Method and device for predicting exhaust emissions of internal combustion engine

Abstract

The invention relates to a device and a method for predicting exhaust emissions of an internal combustion engine, comprising: an exhaust aftertreatment device; a lambda probe arranged downstream of the combustion chamber of the internal combustion engine and upstream of the exhaust gas aftertreatment device; a sensor for measuring nitrogen oxide concentration, ammonia concentration and / or oxygen concentration in the exhaust gas stream, arranged downstream of the exhaust gas aftertreatment device; and an engine control unit in which an emission model having a combustion raw emission model by which raw emissions can be calculated from operating data of the internal combustion engine and an exhaust aftertreatment model by which emissions downstream of the exhaust aftertreatment device can be calculated from raw emission data is integrated. According to the invention, a calculation unit is integrated in the engine control unit, in which a fuel mass can be calculated as an input signal for a combustion raw emission model and / or an exhaust gas aftertreatment model as a function of the measured values of the lambda probe and the measured values of the sensor.

Description

Technical Field

[0001] The present invention relates to a method for predicting exhaust emissions of an internal combustion engine and a device for predicting exhaust emissions of an internal combustion engine, comprising an exhaust gas after-treatment device, the device comprising: a lambda probe, the lambda probe being arranged downstream of a combustion chamber of the internal combustion engine and upstream of the exhaust gas after-treatment device; a sensor for measuring nitrogen oxide concentration, ammonia concentration and / or oxygen concentration in the exhaust gas flow, the sensor being arranged downstream of the exhaust gas after-treatment device; and an engine control unit, in which an emission model is integrated, the emission model comprising a combustion raw emission model and an exhaust gas after-treatment model, the combustion raw emission model being capable of calculating raw emissions based on operating data of the internal combustion engine, and the exhaust gas after-treatment model being capable of calculating emissions downstream of the exhaust gas after-treatment device based on the raw emission data. Background Art

[0002] In order to minimize pollutants as required by law, the exhaust gas after-treatment equipment of modern gasoline internal combustion engines usually has two three-way catalytic converters and a particle filter as waste after-treatment equipment. In addition to the particle filter, diesel engines usually also have an oxidation catalyst and a catalyst for selective catalytic reaction (SCR catalyst) as well as an ammonia slip catalyst. Since it is not possible to continuously measure the individual components of the emissions generated at the outlet of the exhaust gas equipment and released into the environment in the vehicle, it is necessary to calculate the emissions generated based on the available vehicle data in order to comply with legal regulations. For this purpose, models have been developed for simulating the combustion process and the processes in the exhaust gas after-treatment equipment, and these models are used to calculate the components and amounts of exhaust gas released into the environment. Such a combustion raw emission model for calculating emissions generated during combustion is described, for example, in DE 10 2015 207252 A1.

[0003] The prediction quality of the corresponding exhaust gas aftertreatment model depends in particular on the correct measurement or calculation of the fuel and air masses injected into the cylinders per working cycle, as this is crucial for the generation of pollutants during combustion. These can be read out either via a combustion raw emission model (in which the combustion occurring in the gasoline engine is simulated and the data for determining the raw emissions are calculated from this), or directly from the engine control unit, or calculated using measured values from a first lambda probe, which is arranged upstream of the first three-way catalytic converter and downstream of the gasoline engine's combustion chamber.

[0004] Since the measured value of the lambda probe measures the actually existing combustion air ratio, its use generally yields more accurate results when calculating the actually injected fuel mass and thus the emissions calculated in the exhaust gas aftertreatment model.

[0005] However, the problem in this case is that the lambda probe's measured values are not available throughout the entire operating period of the internal combustion engine. They are only used after the lambda probe's operational readiness has been confirmed. However, it has been shown that the lambda probe's measured values may have certain measurement errors due to manufacturing tolerances, installation position, and aging. If these measured values were now used to calculate the fuel mass and, based on this fuel mass, the current emissions, errors could result in the predictions, the magnitude of which would no longer be sufficient to ensure compliance with legal regulations. Summary of the Invention

[0006] The object is therefore to provide a device and a method for time-resolved prediction of exhaust emissions from an internal combustion engine, with which at least uncertainties regarding the correct determination of the exhaust emissions are revealed and made available for further processing. Furthermore, it is desirable not only to identify possible errors but also to achieve the most accurate prediction of the exhaust emissions possible throughout the entire operating period, in particular to avoid calculation errors in the exhaust emissions due to errors in the calculation of the injected fuel mass as an input value for the exhaust emission model.

[0007] The apparatus for predicting exhaust emissions from an internal combustion engine according to the present invention includes an exhaust gas aftertreatment device, which can be configured in various ways depending on the engine type. For example, in gasoline engines, such an exhaust gas aftertreatment device typically consists of a first three-way catalytic converter and a downstream second three-way catalytic converter, with a particle filter positioned between them. In diesel engines, such an exhaust gas aftertreatment device typically consists of at least one SCR catalyst, an oxidation catalyst, and / or an ammonia slip catalyst, and a diesel particle filter. Furthermore, a lambda probe is positioned downstream of the combustion chamber of the internal combustion engine and upstream of the exhaust gas aftertreatment device, which measures the oxygen concentration and, thereby, determines the existing combustion air ratio. Additionally, the apparatus according to the present invention includes sensors for measuring the nitrogen oxide concentration, ammonia concentration, and / or oxygen concentration in the exhaust gas flow, which are positioned downstream of the exhaust gas aftertreatment device. These sensors are typically suitable for determining nitrogen oxide and ammonia concentrations based on measurement signals, as well as for measuring the oxygen concentration in the exhaust gas, due to structurally related cross-sensitivities. These sensors typically include an additional controller that allows for accounting for existing cross-sensitivities, thereby obtaining highly accurate measurement values. Correspondingly, the measured values of all measured exhaust gas components may be taken into account, or only the measured value of one of these exhaust gas components may be taken into account.

[0008] The device according to the present invention also includes an engine control unit, which also controls the internal combustion engine and its actuators. An emissions model is integrated into the engine control unit. This model includes a raw combustion emissions model and an exhaust gas aftertreatment model. The raw combustion emissions model allows calculation of raw emissions based on the operating data of the internal combustion engine. Raw emissions are defined as emissions generated in the combustion chambers of the cylinders and entering the exhaust gas tract of the internal combustion engine without further aftertreatment. The raw combustion emissions model thus serves as a virtual sensor. The raw combustion emissions model models the following physical and chemical processes in combustion based on the control data of the internal combustion engine and the existing geometric conditions, thereby calculating the raw emissions expected based on the existing parameters. The exhaust gas aftertreatment model allows calculation of emissions downstream of the exhaust gas aftertreatment system based on the raw emission data. This is achieved by modeling the physical and chemical processes in the individual components of the exhaust gas aftertreatment system, thereby enabling calculation of emissions downstream of each component of the exhaust gas aftertreatment system. Correspondingly, the exhaust gas after-treatment model is connected to the original combustion emission model to exchange data, and the output data of the original combustion emission model is used as the input data of the exhaust gas after-treatment model.

[0009] Additionally, a calculation unit is integrated into the engine control unit, which can also be part of the emissions model. This calculation unit calculates the fuel mass based on the measured values of the lambda probe and at least one measured value from the sensor. This fuel mass serves as an input signal for the raw combustion emissions model and / or the exhaust aftertreatment model. Accordingly, the exhaust aftertreatment model is not only supplied with virtual values from the raw combustion emissions model, but also uses actual measurements from the internal combustion engine as starting values for calculations. Based on these measurements, the injected fuel mass for each operating cycle and cylinder is calculated. This fuel mass is a decisive factor in the emissions calculation. Consequently, calculation errors in the emissions model caused by errors in determining the injected fuel mass are prevented. This is particularly important because emissions at the outlet of the exhaust aftertreatment system are particularly critical for legislators. Accordingly, this device allows for highly reliable, time-resolved determination of raw emissions, enabling continuous emissions monitoring regardless of prevailing driving conditions. Therefore, by using the actual measured values of the lambda probe and the sensor at the outlet, an input gas composition can be determined that closely resembles the actual gas composition in the exhaust gas before the exhaust aftertreatment system. This ensures that exhaust emissions can be predicted not only with a very high level of quality but also in a very robust manner, as all changes in the actual system are immediately fed back into the emissions model and taken into account there.

[0010] In a method for predicting exhaust gas emissions from an internal combustion engine according to the present invention, the internal combustion engine includes an exhaust gas aftertreatment device having at least two catalytic converters and a particulate filter. A lambda probe is arranged upstream of the exhaust gas aftertreatment device and downstream of a combustion chamber of the internal combustion engine. A sensor is arranged downstream of the exhaust gas aftertreatment device for measuring the nitrogen oxide concentration, ammonia concentration, and / or oxygen concentration in the exhaust gas flow. After the lambda probe and the sensor have reached operational readiness, the lambda probe is used to measure the oxygen concentration in the exhaust gas flow upstream of the exhaust gas aftertreatment device to determine the injected fuel mass. Based on the measured values of the lambda probe, the injected fuel mass is then calculated and used as an input value for an emissions model, which calculates exhaust gas emissions downstream of the exhaust gas aftertreatment device. These calculated exhaust gas emissions specifically include the current values for the nitrogen oxide concentration, ammonia concentration, and oxygen concentration and the resulting combustion air ratio. At least one of these values calculated in the emissions model is then compared with a measured value of the sensor downstream of the exhaust gas aftertreatment device. Next, if at least one value calculated by the emissions model deviates from the corresponding value measured by the sensor, it is inferred that there is an error in the emissions model or at the lambda probe or sensor. In this case, the error should not only be understood as an error or defect in the component, but also as a need for adaptive adjustment of the emissions model. Therefore, the values calculated in the emissions model can be continuously checked by actual values, which makes it possible to adjust the emissions model when needed. In this way, the emissions model can be continuously adjusted to adapt to actual conditions during operation, so that even in the stage where the sensor or lambda probe is not yet in a ready-to-operate state, the emissions model can also determine more accurate emissions values. In a simple embodiment, this comparison can also be used only to output an error by the engine control unit, so that the lambda probe or sensor can be checked, for example, during maintenance. In addition, the calibration and parameterization of the model can be checked, or it can be overwritten or reset to an initial state during the data status update process.

[0011] Correspondingly, if there is an error or deviation which exceeds a defined threshold, the calculated value of the emission model is preferably corrected by taking into account at least one measured value of the sensor. In this way, the emission model can be improved.

[0012] This improvement advantageously occurs by calculating the oxygen, ammonia, and / or nitrogen oxide concentrations downstream of the exhaust aftertreatment device based on the lambda probe's measured values using an emissions model. Subsequently, at least the oxygen, ammonia, and / or nitrogen oxide concentrations measured by the sensor are compared with the corresponding calculated values of the oxygen, ammonia, and / or nitrogen oxide concentrations from the emissions model. If the sensor's measured values deviate from the corresponding calculated values from the emissions model, the lambda probe's measured values are corrected based on the sensor's measured values, and the injected fuel mass is calculated based on the corrected measured values. Subsequently, the exhaust emissions downstream of the exhaust aftertreatment device are calculated using the corrected fuel mass using the emissions model. As a result, the fuel mass used by the emissions model is not the fuel mass derived from the internal combustion engine's control data (i.e., injection valve opening time, injection time, and fuel pressure), but rather the actual fuel mass corresponding to measurements at the combustion chamber outlet. This allows for significantly more accurate determination of the emissions calculated using the emissions model. However, the fuel mass derived from the lambda probe's measured values is not merely referenced; it is also corrected again based on the sensor's measured values. The oxygen content and the combustion air ratio downstream of the combustion chamber of the internal combustion engine and upstream of the exhaust gas aftertreatment device are thus calculated based on the measured values of the lambda probe and the sensor, from which the injected fuel mass used as input for the emissions model is derived. The measured values of the lambda probe are thus corrected again based on the measured values of the sensor, which further improves the quality of the emissions predictions calculated in the emissions model because measurement deviations at the lambda probe that may arise due to aging, inaccurate installation position, or existing manufacturing tolerances are minimized.

[0013] Furthermore, it is advantageous to read the injected fuel mass, air mass, and residual gas content from the engine control data during phases when the lambda probe or sensor is not operationally ready or transmits an unreliable signal. This is necessary during the engine's warm-up phase or when the exhaust system is in an unstable state, such as when the lambda probe and sensor have not yet reached their required operating temperature or have not reached a quasi-stable state. This ensures that emissions can be calculated with high accuracy even during these operating phases. Accordingly, emissions can be determined throughout the entire operation of the internal combustion engine.

[0014] Preferably, the combustion-based emissions model, which constitutes the first part of the emissions model, uses engine geometry data, operating data, engine control data, fuel injection data, air mass flow, and fuel composition of the internal combustion engine as input data. The exhaust gas composition and exhaust gas mass flow downstream of the combustion chamber of the internal combustion engine are calculated with reference to the modeled combustion data. Because gasoline engine operating parameters such as ignition timing, gas exchange element position, injection pressure, and injection timing are available in the engine control unit, the combustion-based emissions model can be used to model the physical and chemical combustion process with good approximation and determine the exhaust gas composition and mass.

[0015] In a preferred embodiment, the output data of the raw combustion emissions model is used as input data for an exhaust gas aftertreatment model, which constitutes the second part of the emissions model. In phases where the lambda probe and sensor are not operationally ready, the fuel mass determined from the engine control data is used as the input value for the raw combustion emissions model. In phases where the lambda probe and sensor are operationally ready, the lambda probe's measured values, corrected using the sensor's measured values, are used as the input values for the exhaust gas aftertreatment model and / or the raw combustion emissions model. This allows, on the one hand, exhaust gas emissions to be determined over the entire operating period, and, on the other hand, allows for highly accurate predictions when the lambda probe is operationally ready by combining the actual measured values of the lambda probe with the actual fuel mass derived therefrom, since the emissions calculated by the exhaust gas aftertreatment model depend particularly on the actual fuel mass used. This allows the actual measured values to be coupled to the emissions model. Consequently, the emissions model can remain unchanged, and only the existing fuel mass can be adjusted.

[0016] It is particularly advantageous to integrate the emissions model, consisting of the exhaust gas aftertreatment model and the combustion-based emissions model, into the engine control unit. This eliminates the need for additional electronic components and allows all calculation steps for determining exhaust emissions to be performed in the engine control unit.

[0017] In the exhaust gas aftertreatment model, the expected concentrations of emission components present in the exhaust gas flow are preferably read or calculated based on the environmental data using a physical model or algorithm. These concentrations are determined experimentally in a known manner. These exhaust gas aftertreatment models are stored or calibrated specifically for each engine type and enable very precise predictions.

[0018] In the combustion-based emissions model, basic data about the combustion process, including exhaust gas composition, exhaust gas mass flow, combustion air ratio, and fuel mass, are preferably calculated as output data of the exhaust gas downstream of the combustion chamber of the internal combustion engine. This data can be used for further processing in the exhaust aftertreatment model and for calibrating existing sensors.

[0019] In the exhaust aftertreatment model, the exhaust gas composition, exhaust gas mass flow, exhaust gas temperature, and oxygen, ammonia, and nitrogen oxide concentrations downstream of each exhaust aftertreatment device are preferably calculated based on the output data of the combustion raw emission model and, when the lambda probe and sensors are in operational readiness, additionally based on calibrated measured values from the lambda probe. These calibrated measured values are generated by comparing the oxygen, ammonia, and / or nitrogen oxide concentrations measured by the sensors with the corresponding oxygen, ammonia, and / or nitrogen oxide concentrations calculated in the emissions model. This allows emissions to be reliably determined at every location in the exhaust aftertreatment system. This can be used, for example, for on-board diagnostics to identify faulty sensors, actuators, or catalytic converters or filters that are no longer functioning properly. However, by continuously determining exhaust gas composition and exhaust gas quality throughout operation, legal regulations are also met.

[0020] Furthermore, it has been found to be advantageous to use the nitrogen oxide concentration measured by the sensor to correct the lambda probe's measured value when determining an oxygen excess, and to use the ammonia concentration measured by the sensor to correct the lambda probe's measured value when determining an oxygen deficiency, as is typical in gasoline engines. The use of these measured values has been shown to improve the emissions model, which in this case achieves particularly realistic results.

[0021] Preferably, at least one of the oxygen, ammonia, and / or nitrogen oxide concentrations downstream of the exhaust aftertreatment device calculated in the exhaust aftertreatment model is used as a process variable, and the corresponding sensor measurement value is used as a setpoint for the PID controller. Based on the sensor control value, an error deviation in the lambda probe's measurement value is inferred, and the measured value is corrected based on the determined sensor control value. This creates a closed control loop that adjusts the calculated oxygen, ammonia, and / or nitrogen oxide concentrations to the actual oxygen, ammonia, and / or nitrogen oxide concentrations, and thus allows for a very precise determination of the actual injected fuel mass based on the sensor measurement value. The model compensates for the effects of measurement errors on the lambda probe due to aging and drift.

[0022] In another embodiment, the measured value of the lambda probe is corrected by a maximum of 5% for calculating the injected fuel mass. This avoids excessive deviations that can be caused, for example, by malfunctions of the lambda probe.

[0023] Thus, a device and a method for predicting exhaust gas emissions of an internal combustion engine are provided, which can meet the legal requirements for determining exhaust gas emissions in all operating states. The values determined by the corresponding model are highly accurate because they are corrected during the calculation by incorporating actual sensor data. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In the attached Figure 1 An embodiment of the apparatus for predicting exhaust emissions of a gasoline internal combustion engine according to the present invention is schematically shown in FIG. 2 and will be described together with a related method according to the present invention with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] The device according to the present invention consists of an internal combustion engine 10, to which air is supplied via an air supply line 12, in which a throttle valve 14 is arranged to control the air quantity. This air enters the combustion chamber 15 of the internal combustion engine 10, into which gasoline is injected from a rail 18 via an injection valve 16. The gasoline is compressed together with the air in the combustion chamber 15 and is expelled again from the combustion chamber 15 by the movement of the piston. From there, the combustion products, which constitute the raw exhaust of the internal combustion engine 10, enter an exhaust gas duct 20. An exhaust gas aftertreatment device 22 is arranged in this exhaust gas duct 20, which catalytically converts and filters pollutants from the raw exhaust to reduce harmful exhaust emissions at the end of the exhaust gas aftertreatment device 22 in accordance with legal regulations.

[0026] In the exhaust gas aftertreatment device 22, a lambda probe 24 is arranged downstream of the combustion chamber 15 of the internal combustion engine 10. This lambda probe is used to determine the combustion air ratio by directly measuring the residual oxygen content. The combustion air ratio can also be used to calculate the injected fuel mass in a known manner. This lambda probe 24 is located upstream of the first three-way catalytic converter 26 arranged closest to the combustion chamber 15 of the internal combustion engine 10. In this first three-way catalytic converter, in addition to further conversion, the carbon monoxide, nitrogen oxides, and unburned hydrocarbons present in the raw exhaust gas are primarily converted into carbon dioxide, nitrogen, and water. Downstream of this first three-way catalytic converter 26, another lambda probe 28 is arranged in the exhaust gas aftertreatment device 22. This oxygen sensor measures the remaining residual oxygen content, thereby enabling the efficiency of the three-way catalytic converter 26 to be monitored. From here, the exhaust gas continues to flow through a particle filter 30, where it filters out soot particles from the combustion process. The particle filter is located upstream of a second three-way catalytic converter 32, where unconverted carbon monoxide, nitrogen oxides, and hydrocarbons are converted into carbon dioxide, nitrogen, and water. The exhaust gas then enters a sensor 34, which in this embodiment is implemented as a multi-gas sensor and measures the oxygen, ammonia, and nitrogen oxide concentrations present therein. However, it is also possible to arrange only one sensor for measuring one of the aforementioned concentrations, or to arrange three separate sensors, each for measuring the concentration of one of the aforementioned exhaust gas components. The exhaust gas then leaves the exhaust system and exhaust gas aftertreatment device 22 via an outlet 35.

[0027] The internal combustion engine 10 is controlled by an engine control unit 36, which controls the position of the throttle valve 14 and the injection pressure and injection time of the injection valve 16 and other actuators. An interface 38 of the engine control unit 36 is connected to the corresponding actuators and sensors via electrical connections and also has a connection 40 to the lambda probes 24, 28 and the sensor 34.

[0028] According to the invention, an emissions model 42 is additionally integrated into the engine control unit 36, by which the resulting emissions are calculated based on the available data using empirical and physical models. This emissions model 42 is calibrated or parameterized by tests on the corresponding engines.

[0029] In this case, emissions model 42 consists of a raw combustion emissions model 44, which calculates raw emissions based on the operating data of internal combustion engine 10, and an exhaust gas aftertreatment model 46, which calculates emissions at outlet 35 of exhaust gas aftertreatment device 22 based on the output data of raw combustion emissions model 44. For this purpose, raw combustion emissions model 44 is supplied with the operating data of internal combustion engine 10. These operating data, used as input data, include engine geometry data, operating data, and engine control data of the internal combustion engine, namely, ignition and injection times as a function of crankshaft angle, operating pressure and temperature, load conditions, crankshaft speed, fuel injection data, air mass flow as a function of throttle valve position and valve control time, data on exhaust gas recirculation (if applicable), compression ratio, and fuel composition. To this end, the injected fuel mass is first calculated based on the fuel pressure and valve opening time of injection valve 16. This is then used, along with the determined combustion data, the air mass, and the residual gas content, in the raw combustion emissions model to calculate the exhaust gas composition and exhaust gas mass flow 44 downstream of combustion chamber 15 of internal combustion engine 10.

[0030] These data are further used as input data for an exhaust gas aftertreatment model 46. In this exhaust gas aftertreatment model, the conversion processes of the two three-way catalytic converters 26, 32 and the particle filter 30 are simulated, so that the exhaust gas composition and the exhaust gas mass flow at the outlet 35 can be calculated based on the raw exhaust gas composition and the raw exhaust gas mass flow, wherein the exhaust gas composition includes, in particular, the values of the nitrogen oxide concentration, the ammonia concentration, and the oxygen concentration.

[0031] Once the lambda probe 24 and sensor 34 are operationally ready and the fine-tuning control of the first three-way catalytic converter 26 and / or the second three-way catalytic converter 32 is activated in the engine control unit 36, the injected fuel mass is corrected for emission calculation. To this end, in a first step, the measured value of the first lambda probe 24 is used to calculate the injected fuel mass, which is indirectly input into the exhaust gas aftertreatment model via the modified raw exhaust gas composition of the combustion raw emission model.

[0032] In a second step, the values for the nitrogen oxide, ammonia, and oxygen concentrations downstream of the exhaust gas aftertreatment device 22 calculated by the exhaust gas aftertreatment model 46 are compared with the corresponding measured values of the sensor 34. If a discrepancy is found, it is inferred that the measured values of the lambda probe 24 deviate from the actually existing values. The measured values of the lambda probe 24 are then corrected in three control loops, particularly by means of a PID controller, so that the modeled values for the nitrogen oxide, ammonia, and oxygen concentrations at the location of the sensor 34 are adjusted toward the measured values of the sensor 34 during the calculation by the exhaust gas aftertreatment model 46. Since three separate control loops are used in this case, which determine different corrected measured values at the lambda probe, a final corrected measured value is then calculated from the corrected measured values of the lambda probe by averaging or weighting the individually obtained corrected measured values. This final corrected measured value is then used as an input value for the emissions model 42. The error correction of the measured values of the lambda probe 24 performed here is limited to, for example, 5%. This correction is carried out in a calculation unit 48 which is integrated in the engine control unit 36 and in which the measured values of the lambda probe 24 are corrected with the aid of the measured values of the sensor 34 and the fuel mass injected per working cycle is calculated with the aid of these corrected values. This fuel mass is then used as a basis for calculating the exhaust gas composition downstream of each exhaust gas aftertreatment device, the exhaust gas mass flow downstream of each exhaust gas aftertreatment device 26 , 30 , 32 , the exhaust gas temperature, the nitrogen oxide concentration, the ammonia concentration, and the oxygen concentration.

[0033] This method thus ensures that once continuous lambda control and catalyst fine-tuning control begin, the emissions model operates with very accurate input data, because the fuel mass used in the model corresponds largely to reality, thereby compensating for errors in the fuel mass calculation in the model. Furthermore, any deviations of the lambda probe that may occur due to aging, manufacturing tolerances, or installation inaccuracies are compensated.

[0034] It should be understood that the design of the internal combustion engine can vary from that shown, as can the exhaust gas aftertreatment system. Thus, the method is applicable to both gasoline and diesel engines, whose exhaust gas aftertreatment systems typically consist of at least one SCR catalyst, an oxidation catalyst, an ammonia slip catalyst, and a diesel particulate filter. Of course, additional components may also be present, such as exhaust gas recirculation or other catalysts and filters, secondary air systems, and other components, which necessitate the use of different emissions models or the calibration or parameterization of the emissions model for each engine. It should also be noted that only one of the determined concentrations can be used to calibrate the first lambda probe and the exhaust gas aftertreatment model, particularly by adapting the parameters to account for aging. In particular, it has been shown that when using diesel engines, and therefore when oxygen is often present, the measured nitrogen oxide concentration can be advantageously used to calibrate the lambda probe's measured value. In contrast, in gasoline engines, as described herein, which typically operate under oxygen deficiency conditions, the measured ammonia concentration is particularly suitable for calibrating the lambda probe's measured value, thereby allowing only one sensor's measured value to be used for calibration, if necessary. Multiple individual sensors can also be used instead of a multi-gas sensor.

Claims

1. A method for predicting exhaust gas emissions of an internal combustion engine (10), the internal combustion engine having an exhaust gas aftertreatment device (22), the exhaust gas aftertreatment device having at least two catalytic converters (26, 32) and a particle filter (30), wherein: A lambda probe (24) is arranged upstream of the exhaust gas aftertreatment device (22) and downstream of the combustion chamber (15) of the internal combustion engine (10), and a sensor (34) is arranged downstream of the exhaust gas aftertreatment device (22), the sensor being used to measure the nitrogen oxide concentration, the ammonia concentration and / or the oxygen concentration in the exhaust gas flow, wherein after the lambda probe (24) and the sensor (34) reach an operational readiness state, the following operations are performed: - measuring the oxygen concentration in the exhaust gas flow by means of the lambda probe (24) in order to determine the injected fuel mass, - calculating the injected fuel mass from the measured values of the lambda probe (24), said injected fuel mass being used as an input value for an emissions model (42), - calculating the exhaust gas emissions downstream of the exhaust gas aftertreatment device (22) with the aid of the emission model (42), - comparing at least one of the measured values of the sensor (34) with the nitrogen oxide concentration, the ammonia concentration and the oxygen concentration calculated in the emission model (42), If at least one value calculated by the emission model (42) deviates from the corresponding value measured by the sensor (34), an error is inferred in the emission model (42) or in the lambda probe (24) or the sensor (34).

2. The method for predicting exhaust gas emissions of an internal combustion engine (10) according to claim 1, It is characterized by: If an error exists, the calculated value of the emission model (42) is corrected by taking into account the at least one measurement value of the sensor (34).

3. The method for predicting exhaust gas emissions of an internal combustion engine (10) according to claim 1 or 2, It is characterized by: calculating, by means of the emission model (42), an oxygen concentration, an ammonia concentration and / or a nitrogen oxide concentration downstream of the exhaust gas aftertreatment device (22) which is to be expected based on the measured values of the lambda probe (24), Subsequently, at least the measured value of the oxygen concentration, the ammonia concentration and / or the nitrogen oxide concentration measured by means of the sensor (34) is compared with the corresponding calculated value of the oxygen concentration, the ammonia concentration and / or the nitrogen oxide concentration from the emission model (42), If the measured value of the sensor (34) deviates from the corresponding calculated value of the emission model, the measured value of the lambda probe (24) is corrected according to the measured value of the sensor (34), and the injected fuel mass is calculated according to the corrected measured value, The exhaust gas emissions downstream of the exhaust gas aftertreatment device (22) are subsequently calculated by the emission model (42) using the corrected fuel mass.

4. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: In a phase in which the lambda probe (24) or the sensor (34) is not operationally ready or transmits an implausible signal, the injected fuel mass as well as the air mass and the residual gas content are read in from engine control data.

5. The method for predicting exhaust gas emissions of an internal combustion engine (10) according to claim 4, characterized in that In a combustion raw emission model (44) constituting the first part of the emission model (42), engine geometry data, operating data and engine control data, fuel injection data, air mass flow and fuel composition of the internal combustion engine (10) are used as input data, and the exhaust gas composition and exhaust gas mass flow downstream of the combustion chamber (15) of the internal combustion engine (10) are calculated with reference to the modeled combustion data.

6. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: Output data of the combustion raw emission model (44) is used as input data of an exhaust gas aftertreatment model (46) constituting a second part of the emission model (42), wherein, in a phase in which the lambda probe (24) and the sensor (34) are not in an operationally ready state, the fuel mass determined from the engine control data is used as an input value of the combustion raw emission model (44), and in a phase in which the lambda probe (24) and the sensor (34) are in an operationally ready state, the measured value of the lambda probe (24) corrected by the measured value of the sensor (34) is used as an input value of the exhaust gas aftertreatment model (46) and / or the combustion raw emission model (44).

7. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: The emission model (42) composed of the exhaust gas aftertreatment model (46) and the combustion raw emission model (44) is integrated into an engine control unit (36).

8. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: In the exhaust gas aftertreatment model (46), the expected concentrations of emission components occurring in the exhaust gas flow are read or calculated based on environmental data by means of a physical model or an algorithm.

9. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: In the combustion raw emission model (44), basic data of the combustion process, the exhaust gas composition of the exhaust gas downstream of the combustion chamber (15) of the internal combustion engine (10), the exhaust gas mass flow, the combustion air ratio and the fuel mass are calculated as output data.

10. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: In the exhaust gas aftertreatment model (46), the exhaust gas composition, exhaust gas mass flow, exhaust gas temperature and oxygen concentration, ammonia concentration and nitrogen oxide concentration downstream of each exhaust gas aftertreatment device (26, 30, 32) are calculated based on the output data of the combustion raw emission model (44) and when the lambda probe (24) and the sensor (34) are in an operationally ready state, additionally based on the corrected measured values of the lambda probe (24), the corrected measured values being generated by comparing the oxygen concentration, the ammonia concentration and / or the nitrogen oxide concentration measured by means of the sensor (34) with the corresponding oxygen concentration, ammonia concentration and / or nitrogen oxide concentration calculated in the emission model (42).

11. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: When determining an oxygen excess, the nitrogen oxide concentration determined from the measured values of the sensor (34) is used to correct the measured values of the lambda probe (24).

12. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: When oxygen deficiency is determined, the ammonia concentration determined from the measured values of the sensor (34) is used to correct the measured values of the lambda probe (24).

13. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: At least one value of the oxygen concentration, the ammonia concentration and / or the nitrogen oxide concentration downstream of the exhaust gas aftertreatment device (22) calculated in the exhaust gas aftertreatment model (46) is used as a process variable, and the corresponding measured value of the sensor (34) is used as a set value for a PID controller, and an error deviation of the measured value of the lambda probe (24) is inferred based on the control value of the sensor (34), and the measured value of the lambda probe is corrected based on the determined control value of the sensor (34).

14. Method for predicting exhaust gas emissions of an internal combustion engine (10) according to one of the preceding claims, It is characterized by: To calculate the injected fuel mass, the measured value of the lambda probe (24) is corrected by a maximum of 5%.

15. A device for predicting exhaust gas emissions of an internal combustion engine (10), the device comprising: an exhaust gas aftertreatment device (22); a lambda probe (24), the lambda probe being arranged downstream of a combustion chamber (15) of the internal combustion engine (10) and upstream of the exhaust gas aftertreatment device (22); a sensor (34), the sensor being used to measure a nitrogen oxide concentration, an ammonia concentration and / or an oxygen concentration in an exhaust gas flow, the sensor being arranged downstream of the exhaust gas aftertreatment device (22); and an engine control unit (36), wherein an emission model (42) is integrated in the engine control unit, the emission model comprising a combustion raw emission model (44) and an exhaust gas aftertreatment model (46), wherein the combustion raw emission model can be used to calculate raw emissions based on operating data of the internal combustion engine (10), and the exhaust gas aftertreatment model can be used to calculate emissions downstream of the exhaust gas aftertreatment device (22) based on the raw emission data. It is characterized by: A calculation unit (48) is integrated into the engine control unit (36), in which a fuel mass can be calculated based on the measured values of the lambda probe (24) and the measured values of the sensor (34), and the fuel mass is used as an input signal for the combustion raw emission model (44) and / or the exhaust gas aftertreatment model (46).

Images