DE102018127711A1 - Method for developing an internal combustion engine - Google Patents

Abstract

To assess compliance with RDE pollutant limits (E _{i_Grenz} ) of an internal combustion engine (2) is provided that from frequency distributions (4, 5) of reference operating points (P _Bi ) and changes (P _BAi ) and stationary test bench measurements a stationary _{degree of} fulfillment (EG _stat ) of an emission limit _value (E _{i_Grenz} ) is determined that from transient test bench measurements of the operating point changes (P _BAi ) pollutant critical operating point changes (P _BAi ) are determined that from pollutant emissions ( E _{i_Zyklus} ) a test cycle '(Z) and quasistationary pollutant emissions (E _{i_quasi} ) a quasi-stationary test cycle' (Z _quasi ) a dynamic degree of fulfillment (EG _dyn ) is determined that test bench test runs (P _Li ) with different virtual drivers (D _v ), various virtual vehicles (F _v ), various virtual environments (U _v ) under measurement of pollutant emissions (E _i ) durchgefü be carried and from a robustness performance level (EC _Robust) is determined that it becomes a Repräsentativprüflauf (15) is determined and that the Repräsentativprüflauf (15) to optimize the pollutant emissions (E _i) is used at the test.

Description

The invention relates to a method for developing an internal combustion engine of a given engine concept with a control unit with variable control parameters for controlling the internal combustion engine.

There are legal requirements for permissible pollutant emissions of vehicles (in particular CO ₂ , CO, NO _x and particle number), eg Regulation (EC) No 715/2007 of the European Parliament and of the Council, in which the Euro 5 and Euro 6 Standard are defined are. Compliance with these legal requirements by vehicles has so far been tested on test rigs by means of standardized test cycles (such as the New European Driving Cycle (NEDC)). For this purpose, the exhaust gas generated during the test cycle is taken from the test bench and examined. The problem here is that the conditions on the test bench under the standardized test cycle are not comparable to the real conditions when the vehicle is moved on a real track. This allows a vehicle to comply with the legal requirements on the test bench, but still exceed them in real operation.

In order to prevent this there is a desire of the legislator to shift the review of the prescribed pollutant emission limit values from the test bench to the real road. This requires pollutant emissions to be measured and verified during a real vehicle journey on a real track using Portable Emission Measurement Systems (PEMS). There is thus no longer a standardized test cycle, because driving on a public road with normal traffic is always subject to random influences. The legislator's aim is for a vehicle to comply with the limits of pollutant emissions under normal operating conditions, and not just at the test bench. Also for the evaluation of pollutant emissions after the test drive are made by the legislator, e.g. the use of certain data analysis tools.

The legislator defines a Real Driving Emissions (RDE) test procedure. It only specifies certain specifications for the vehicle mass, the ambient temperature and the geographical altitude under which the test drive must take place. In addition, it is still defined to what proportions different driving situations must be included in the test procedure, for example, 33% ± 10% division on city, country and highway but at least 16km each, vehicle speed country in the range of 60-90km / h, length of test drive 90 - 120min, etc. After this review is to take place on a public road, the respective test drive is also subject to random influences, such as foreign traffic, traffic lights, etc. It is immediately apparent that a real test drive is not reproducible, but one more or less represents random event chain.

This paradigm shift also has a direct impact on vehicle manufacturers in developing new vehicles and especially new engines.

So far, each development step could be checked by means of the standardized test cycles on the test bench. For this purpose, only the respective test specimen had to be subjected to the test cycle after each development step, and the pollutant emissions investigated. This is no longer possible with the new RDE test procedure, since it is fundamentally unpredictable whether a newly developed vehicle will survive the RDE test procedure at the end of development by adhering to the limit values for pollutant emissions. Only the finished vehicle can be moved on the real road, i. it can only be done at the very end of the development of the RDE test procedure. If the vehicle does not stand up to this review, it would be understandable that this would have enormous effects on the vehicle manufacturer, which in extreme cases would at least partially have to reopen a year's development at enormous cost and effort.

The use of the previous standardized test cycles during vehicle development also does not help here since compliance with the limit values of the pollutant emissions using such standardized test cycles does not automatically ensure compliance with these limit values under the RDE test procedure.

One could now merge all possible operating states of a vehicle in a test scenario and use this test scenario to check each development step. However, this is hardly expedient, because the implementation of such a test scenario on a test bench would take a very long time, which would delay development, which would increase expensive test bench times and would be very complex overall. An arbitrary creation of a test scenario is also not effective as it does not It could be ensured that compliance with the legal requirements will be achieved under the RDE test procedure.

In addition, not every driving maneuver, e.g. Acceleration from a low speed, overtaking on a country road, turning in city traffic, etc., must have the same effect on pollutant emissions in every vehicle. This means that one test scenario may be appropriate for a particular vehicle but not for another vehicle.

Basically, the above applies equally to other targets of the development of a vehicle, for example, for the consumption of a vehicle, although there are no (at least not yet) legal requirements. Consumption, however, is usually also a development goal in the development of a vehicle, which is why the goal of achieving the desired consumption, for example in the RDE test procedure, is also sought here.

It is therefore an object of the invention to provide a method which makes it possible, even in early development phases of an internal combustion engine, to obtain meaningful results on the compliance with prescribed limit values of pollutants by the new RDE legislation without testing the internal combustion engine in a vehicle.

The object is achieved in that reference operating data for the internal combustion engine are determined that from a time-based and / or wegbasierte frequency distribution of operating points and a frequency distribution of operating point changes of the reference internal combustion engine is determined that stationary measurements of the internal combustion engine in the determined operating points of the Reference internal combustion engine are carried out on a test bench, wherein at least emissions of a pollutant are measured, that from the frequency distributions of the operating points and the measured emissions of the pollutant at the respective operating points, a stationary degree of fulfillment of a legally defined emission limit value of the pollutant is determined that transient measurements of the determined Operating point changes are performed with the internal combustion engine on the test bench, wherein at least the emissions of the pollutant are measured and ei ne plurality of, for the emissions of the pollutant critical operating point changes is determined that at least a predetermined test cycle is performed with the engine at the test bench, the at least one emission of the pollutant of the test cycle is measured and determines therefrom a dynamic degree of cyclic efficiency of the statutory emission limit of the pollutant that a quasi-stationary test cycle of the given test cycle is simulated from the given test cycle and the stationary measurements, whereby quasi-stationary emissions of the pollutant of the quasi-stationary test cycle are determined and a steady-state cycle satisfaction level of the legislated emission limit value of the pollutant determined therefrom, that of the emissions of the pollutant of the given test cycle and the quasi-stationary emissions of the pollutant of the quasi-stationary test cycle we determine a dynamic degree of fulfillment d, that a plurality of test runs of the internal combustion engine with a plurality of different virtual drivers and / or different virtual vehicles and / or different virtual environments is performed on a test bench, wherein at least the emissions of the pollutant are measured in the test runs and from a robustness Fulfillment level of the statutory emission limit value of the pollutant is determined that at least from the majority of testing a, for the emissions of the pollutant critical, representative test run is determined and that the representative test run is used to optimize the at least one emission of the pollutant on the test bench. By statistical analysis and determination of simple quantifiable characteristics of the internal combustion engine, a representative test run can be generated, which makes it possible to make a statement about the compliance with RDE limit values already in an early development phase of the internal combustion engine.

Preferably, the reference operating data are determined by at least one reference vehicle with a reference engine with a comparable with the engine engine concept that at least one reference distance is set, that at least one reference driver is determined that with the at least one reference vehicle and the at least one Reference driver a number of real test drives on the at least one reference distance is performed and from the relevant reference operating data of the reference internal combustion engine and the reference vehicle are detected. It is particularly advantageous if at least four different reference paths and at least three different reference drivers are provided in order to generate a sufficiently large and meaningful database.

Preferably, the steady-state degree of satisfaction and / or the dynamic degree of cycle satisfaction and / or the stationary degree of cyclic fulfillment and / or the dynamic degree of fulfillment and / or the Robustness degree of satisfaction optimized by iterative changes in the control parameters of the control unit of the internal combustion engine are performed and / or components of the internal combustion engine to be changed. This results in an iterative improvement process and the internal combustion engine can be optimized in the course of the process.

Hydrocarbons and / or nitrogen oxide / dioxide and / or a particle number and / or carbon monoxide are advantageously provided as pollutants. As a result, the emissions of the legally limited pollutants are used as development variables.

Preferably, the robustness degree of satisfaction is determined by forming a sum of a mean and a standard deviation of the measured emission of the pollutant from the plurality of test runs and comparing the sum with the legally determined emission limit value of the pollutant. As a result, a meaningful characteristic value can be determined with simple statistical and mathematical relationships.

• 1 den Ablauf des erfindungsgemäßen Verfahrens anhand eines Blockschaltbildes,
• 2 Diagramme mit Häufigkeitsverteilungen von Betriebspunkten,
• 3 einen Vergleich zwischen transienten und quasistationären Betriebspunktrampen in einem Diagramm,
• 4 einen Vergleich von Messergebnissen eines vorgegebenen Messzyklus und eines quasistationären Messzyklus,
• 5 Aufbau und Funktionsweise eines EiL-Prüfstands,
• 6a-6c Jeweils ein Diagramm einer Schadstoffemission über einer Anzahl von Prüfläufen aufgetragen,
• 7 die Ermittlung von Wechselwirkungsdiagrammen,
• 8 ein Blockschaltbild zur Ermittlung eines Repräsentativprüflaufs.

The subject invention will be described below with reference to the 1 to 7 explained in more detail, the exemplary, schematic and not limiting advantageous embodiments of the invention show. It shows

• 1 the sequence of the method according to the invention with reference to a block diagram,
• 2 Diagrams with frequency distributions of operating points,
• 3 a comparison between transient and quasi-stationary operating point lamps in a diagram,
• 4 a comparison of measurement results of a given measurement cycle and a quasi-stationary measurement cycle,
• 5 Structure and function of an EiL test bench,
• 6a-6c Plotted a diagram of a pollutant emission over a number of test runs,
• 7 the determination of interaction diagrams,
• 8th a block diagram for determining a representative test run.

In 1 is the basic sequence of the method according to the invention 1 for the development of an internal combustion engine 2 illustrated by a block diagram. The procedure 1 For example, it is essentially divided into five main modules HM ₁ - HM ₅ and for example three additional modules ZM ₁ - ZM ₃ , The arrows symbolize the flow of information between the individual modules. The individual main modules HM _i (Index i denotes the number of main modules HM) of the method 1 build on each other, that is, each module HM _i , obtained information from the previous main module HM _i-1 uses and processes. The additional modules ZM _i provide information about necessary boundary conditions of the procedure 1 to disposal.

In the main module HM ₁ become stationary test bench measurements of the combustion engine to be developed 2 performed on an engine test bench. The test bench measurements are carried out in a known manner, which is why the details of the design and mode of operation of a test stand are not discussed in more detail here (the basic structure of an engine test bench is described in US Pat 5 illustrated by a special implementation of an engine-in-the-loop test rig (EiL)). In stationary measurements, certain operating points of the internal combustion engine 2 set on the test bench and certain sizes are measured. An operating point P _Bi of the internal combustion engine 2 is essentially due to a speed N and a torque M (generally a load) defined. As a rule, a large number of measured variables are measured on the test bench or certain calculated variables are calculated from the measured variables measured, for example fuel consumption, temperatures, mass flows, pressures and, in particular, pollutants E _i , Of particular interest are pollutants E _i because these are limited by law, as described above. The development of internal combustion engines 2 Accordingly, it is based on compliance with legally prescribed pollutant limit values E _iGrenz , For example, limit values for vehicles E _iGrenz the emission of nitrogen oxides NOx, carbon monoxide CO, hydrocarbon HC, PM particles set (the index i stands for these different pollutants E _i ). A vehicle manufacturer must demonstrate that a manufactured vehicle also complies with these specified limits in order to obtain approval for the vehicle in question in a particular country or region, eg Europe, USA, Japan, etc.

To _verify compliance with the limits E _iGrenz , certain test cycles Z that a vehicle had to complete on a _{chassis dynamometer} were previously required. In Europe, for example, the so-called "New European Driving Cycle" (NEDC) was used, which essentially specified a speed profile of the vehicle. However, the exhaust gas legislation was not harmonized so far, which is why it also has different test cycles for different countries Z gave. With these fixed cycles Z, the manufacturers had a relatively simple and, in particular, fixed tool, with which they can drive their vehicles and in particular internal combustion engines 2 could develop. The vehicles and thus also the combustion engines 2 were thus optimized so that they met exactly the specifications of the NEDC or the respective test cycle Z. However, the driving profile of the test cycles Z does not correspond to the real operation on the road, which is why there are clear deviations between the real emissions of pollutants E _i during operation of the vehicle on the road and the operation on the test bench gave.

Due to ever greater deviations between the pollutant emissions of the real operation and test bench measurements, the new Real Driving Emission legislation (RDE) was introduced. The manufacturers are thus forced to a much larger operating range of the vehicle and thus also of the internal combustion engine 2 to be considered in the development, since a vehicle essentially no longer has to be designed for a known test cycle Z but for any unknown test cycle. With the method according to the invention 1 It is possible, even in a very early development phase of the internal combustion engine 2 to prepare the motor for RDE operation robustly without a vehicle and to make a basic statement about a later compliance with the RDE limit values.

In order to be able to take the best possible account of the unknown constraints of RDE legislation without uneconomically increasing the development effort and the associated costs, it is necessary to set certain boundary conditions for the procedure 1 set. In the present case, this happens in the additional module ZM ₁ using statistical methods. For this purpose, preferably at least one reference vehicle F _R , several reference drivers D _R and several reference links S _R defined and every reference driver D _R preferably runs several times with each reference vehicle F _R on every reference track. The reference vehicle F _R is chosen so that it is essentially the future application of the engine to be developed 2 corresponds, as should the engine concept of the reference engine 2 _R of the reference vehicle F _R preferably similar to the concept of the engine to be developed 2 his. As similar example, engines with similar displacement, similar charge, similar torque / medium pressure, etc. to see or reference vehicles FR with similar vehicle mass m, similar planned application area, similar powertrain (manual transmission, automatic transmission, all-wheel drive).

reference vehicles F _R For example, vehicles of different vehicle classes with the same internal combustion engine 2 be, as is common in many cases today. In particular, the vehicle mass of the various reference vehicles F _R should differ to a wide operating range of the internal combustion engine 2 cover. The reference driver D _R should have a wide range of driving behavior, from aggressive driving with many acceleration phases, high engine speeds N and higher vehicle speeds v to passive drivers with low vehicle speeds v, few low acceleration phases and low engine speeds N. The reference distances S _R should preferably also differ significantly and should have different urban, interurban and motorway sections. The proportions of the urban, interurban and motorway sections along the entire route are preferably based on the specifications of the RDE legislation. During the many reference runs certain reference operating data of the reference vehicles F _R and reference motors 2 _R recorded, these are in particular time t, vehicle speed v, engine speed N, engine load or torque M, oil temperature, coolant temperature, etc. If the reference runs with Portable Emission Measurement Systems (PEMS), also pollutants E _i be measured. The reference operating data can for example be measured directly or from a control unit of the reference vehicle F _R and / or the reference internal combustion engine 2 _R be determined. Preferably, however, not for every implementation of the method 1 new reference runs are performed, but it is preferably a database 3 created with reference operating data of the measured reference runs and depending on the combustion engine to be examined 2 and the planned vehicle application, desired reference journeys can be selected with corresponding reference operating data and for the method according to the invention 1 be used. By measuring, recording and storing many reference runs, a steadily growing database is generated, which increases the statistical significance.

In the specific embodiment, ninety-two reference measurements will be made with four different reference paths S _R , two different reference vehicles F _R and three different reference drivers D _R used. The reference runs are preferably both cold and warm of the reference internal combustion engine 2 _R performed to also cover the influence of various operating conditions (cold, warm). From the reference measurements, the reference operating data are determined as a function of the time t and the distance traveled s and become a temporal and a path-based frequency distribution 4 . 5 the reference operating data determined, in particular the operating points P _Bi of the reference internal combustion engine 2 as in 2a and 2 B is shown. In addition, a frequency distribution 6 the operating point changes P _BAi determined as in 2c is apparent. In the diagram of the temporal frequency distribution 4 in 2a are the operating points P _Bi as a function of time t (P _Bi = P _Bi (t)) above a normalized speed N _N and a normalized torque M _N of the reference internal combustion engine 2 _R Plotted in percent. Normalized means that the respective torque M and the respective speed N of an operating point P _Bi to the rated speed N _max and the nominal torque M _{max of} the reference internal combustion engine 2 _R are related. The size of the circles of the operating points P _Bi (t) symbolizes the respective temporal part of an operating point P _Bi (t) at the total time t of all reference runs. Analogous in the diagram are the path-based frequency distribution 5 in 2 B the operating points P _Bi as path-based operating points P _Bi (s). The size of the circles show the proportion of the respective operating point P _Bi (s) on the total distance traveled s of all reference runs. In the diagram in 2c is the frequency distribution 6 the operating point changes P _BAi over the speed change D _N in rpm per second and the torque change D _M shown in percentages per second. From this statistically meaningful results can be derived, as long as certain operating points P _Bi of the internal combustion engine 2 during the ninety-two reference trips occur and which distance s is covered. The frequency distribution 6 the operating point changes P _BAi in 2c provides information about the transient operation of the reference internal combustion engine 2 _R , It can be seen that in the example shown in about 72% of all ninety-two reference runs, the change of the load or torque M of the reference internal combustion engine 2 _R less than about 10% per second and the change in the speed of the internal combustion engine 2 less than about 100 revolutions per second. From this it can be concluded that the real driving operation covered by the homing operations has for the most part a stationary characteristic. This already provides an important information gain for the further development of the internal combustion engine 2 there and means that the stationary optimization of the, from the homing determined, operating points P _Bi a significant importance in terms of compliance with RDE pollutant limits E _iGrenz due.

The additional module ZM ₂ serves to ensure a complete database describing system behavior in all subsequent modules of the procedure 1 , In concrete terms, this means the collection of all the necessary measured variables in a measurement plan, the so-called system state structure, which is the result of the subsequent measurements in the main modules HM _i should be used. The system health structure includes all metrics that are conceptually specific to allow for emission scores in the impact modules and any technical cause attributions for identified emission problems.

der den gemessenen stationären Schadstoffausstoß E_iStat des jeweiligen Schadstoffes E_i dem gesetzlich festgelegten Grenzwert E_iGrenz gegenüberstellt. Der stationäre Erfüllungsgrad EGstat ist dimensionslos, was aus der Division zweier identer Einheiten g/km resultiert. Vorzugsweise werden die stationären Betriebspunkte P_Bi des Verbrennungsmotors 2 am Prüfstand solange optimiert, bis ein Erfüllungsgrad EGstat < 1 erzielt wird. Dabei können beispielsweise die Kalibrierdaten des Verbrennungsmotors 2, über die ja bekanntermaßen der Betrieb des Verbrennungsmotors 2 gesteuert wird, adaptiert werden. Ist der Erfüllungsgrad EG_Stat für einen Schadstoff E_i kleiner eins, ist der Verbrennungsmotor 2 hinsichtlich seines stationären Betriebsverhaltens im Wesentlichen zur Erfüllung des RDE-Grenzwerts des entsprechenden Schadstoffs E_i ausreichend kalibriert. Der stationäre Erfüllungsgrad EGstat könnte aber auch unabhängig vom gesamten Verfahren 1 z.B. zum Vergleich verschiedener Konzepte oder verschiedener Entwicklungsstände des Verbrennungsmotors 2 verwendet werden.Based on the information gained from the additional modules ZM ₁ and ZM ₂ be in the main module HM ₁ stationary test bench measurements of the internal combustion engine 2 carried out on the test bench. This will be all operating points P _Bi of the internal combustion engine 2 , Preferably by means of a characteristic map of the load or the torque M over the speed N of the internal combustion engine 2 , are set on the test bench and all measured variables specified in the additional module ZM2 (system status structure) are measured, in particular the emission of the relevant pollutants E _i , About the temporal and path-based frequency distribution 4 . 5 can already be a first comparison with the statutory limits E _iGrenz the pollutants E _i made in grams per kilometer [g / km]. In doing so, the emissions of pollutants E _i , which are typically measured in grams per second [g / s], using temporal and path-based frequency distributions 4 . 5 in grams per kilometer [g / km] converted to the respective statutory limits E _iGrenz to be comparable. This makes it possible, even in this very early development phase, to use purely stationary test bench measurements of the internal combustion engine 2 (without vehicle) a prognosis regarding compliance with the RDE limit values of a vehicle with the internal combustion engine 2 close. The quantification takes place via a so-called stationary degree of fulfillment $e{G}_{Stat}=\frac{e_{tStat}}{e_{iGrenz}}\frac{\left[G\text{/}km\right]}{\left[G\text{/}km\right]}.$

the measured steady-state pollutant emissions E _iStat of the respective pollutant E _i the statutory limit E _iGrenz faces. The stationary degree of fulfillment EGstat is dimensionless, which results from the division of two identical units g / km. Preferably, the stationary operating points P _Bi of the internal combustion engine 2 optimized at the test stand until a degree of fulfillment EGstat <1 is achieved. In this case, for example, the calibration data of the internal combustion engine 2 , about the known yes the operation of the internal combustion engine 2 is controlled, adapted. Is the degree of fulfillment EC _Stat for a pollutant E _i less than one, is the combustion engine 2 essentially with regard to its stationary operating behavior to meet the RDE limit of the corresponding pollutant E _i sufficiently calibrated. The inpatient degree of fulfillment EGstat could also be independent of the entire procedure 1 eg to compare different concepts or different development statuses of the internal combustion engine 2 be used.

In the next step of the procedure 1 become transient measurements of the internal combustion engine 2 performed on the test bench, as by the main module HM ₂ in 1 is symbolized. In the process, the influence of the dynamic operation and different operating states of the internal combustion engine becomes 2 on the emissions of pollutants E _i determined. Various operating states are, for example, a cold start, a warm-up phase and a regular engine operation at a certain operating temperature of the internal combustion engine 2 , which is usually in the range between 85 ° C and 90 ° C coolant temperature or approximately 80-120 ° C oil temperature. Test runs during cold start and in the warm-up phase are therefore preferably performed in order to cover the influences of any correction functions in the calibration data on the emission behavior, which in the control unit of the internal combustion engine 2 can be deposited. Such correction functions change, for example, certain parameters of the ignition, the fuel injection, etc. in cold operating condition, for example, to bring an exhaust aftertreatment system faster to a desired operating temperature. In particular, when the exhaust aftertreatment system has not yet reached the desired operating temperature, the internal combustion engine abuts 2 usually higher pollutants E _i out.

The transient measurements are carried out on the test bench by means of operating point lamps R _PBi , so-called operating phases. By means of statistical Design of Experiments (DoE) certain operating phases are defined, by means of which the test bench measurements are carried out. Experiments planned using DoE have the advantage that the scope of the measurements is not too large and the results nevertheless have a high significance. For the invention it can be assumed that the operating phases are predetermined. The result of the transient measurements are the most critical operating point lamps R _PBi or operating phases, ie those transient operating phases that lead to the highest emission of a pollutant E _i to lead. From this one can draw conclusions about the technical causes and possibly changes to components of the internal combustion engine 2 and / or make the settings of the control unit.

To identify the most critical operating phases, the measurement results of emissions of pollutants E _i , the operating phases measured at the test bench, compared to quasi - stationary operating phases, as in 3 based on an operating spot ramp P _BAi between a time t ₁ and a time t ₂ is shown. This shows the torque M (in percent of the nominal torque), the speed N in rpm and the emission of the pollutant CO in grams per second. It can be seen that the speed N and the torque M on the operating spot ramp P _BAi fall between the time t ₁ and the time t ₂ . The quasi-stationary operating phases become from the stationary measurements of the first main module HM ₁ determined, thus essentially form the transient operating point R _PBi with stationary operating points P _Bi and the corresponding stationary emissions of the pollutant E _i from. It is calculated as the difference between transient and quasi-stationary pollutant emissions (hatched area 7 ) for each operating spot ramp R _PBi an average deviation is determined and it becomes those operating spot lamps R _PBi which have the largest mean deviations between transient and quasi-stationary emission of pollutants. In the example shown in FIG 3 the deviation is shown as a hatched area between the transient pollutant emissions CO _t (solid line) of the pollutant CO and the quasi-stationary pollutant emission CO _quasi (dashed line) of the pollutant CO.

From the selected operating spot lamps R _PBi with the biggest deviations and, in the add-on module ZM ₁ determined, statistical frequency distribution 6 the operating point changes P _BAi The reference runs become a so-called ramp weight G _egg for the emissions of pollutants E _i determined according to the following equation. $$G_{ei}=\frac{\mathrm{\Delta }{\dot{m}}_{ei}}{\mathrm{\Delta }{\dot{m}}_{ei\_\text{Max}}}*\frac{H_4}{{\mathrm{<figure-callout\; id="4"\; label="H"\; filenames="DE102018127711A1\_0006.png"\; state="\{\{state\}\}">H</figure-callout>}}_{4\_\text{Max}}}*\frac{H_6}{{\mathrm{<figure-callout\; id="5"\; label="H"\; filenames="DE102018127711A1\_0005.png,DE102018127711A1\_0006.png"\; state="\{\{state\}\}">H</figure-callout>}}_{5\_\text{Max}}}*\frac{1}{G_{ei\_\text{Max}}}$$

In this case, Δṁ _Ei denotes the mean deviation between transient and quasi-stationary mass flow of a pollutant emission (average deviation of the hatched area 7 in 3 ) for an operating spot ramp R _PBi . h ₄ the from the frequency distribution 4 in the main module HM ₁ determined frequency of the respective operating point P _Bi at the beginning of an operating spot ramp R _PBi and h ₆ the from the frequency distribution 6 in the main module HM ₁ determined frequency operating point changes P _BAi , The values are normalized by placing them on the, between the time t ₁ and time t ₂ measured maximum value.

The higher the ramp weight G _egg a service point lamps R _PBi , the higher the mean deviation 7 between transient and quasi-stationary pollutant emissions of the considered operating spot ramp R _PBi and the higher is the statistical relevance, ie the frequency of occurrence of the corresponding operating point lamps R _PBi in all homed tours.

Preferably, the described determination of the most critical operating point lamps takes place R _PBi automatically by means of a suitable test bed automation system 8th (such as in 5 is shown). From the operating spot lamps R _PBi with the highest ramp weight G _egg can be a desired number of operating spot lamps R _PBi be selected and the transient emission behavior of the internal combustion engine 2 based on these selected operating spot lamps R _PBi be optimized, for example by adaptation of the calibration data. But it can also operate spot lamps R _PBi with a high ramp weight G _egg despite the high ramp weight G _i in the optimization and further development of the internal combustion engine 2 be neglected, for example, when these operating spot lamps R _PBi be avoided by a suitable switching strategy of an automatic transmission, ie in practical operation of the internal combustion engine 2 do not occur at all.

The results of the second main module HM ₂ (transient measurements) flow together with the results of the first main module HM ₁ (stationary measurements) in a third additional module ZM ₃ a predictive model for determining the CO ₂ Emissions and emissions of pollutants E _i provides. By means of the prognosis model can be a emission of a pollutant E _i Any predetermined test cycle Z can be determined. Thus, the empirical data from the reference runs are used to make an emissions forecast of an unknown test cycle Z. This prognosis model will be in the following third main module HM ₃ and fifth main module HM ₅ used as indicated by the arrows for information flow in 1 is apparent.

As at the arrows in 1 As can be seen, the results flow from the second main module HM ₂ and the third add-on module ZM ₃ in the third main module HM ₃ - cycle validation - one. In the third main module HM ₃ becomes one or more known test cycles Z with the internal combustion engine 2 carried out on the test bench. The advantage of a predetermined test cycle Z is that this is reproducible feasible, creating a quick and easy comparison of different concepts, development stages, components, etc. of the engine 2 can be carried out. In essence, this corresponds to the previous development method, where an internal combustion engine 2 has been studied and developed on the basis of a defined, legally prescribed test cycle Z.

Preferably, a Worldwide Harmonized Light Vehicles Test Cycle (WLTC) cycle and an RDE compliant test cycle Z are used. RDE compliant means that the test cycle Z can be essentially arbitrary, for example derived from a recorded driving test of a vehicle on the road, but that certain boundary conditions of the RDE legislation are complied with, eg the proportion of urban, interurban and motorway travel , However, other known cycles could be used, such as the NEDC, FTP75, etc. From the measured WLTC cycle and the RDE-compliant test cycle Z are the emissions of the pollutants E _i recorded over time. From the stationary measurements of the main module HM ₁ is a so-called quasi-stationary comparison cycle of the measured cycles (here WLTC, RDE compliant) is formed and with the additional module ZM ₃ (Emissions forecasting model) will be corresponding quasi-stationary emissions of pollutants E _i of the quasi-stationary test cycle Z, so to _speak determined. The quasi-stationary test cycle Z, so to _speak is therefore one, in the form of stationary operating points P _BI simulated test cycle Z, which means that to the operating points P _Bi of the test cycle Z the measured values of the pollutants E _i the corresponding stationary operating points P _Bi be used. This is preferably done automatically, for example in a suitable evaluation software.

The actual results measured at the test stand are compared with the results of the emission forecast, as in 4 represented by a diagram. The course of the measured emissions E _{i_cycle} the pollutants E _i and other desired magnitudes of the test cycle Z are plotted over time. The curves of the transient test cycle Z measured on the test bench are shown as solid lines and the quasi-stationary curves E _{i_quasi} of the emission prediction model are shown as dashed lines. In addition to the pollutants E _i (here in the form of the pollutant CO _cycle , CO _quasi ) are preferably all in the additional module ZM ₂ set measured variables, for example, a firing angle ZW _cycle . ZW _quasi , a fuel injection quantity, an air ratio (λ _cycle , λ _quasi ), etc., which provide information about technical or physical causes of a deviation of a pollutant E _i can give. Certain parameters of any exhaust aftertreatment system or operating conditions (coolant temperature T _{KM cycle} , T _KMquasi ) can also be analyzed.

From the measured progressions E _{i_cycle} of pollutants E _i , the corresponding quasi-stationary curves E _{i_quasi} of pollutants E _i as well as the statutory limits E _{i_limit of} the individual pollutants E _i For example, three further dimensionless _{degrees of} fulfillment EG _{dyn_cycle} , EG _{stat_cycle} and EG _dyn are calculated as shown below. $$e{G}_{dyn\_Zyklus}=\frac{e_{i\_Zyklus}}{e_{i\_Grenz}}\frac{\left[G\text{/}km\right]}{\left[G\text{/}km\right]}.e{G}_{stat\_Zyklus}=\frac{e_{i\_quasi}}{e_{i\_Grenz}}\frac{\left[G\text{/}km\right]}{\left[G\text{/}km\right]}.e{G}_{dyn}=\frac{e_{i\_Zyklus}}{e_{i\_quasi}}\frac{\left[G\text{/}km\right]}{\left[G\text{/}km\right]}$$

The _degree of fulfillment EG _{dyn_cycle describes} the relationship between the measured emission of a pollutant E _i of the transient test cycle in [g / km] (here WLTC, compliant with RDE) and the statutory limit E _{i_ limit of} the corresponding pollutant E _i in [g / km]. The _{degree of} fulfillment EG _{stat_cycle} describes the relationship between _,, from the third additional module ZM ₃ (Emissions forecast model) determined emission of a pollutant E _i a quasi-stationary cycle and the legally established limit E _{i_ limit of} the corresponding pollutant E _i in [g / km]. EC _{stat_cycle} says how good the stationary behavior of the internal combustion engine 2 in terms of compliance with the legal limit E _{i_ limit of} a pollutant E _i is designed and is together with that, in the first main module HM ₁ determined stationary degree of fulfillment EG _stat an important parameter for stationary optimization of the internal combustion engine 2 , The quasi-stationary cycle is the transient cycle (in this case WLTC, RDE-compliant), represented in the form of stationary operating points.

The _degree of fulfillment EG _dyn sets the measurement result of an emission E _{i_cycle of} a pollutant E _i of the cycle (WLTC, RDE-compliant) in [g / km] with the quasi-stationary result E _{i_quasi} into the ratio, that from the third additional module ZM ₃ (Emission forecasting model) is determined. The degree of fulfillment EG _dyn thus describes the proportion of transient operation of the internal combustion engine 2 at the emission of a pollutant E _i , For example, if EG _{stat_cycle is} small compared to EG _{dyn_cycle} , then EG _{dyn is} greater than one, which means that the actual measured emissions of a pollutant E _i of the cycle to a greater extent by the transient behavior of the internal combustion engine 2 be influenced, as by the stationary behavior. Thus EG _{dyn is} an important parameter as a basis for decision making for a transient or stationary optimization of the internal combustion engine 2 , For example, it would be at a dominant influence of the stationary behavior of the internal combustion engine 2 on the emission of a pollutant E _i not effective to optimize the transient behavior, as this is responsible only to a small degree for the compliance with the statutory limit E _{i_Grenz} .

In the next step of the procedure 1 will be in a main module HM ₄ , as in 1 can be seen, carried out a robustness and sensitivity analysis. This is the internal combustion engine 2 on a suitable test bench, eg a so-called engine-in-the-loop test bench 9 (EiL), in a model-based virtual environment U _v with a virtual vehicle F _v and a virtual driver D _v operated as schematically in 5 is shown. At the same time, the emission behavior of the internal combustion engine becomes 2 when used in a virtual vehicle F _v simulated, so it is not just the emission behavior of the internal combustion engine 2 itself examined, as in previous modules of the procedure 1 but the interaction between internal combustion engine 2 , virtual vehicle F _v and virtual driver D _v in a virtual environment U _v , By doing so, you gain information that is otherwise only available through the physical installation of the internal combustion engine 2 could win in one or more real vehicles and completion of many different test drives with different drivers (similar to the additional module ZM1), which would, however, be associated with considerable effort and consequently high development costs.

At EiL test bench 9, the internal combustion engine 2 in a known manner via a connecting shaft 10 with a loading machine 11 to drive or load the internal combustion engine 2 connected. The internal combustion engine 2 is via a test bench automation system 8th a certain desired value, for example a desired torque M _soll , predetermined by the internal combustion engine 2 is adjusted. Likewise, the loading machine 11 a certain desired value, for example a speed N _soll , predetermined. For this purpose, from the target size of a control unit 12 of the internal combustion engine 2 calculated a manipulated variable, for example, injection quantity, injection timing, setting one exhaust gas recirculation system, etc. The actual variables required for control, for example, in the form of an actual torque Mist and an actual speed N _is at the connecting shaft 10 measured and the test bench automation system 8th to hand over. With the loading machine 11 is usually the speed N of the internal combustion engine 2 regulated, wherein a target speed N _{soll is} specified. Via a suitable interface 13 become the test bench automation system 8th the set values from the simulation of the drive of the vehicle by means of the virtual driver D _v , the virtual environment U _v and the virtual vehicle F _v specified.

For virtual modeling of various virtual vehicles F _v , virtual environments U _v and virtual driver D _v become certain model parameters M _i determined and then varied. Examples of modeling different virtual vehicles F _v are the change of a drag coefficient (cw value) and a vehicle mass. The modeling of different virtual environments U _v can by variation of a slope or a slope β _{S of} a road, variation of a curve radius R _K , a curve angle δ _K or a _{stop time} t _stop (eg at a traffic light). Various virtual drivers D _v can be modeled, for example, by the maximum longitudinal and lateral accelerations a _l . a _q be varied of the virtual vehicle or different speeds v are given. Of course, this is just an example and other modeling parameters would be possible. Preferably, the information of the additional module ZM ₁ analyzed and it becomes certain boundary conditions for the definition and variation of the model parameters M _i established. This ensures that no model parameters M _i be given, which in the real application of the internal combustion engine 2 would not occur.

For every variation of a model parameter M _i becomes a test run PL _{i of} the internal combustion engine 2 carried out at the EiL test bench and it will be the emissions of pollutants E _i measured and recorded. Preferably, the variation of the model parameters takes place M _i and the execution of the test runs PL _i automated. As a result, you get for every pollutant E _i an emission in [g / km] plotted over the number of test runs PL _i (virtual test drives). In order to reduce the costly test bench time, the duration of the individual test runs PL _{i are} preferably chosen to be relatively short and do not correspond to the RDE-compliant duration of two hours. In the 6a-6c are exemplary the emissions of the pollutants nitrogen oxide / dioxide NO _x , hydrocarbon HC and carbon monoxide CO over the number A _PLi the executed test runs PL _{i shown} in a bar chart. In the concrete example, 358 test runs PL _{i were carried out} with a duration of between three and 12 minutes. For comparison, the legally prescribed limit E _{i_Grenz_EU6} each pollutant E _i the Euro 6 emission standard is shown as a dashed line. As it is in this specific case an internal combustion engine 2 the Euro- 4 Emission standard, the limit values are quite clearly exceeded. However, this plays to explain the sequence of the method according to the invention 1 not matter.

One can clearly see the large deviations in the emissions between the individual test runs PL _i or virtual test drives. It can be deduced that the design of the internal combustion engine 2 (here Euro- 4 ) is not yet sufficiently robust against real environmental influences. It can be seen how complex the requirements of the RDE legislation are, since no test drive is the same as the others, and the development of the internal combustion engine 2 to prepare for an unknown homologation test. It is therefore important to consider as many critical driving conditions as possible which have a negative impact on the emissions of the individual pollutants E _i impact.

eingeführt. Dabei wird aus den Ergebnissen aller durchgeführten Prüfläufe PL_i (virtueller Testfahrten) ein Mittelwert MW_Ei und eine Standardabweichung σ_Ei einer Emission eines Schadstoffs E_i in [g/km] gebildet und die Summe des Mittelwerts MW_Ei und der Standardabweichung σ_Ei wird mit dem gesetzlich festgelegten Grenzwert E_{i_Grenz} des Schadstoffs E_i in [g/km] ins Verhältnis gesetzt. Daraus lässt sich ableiten, dass eine hohe Varianz bzw. Standardabweichung σ_Ei nur bei einem geringen Mittelwertniveau zulässig ist und umgekehrt, dass bei einem hohen Mittelwertniveau nur eine geringe Standardabweichung σ_Ei zulässig ist. Allgemein ist es das Entwicklungsziel bei der Entwicklung des Verbrennungsmotors 2, einen Robustheits-Erfüllungsgrad EG_Robust < 1 zu erreichen. Beispielsweise kann dazu eine iterative Optimierung des Verbrennungsmotors 2 im Hauptmodul HM₂ (transiente Messung) erfolgen.In order to quantify the different influences of the real operation, a robustness degree of fulfillment is used as a parameter $e{G}_{RObust}=\frac{M{W}_{e_i}+{\sigma }_{e_i}}{e_{iGrenz}}\frac{\left[G\text{/}km\right]}{\left[G\text{/}km\right]}$

introduced. From the results of all test runs PL _i (virtual test drives), an average value MW _Ei and a standard deviation σ _{Ei of} an emission of a pollutant are calculated E _i is formed in [g / km] and the sum of the mean value MW _Ei and the standard deviation σ _Ei becomes the legally stipulated limit E _{i_limit of} the pollutant E _i in [g / km]. It can be deduced from this that a high variance or standard deviation σ _Ei is permissible only at a low mean value level and conversely that only a small standard deviation σ _{Ei is} permissible for a high mean value level. Generally, it is the development goal in the development of the internal combustion engine 2 To achieve robustness degree of fulfillment EC _Robust <first For example, an iterative optimization of the internal combustion engine 2 in the main module HM ₂ (transient measurement).

Finally, in the main module HM ₄ carried out a sensitivity analysis, whereby a structured identification of technical causes for certain emission maxima of a pollutant E _i is carried out. For this purpose, the temporal progressions of the measured quantities and each variation of a model parameter are used M _i an interaction diagram 14 creates the influence of the model parameter M _i visualized on the specific measurand. In 7 is an example of the influence of the variation of the virtual driver D _v on the CO ₂ Output. The virtual driver is as a whole of all the virtual driver D _v relevant changes of the model parameters M _i for each of which a test run PL _{i has been} carried out, as represented by the diagrams arranged one behind the other in FIG 7 is indicated. For modeling various virtual drivers D _v were in the specific example the longitudinal acceleration a _l , the lateral acceleration a _q , the vehicle speed v varies. To model the virtual vehicle F _v was additionally the Vehicle mass m varies. Every point in the interaction diagram 14 essentially represents the measured CO ₂ Value of a test run PL _i with a specific value of a model parameter M _i (Longitudinal acceleration a _l , Lateral acceleration a _q , Vehicle speed v, vehicle mass m). For this purpose, an average value from the respective time profile of the CO ₂ Values of a test run PL _i and into the interaction diagram 14 transfer. Preferably, for each virtual vehicle F _v , every virtual driver D _v and every virtual environment U _v such an interaction diagram 14 created desired variables. Based on this, the expert can derive statements about technical causes for emission maxima or maxima of another relevant measurand. In addition, parameter combinations of model parameters can be used M _i that were not measured, mapped in an interpolating regression model and also evaluated in the interaction diagram. This can be done for example by means of a suitable software.

The results of the fourth main module HM ₄ and implicitly, the results of all preceding main modules HM _i flow into the fifth and final main module HM ₅ of the procedure 1 a, like in 1 is shown. In the fifth main module HM ₅ In essence, the information obtained is the main module HM ₄ to a single representative test run 15 condensed, which contains all identified critical influences on the pollutant emissions. An internal combustion engine 2 Having this critical representative test run 15 in compliance with the legal emission limit values E _{i_Grenz} , it will then very likely be suitable to pass an RDE homologation in a vehicle with any unknown test cycle Z.

In addition, the generated representative check run is 15 preferably RDE compliant, so that, for example, in the known post-processing method of the RDE test, EMROAD no weighting factor F _G <1 is used. This would be the case, for example, if sections of the representative test run were involved 15 too high or too low CO ₂ Output would have. For checking, for example, the CO ₂ -Emission of the four sections of the known WLTC cycle used as a basis for comparison and when exceeding or falling below a certain CO ₂ Output, a weighting factor is applied. The main purpose of this is to ensure that, in practice, no too demanding (or too less demanding) test runs are evaluated. However, the legal boundary conditions of the RDE test are known to the person skilled in the art and will not be discussed in detail here.

For example, you could also do without the representative test run 15 RDE compliant, eg if you have one, with regard to a certain pollutant E _i particularly critical test run in the context of the development of the internal combustion engine 2 want to use.

By means of the additional module ZM ₃ (Emissions forecast) will give a statement about the CO ₂ Output of the generated representative check run 15 met. In this way, the comparison described above can be made with the legal requirements derived from the WTLC cycle. In addition, the CO ₂ Emissions are not a legally limited pollutant E _i , however, the correlates CO ₂ Emissions with fuel consumption and is therefore also an important developmental variable that is trying to minimize. 8th shows the detailed procedure for determining the representative check run 15 ,

In block A, a so-called maneuver database with critical virtual driving maneuvers FM _{x is} created. Critical driving maneuvers FM _x can, for example, certain sequential driving conditions such as acceleration, a stop, a constant driving in certain boundary conditions of the virtual vehicle F _v and the internal combustion engine 2 his. Different boundary conditions can eg speed and load of the internal combustion engine 2 , Slope of the virtual route, operating state of the internal combustion engine 2 (cold, warm), etc. The maneuver database is at least out of the test runs P _Li of the fourth main module HM ₄ fed, but may also have other sources. For example, real recorded driving maneuvers FM _{x of} the reference trips of the additional module ZM ₁ be used. Critical driving maneuvers FM _x must, however, not actually recorded real driving maneuvers FM _x (from test runs P _Li or homing R _i ), but can also be taken from literature 16 or from experience 17 about a particular concept of an internal combustion engine 2 come. Critical driving maneuvers FM _x can also take off from the fourth main module HM ₄ carried out sensitivity analysis. In any case, the maneuver database should be a concept-specific collection of critical driving conditions. Preferably, the data of all information sources in a suitable common data management module 18 saved.

In block B, all or selected critical driving maneuvers FM _{x of} the maneuver database (block A) are assembled into a long test run PL _long . This test run PL _long now essentially includes all possible critical driving conditions of the internal combustion engine 2 in the virtual environment U _v , For practical use in many iterations, eg optimization on a test stand, this long test run PL is _long but not expedient, on the one hand because of its length, which would lead to long test bench times, and secondly because many critical driving maneuvers FM _x can be present, which differ strong on emissions of pollutants E _i impact. If the available test time is limited, it may possibly be the goal to consolidate this long test run PL _{long short} on the shortest possible critical test run PL, with the respect. Pollutant emissions are to be retained critical driving maneuvers FM _x and the uncritical or less critical driving maneuver FM _x should be removed.

This shortened test run PL _short is subjected to a simulation in block C, in which it is checked whether the generated shortened test run PL _short , which is composed of individual driving maneuvers FM _x , generally error-free feasible. This may be a suitable simulation environment 19 such as InMotion from IPG Car Maker. In addition, the emissions forecasting model can be derived from the add-on module ZM ₃ used to make a statement about emissions of pollutants E _i and also a statement about the CO ₂ -Emission of the internal combustion engine 2 when passing through the shortened test cycle PL _to meet _briefly . Depending on the development goal, however, a shortened test run PL could be made _short that does not meet the RDE requirements, for example, the city shares may be higher if the engine 2 specifically with regard to the operating behavior in urban operation is to be optimized. On the basis of the simulation results, the shortened test run PL can be adapted _briefly (replacement, removal, addition of critical driving maneuvers FM _x ) or reused as it is, if the results meet the desired requirements. Since the individual critical driving maneuvers FM _{x of} the shortened test run PL _briefly different operating states of the internal combustion engine 2 may have, for example, in terms of the operating temperature (coolant temperature T _KM or oil temperature T _oil ), it should preferably be ensured that no sudden changes in the operating conditions of a driving maneuver FM _x to the following driving maneuver FM _x during the passage of the shortened test run PL is _short , as this is not possible in real operation and the shortened test run Accordingly, PL is not _short of an internal combustion engine 2 could be completed on the test bench. Critical driving maneuvers FM _x at cold start or not yet warm combustion engine 2 should therefore be arranged _shortly at the beginning of the shortened test run PL. However, it is also possible to install conditioning track sections on which an adaptation of certain state variables (eg temperatures) takes place.

In block D is the internal combustion engine 2 subjected to the shortened test run PL _briefly on EiL test stand, the emissions of pollutants E _i measured and recorded. In the next block E, an evaluation of the recorded in block D measurement data and, if necessary, removal of non-critical sections of the shortened test run PL _just done to finally create the most efficient shortened test run PL _short of a so-called Repräsentativprüflauf 15 as the end result of the process according to the invention 1 in block F is available. For the representative test run 15 to be able to use on test benches with a setpoint-based control are preferably appropriate setpoints 20 from the representative test run 15 determined as symbolized by the optional block G.

With the representative test run 15 This creates a simple tool with which an efficient development of an internal combustion engine 2 a specific approach to comply with the new RDE legislation. The development of the internal combustion engine 2 This can be done essentially in a conventional manner based on a fixed cycle, as previously with the legally prescribed cycles NEDC, FTP75, etc. The advantage over the conventional cycles, however, is that the internal combustion engine 2 adherence to RDE limits when conducting the representative test run 15 with very high probability, will complete every unknown RDE homology cycle positively.

The description of the procedure 1 Naturally, the described main modules HMi and additional modules ZMi could also be used alone, ie independently of the entire method, or in suitable combinations from the sequence of the method 1 removed, used as a separate method.

For the purposes of the invention, the motor concept is also to be understood as meaning conventional drive trains and electrified drive trains (so-called hybrid drive trains), e.g. Combinations of an internal combustion engine with automatic transmissions, manual transmissions and electric motors.

Claims (6)

Method for developing an internal combustion engine (2) with a given engine concept (K) for use in a vehicle (F), wherein the internal combustion engine (2) comprises at least one control unit (12) with variable control parameters for controlling the internal combustion engine (2), characterized in that reference operating data for the internal combustion engine (2) are determined that from a time-based and / or path-based frequency distribution (4, 5) of operating points (P _Bi ) and a frequency distribution (6) of operating point changes (P _BAi ) of the reference internal combustion engine (2 _R ) is determined that stationary measurements of the internal combustion engine (2) in the determined operating points (P _Bi ) of the reference internal combustion engine (2 _R ) on a Test stand are performed, wherein at least emissions of a pollutant (E _i ) are measured that from the frequency distributions (4, 5) of the operating points (P _Bi ) and the measured emissions of the pollutant (E _i ) in the respective operating points (P _Bi ) a stationary _{degree of} fulfillment (EG _stat ) of a legally defined emission limit value (E _{i_limit} ) of the pollutant (E _i ) is determined that t ransiente measurements of the determined operating point changes (P _BAi ) are performed with the internal combustion engine (2) at the test bench, wherein at least the emissions of the pollutant (E _i ) are measured and a plurality of, for the emissions of the pollutant (E _i ) critical operating point changes (P _BAi ) is determined that at least one predetermined test cycle (Z) with the internal combustion engine (2) is performed at the test stand, wherein the at least one emission (E _{i_Zyklus} ) of the pollutant (E _i ) of the test cycle (Z) is measured and from a dynamic _{Cycle satisfaction level} (EG _{dyn_cycle} ) of the legally defined emission limit value (E _{i_Grenz} ) of the pollutant (E _i ) is determined that from the given test cycle (Z) and the steady state measurements a quasi-stationary test cycle (Z _quasi ) of the given test cycle (Z) is simulated whereby quasi-stationary emissions (E _{i_quasi} ) of the pollutant (E _i ) of the quasi-stationary test cycle (Z _quasi ) and from this a stationary cycle _{satisfaction level} (EG _{stat_zyklus} ) of the legally stipulated emission limit value (E _{i_limit} ) of the pollutant (E _i ) is determined that _comprises the emissions (E _{i_cycle} ) of the pollutant (E _i ) of the given test cycle (Z) and the quasi-stationary emissions (E _{i_quasi} ) of the pollutant (E _i ) of the quasi-stationary test cycle (Z _quasi ) a dynamic degree of fulfillment (EG _dyn ) is determined that a plurality of test runs (P _Li ) of the internal combustion engine (2) with a plurality of different virtual drivers (D _v ) and / or different virtual vehicles (F _v ) and / or different virtual environments (U _v ) is performed on a test bench (9), wherein at least the emissions of the pollutant (E _i ) in the test runs (P _Li ) are measured and therefrom a robustness performance level (EC _Robust) of the statutory emission limit value (e _{i_Grenz)} of the pollutant (e _i) it is determined that the one of the plurality of tests (P _Li ) for which emissions of the pollutant (E _i ) critical, representative test run (15) is determined and that the representative test run (15) for optimizing the at least one emission of the pollutant (E _i ) on the Test stand is used. Method according to Claim 1 , characterized in that the reference operating data are determined by at least one reference vehicle (F _R ) with a reference motor (2 _R ) with a, the internal combustion engine (2) comparable engine concept (K) is determined that at least one reference distance (S _R ) is determined that at least one reference driver (D _R ) is determined that with the at least one reference vehicle (F _R ) and the at least one reference driver (D _R ) a number of real test drives on the at least one reference distance (S _R ) is performed and from this the relevant reference operating data of the reference internal combustion engine (2 _R ) and the reference vehicle (F _R ) are detected. Method according to Claim 2 , characterized in that at least four different reference paths (S _R ) and at least three different reference drivers (D _R ) are provided. Method according to Claim 1 or 2 , characterized in that the stationary _{degree of} fulfillment (EG _stat ) and / or the dynamic _{degree of} fulfillment of the cycle (EG _{dyn_cycle} ) and / or the stationary _{degree of completion} (EG _{stat_clause} ) and / or the dynamic _{degree of} fulfillment (EG _dyn ) and / or the degree of robustness fulfillment ( EC _Robust ) can be optimized by iterative changes of the control parameters of the control unit (12) of the internal combustion engine (2) are performed and / or components of the internal combustion engine (2) to be changed. Method according to one of Claims 1 to 4 , characterized in that as pollutants (E _i ) hydrocarbon (HC) and / or nitrogen oxide / dioxide (NO _x ) and / or a particle number (PM) and / or carbon monoxide (CO) are provided. Method according to one of Claims 1 to 5 , characterized in that the robustness degree of fulfillment (EC _Robust ) is determined by a sum of an average value (MW _Ei ) and a standard deviation (σ _{Ei of} the measured emission of the pollutant (E _i ) from the plurality of test runs (P _Li ) is formed and the sum with the legally defined emission limit (E _{i_Grenz} ) of the pollutant (E _i ) is compared.

Images