FR2885392A1 - Petrol or diesel injector`s characteristics representing values defining method for internal combustion engine, involves modifying parameter values if results of digital simulation are far from experimental data or stopping choice of values - Google Patents

Abstract

The method involves carrying out test and measurement (B1) at different operating points to obtain significant experimental data (B4) of characteristics of injector in hot and cold conditions. A digital simulation (B3) with chosen set of parameter values (B2) is carried out for each operating points and results of each simulation is compared with the data. The values (B2) are modified if the results are far from the data and the choice of values (B2) is stopped if the results are sufficiently near the data.

Description

The invention relates to the design of engine combustion systems

  heat pumps comprising a combustion chamber equipped with a fuel injector.

  The optimization of such a combustion system is highly dependent on the coupling existing between the injection device and the combustion chamber. the injection conditions the mixture of air and fuel, which determines the combustion conditions, and therefore in particular the formation of pollutants.

  The design of such a system is carried out with numerical simulation methods associated with tests. These tests make it possible to validate the results of numerical simulation, and they provide experimental data allowing, if necessary, to recalibrate the numerical model.

  In the numerical model used, the combustion chamber is modeled by discretization in elementary volumes, and the injector is defined on the basis of existing physical models that are parameterized from the experimental data.

  A modification of the geometry of the combustion chamber can be easily transcribed in the model to perform a numerical simulation of the behavior of the modified combustion chamber.

  But the injector model used is defined and parameterized from experimental data from a test that is performed at a point of operation considered.

  A simulation at another operating point, according to other injection parameters, or simply with a different orientation of the injector requires corresponding tests to define an injector model that correctly translates these new conditions.

  Specifically, a change in the orientation or position of the injector requires the manufacture of a prototype cylinder head suitable for testing to examine the operation of the combustion system.

  Similarly, a modification of the injection pressure or another injection parameter involves manufacturing an injector prototype and performing physical tests of that prototype to define and use a valid numerical model.

  This situation is due to the complexity of the physical phenomena involved in the injection process, which are not completely solved by the mathematical models used.

  Fuel injection modeling uses a Eulerian formulation of the continuous gas phase Navier Stokes equations, and a Lagrangian formulation of the dispersed liquid phase equations of motion and energy balances.

  But this modeling does not solve the equations of the physics of the liquid flow in the immediate vicinity of the nozzle, that is to say about one millimeter thereof.

  It is therefore particularly necessary to initialize the Lagrangian equations describing the liquid, which is done from experimental data from tests of the injector at the point of operation considered.

  These tests make it possible to determine and introduce into the numerical model such characteristics as the size of the drops and their speed, as a function of time and space.

  The design and optimization of a combustion system therefore require many physical tests, which involves making as many prototypes, inducing a very significant development cost.

  The object of the invention is to propose a method for determining a numerical injector model making it possible to easily make variations of position, orientation, injection pressure, type of injector, for a low cost .

  For this purpose, the subject of the invention is a method for defining significant values of the behavior of a fuel injector in an internal combustion engine combustion chamber, for reproducing by numerical simulation the behavior of this multi-point injector. of operation, these values being parameters of a physical model of evaporation, drag and atomization implemented in the numerical simulation, this method comprising the following steps.

  a) perform tests and measurements at different operating points to obtain significant experimental data of the injector behavior under cold thermodynamic conditions and hot thermodynamic conditions; b) select a set of parameter values and perform a numerical simulation with the same set of parameter values for each operating point that has been tested, and then compare the results of each numerical simulation with the experimental data; c) modify the parameter values if the results of the numerical simulations are far from the experimental data and repeat in b), and stop the choice of parameter values if the results of the simulations are sufficiently close to the experimental data.

  According to one characteristic of the invention, the tests and measurements are carried out on the one hand under cold thermodynamic conditions at three injection pressures and at three injection times corresponding to two extreme points and at a midpoint of operation of the engine, and secondly in hot thermodynamic conditions at three injection pressures and three pairs of temperature and pressure of the injection chamber corresponding to two extreme points and a midpoint of engine operation.

  According to another characteristic of the invention, the chosen parameter values include initial injection conditions comprising an initial size value of the drops, an initial speed value of the drops, and at least one insertion angle value of the drops. drops, to initialize a liquid phase Lagrangian model that is implemented in numerical simulation.

  According to another characteristic of the invention, the quantities compared in b) are liquid penetration values in the combustion chamber, steam penetration and evaporation rate values under hot thermodynamic conditions, and size values. drops at a predetermined distance from the injector.

  According to another characteristic of the invention, steps b) and c) are implemented in an automated manner by implementing a generic algorithm for optimizing parameters.

  The invention will now be described in more detail and with reference to the accompanying drawings which illustrate an embodiment thereof by way of non-limiting example.

  The invention will be better understood, and other objects, features, details and advantages thereof will appear more clearly in the following explanatory description made with reference to the accompanying schematic drawings given solely by way of example illustrating a embodiment of the invention and in which: Figure 1 is a schematic representation of a combustion system seen in section; Figure 2 is a flowchart representative of the method of the invention; Figure 3 is a flowchart representative of a prior art method.

  Figure 1 shows one example among others of combustion system for gasoline direct injection engine. The geometric arrangement of the various components can vary significantly depending on the combustion concept used, but the problem of injection remains common to all these concepts.

  In the example of FIG. 1, a combustion system 1 comprises a combustion chamber delimited by walls 2 and by a piston head 3. An injector 4 and a spark plug 5 are mounted in the upper part of this chamber. combustion, to introduce therein fuel, and to start its combustion.

  The injector 4 introduces into the combustion chamber fuel 7, in a substantially dispersed distribution, called spray. At the injector outlet, that is to say about one millimeter from the injection nozzle, the fuel is in a so-called dense liquid phase, identified by 8 for which there is no relevant physical modeling, and then it is distributed in a phase called dispersed liquid, marked by 9, and in which it is in the form of a multitude of drops, which then evaporate to give a so-called gas phase.

  The behavior of the fuel jet that is injected into the combustion chamber is strongly conditioned by the operating point considered. We distinguish the so-called cold thermodynamic conditions, leading to a distribution of fuel in the room called cold bomb, and hot thermodynamic conditions corresponding to the operation of the engine in real conditions and giving a distribution of the fuel called hot bomb.

  The method according to the invention is based on physical tests and makes it possible to characterize the spray by means of a numerical simulation of two-phase flows using a Lagrangian approach for the dispersed liquid phase and an Eulerian approach for the continuous gas phase.

  It leads to a digital injector model that is called a virtual injector because it does not contain a numerical model of the geometry of the injector, unlike the combustion chamber whose geometry is discretized in elementary volumes.

  The virtual injector is a set of numerical quantities allowing to initialize the Lagrangian model of the liquid phase. These quantities include parameters such as the speed and size of the drops as a function of time and space, the temperature of the liquid, the angular orientation of the liquid jets, and the characteristic atomization rate.

  This method includes a first stage of tests to determine whether the injector is suitable for numerical modeling, a second stage of tests to determine experimental data characteristic of the injector, a third stage of choice of physical models to be put in numerical simulation, a fourth numerical simulation calculation parameter adjustment step, and a fifth adjustment step of physical model parameters and initial fuel introduction conditions to obtain close simulation results experimental data.

  The injector is primarily studied with regard in particular to the orientation of the nozzle and its static flow and others to define the computational domain that is to say the overall geometry of the bomb, and to predetermine the physical models to implement.

  We first use the geometric parameters of the injector, including the orientation of the nozzle and the number of holes that it comprises to define the computational domain, that is to say the overall geometry of the bomb. Then, the static flow information is used to feed the parameters of the physical models.

  The first step is a test to ensure that the injector considered has a repeatable operation, because the numerical modeling adopted treats an average case in the statistical sense of the term.

  This first step consists in testing the injector to visualize the injection in order to verify that quantities such as the liquid penetration and the characteristic angles of the spray have sufficiently small variations on a significant number of events. Advantageously, it is ensured that these quantities have variations of less than five percent, relative to an average value which is determined on the basis of all the events displayed during the test.

  This test, which appears in block B1 of FIG. 2, is carried out at several points of operation in hot high-pressure bombs. These different operating points correspond to three injection pressures, and three pressure / temperature pairs, which correspond to two extreme points and a middle point of engine operation.

  The second step is to perform tests, which also appear in block B1, to obtain significant experimental data from the spray coming out of the injector, which data is then used to establish and validate the numerical model. These are external geometrical characteristics of the injector, and characteristics of the spray under cold and hot thermodynamic conditions.

  The external geometrical aspects of the injector to be identified are mainly the dimensions and the characteristic opening angles or the fuel outlet nozzle.

  The characteristics of the spray are initially determined under cold thermodynamic conditions, for three injection pressures and for three injection times, which correspond to the extreme points and the mid-point of engine operation.

  The quantities identified are a fuel introduction rate, a liquid penetration value, a spray opening angle value, the liquid field rate, and the size and speed of the drops.

  The rate of introduction is determined by measurement of unsteady hydraulic flowmetering. The penetration of liquid, the opening angle of the spray, and the pace of the liquid field are defined by unsteady visualization in a high pressure bomb. The size and speed of the drops are measured in two planes perpendicular to the axis of the spray, about twenty-five millimeters apart, in a range of ten to sixty millimeters from the nozzle. The plane closest to the nozzle to be chosen not to be disturbed by the density of liquid. These measurements are performed by phased doppler particle analyzer, known by the acronym PDPA (Phase Doppler Particle Analyzer).

  In a second step, tests are carried out to obtain the characteristics of the spray under hot thermodynamic conditions, that is to say representative of those present in the combustion chamber in operation. This is done for three injection pressures and three pressure / temperature pairs corresponding to the extreme points and the midpoint of engine operation.

  The quantities identified are a liquid penetration value, a vapor penetration value, the vapor field rate, and an overall evaporation time value. These quantities are determined for a high pressure bomb by appropriate optical diagnosis such as darkening or addition, all acquisitions being made for a sufficient number of injections in order to obtain significant average values.

  A close-up view of the nozzle of the injector is also performed to detect if steam is coming out directly from the injector, that is to say if there are cavitation phenomena.

  The third step, which appears in block B2, consists of choosing physical models from different existing models, the numerical model of injector being defined by several different physical models.

  These physical models are more particularly a model of drag, an evaporation model and a model of atomization. The drag model and the evaporation model are chosen from the Mach number based on the speed difference between a drop and the gas flow around it. The atomization model is chosen to be adapted to the dominant instability mode of the liquid form which atomizes into small drops.

  The fourth step that appears in the B3 block is to choose numerical simulation calculation parameters that are optimal with respect to the optimal ratio between the accuracy or stability of the results of the numerical simulation and computation time.

  These numerical calculation parameters are the cell size of the mesh discretizing the geometric domain and a time increment value. The cell size must vary within a limited range at the limits compatible with the Lagrangian modeling approach of the dispersed liquid phase. Numerical simulation includes a three-dimensional computational code based on Navier-Stokes mathematical models or using a RANS modeling approach.

  This step is iterative and consists of performing a simulation, reducing the values of the calculation parameters, performing another simulation with the new calculation parameters, and comparing the results of the successive simulations. The choice of the calculation parameters is defined when the result provided by the numerical simulation no longer changes to a variation of the numerical parameters. This step can be performed by dividing the time step and cell size by two at each iteration. When the calculation parameters of the numerical simulation are chosen, they are used throughout the rest of the method.

  The fifth step consists in determining the physical parameters of each physical model chosen in the third step, and the parameters of the initial conditions of the flow of the liquid phase, all these quantities appearing in the block B2. The initial conditions are the initial size of the drops and their initial velocity as a function of time and space, as well as the angles of introduction of the drops.

  The search for a set of values for the single physical parameters for a whole range of operation makes it possible to define a model of global virtual injector that can be used whatever the point of engine operation.

  The choice of the parameters of each physical model and the initial conditions is performed by comparing the numerical simulation results with the experimental data from the first step, which corresponds to the loop linking the block B4 to the block B2, these parameters being chosen for meet a desired accuracy criterion. From the same set of physical parameters of the injector model, a numerical simulation is performed for each operating point which was tested in the first step.

  The comparison between simulation results and experimental data is then made for the following quantities: the value of liquid penetration, the shape of the liquid spray, the value of vapor penetration, the spatial distribution of vapor and liquid, the time required for Complete evaporation of the injected liquid, the probability density function of size and speed of drops in the planes at 30 and 50 millimeters from the nozzle of the injector. This comparison is made for each of these magnitudes, and at each operating point, on the basis of the same set of physical parameters.

  Comparisons from visualizations are made on the basis of images representing converged averages, the number of images required to obtain an average that may vary depending on the case considered. Advantageously, the comparison is preceded by a digital processing adapted to improve the accuracy of the comparisons, this treatment being for example a digital shadow processing.

  The search for the values of the parameters is made iteratively, by specifying in the numerical simulation input the quantities of the block B2, that is to say the speed of introduction, the size of the drops, the angle of the spray, the physical models. , drag, evaporation. The numerical simulation then provides results concerning the penetration of liquid and vapor, the size of the drops in the spray, the evaporation rate and the spatial and temporal distribution of the liquid and the vapor.

  The results of each simulation are compared with the experimental data, and if each comparison gives sufficient accuracy it means that the set of physical parameters is satisfactory.

  If the precision is insufficient, the physical parameters are modified as long as the difference between the numerical simulation results and the data from the tests is too great.

  The choice of new physical parameters to improve the accuracy of the results on all the comparisons is done either manually or using mathematical tools such as optimization algorithms such as genetic algorithm or other.

  This method makes it possible to define a virtual injector from a real injector independently of the injection technology and the internal geometry of the injector.

  The comparison of the results of the numerical simulations with the experimental data in high pressure hot and cold bombs ensures a validity of the virtual injector in homogeneous and stratified combustion.

  The virtual injector obtained gives precise results as regards the size of the drops, the distribution of liquid in the space, the vapor distribution, the amount of impact on the walls of the chamber, as well as the evaporation rate. .

  Since the definition of the virtual injector is made over ranges of injection conditions and wide thermodynamic conditions, the virtual injector can be used without any other adjustment than the transcription of the injection parameters such as the pressure and the duration of the injection. injection, over the entire operating range of the engine. This digital model or virtual injector can thus be integrated into a complete numerical model of the combustion system.

  The flowchart of a state-of-the-art approach is given in FIG. 3. This approach results in a numerical injector model that is valid for certain operating points, but not in a wide range of operation, in particular the amount of experimental data is insufficient, such data being limited to cold bomb tests, so that it is not possible to define a single set of physical parameters which is also valid at other points of operation, such as hot bomb.

  The method according to the invention constitutes a pragmatic solution of simulation at a macroscopic level which makes it possible to represent the overall behavior of the injector with a sufficient precision, while maintaining acceptable calculation times.

  This method is independent of the internal and external geometry of the injector: only the characteristic dimensions of the injector nozzle are necessary. It offers a systematic approach to define a virtual injector regardless of the technology of the injector.

  The injector defined according to this method offers a high precision as regards the size of the drops, the distribution of liquid in the space, the distribution of vapor, the amount of impact on the walls of the chamber, and the speed of 'evaporation. The virtual injector thus defined is valid for the preparation of the mixture and the liquid impact on the walls of the combustion chamber for the entire range of operating points engine.

  The method according to the invention ensures that sufficient validations are carried out before simulation, which also makes it possible to reduce the overall cost of the numerical simulations necessary for the development of a combustion system. Indeed, if one applies an insufficiently validated spray calibration method such as that illustrated on the flowchart of FIG. 3, it is necessary to modify the parameters of the model iteratively, which is particularly expensive.

  The invention applies to multiple types of internal combustion engines such as piston engines, rotary engines, used in the automotive field, motorcycles, light aircraft or construction equipment.

Claims (5)

  A method for defining significant values of the behavior of a fuel injector (4) in a combustion chamber (2) of an internal combustion engine, for reproducing by numerical simulation the behavior of this injector (4) at several points of operation, these values being parameters of a physical model of evaporation, drag and atomization implemented in the numerical simulation, this method comprising the following steps: a) performing tests and measurements (B1) at different operating points to obtain significant experimental data of the behavior of the injector (4) in cold thermodynamic conditions and in hot thermodynamic conditions; b) select a set of parameter values (B2) and perform a numerical simulation (B3) with the same set of parameter values for each operating point that has been tested (B1), and then compare the results each numerical simulation with the experimental data (B4); c) modify the parameter values (B2) if the results of the numerical simulations (B3) are far from the experimental data (B4) and repeat in b), and stop the choice of the parameter values if the results of the simulations are sufficiently close to the experimental data.   2. Method according to claim 1, wherein the tests and measurements (Bl) are carried out on the one hand under cold thermodynamic conditions at three injection pressures and three injection times corresponding to two extreme points and a midpoint. operating the engine, and secondly in hot thermodynamic conditions at three injection pressures and three pairs of temperature and pressure of the injection chamber corresponding to two extreme points and a midpoint of engine operation.   The method of claim 1 or 2, wherein the selected parameter values (B2) include initial injection conditions including an initial drop size value, an initial drop velocity value, and at least one d-value. angle of introduction of the drops, to initialize a Lagrangian model of liquid phase which is implemented in the numerical simulation.   4. Method according to one of claims 1 to 3, wherein the compared values (B4) in b) are liquid penetration values in the combustion chamber (2), steam penetration and velocity values d evaporating under hot thermodynamic conditions, and drop size values at a predetermined distance from the injector (4).   5. Method according to claim 4, wherein steps b) and c) are implemented in an automated manner by implementing a generic algorithm for optimizing parameters.