Method for controlling the nox concentration in the exhaust gas of an internal combustion engine
In a method for controlling the NOx concentration in the exhaust gas of an internal combustion engine in which an NOx control deviation is calculated from an actual NOx value and a desired NOx value, a control value is calculated based on the NOx control deviation via an NOx controller and by means of the control value at least one condition of the cylinder inlet flow volume of the internal combustion engine is adjusted in that, additionally, an adaption injection begin is determined in dependence on the control value of the NOx controller, and the injection begin is changed by the adaption injection begin.
1. A method for controlling the NOx concentration in the exhaust gas of an internal combustion engine ( 1 ), comprising the steps of:
calculating an NOx control deviation (dNOx) of a momentary NOx value (NOx(IST)) from a desired NOx value (NOx(SL)),
calculating based on the NOx control deviation (dNOx) by way of a NOx controller ( 20 ) a control value (SG) needed for minimizing the NO x control deviation (d(NO x )) and adjusting, via the control value (SG), at least one of a charging pressure (p 5 ) and also a recirculated exhaust gas flow volume (AGR) of the internal combustion engine ( 1 ),
additionally determining a first adaption injection begin (SB(A)) by an evaluation of the control value (SG) of the NOx controller ( 20 ) based on limit values and selecting, in dependence on this evaluation, a first adaption function ( 28 ), a second adaption function ( 29 ) or a normal function 27 , and
with the first adoption function activated, retarding the injection begin (SB) by the adaption injection begin (SB(A)), with the second adaption function ( 29 ) activated, advancing the injection begin (SB) by way of the adaption injection begin (SB(A)) and, with the normal function ( 27 ) activated, keeping the injection begin (SB) unchanged.
2. The method according to claim 1 , wherein the first adaption function is activated when the recirculated exhaust gas flow volume (ACR) is greater than, or equal to, a flow volume maximum value (MAX1) or when the charging pressure (p 5 ) is greater than, or equal to, a maximum pressure value (MAX3).
3. The method according to claim 2 , wherein the first adaption function ( 28 ) is reset and the normal function ( 27 ) is reactivated when the recirculated exhaust gas flow volume (AGR) is smaller than the flow volume maximum value (MAX1) reduced by a flow hysteresis value (HY1) or when the charging pressure (p 5 ) is smaller than the maximum pressure value (MAX3) reduced by a pressure hysteresis value (HY3) or when the adaption injection begin (SB(A)) is greater than, or equal to, a maximum injection begin value (MAX2).
4. The method according to claim 1 , wherein the second adaption function ( 29 ) is activated when the recirculated exhaust gas volume (AGR) is smaller than the flow volume maximum value (MAX1) reduced by the low hysteresis value (HY1) or when the charging pressure (p 5 ) is smaller than the maximum pressure value (MAX3) reduced by the pressure hysteresis value (HY3) and when additionally the adaption injection begin (SB(A)) is greater than a minimum value (MIN2).
5. The method according to claim 4 , wherein the second adaption function ( 29 ) is reset and the normal function ( 27 ) is reactivated when the recirculated exhaust gas flow volume (ACR) is greater than, or equal to, the flow volume maximum value (MAX1) or when the charging pressure (p 5 ) is greater than, or equal to, the maximum pressure value (MAX3) or when the adaption injection begin (SB(A)) is smaller than, or equal to, the minimum injection begin value (MIN2).
6. The method according to claim 1 , wherein the adaption injection begin (SB(A)) is calculated by assigning to the control value (SG) of the NOx controller ( 20 ) via the logic ( 24 ) a control value, the control value is added via an integrator ( 25 ) and is limited by a limiter ( 26 ).
7. The method according to claim 6 , wherein the adaption injection begin (SB(A)) is calculated by way of a controller with at least PI behavior rather than via the integrator ( 25 ).