Method for optimizing nitrogen oxide emissions and carbon dioxide emissions of a combustion engine

Abstract

A method for simultaneous optimization of nitrogen oxide emissions and carbon dioxide emissions of a combustion engine with an exhaust gas aftertreatment system of a motor vehicle. The method comprises the following steps: at the start a prediction horizon (PH) is selected (100), then a nitrogen oxide limit value (m.sub.NOx_max) is specified (101). Minimisation (102) of a cost function (K) comprising the nitrogen oxide emissions and the carbon dioxide emissions is carried out, wherein the nitrogen oxide limit value (m.sub.NOx_max) is complied with. Then actuators of the combustion engine are set (105) to a setpoint value (S) that is determined when minimizing (102) the cost function (K). Finally, the steps of the method are repeated.

Claims

1. A method for simultaneous optimization of nitrogen oxide emissions and carbon dioxide emissions of a combustion engine (2) with an exhaust gas aftertreatment system (10) of a motor vehicle comprising the following steps: Selecting (100), via a computer, a prediction horizon (PH, PH1, PH2); Specifying (101), via a computer, a nitrogen oxide limit value (m.sub.NOx_max); Minimizing (102), via a computer, a cost function (K) comprising the nitrogen oxide emissions and the carbon dioxide emissions, wherein the nitrogen oxide limit value (m.sub.NOx_max) is complied with, and wherein a violation of the nitrogen oxide limit value (m.sub.NOx_max) is taken into account in the form of a penalty term () in the cost function (K) that adopts a first value if the nitrogen oxide limit value (m.sub.NOx_max) is not exceeded and adopts a second value of finite size if the nitrogen oxide limit value (m.sub.NOx_max) is exceeded; and Adjusting (105) actuators of the combustion engine (2) to a setpoint value (S) determined when minimizing (102) the cost function (K).

2. The method according to claim 1, wherein models (200-208) of the route to be controlled are incorporated in the minimization (102) of the cost function (K).

3. The method according to claim 1, wherein a weighting factor () is used in order to vary a weighting between the nitrogen oxide raw emissions and the carbon dioxide emissions.

4. The method according to claim 3, wherein during the minimization (102) of the cost function (K) a weighting factor () is determined that is used when adjusting (105) the actuators of the combustion engine (2).

5. The method according to claim 1, wherein during the minimization (102) of the cost function (K) a correction factor (f.sub.NSC) for adjustment of the regeneration strategy for a nitrogen oxide storage catalytic converter (11) is determined that is used when adjusting (105) the actuators of the combustion engine (2).

6. The method according to claim 1, wherein during the minimization (102) of the cost function (K) a heating strategy for the catalytic converters (11, 12, 13) is determined that is used when adjusting (105) the actuators of the combustion engine (2).

7. The method according to claim 1, wherein the minimization (102) of the cost function (K) is carried out using at least one selected from the group consisting of a policy iteration, a value iteration, dynamic programming, a rollout algorithm, a shooting method.

8. The method according to claim 1, wherein the prediction horizon (PH1) is based on a specifiable period of time.

9. The method according to claim 1, wherein a prediction horizon (PH2) is based on a specifiable length of a route.

10. The method according to claim 9, wherein the route-based prediction horizon (PH2) is converted into a period of time by using the average speed of the motor vehicle and/or by using a speed to be expected over the route.

11. The method according to claim 1, wherein from a plurality of prediction horizons (PH1, PH2) the prediction horizon is selected that ends furthest into the future.

12. A non-transitory, computer-readable medium comprising a computer program that, when executed by an electronic control unit, causes the electronic control unit to select a prediction horizon (PH, PH1, PH2); specify a nitrogen oxide limit value (m.sub.NOx_max); minimize a cost function (K) comprising the nitrogen oxide emissions and the carbon dioxide emissions, wherein the nitrogen oxide limit value (m.sub.NOx_max) is complied with, and wherein a violation of the nitrogen oxide limit value (m.sub.NOx_max) is taken into account in the form of a penalty term () in the cost function (K) that adopts a first value if the nitrogen oxide limit value (m.sub.NOx_max) is not exceeded and adopts a second value of finite size if the nitrogen oxide limit value (m.sub.NOx_max) is exceeded; and adjust actuators of the combustion engine (2) to a setpoint value (S) determined when minimizing (102) the cost function (K).

13. An electronic control unit (3) configured to select a prediction horizon (PH, PH1, PH2), specify a nitrogen oxide limit value (m.sub.NOx_max); minimize a cost function (K) comprising the nitrogen oxide emissions and the carbon dioxide emissions, wherein the nitrogen oxide limit value (m.sub.NOx_max) is complied with, and wherein a violation of the nitrogen oxide limit value (m.sub.NOx_max) is taken into account in the form of a penalty term () in the cost function (K) that adopts a first value if the nitrogen oxide limit value (m.sub.NOx_max) is not exceeded and adopts a second value of finite size if the nitrogen oxide limit value (m.sub.NOx_max) is exceeded; and adjust actuators of the combustion engine (2) to a setpoint value (S) determined when minimizing (102) the cost function (K).

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) Exemplary embodiments of the invention are described in detail in the drawings and in the following description.

(2) FIG. 1 shows a schematic representation of a combined exhaust gas aftertreatment system that can be controlled by means of an exemplary embodiment of the method according to the invention.

(3) FIG. 2 shows the relationship between nitrogen oxide emissions and carbon dioxide emissions in a diagram.

(4) FIG. 3 shows a flow chart of an exemplary embodiment of the method according to the invention.

(5) FIGS. 4a to 4c each show two prediction horizons on a timeline.

DETAILED DESCRIPTION

(6) FIG. 1 shows schematically an exhaust system 1 that is connected to a combustion engine 2 for a motor vehicle that is not shown. The combustion engine produces exhaust gas that is discharged through the exhaust system 1. Furthermore, FIG. 1 represents a combined exhaust gas aftertreatment system 10 that is disposed in the exhaust system 1 and that is arranged to treat the exhaust gas in the exhaust system 1. For this purpose, the combined exhaust gas aftertreatment system 10 comprises a nitrogen oxide storage catalytic converter 11 that stores oxides of nitrogen (NOx), a SCR catalytic converter 12 that reduces the oxides of nitrogen in the exhaust gas, and a diesel oxidation catalytic converter 13. Here it should be remarked that in other exemplary embodiments the exhaust gas aftertreatment system 10 can comprise fewer catalytic converters and the order of the catalytic converters 11, 12, 13 can vary. The exact manner of operation of said components will not be described in detail here in order not to divert from the focus of this invention. The exhaust gas flows out of the combustion engine 2 through the exhaust system 1 into the combined exhaust gas aftertreatment system 10, where the nitrogen oxide component is reduced by the nitrogen oxide storage catalytic converter 11 and the SCR catalytic converter 12, and then exits the exhaust gas aftertreatment system 10 via the exhaust system 1.

(7) Moreover, two nitrogen oxide sensors 30 and 31 are disposed in the exhaust system 1. A first nitrogen oxide sensor 30 is disposed between the combustion engine 2 and the exhaust gas aftertreatment system 10 and is arranged to measure the nitrogen oxide raw emissions of the combustion engine 2. A second nitrogen oxide sensor 31 is disposed downstream of the exhaust gas aftertreatment system 10 and measures a nitrogen oxide value of the treated exhaust gas there. The two nitrogen oxide sensors 30 and 31 are connected to a control unit 3 and provide thereto information about the NOx content of the exhaust gas. Moreover, the control unit 3 is arranged to control the combined exhaust gas aftertreatment system 10 and the combustion engine 2.

(8) FIG. 2 shows a diagram, in which nitrogen oxide raw emission (NOx) are plotted against carbon dioxide emissions (CO2). An improvement in the emissions always causes a change of the respective other emissions. In the diagram, a pareto front 50 of the nitrogen oxide emissions and the carbon dioxide emissions is shown that represents the optimum states for both emissions. According to the invention, a weighting factor is provided, by means of which the weighting between the nitrogen oxide emissions and the carbon dioxide emissions is varied. With the change in the weighting by the weighting factor , it is to be noted that the pareto front 50 is not exited in order to ensure that an optimum state is achieved. The weighting factor can adopt any value between zero and one. In FIG. 2, individual optimum states on the pareto front 50 with the weighting factors =0, =0.25, =0.5, =0.75 and =1 are emphasised by way of example.

(9) In FIG. 3, a flow chart of an exemplary embodiment of the method according to the invention is represented. In a first step, a prediction horizon PH is selected 100. The prediction horizon PH is a time window, within which information about future values and/or influencing variables can be obtained, such as for example the nitrogen oxide emissions and the carbon dioxide emissions and/or other harmful emissions, the temperature of the combustion engine 2, the temperature of the exhaust gas and/or the temperatures of the catalytic converters 11, 12, 13, an exhaust gas mass flow {dot over (m)}, the speed of the motor vehicle etc. For a detailed description of the selection 100 of the prediction horizon PH, refer to FIG. 4. In a further step, a nitrogen oxide limit value m.sub.NOx_max is specified 101 using legal limit values for the nitrogen oxide emissions during the prediction horizon PH. In said exemplary embodiment, the nitrogen oxide limit value m.sub.NOx_max is 80 mg/km.

(10) For the current prediction horizon PH, a minimization 102 of a cost function K is carried out as a function of an emission weighting factor (), a correction factor f.sub.NSC for the adjustment of a regeneration strategy for the nitrogen oxide storage catalytic converter 11 and an indicator I.sub.T for requesting heating measures for the catalytic converters 11, 12, 13. In order to estimate the effect of the various measures, individual models, a plurality of models or possibly all of the models 200-208 that are described below of the route to be controlled are used: A model 200 for the carbon dioxide emissions, which is stored in the form of characteristic fields as a function of a load and a revolution rate of the combustion engine 2 and the weighting factors ; a model 201 for the nitrogen oxide emissions, which is stored in the form of characteristic fields as a function of the load and the revolution rate of the combustion engine 2 and the weighting factors ; an oxygen-based model 202 for the nitrogen oxide emissions, which is obtained phenomenologically (alternative to the model 201); a model 203 for the exhaust gas mass flow {dot over (m)}, which is stored in the form of characteristic fields as a function of a load and a revolution rate of the combustion engine 2 and the weighting factors ; a turbo charger model 204, which is determined from an energy balance; a model 205 for the exhaust gas temperature, which phenomenologically determines the temperature at the output of the combustion engine 2 as a function of the load and the revolution rate of the combustion engine 2, the temperature of the air/fuel mixture at the input of the combustion engine 2, the temperature of the combustion engine 2, the exhaust gas mass flow {dot over (m)} and the weighting factors ; a model 206 for calculating the temperature of the catalytic converters 11, 12, 13 based on an energy balance, wherein exothermal reactions are also formed, for example a conversion of unburnt hydrocarbons, as a function of the exhaust gas temperature, the exhaust gas mass flow {dot over (m)}, the speed of the vehicle, and a mass of the hydrocarbons in the exhaust gas, wherein the model 205 in said exemplary embodiment is an explicitly formulated model, i.e. a solution of a basic differential equation; a model 207 of a nitrogen oxide conversion rate for the nitrogen oxide storage catalytic converter 11; a model 208 for a nitrogen oxide conversion rate for the SCR catalytic converter 12.

(11) Route information 210 is incorporated in the models 200-208. Said information is determined from navigation data, traffic information and map data, for example. In the present exemplary embodiment, the models 200-208 pass through a low pass filter 210 before being used in the minimization 102 of the cost function K, so that the described variables are incorporated as statistical expected values.

(12) The cost function K is expressed by the following equation 1:

(13) $\begin{array}{cc}K\left(u\left(t\right),t\right)={}_{t=0}^{\mathrm{PH}}\left({\stackrel{.}{m}}_{{\mathrm{CO}}_2}\left(u\left(t\right),t\right)+*m_{\mathrm{NOx}}\left(u\left(t\right),t\right)*{}_{\mathrm{NSC}}\left(u\left(t\right),t\right)+\left(u\left(t\right),t\right)\right)\mathrm{dt}& \left(\mathrm{Equation}1\right)\end{array}$

(14) In this case, u(t) is a control vector of the combustion engine 2 comprising as inputs the weighting factor , the correction factor f.sub.NSC for adjustment of the regeneration strategy for the nitrogen oxide storage catalytic converter 11 and the indicator I.sub.T for requesting heating measures for the catalytic converters 11, 12, 13 as a function of the time t. {dot over (m)}.sub.CO.sub.2 and {dot over (m)}.sub.NOx represent the exhaust gas mass flow of carbon dioxide and of nitrogen oxide of the combustion engine 2. .sub.NSC represents the storage efficiency of the nitrogen oxide storage catalytic converter 11. is a factor for assessing the regeneration of the nitrogen oxide storage catalytic converter 11 that must be carried out at a later point in time after nitrogen oxide has been stored therein, and then results in an increase in the carbon dioxide emissions. is calculated by forming a moving average from preceding regenerations of the nitrogen oxide storage catalytic converter 11 by setting the carbon dioxide emissions applied during the regeneration in relation to the converted nitrogen oxide mass. represents a penalty term, by means of which the nitrogen oxide limit value m.sub.NOx_max is taken into account as a boundary condition in the equation 1. Said boundary condition of the nitrogen oxide restriction can be expressed by the following equation 2:

(15) $\begin{array}{cc}{}_{t=0}^{\mathrm{PH}}({\stackrel{.}{m}}_{{\mathrm{CO}}_2}\left(u\left(t\right),t\right)*(1-{}_{\mathrm{NSC}}\left(u\left(t\right),t\right)*\left(1-{}_{\mathrm{SCR}}\left(u\left(t\right),t\right)\right)\mathrm{dt}{m}_{\mathrm{NOx\_max}}-m_{\mathrm{NOx\_act}}& \left(\mathrm{Equation}2\right)\end{array}$
.sub.SCR is analogous to an efficiency of the SCR catalytic converter 12. m.sub.NOx_act represents the current nitrogen oxide mass and is calculated by continuous integration of the measurement values of the second nitrogen oxide sensor 31 or by integration of model values. A possible violation of the boundary condition specified in the equation 2 is transferred by means of a transfer function into the penalty term in the equation 1. If the boundary condition specified in the equation 2 is satisfied, the penalty term in the equation 1 is selected as zero, for example. If the equation 2 is not satisfied, then the penalty term in the equation 1 adopts a finite positive value that is so large that the corresponding control strategy for the minimization 102 of the cost function K is reliably excluded. For example, the penalty term can adopt a value that exceeds the other values from equation 1 by a factor of ten.

(16) The minimization 102 of the cost function is carried out using a shooting method or one or more of the following algorithms based on the Bellman optimality principle: Policy iteration; Value iteration; Dynamic programming; and/or a rollout algorithm.

(17) The minimization 102 of the cost function K and the associated determination of an optimum control vector (t), is carried out according to equation 3. The optimum control vector (t) contains as entries the optimum weighting factor and the correction factor f.sub.NSC for adjustment of the regeneration strategy for the nitrogen oxide storage catalytic converter 11 and the indicator I.sub.T for requesting heating measures for the catalytic converters 11, 12, 13, which are implemented in the form of special engine operating modes.

(18) $\begin{array}{cc}\hat{u}\left(t\right)=\arg \underset{uU}{\min }k\left(u\left(t\right),t\right)& \left(\mathrm{Equation}3\right)\end{array}$

(19) In the further method, setpoint values S for actuators of the combustion engine 2 are derived 103 from the optimum weighting factor and the correction factor f.sub.NSC for adjustment of the regeneration strategy for the nitrogen oxide storage catalytic converter 11 and the indicator I.sub.T for requesting heating measures for the catalytic converters 11, 12, 13. An additional correction 104 of the setpoint values S is carried out in order to achieve a desired system behaviour under specifiable boundary conditions. Finally, the actuators of the combustion engine 2 are adjusted 105 by means of the setpoint values S. The method is subsequently repeated based on the new system state starting with the selection 100 of the prediction horizon PH in order to optimize the nitrogen oxide emissions and the carbon dioxide emissions.

(20) In the FIGS. 4a to 4c, the selection 100 between a first prediction horizon PH1 and a second prediction horizon PH1 for a respective new repetition of the method according to the invention is described using timing diagrams. For the first prediction horizon PH1, a fixed period of time is specified, i.e. it is time-based. The point in time t.sub.PH1a, t.sub.PH1b, t.sub.PH1c at which the first prediction horizon PH1 ends is equally far from the point in time t.sub.1, t.sub.2, t.sub.3 at which the prediction horizons PH1, PH2 are selected in all timing diagrams. Consequently, the period of time for the prediction horizon during the journey always remains the same. For the second prediction horizon PH2, a length of a route to be traveled is specified, i.e. it is route-based. In order to determine the length of the route, the route information 210, navigation data and/or data stored over preferred routes can be used. Alternatively, a route length can be applied. The route-based prediction horizon PH2 is converted into a period of time by using the average speed of the motor vehicle and/or by using a speed to be expected over the route that is also determined using the aforementioned data. Accordingly, the point in time t.sub.PH2a, t.sub.PH2b, t.sub.PH2c at which the second prediction horizon PH2 ends remains the same and the period of time is different for each sub diagram, because the route that is still to be traversed is shorter. With one embodiment of the method according to the invention, it is provided to select the prediction horizon PH1, PH2 that ends later. In the cases represented in the timing diagrams 4a and 4b, the route-based second prediction horizon PH2 is therefore selected, because the point in time t.sub.PH2a, t.sub.PH2b at which the second prediction horizon PH2 ends lies further into the future than the points in time t.sub.PH1a, t.sub.PH1b at which the first prediction horizon PH1 ends. In the sub diagram 4c, the opposite case is represented, in which the time-based first prediction horizon PH1 is selected because the point in time t.sub.PH1c at which the first prediction horizon PH1 ends is now further into the future than the point in time t.sub.PH2c at which the second prediction horizon PH2 ends.