest possible fuel consumption, which have a tendency to auto-ignition at typical operating temperatures, such as occur in combustion chambers of diesel or gasoline engines.

Particularly in internal combustion engines with spark ignition, for example in the manner of a conventional gasoline engine, which are operated in conjunction with a turbocharger for compressing the air to be introduced into the combustion chamber, such unwanted, stochastically occurring autoignition of the fuel, which is also known under the term "supercharged". Knock "or" mega-knock "is known to lead to damage to the crankshaft or other moving parts of the internal combustion engine to its complete destruction.

In contrast to classic ignition processes, such as occur in diesel engines, the probability of such undesirable "super-knocking" or "mega-knocking" can not be predetermined in internal combustion engines with spark ignition.

Also, the use of conventional knock sensors, by means of which the classic knocking in the internal combustion engine can be reduced, is not suitable for reducing super-knocking in spark-ignition internal combustion engines. In order to avoid the occurrence of undesired super-knocking or "mega-knocking" in modern internal combustion engines, therefore, a precise understanding of the complicated physical processes during the combustion process of the fuel-air mixture in the combustion chamber of the internal combustion engine is required.

The DE 10 2008 038 102 A1 relates to a method for preventing a pre-ignition of a fuel-air mixture in a cylinder chamber of an Otto engine operated high compression ratio or with supercharging. For this purpose, an associated internal cylinder pressure and from this a combustion center position of the respective cylinder is determined for the respective current working cycle on the basis of the detected angular velocity profile of the crankshaft of the internal combustion engine. From this data, the anticipated pre-ignition tendency is calculated for the subsequent work cycle. In the event that an exceeding of an applicable angle threshold of the combustion shear point position for this previous work cycle is determined, an additional amount of additional fuel is injected into the respective cylinder in the current working phase.

The DE 10 2010 003 305 A1 discloses a method for reducing the risk of future irregular combustion, such as in the form of a super-knock or glow-ignition, in an internal combustion engine. In a first step of the method, one or more influencing factors are monitored prior to the possible occurrence of irregular combustion, of which the probability of the future occurrence of irregular combustion is dependent. Such influencing factor may be, for example, the deposition behavior in the combustion chamber of the internal combustion engine. In a further, second step, at least one measure for reducing the probability of the future occurrence of irregular combustion depending on the monitoring carried out. Such a measure may include, for example, changing engine operating parameters or a deposit cleaning of the combustion chamber of the internal combustion engine.

The US 6,006,157 relates to a method for predicting irregular combustion in an internal combustion engine, according to which a Karlovitz number is determined as a function of various operating parameters of the internal combustion engine. This is compared with a threshold at which irregular combustion occurs. Said threshold value is determined based on models and stored in a control unit controlling the internal combustion engine. In the event that the determined Karlovitz number exceeds the threshold, an error message is generated.

The DE 11 2005 003 527 T5 deals with a method for controlling an internal combustion engine with the aid of a motor control unit, which evaluates real-time sensor data. From the real-time sensor data, a prior is used to derive at least one combustion parameter estimate and / or at least one combustion parameter measurement. The combustion parameter estimation or the combustion parameter measurement is used to control the internal combustion engine by the engine control unit.

The present invention therefore deals with the problem of finding new ways of determining the likelihood of "super-knocking" in fuel-air mixtures.

This problem is solved by the subject matter of the independent claims. Preferred embodiments are subject of the dependent claims.

The invention is based on the general idea of simulating the occurrence of super-knock in an air-fuel mixture used in a combustion chamber of an internal combustion engine as part of a numerical simulation method taking into account local turbulence effects in the fuel-air mixture, thus providing a highly accurate probability statement allows whether and with what probability under the simulated conditions super-knocking can occur in the fuel-air mixture.

The consideration of localized turbulence effects is further based on the idea that, when super-knocking occurs in a combustion chamber of an internal combustion engine, there are local areas in which a resonance occurs between the reaction front velocity of auto-ignition along a temperature gradient and the speed of sound. To determine the likelihood of these regions being present, a local probability density of a spatial temperature gradient in a fuel-air mixture and a length associated with that spatial temperature gradient over which the temperature gradient is substantially constant is determined.

From the local probability density, the sought local probability is then determined by integrating the local probability density into a local knock probability.

Thus, according to the invention, as already explained, the local joint probability density is determined taking into account local turbulence effects of the fuel-air mixture. By using stochastic methods, the occurrence of super-knocking in the internal combustion engine can be predicted in this way with particularly high accuracy. The local probability density is determined by a so-called direct numerical simulation (so-called "DNS") using a theoretical model, which takes into account local turbulent conditions of the temperature gradient of the space considered, whereby in the direct numerical simulation the partial derivatives into the Navier Stokes equations and in the equation for the temperature and / or for the mixture break by difference formulas and / or by a Fourier transform are replaced and the resulting equation system is solved numerically. This allows a much more accurate prediction of the occurrence of undesired super-knocking effects than is possible with conventional turbulence models known in the art.

By means of the method according to the invention, therefore, a precise probability statement regarding the occurrence of knocking processes, in particular super-knocking, can be made in a defined spatial region of the combustion chamber by integration of the probability density within the integration limits to be determined into a detonation probability, depending on various thermodynamic parameters such as As temperature, pressure, stoichiometric composition of the fuel-air mixture, the speed of sound and the kinematic viscosity, but also in response to various reaction kinetic parameters such. B. ignition delay and their Derivation according to the temperature, as well as the sound excitation time of self-ignition and also depending on various turbulence parameters such. B. kinetic energy of turbulence, their dissipation and the temperature variance are taken.

The integration limits for the integration are determined as a function of at least one local thermodynamic or / and at least one local turbulence parameter. The consideration of so-called self-ignition conditions in the space considered is carried out by taking into account the integral time scale of the turbulence in the fuel-air mixture.

In a preferred embodiment, the common probability density can be determined taking into account local turbulence effects of the fuel-air mixture, in particular by a so-called direct numerical simulation ("DNS") in an idealized turbulent flow, for example in the manner of a homogeneous shear turbulence. Since such turbulence effects have a decisive influence on whether super-knocking effects can occur in the considered fuel-air mixture, they are of decisive importance in the method according to the invention for the occurrence of super-tapping. The consideration of local turbulence effects allows a greatly improved probability statement for the occurrence of super-tapping compared to conventional methods.

Preferably, the integration according to step b) takes place by integration within predetermined integral limits, the integral limits depending on the at least one local thermodynamic and at least one local reaction kinetic parameter, in particular the ignition delay time, and at least one local turbulence parameter. In this way, mathematically, for certain value ranges of the at least one local thermodynamic or / and at least one local reaction kinetic parameter for which it is known, for example from experimental investigations or theoretical modeling, that no super-beating can occur, a "zero value" for the probability density in the integration to the sought probability are generated.

In a further embodiment, in step a) additionally a local probability density of a spatial mixture fracture gradient of the fuel-air mixture and of a length assigned to this mixture fracture gradient, over which the mixture fraction gradient is respectively substantially constant, can be determined. Also in this case, the local probability density of the mixture fraction gradient depends on at least one local turbulence parameter and at least one local thermodynamic parameter. According to this embodiment, in an additional step a1) of the method according to the invention, the local probability densities of the temperature gradient and the mixture fraction are combined to form a combined combined probability density of a gradient of temperature gradient and mixture fraction gradient of equal length for both gradients. By additionally considering the mixture break, the prediction for the occurrence of super-beating can be further improved.

• a) die kinetische Energie der Turbulenz k,
• b) die Dissipation der kinetischen Energie der Turbulenz <ε>,
• c) die Varianz des lokalen Temperaturfluktuation <T'²>,
• d) die Varianz der Mischungsbruchfluktuation <Z'²>.

In a particularly preferred embodiment, the at least one local reaction kinetic parameter may be an ignition delay time of an ignition process of the fuel-air mixture and / or the derivative of this Zündverzugszeit after the temperature or the Mischungsbruchs and / or a sound excitation time of this ignition process, and / or the at least one local thermodynamic parameters are the local temperature or the local mixture fraction of the fuel-air mixture and / or the local kinematic viscosity of the fuel-air mixture and / or the local pressure of the fuel-air mixture and / or the speed of sound, and / or the at least one local turbulence parameter one of the following local parameters:

• a) the kinetic energy of the turbulence k,
• b) the dissipation of the kinetic energy of the turbulence <ε>,
• c) the variance of the local temperature fluctuation <T ' ² >,
• d) the variance of the mixture break fluctuation <Z ' ² >.

In a particularly expedient embodiment, the length, over which the temperature or the mixture gradient is constant, as a linear distance l between a first space point of the fuel-air mixture with maximum temperature T _max and a second space point with minimum temperature T _min , or . be defined a first point in space of the fuel-air mixture with maximum mixing fraction Z _max and a second space point with minimum mixing fraction Z _min .

The temperature gradient ∂T / ∂x is approximated as follows: ∂T / ∂x α ΔT / l = (T _max - T _min ) / l.

Accordingly, the mixing gradient is approximated as follows: ∂Z / ∂x α ΔZ / l = (Z _max - Z _min ) / l.

The mixture fraction Z is with the air ratio λ over the relationship λ = Z _st / Z · ((1-Z) / (1-Z _st )) connected. In this case, Z _{st is} the value of the mixture fraction Z at stoichiometric mixture λ = 1. The mixture fraction is thus an alternative to the air ratio λ definition of the local air-fuel ratio compared to the stoichiometric mixture. As a mass ratio, it is more appropriate than the former for numerical simulations of the mixing ratios in the combustion chamber.

In a particularly expedient embodiment, a respective dissipation element is defined by the length l and the temperature difference (.DELTA.T = T _max - T _min ) or the mixing fraction difference (.DELTA.Z = Z _max - Z _min ).

• A) Definieren von wenigstens einer Zone in der Brennkammer,
• B) Durchführen des Verfahrens mit einem oder mehreren der vorangehend genannten Merkmale für jede der in Schritt A) definierten Zonen, so dass jeder Zone eine lokale Wahrscheinlichkeit zugeordnet ist.

The invention further relates to a method for determining a spatial probability distribution for the occurrence of super-knocking processes in a combustion chamber of an internal combustion engine, in particular a supercharged gasoline engine, comprising the following steps:

• A) defining at least one zone in the combustion chamber,
• B) performing the method with one or more of the foregoing features for each of the zones defined in step A), such that each zone is assigned a local probability.

Such a method, by applying the aforementioned method for determining a local probability for the occurrence of super-knocking in a plurality of grid points in the fuel-air mixture, when it is arranged, for example, in the combustion chamber of the internal combustion engine, allows the calculation of a spatial probability distribution for the occurrence of super-knock in the fuel-air mixture in the combustion chamber. For this purpose, the combustion chamber can be divided into the plurality of grid points in the manner of a three-dimensional grid, and the method according to the invention for calculating the local probability for the occurrence of super-beating can be carried out successively for each grid point.

The method for determining the spatial probability distribution is particularly suitable in connection with the development of gasoline engines, if the combustion chamber of such a gasoline engine is to be designed such that no unwanted super-beating, which can lead to the destruction of the gasoline engine may occur. However, of course, numerous other fields of application for the two methods according to the invention in connection with the development of internal combustion engines are conceivable. For example, it may be thought to use the inventive method when the combustion chamber of the internal combustion engine is to be geometrically dimensioned.

In a preferred embodiment, the method according to the invention is carried out by numerical simulation for a plurality of zones or grid points in the combustion chamber. Here, the common probability density of a spatial temperature gradient and the mixing gradient of the fuel-air mixture and this spatial temperature or mixing gradient associated with the same length over which the temperature or mixing gradient is substantially constant, taking into account local turbulence effects of the fuel-air mixture in a table , ie in the form of a "lookup table". The same applies to the thermodynamic and reaction kinetic parameters. In this way, it can be determined with what probability super-knocking processes can occur in different zones or grid points of the combustion chamber of the internal combustion engine.

• C) Ermitteln des Maximalwerts P^max _Klopf aus den in Schritt B) für jede Zone bzw. jeden Gitterpunkt berechneten lokalen Wahrscheinlichkeiten. Auf diese Weise lässt sich die erzeugte Datenmenge deutlich reduzieren, was von Bedeutung sein kann, wenn die berechneten Maximalwerte in einem Steuergerät eines Kraftfahrzeugs in der Art eines ”Lookup-Tables” zur Ansteuerung der Brennkraftmaschine verwendet werden soll und in dem Steuergerät nur ein begrenzter Speicherplatz zum Speichern eines solchen ”Lookup-Tables” zur Verfügung steht. Ein solches ”Lookup-Table” kann dazu verwendet werden, um zu überprüfen, ob bei bestimmten Werten der Betriebsparametern der Brennkraftmaschine die Gefahr des Auftretens von unerwünschtem ”Super-Klopfen” besteht, so dass das Steuergerät in einem solchen Fall die eingestellten Werte der Betriebsparameter in einen Wertebereich ändert, in welchem das Auftreten von Super-Klopfen ausgeschlossen werden kann.

In a particularly preferred embodiment, the process may comprise the following additional step C):

• C) Determining the maximum value P ^max _knock from the local probabilities calculated in step B) for each zone or grid point. In this way, the amount of data generated can be significantly reduced, which can be of importance if the calculated maximum values in a control unit of a motor vehicle in the manner of a "lookup tables" to control the internal combustion engine to be used and in the controller only a limited space to store such a "lookup" Tables "is available. Such a "look-up table" can be used to check whether there is a risk of undesirable "super-knocking" occurring at certain values of the operating parameters of the internal combustion engine, so that the control unit in such a case the set values of the operating parameters changes to a range in which the occurrence of super-knocking can be excluded.

The invention further relates to a control device for an internal combustion engine having a combustion chamber, in particular for a gasoline engine, wherein the control device is set up / programmed for carrying out the above-explained method for calculating the spatial probability distribution for the occurrence of super-beating. For this purpose, the control device can have a calculation unit, by means of which a maximum probability of the occurrence of super-knocking processes in a fuel-air mixture introduced into the combustion chamber is determined, this determination taking into account global turbulence effects of the fuel-air mixture in the combustion chamber. Alternatively or additionally, the control device may comprise a memory unit in which the spatial probability distribution determined by means of the previously explained method can be stored.

This means that in the control device, the method according to the invention can be carried out in real time in order to be able to analyze the fuel-air mixture contained in the combustion chamber with regard to the occurrence of super-knocking.

Such a control device may for example be part of a control device of the internal combustion engine, which controls the internal combustion engine operationally when used in a motor vehicle. The method according to the invention can then be carried out in the control device in real time, if it is to be checked whether control parameters set in the internal combustion engine, such as, for example, are set. As the engine speed or the temperature of the fuel-air mixture favor the occurrence of super-tapping. For this purpose, the control unit may have a calculation unit, by means of which a spatial probability distribution for the occurrence of super-knocking processes in an introduced into the combustion chamber air-fuel mixture is determined by at least one zone for each local probability for the occurrence of super-knocking in the is determined in the combustion chamber introduced fuel-air mixture, this determination takes place taking into account local turbulence effects of the fuel-air mixture in the combustion chamber.

By means of the method according to the invention, a super-knocking probability can be calculated as a function of a crank angle of the cylinder piston of the internal combustion engine and in dependence on the above-mentioned control parameters. In the event that it has been determined by means of the method according to the invention that the determined probability for the occurrence of super-knocking exceeds a predetermined maximum upper limit value, the control unit or its calculation unit can modify one or more operating parameters by applying a suitable control algorithm in order to adjust the Avoid occurrence of super-tapping. In this context, numerous applications of the method according to the invention are conceivable. For example, it may be thought that in the control device, a suitable control or regulation strategy is implemented to possibly certain operating parameters such. B. to change the speed of the internal combustion engine, the pressure or the temperature of the fuel-air mixture o. Ä. To avoid super-knock. In this case, the method according to the invention can be used for checking to ensure that a certain choice of the values of different operating parameters is associated with a sufficiently small probability of the occurrence of super-knocking. For this purpose, a predetermined maximum permissible upper limit for the probability can be defined in the control device.

As an alternative to an execution of the method in real time, this can also be carried out "offline" and the maximum knocking probability generated in this way in the control device / the control unit in the manner of a "lookup table" are stored. For this purpose, the control device may comprise a memory unit in which the determined spatial probability distribution can be stored. To reduce the amount of data can in advance also from the for certain values of the control parameters such. B. engine speed, pressure of the fuel-air mixture, temperature of the fuel-air mixture generated spatial probability distribution only the maximum value of the spatial probability distribution for the occurrence of super-knocking in the "lookup table" taken and stored in the memory unit ,

In a further embodiment, the control device can use a spatial probability distribution for the occurrence of super-knocking processes in different zones of the combustion chamber in the manner of a so-called multi-zone model, by using the method according to the invention for the different zones in each case a local probability of occurrence is determined by super-knocking processes in the fuel-air mixture introduced into the combustion chamber, this determination again taking into account local turbulence effects and thermodynamic parameters in the combustion chamber. As a special case, in particular only one zone can be used. This means that the control device can operationally, ie during operation of the internal combustion engine, determine the probability of the occurrence of super-tapping as a function of various operating parameters, which can be set at least partially by the control device, in the combustion chamber.

Preferably, the turbulence parameters are determined from the instantaneous engine speed of the internal combustion engine and from the average piston speed of the cylinder piston, so that the engine speed or / and the average piston speed can be used as input parameters for carrying out the method according to the invention. The variance of the local temperature fluctuation may be approximated by the difference between an intake temperature of the air or exhaust gas recirculation of the air-exhaust gas mixture after closing the intake valves and a characteristic wall temperature of a wall of the combustion chamber of the internal combustion engine.

The invention also relates to a motor vehicle having an internal combustion engine having a combustion chamber and to a control device having one or more of the features explained above, the control device determining the internal combustion engine as a function of the maximum knocking probability determined as a function of the spatial probability distribution determined by the calculation unit or stored in the memory unit controls. Preferably, the activation can take place as a function of the spatial probability distribution which has been explained above and which is reduced to respective maximum probability values. To reduce the amount of data can in advance also from the for certain values of the control parameters such. B. engine speed, pressure of the fuel-air mixture, temperature of the fuel-air mixture generated spatial probability distribution only the maximum value of the spatial probability distribution for the occurrence of super-knocking in the "lookup table" taken and stored in the memory unit ,

Other important features and advantages of the invention will become apparent from the dependent claims, from the drawings and from the associated figure description with reference to the drawings.

Preferred embodiments of the invention are illustrated in the drawings and will be described in more detail in the following description, wherein like reference numerals refer to the same or similar or functionally identical components.

It show, each schematically:

1 a block diagram of the method according to the invention,

2a / b a two-dimensional temperature distribution and its spatial decomposition into so-called dissipation elements,

3 an area diagram for the prediction of super-tapping,

4 an example of a normalized common probability density of the temperature gradient ΔT / l and the length l,

5 an exemplary ignition delay time temperature diagram,

6a / b is a rough schematic representation of a combustion chamber of an internal combustion engine,

7 a maximum probability for the occurrence of super-beating as a function of a current crank angle of the internal combustion engine calculated by means of the method according to the invention,

8th a rough schematic representation of a control device according to the invention for carrying out the method according to the invention.

In the 1 the method according to the invention for determining a local probability for the occurrence of super-knocking processes in a fuel-air mixture introduced into a combustion chamber is illustrated in a schematic flowchart.

In a first step S1 of the method, the temperature T, the pressure P, and the local air ratio λ of the considered fuel-air mixture are provided as thermodynamic parameters. Where τ _{e is} the sound excitation time and τ is the ignition delay time. Furthermore, in step S1, the reaction kinetic parameters τ (T, p, λ) and τ _e (T, p, λ) are provided as a function of the temperature T and the pressure p and the air ratio λ. As already explained, the mixture fraction Z is linked to the air ratio λ via the relationship λ = Z _st / Z * ((1-Z) / (1-Z _st )). In this case, Z _{st is} the value of the mixture fraction Z at stoichiometric mixture λ = 1. The mixture fraction is thus an alternative to the air ratio λ definition of the local air-fuel ratio compared to the stoichiometric mixture.

The functional dependencies τ (T, p, λ) and τ _e (T, p, λ) for providing the reaction kinetic parameters can be determined by numerical simulation calculations. In the 5 By way of example, a dependence τ (T, p) calculated by simulation is shown, namely for a stoichiometric (λ = 1) PRF mixture with 85% iso-octane and 15% n-heptane at three different pressures p ₁ , p ₂ , p ₃ .

Consequently, the parameters τ (T, p, λ) and τ _e (T, p, λ) required for the calculation of the desired probability can be determined from the known thermodynamic parameters p, T and λ of the PRF mixture of the fuel-air mixture used , Of course, mixtures of other chemical components for the representation of gasoline fuels, for which also ignition delay times and the sound excitation time can be calculated and approximated, can be used.

Finally, the turbulence parameters k, <ε>, <T ' ² > and <Z' ² > describing the local turbulence conditions in the fuel-air mixture are also provided in step S1. Here, k is the kinetic energy of the turbulence of the fuel-air mixture in the space considered, <ε> its dissipation, <T ' ² > the variance of the temperature fluctuations and <Z' ² > the variance of the mixture break fluctuation.

In order to ensure the highest possible prediction accuracy of super-knocking processes in different spatial regions of the combustion chamber of the internal combustion engine even under real conditions, which is the occurrence of turbulence-related strong local fluctuations of the temperature gradient ∂T / ∂x or the mixture break gradient ∂Z / ∂x of the fuel-air mixture In the method according to the invention, in contrast to methods known from the prior art, local turbulence effects of the temperature gradient ∂T / ∂x and / or the mixing fraction gradient ∂Z / ∂x and of the lengths l over which they are respectively substantially constant , considered in the fuel-air mixture.

In the method according to the invention, for this purpose a simulated fuel-air mixture assumed to be homogeneous is decomposed into a plurality of subregions in which the temperature gradient ∂T / ∂x and / or the mixture fraction gradient ∂Z / ∂x of the fuel-air mixture are substantially constant. For this purpose, a so-called. Direct numerical simulation ("DNS") of the flow field and the temperature field and / or the mixture fraction field of the fuel-air mixture is carried out for this purpose, the Navier-Stokes equations known to the expert are used and the partial derivatives in the Navier-Stokes Equations and in the equation for the temperature and / or for the mixture break replaced by difference formulas and / or by a Fourier transform, and then solved the resulting equation system numerically. In this case, the considered physical space can be divided, for example, into 1024 ³ , 2048 ³ or 4096 ³ grid points. In order to be able to resolve even the smallest occurring lengths of turbulence, only sections of the entire flow with a spatial size of about 1 cm ³ are typically considered. Based on the numerical solution of the Navier-Stokes equations, a dissipation element analysis is then performed. This is in the flow chart of 1 summarized under the method step S2.

By means of the dissipation element analysis according to step S2, the flowchart describing the method according to the invention (cf. 1 ) determines a combined joint probability density P (l, g) of an approximated gradient ∂φ / ∂x α Δφ / l = g and a length l assigned to this gradient. In this case, the gradient ∂φ / ∂x can be determined as a combination of temperature gradient ∂T / ∂x and mixture fraction gradient ∂Z / ∂x, which will be described below in connection with equation (5) will be explained in more detail. An analogous relationship applies to the combination of the approximated gradients ΔT / l and ΔZ / l to Δφ / l according to equation (6).

After performing the dissipation element analysis, a common local probability density P (l, g) with the gradient g = Δφ / l and the length l can thus be determined according to step S3. For the simplified case of a constant mixing break, the approximated temperature gradient ΔT / l should be used instead of the combined gradient. By normalizing with corresponding mean values, the calculation method is based on a dimensionless and therefore universal representation (cf. 4 ).

The dissipation element analysis mentioned in connection with step S2 will now be described below with reference to FIG 2 and 3 explained in more detail. In the dissipation element analysis, so-called dissipation elements are defined as space-filling structures in a scalar field, for example the temperature field of the fuel-air mixture in the combustion chamber. The dissipation elements can be generated by using so-called gradient trajectories. If you follow z. B. in a temperature field of such a gradient trajectory from any point in space in the rising direction, one reaches a point in space with a maximum temperature value. Accordingly, one reaches a temperature minimum value from this arbitrary point in space, if one follows a gradient trajectory in descending direction. A dissipation element is defined as a spatial region in the sense of a set of spatial points, from which in each case the same spatial point with maximum value or minimum value is achieved via gradient trajectories. Within a dissipation element, the temperature gradient thus has a substantially constant course. The same applies mutatis mutandis for the determination of Dissipationselementen in a mixed fraction field, which for clarity in the 3 not shown.

In the 3a is an example of an area 9 represented with a two-dimensional turbulent temperature distribution. The with the reference number 5 designated points are places in which the temperature T assumes a temperature minimum value, denoted by the reference numeral 6 designated points are places where the temperature T assumes a maximum temperature value. With 7 labeled lines represent boundary lines between the various dissipation elements 8th , From the 3a can be inferred that the area 9 by means of the dissipation elements 8th can be subdivided into a plurality of smaller regions, each of which is amenable to a more detailed analysis for the occurrence of super-tapping.

The length of a dissipation element 8th is typically 2 Taylor lengths λ _T = (10 · ν · k / <ε>) ^1/2 . Only within a respective dissipation element 8th in which the temperature gradient is substantially constant, the already explained resonance between the reaction front velocity u and the sound velocity a can occur and produce a knocking process up to the super-beating.

Looking at the 3a that with 10 designated dissipation element, which in the 3b is shown separately and enlarged, and the point arranged therein at maximum temperature 11 as well as the point 12 with minimum temperature, the quotient of the temperature difference between maximum and minimum temperature and the distance of the two points denoted by l 11 . 12 a measure of the temperature gradient ∂T / ∂x in the dissipation element 10 ,

Assuming now that the dissipation element 10 is a filled with a fuel-air mixture local spatial portion of the combustion chamber of the internal combustion engine, so is in the dissipation element 10 the following scenario conceivable: During a compression phase in the combustion chamber, the temperature of the fuel-air mixture increases, causing the self-ignition or pre-ignition of the same in the region of the point 11 with maximum temperature. In this area, strong acoustic wavefronts are generated, which propagate in all directions. In the event that the point 11 adjacent dissipation elements have a suitable temperature gradient or mixture break gradient, the above-mentioned acoustic coupling can take place. By means of such an acoustic coupling, an auto-ignition of the fuel-air mixture in other dissipation elements of the area 9 be generated, which can ultimately lead to the undesirable super-tapping.

In order to be able to make a probability statement for the occurrence of super-knock on the basis of the previously explained decomposition of the area to be examined, ie the combustion chamber of the internal combustion engine, it is assumed in the method according to the invention that the section detected by the direct numerical simulation according to method step S2 in each case the area between several grid points in the vicinity of the grid point x → _i describes, if, as will be explained in more detail below, the combustion chamber into a three-dimensional grid with a plurality of grid points x → _i is decomposed, and thus that area contains a plurality of dissipation elements. Every dissipation element 8th in the respective field is characterized by means of two independent parameters, on the one hand by means of the temperature gradient ∂T / ∂x, which is approximated by ΔT / l, on the other hand by means of the mixture breaking gradient ∂Z / ∂x, which is approximated by ΔZ / l on the other hand by means of the respective linear distance 1 , which is responsible for the temperature field between the points 11 . 12 with maximum and minimum temperature, for the mixing field between the maximum and minimum mixing fraction extends. The respective lengths are chosen to be the same for the statistical statement. Furthermore, a single gradient g = Δφ / l representing a combination of the approximated temperature gradient ΔT / l and the approximated mixture fraction gradient ΔZ / l (see Equation 5) is taken as the basis of the probability statement. In a simplified variant, in the context of the dissipation element analysis, it is also possible to dispense with a consideration of the mixture field or mixture fracture gradient and only the temperature gradient can be used to determine the dissipation elements.

berechnet werden kann. Entscheidend ist erfindungsgemäß hierbei, dass bei der Integration nur derjenige Teil des funktionalen Raums von l und g berücksichtigt wird, in welchem überhaupt ein Klopfvorgang auftreten kann. Aus der theoretischen Modellierung ist nämlich bekannt, dass für bestimmte Wertebereiche von l und g keinerlei Super-Klopfeffekte auftreten können.Looking again at the flow chart of 1 , it can be seen that the sought probability P _knock for the occurrence of a knocking operation, in particular of super-knocking, in the considered local space area x → _i the combustion chamber according to step S4 of the flowchart by integration of the probability density P (l, g) over l and over the gradient g, ie

can be calculated. Decisive in this case according to the invention is that in the integration only that part of the functional space of l and g is considered, in which a knocking can occur at all. For it is known from theoretical modeling that for certain value ranges of 1 and g, no super-knocking effects can occur.

A so-called "area diagram", which is known to the person skilled in the art and in which 3 for which values of the reaction front velocity u, the sound velocity a, the so-called sound excitation time τ _{e of} the fuel-air mixture, and a length l, over which the temperature gradient ∂T / ∂x and / or the Mixture fracture gradient ∂Z / ∂x in the fuel-air mixture is substantially constant, knocking can occur up to the super-tapping. The sound excitation time τ _e is defined here as the time duration which elapses until the relative heat release rate in the auto-ignition of the fuel-air mixture rises from 5% to a maximum value.

In connection with the area diagram 1 of the 2 is generally known to those skilled in that for the occurrence of super-knocking a self-ignition of the fuel-air mixture at hot spots in the combustion chamber of an internal combustion engine is a necessary, but not sufficient condition. In this connection, theoretical models are known which deal with the occurrence of undesirable super-knocking effects from the hot spots of auto-ignition of the reactive fuel-air mixture taking into account different types of fuel. Such a theoretical treatment of the auto-ignition of the air-fuel mixture and a possible occurrence of super-knocking is based on an analysis of the propagation modes of reaction fronts of the reactive fuel-air mixture with temperature gradient ∂T / ∂x and / or with mixing fraction gradient ∂Z / ∂x in the Final phase of the ignition of the fuel-air mixture.

In such a theoretical treatment, the occurrence of super-knock is described as an acoustic resonance effect between the reaction front velocity u and the sound velocity a. For the occurrence of a super-knocking process, it is therefore necessary that the derivation of the ignition delay time τ after the temperature T, ie ∂τ / ∂T, locally with the spatial temperature gradient ∂T / ∂x, as well as the derivation of Zündverzugszeit after the mixture break ∂ τ / ∂Z with the mixture fraction gradient ∂Z / ∂x, coupled in the propagation direction of the reaction front such that a reaction wave results in a reaction front velocity u, which is in the order of magnitude of the speed of sound a. In this case, the occurrence of super-knocking is possible.

When using an "area diagram" in conventional simulation methods known from the prior art, on the one hand the influences of the mixture rupture gradient and on the other hand the turbulent conditions determining the local temperature gradient ∂T / ∂x or the local mixture rupture gradient ∂Z / ∂x remain as they occur in the combustion chamber of an internal combustion engine, for the latter, fixed assumed numerical values are used, which may lead to erroneous or only inaccurate prediction of the occurrence of super-tapping.

The use of the area diagram of the 3 so alone is insufficient to predict the occurrence of super-tapping; In contrast, the area diagram in the method according to the invention is used only as a partial aspect, namely in the integration of the probability density, which is calculated taking into account local turbulence effects, to the sought probability in order to establish integration limits during the integration.

For the following explanations, the more complex case is explained in which both the mixture rupture gradient and the temperature gradient (∂T / ∂x) of the fuel-air mixture (dZ / ∂x) are considered, and for each size a separate probability density is calculated, which can then be combined in an additional step of the method according to the invention to a common probability density. In a simplified variant, the mixture fraction and the mixture fraction gradient can be completely disregarded, i. H. for the following equations, in the simplified case ∂Z / ∂x = 0, ∂τ / ∂Z = 0.

In connection with the foregoing, the reaction front velocity is u = 1 / (∂τ / ∂T · ∂T / ∂x + ∂τ / ∂Z · ∂Z / ∂x), (2)

In the area diagram of the 3 the dimensionless parameter ε is plotted on the x-axis and the dimensionless parameter ξ on the y-axis. The parameters ξ and ε are defined as follows: ξ = a / u = a · (∂τ / ∂T · ∂T / ∂x + ∂τ / ∂Z · ∂Z / ∂x) (3)
ε = 1 / (a · τ _e ). (4)

Here, a is the speed of sound and, as already explained, the length assumed for dissipation elements from both fields, the temperature field and the mixture rupture field, over which the temperature gradient ∂T / ∂x or the mixture rupture gradient ∂Z / ∂x respectively crosses is essentially constant.

Due to the locally given thermodynamic quantities ∂τ / ∂T and ∂τ / ∂Z as well as the local gradients ∂T / ∂x and ∂Z / ∂x the gradient remains ∂φ / ∂x = ∂τ / ∂T · ∂T / ∂x + ∂τ / ∂Z · ∂Z / ∂x (5) and the approximated gradient g = Δφ / l = ∂τ / ∂T · ΔT / l + ∂τ / ∂Z · ΔZ / l (6) Are defined.

From the theoretical modeling, it is now known to the person skilled in the art that only in that with the reference character 1 designated area of the area diagram is an occurrence of detonation, and thus in particular of super-knocking is possible. The area 1 is in the ξ-direction through the lines with the reference numerals 2 and 3 limited, which represent an upper limit ξ _u (ε) and lower limit ξ _l (ε). The integration _limits required for the integration of P (l, g) into P _knock are g _u = ξ _u / a (7)
g _l = ξ _l / a (8)

In Equation (9), the required relationship between the dimensionless parameter ε and the distance l results accordingly l = ε · a · τ _e . (9)

Thus, from the area diagram of 2 the integration limits according to step S4 of the flowchart of 1 remove.

In the diagram of 4 is an example according to step S3 of the flowchart of 1 is calculated by DNS local normalized joint probability density P (l / l _m , ΔT / l / (ΔT / l) m) for the simplified special case λ = 1 shown, in which the mixture field is homogeneous and thus the mixture break gradients disappear, so assume a zero value , The values for ΔT / l and l are normalized with the mean value of the scalar gradient (ΔT / l) _m or with the mean value of the linear distance l _m . This value range is in the example diagram of 4 represented by a gray scale representation. In this case, gray levels in the center (ie at g / g _m = 1, l / l _m = 0.8) of the L-shaped area mean a higher value of the probability density P than gray levels for larger or smaller values of g / g _m and l / l _m , ie from the center (area "I" in the 4 ), the probability density P decreases in all directions. The area denoted by "I" (see dashed line in FIG 4 ) thus has a higher probability density P than the region designated by "II", which in turn has a higher probability density P than the region designated by "III".

For the more complex case in which the mixture rupture gradient is taken into account, the gradient g should be used instead of ΔT / l, where g is determined by the combination of the gradients ΔT / l and ΔZ / l from the respective dissipation elements according to Eq. (6) are to be determined. A normalized joint probability density P (1/1 _m , g / g _m ) results from a convolution of the joint probability densities P (1/1 _m , ΔT / 1 / (ΔT / 1) according to the linear relationship according to equation (6). _m ) and P (l / l _m , ΔZ / l / (ΔZ / l) _m ).

The mean values (ΔT / l) _m , (ΔZ / l) _m and l _m used in the normalization can be determined as a function of various turbulence parameters describing the local turbulence conditions in the fuel-air mixture, such as, for example, K, <ε>, <T ' ² > and <Z' ² > for each grid point x → _i calculate, as explained below, the combustion chamber into a three-dimensional grid with a plurality of grid points x → _i is decomposed. Here, k, as already explained, is the kinetic energy of the turbulence of the fuel-air mixture in the space considered, its dissipation, T ² the variance of the temperature fluctuations and Z ² the variance of the mixture break. The scalar dissipation of the temperature χ _T is proportional to k / ε · <T ' ² >, the scalar dissipation of the mixture fraction χ _Z is proportional to k / ε · <Z' ² >.

From these quantities, (ΔT / l) _m , (ΔZ / l) _m and l _m can be calculated as follows: (ΔT / l) _m α χ _T ^1/2 / (l _m ^2/3 · <ε> ^1/6 ), (10)
(ΔZ / l) _m α χ _Z ^1/2 / (l _m ^2/3 · <ε> ^1/6 ), (10 ')
l _m α λ _T , (11) in which λ _T = (10 · ν · k / <ε>) ^1/2 (12) applies. In this case, λ _{T is} the so-called micro-length or "Taylor length" of the turbulence and ν the kinematic viscosity of the fuel-air mixture.

The method according to the invention for calculating the local probability P _knock can now be used in different local space regions of the combustion chamber, that is to say the combustion chamber of the internal combustion engine, in order to obtain spatially resolved probability statements in this way. For this purpose, it makes sense to the combustion chamber in a three-dimensional grid with a plurality of grid points x → _i to disassemble, and for each grid point x → _i to calculate.

The method according to the invention then becomes loop-like for each grid point x → _i repeated. Such a grid point x → _i can also be used to define a particular zone in the combustion chamber of the internal combustion engine. In the flowchart of 1 is such a loop-like, i _max -repeated embodiment of the method according to the invention by the arrow with the reference numeral 1 indicated. I _max denotes the number of grid points x → _i , the parameter i is used to index the grid points x → _i and local probability P ^' _knock . The process is repeated until the indexing parameter i reaches the maximum value i _max and then ends, which is expressed in the flowchart by the method step S5.

Possible values for the parameter i _max for the spatial simulation of the entire combustion chamber are between 50,000 and 500,000. By means of an i _max- fold implementation of the method according to the invention, it is thus possible to calculate various local probabilities P ' _knock . In the event that by means of the grid point x → _i a respective zone area is defined, but i _max can also assume much lower values. In the simplest case, the method is applied to only a single zone, i _max = 1.

From the flowchart of 1 It can be seen that the method according to the invention for calculating the spatial probability distribution P ' _knock for reducing the amount of data generated can have an optional method step S6, which is described in US Pat 1 is shown in dashed lines. '(1 = 1 ... i _{max knock} in this process step S6 P P) _knock the maximum value P ^max _knock (i _max for 1 = 1 ...) is determined, so P ^max = max _knock' is from all calculated values , Such an optional method step S6 is suitable for the purpose of reducing the amount of data generated, in particular when the spatial probability distribution determined by the method according to the invention is to be used in the manner of a "lookup table" in a control device of the internal combustion engine. In this case, only the associated maximum value P ^max _{knock is} stored for a specific value of respective input parameters T, p, λ.

In the presentation of the 6 is now an example of a combustion chamber 20 an internal combustion engine in a side view ( 6a ) or a top view ( 6b ). By means of a loop-like application of the method according to the invention for different spatial regions of the combustion chamber 20 can be determined areas in which the probability P _knock for the occurrence of super-tapping is particularly high. These areas are marked in the figure and with the reference numeral 21 designated.

Since in pressure p and temperature T of the fuel-air mixture change in dependence on a position of the crankshaft in the cylinder of the internal combustion engine, the inventive method also also allows a probability statement for the occurrence of super-tapping depending on the position of the piston (or. the crankshaft) in the cylinder. This is exemplary in the representation of 7 shown.

In the diagram of 7 shows the reference numeral 30 2, the graph indicates the average temperature T of the fuel-air mixture in the cylinder of an internal combustion engine as a function of the crank angle ("CA") of the piston. At the same time, the reference number indicates 31 Graph referred to the calculated by the method according to the invention probability of the occurrence of a knocking operation as a function of the crank angle.

In the 8th is roughly schematic a control device according to the invention 100 for a a combustion chamber 101 having internal combustion engine 102 shown. The internal combustion engine 102 can be a gasoline engine. In the combustion chamber 101 is a linearly movable cylinder piston 103 arranged and with a crankshaft 104 connected. In the combustion chamber 101 is a fuel-air mixture 105 arranged. The exact technical structure of the combustion chamber or the internal combustion engine is known in the art, so that a detailed explanation is omitted.

The control device according to the invention 100 may be formed in the manner of a conventional microcontroller and be part of a control device of a motor vehicle, by means of which the internal combustion engine 102 is controlled. Alternatively, the control device 100 be designed as a separate component. The control device 100 can be a control unit 106 (ECU) and a memory unit 107 exhibit. The control device 100 is set up / programmed to carry out the method according to the invention. The inventive method can in a first variant in real time in the control device 100 be performed so that in the combustion chamber 101 contained fuel-air mixture 105 in terms of the occurrence of super-tapping can also be analyzed in real time. The method according to the invention can thus be used for checking whether those in the internal combustion engine 102 set current values of certain control parameters such. B. the engine speed n or the temperature T of the fuel-air mixture favor the occurrence of super-tapping. By means of the method according to the invention, for example, a super-knock probability as a function of a current crank angle of the cylinder piston 103 the internal combustion engine 102 , of which in turn temperature T and pressure p of the fuel-air mixture 105 in the combustion chamber, be calculated.

In the event that it has been determined by means of the method according to the invention that the ascertained probability for the occurrence of super-knocking exceeds a predetermined maximum upper limit value, the control device can 100 or the control unit of the internal combustion engine by application of a suitable control algorithm to modify one or more operating parameters of the internal combustion engine to the occurrence of super-knock, resulting in destruction of the combustion chamber 101 and / or the cylinder piston 103 can lead to avoid. In this context, numerous applications of the method according to the invention are conceivable. For example, it may be thought that in the control device 100 or the control unit of the internal combustion engine 102 an appropriate control strategy is implemented to optionally set certain operating parameters, such as B. the speed n of the piston of the internal combustion engine 102 , the pressure p or the temperature T of the fuel-air mixture 105 o. Ä. To avoid super-knocking to change. This can be done for example by changing the fuel supply and the injection timing and by multiple injections of fuel into the combustion chamber. Another way to influence the mentioned operating parameters may be to change a temperature of cooled exhaust gas recirculated into the combustion chamber.

The method according to the invention may be applied in the manner of a verification method to ascertain whether a particular choice of the values of the various operating parameters is associated with a sufficiently small probability of the occurrence of super-tapping. For this purpose, a predetermined maximum permissible upper limit for this probability can be stored in the control device.

Alternatively to an embodiment of the method in real time according to the first variant, the method can also be carried out "offline" in a second variant and the spatial probability distribution P 'generated in this way is _knocked in the memory unit 107 the control device 100 or of the control device of the motor vehicle in the manner of a "lookup table" are stored.

To reduce the amount of data generated, it is advisable that in advance (engine speed n z. B., pressure p, temperature T, etc.) and from the various values of the control parameters generated spatial probability distribution P _'knock only in step S6 determined maximum value P ^max knock of the spatial probability distribution P ' _knock for the occurrence of super-knocking in the "lookup table" is adopted. Thus, in the look-up table for respective values of the control parameters n, p, T, a probability P ^max _knock = P (n, p, T) is available for the occurrence of super-knocking, which consequently results in the control of the internal combustion engine 102 can be taken into account.

Claims (10)

Method for determining a local probability (P ' _knock ) for the occurrence of super-knocking processes in a fuel-air mixture introduced into a combustion chamber of an internal combustion engine, in particular a supercharged gasoline engine, comprising the following steps: a) Determining the local probability density (P) a spatial temperature gradient (∂T / ∂x) of a considered fuel-air mixture and of this spatial temperature gradient (∂T / ∂x) associated length (l) over which the temperature gradient (∂T / ∂x) is each substantially constant wherein the probability density (P) depends on at least one local turbulence parameter (k, <ε>, <T ' ² >, <Z' ² >) and at least one local thermodynamic parameter (p, T, λ), b) determining the local probability (P ' _knock ) by integrating the local probability density (P) into a local knock probability. A method according to claim 1, characterized in that - in step a) in addition a local probability density of a spatial mixture fractional gradient (∂Z / ∂x) of the fuel-air mixture and a mixing fraction gradient (∂Z / ∂x) associated length (l), about where the mixture fraction gradient (∂Z / ∂x) is each substantially constant, the local probability density of the mixture fraction gradient of at least one local turbulence parameter (k, <ε>, <Z ' ² >) and at least one local thermodynamic parameter (p, T, λ), in an additional step a1) following step a), the local probability densities of the temperature gradient (P) and the mixture fraction to a combined combined probability density of a temperature gradient (∂T / ∂x) and mixture fraction gradient (∂Z / ∂x) combined gradient of equal length (l) for both gradients be iniert. A method according to claim 1 or 2, characterized in that the local probability density (P) of the spatial temperature gradient (∂T / ∂x) or / and mixing fraction gradient (∂Z / ∂x) in step a) taking into account local turbulence effects of the fuel-air mixture be determined. Method according to one of claims 1 to 3, characterized in that the integration after step b) by integration within predetermined integral limits (g _u , g _l ), wherein the determination of the integral limits (g _u , g _l ) depending on the at least a local thermodynamic parameter and at least one local reaction kinetic parameter takes place. A method according to claim 4, characterized in that - the at least one local reaction kinetics parameter ignition delay time (τ) of an ignition process of the fuel-air mixture and / or the derivative of this Zündverzugszeit after the temperature (∂τ / ∂T) and / or after the mixture break (∂τ / ∂Z) and / or a sound excitation time of this ignition process (τ _e ), - the at least one local thermodynamic parameter is one of the following local parameters: - local temperature (T), - local mixture fraction (Z), - local kinematic viscosity (v), - local pressure (p), - local sound velocity (a), - local air ratio (λ), - the at least one local turbulence parameter is one of the following local parameters: - kinetic energy of the turbulence (k); - dissipation of the kinetic energy of turbulence (<ε>); Variance of the local temperature fluctuations <T ' ² >; Variance of the local mixing fraction <Z ' ² >; Method according to one of the preceding claims, characterized in that - the length (l), over which the temperature or Mischungsbruchgradient is constant, as a linear distance between a first space point of the fuel-air mixture with maximum temperature (T _max ) and a second spatial point with minimum temperature (T _min ) or between a first spatial point with maximum mixing fraction (Z _max ) and a second spatial point with minimal mixing fraction (Z _min ) are approximated, - the temperature gradient (∂T / ∂x) and the mixture fraction gradient ( ∂Z / ∂x) are approximated as follows: ∂T / ∂x α ΔT / l = (T _max - T _min ) / l or ∂Z / ∂x α ΔZ / l = (Z _max - Z _min ) / l. Method for determining a spatial probability distribution for the occurrence of super-knocking processes in a combustion chamber of an internal combustion engine, in particular a supercharged gasoline engine, comprising the following steps: A) defining a plurality of zones or a grid having a plurality of grid points in the combustion chamber, B) performing the method according to any one of the preceding claims for each zone or each grid point in the combustion chamber, so that each zone or each grid point is assigned a local probability. A method according to claim 7, characterized in that the method comprises the additional step C): C) determining the maximum value P ^max _knock from the local probabilities calculated in step B) for each zone or grid point. Control device ( 100 ) for a combustion chamber ( 101 ) having internal combustion engine ( 102 ), in particular a gasoline engine, characterized in that - the control device ( 100 ) is set up / programmed to carry out the method according to claim 7 or 8, wherein the control unit ( 100 ), in particular a calculation unit ( 106 ), by means of which a maximum probability of the occurrence of super-knocking processes in an introduced into the combustion chamber air-fuel mixture is determined, this determination taking into account global turbulence effects of the air-fuel mixture in the combustion chamber, or / and that - the control device ( 100 ) a storage unit ( 107 ) in which the spatial probability distribution determined by the method according to claim 7 or 8 can be stored. Motor vehicle, - with a combustion chamber ( 101 ) having internal combustion engine ( 102 ), - with a control device ( 100 ) according to claim 9, wherein the control device ( 100 ) the internal combustion engine ( 102 ) as a function of the calculation unit ( 106 ) or in the storage unit ( 107 ) stores the spatial probability distribution stored.

Images