Title of Invention	METHOD OF SETTING THE DOSING OF THE REDUCING AGENT IN SELECTIVE CATALYTIC REDUCTION
Abstract	A method for use in conjunction with an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, wherein the control unit admeters the amount of reducing agent dependent on a stored model, and during the operation of the internal combustion engine the control unit, by comparison of a variable stored for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system in the control unit of the internal combustion engine which is proportional to the desired emission or the desired conversion with a variable detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, determines a deviation between actual emission and desired emission or actual conversion and desired conversion, and determines a correction value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this correction value, and wherein the determination of the actual emission or the actual conversion takes place such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor and/or of an HNCO sensor and/or a lambda sensor arranged downstream of the SCR catalyst or the actual conversion determined with the aid of at least one of these sensors and the untreated emission, for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integrating at least one operating parameter, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges-being determined by variables stored in the control unit.

Title of Invention

METHOD OF SETTING THE DOSING OF THE REDUCING AGENT IN SELECTIVE CATALYTIC REDUCTION

Abstract

A method for use in conjunction with an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, wherein the control unit admeters the amount of reducing agent dependent on a stored model, and during the operation of the internal combustion engine the control unit, by comparison of a variable stored for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system in the control unit of the internal combustion engine which is proportional to the desired emission or the desired conversion with a variable detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, determines a deviation between actual emission and desired emission or actual conversion and desired conversion, and determines a correction value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this correction value, and wherein the determination of the actual emission or the actual conversion takes place such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor and/or of an HNCO sensor and/or a lambda sensor arranged downstream of the SCR catalyst or the actual conversion determined with the aid of at least one of these sensors and the untreated emission, for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integrating at least one operating parameter, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges-being determined by variables stored in the control unit.

Full Text	Method of setting the dosing of the reducing agent in selective catalytic reduction The invention is directed at a method according to the preamble of Claim 1. In addition to solids particles, nitrogen oxides are one of the limited exhaust constituents which are produced during combustion processes and the permitted emissions of which are being reduced further and further. Various methods are nowadays used to minimise these exhaust constituents in internal combustion engines operated in motor vehicles. The reduction of the nitrogen oxides takes place mostly with the aid of catalysts. In oxygen-rich exhaust gas a reducing agent is additionally necessary in order to increase the selectivity and the NOx conversions. These methods have become known under the collective term "SCR methods", SCR standing for "selective catalytic reduction". They have been used for many years in the power station industry and recently also in internal combustion engines. A detailed account of such methods can be gathered from DE 34 28 232 A1. Mixed oxides containing V205, for example in the form V2O5AA/O3/T1O2, can be used as SCR catalysts. Typical V205 contents are in that case between 0.2 and 3%. The use of iron-containing and/or copper-containing zeolites is also conceivable. In practical application, ammonia or compounds which split off ammonia, such as urea or ammonium formate, in solid or solution form, are used as reducing agents. Thus the urea decomposes at high temperatures to form isocyanic acid and ammonia. (Formula Removed) The isocyanic acid hydrolyses further with water contained in the exhaust to form NH3 and C02. (Formula Removed) With complete hydrolysis of one mole of urea, thus two moles of ammonia and one mole of carbon dioxide are produced. (Formula Removed) Owing to the hydrolysis of urea, this makes available the same proven reducing agent as in the power station industry, namely ammonia. One mole of ammonia is in that case necessary for reacting one mole of nitrogen monoxide. (Formula Removed) With an ideal catalyst this means that, for a feed ratio of one, all the nitrogen oxides are reduced, i.e. a 100% NOx conversion is achieved, since XNOx applies for the NOx conversion: (Formula Removed) If the SCR catalysts are preceded by a platinum-containing NO oxidation catalyst for the formation of N02, (Formula Removed) the SCR reaction can be considerably accelerated and the low-temperature activity can be noticeably increased. (Formula Removed) However, in the presence of N02 increased nitrous oxide emissions in accordance with the following reaction also have to be expected: (Formula Removed) With internal combustion engines operated in vehicles, the reduction of nitrogen oxides with the aid of the SCR method is difficult because changing operating conditions, such as fluctuating exhaust temperatures, amounts of exhaust and untreated NOx emissions, prevail therein, which makes it more difficult to admeter the quantities of reducing agent. Although on one hand the highest possible conversion of nitrogen oxides should be achieved, on the other hand it should be ensured that there is no emission of nitrous oxide, isocyanic acid or unspent ammonia. Currently, two methods for determining the correct dosing amount of reducing agent are used when admetering the reducing agent for the SCR method in vehicles. Firstly, this is a pure control without check-back of sensors for determining the actual emissions after the catalyst system. The dosing amount in this case is determined with the aid of models, from data which are stored and/or are determined in a memory of an electronic engine control unit of an internal combustion engine in the form of tables or curves, sets of characteristic curves or functions, and optionally with the aid of sensors for determining the catalyst temperature, the amount of NOx and amount of exhaust. Thus for example the untreated emission of the engine is calculated from the injection quantity, the engine speed, the injection pressure, the fuel/air ratio etc. The possible NOx conversions and the dosing amounts of reducing agent necessary therefor in turn depend on the catalyst temperature, the untreated NOx emission, the amount of exhaust etc. The actual emissions after the system are not detected and thus have no influence on the dosing amount (DE 43 15 278 A1, DE 195 36 571 A1, DE 199 06 344 A1, EP 898 061 A1). The disadvantage of this method is that faults, defects or environmental influences can scarcely be compensated for owing to the lack of check-back about the actual emissions. The second possibility of admetering the reducing agent is to construct a conventional closed control loop with the aid of HNCO, N20, NOx and/or NH3 sensors after the system. For this, the actual values currently supplied by the sensors are compared with the desired values and the dosing amount is constantly adapted. However, there is the problem of permanent regulation in the inertia of the system and the sensors and also of the simultaneously highly dynamic operation of the internal combustion engine in vehicles. Thus for example during acceleration operations or additional loads on engines on exhaust-charged internal combustion engines the NOx emissions may • increase by a factor of 10 within one second. In the case of aspirated engines the increase is even faster owing to the lack of inertia of the exhaust turbocharger. The same applies in the event of loss of load or upon transition into the overrun condition. The sensors for determining the emissions are not capable of detecting these highly-dynamic operations. This is due firstly to the inertia of the sensors, the typical t90 time of which, i.e. the time at which 90% of the final value is reached, should be settled at 300-500 ms, and secondly to the necessary positioning of the sensors after the catalyst system. Thus merely the gas travel time from emerging from the cylinder head to emerging from the catalyst system, depending on exhaust-gas flow rate and volume of the exhaust system, is 200-2000 ms. In order to remedy this problem, a generic exhaust aftertreatment system and a method for controlling this exhaust aftertreatment system, in particular for an internal combustion engine, which partially solves this problem, is proposed in DE 101 00 420 A1. The exhaust aftertreatment system, which comprises at least one catalyst, is supplied with a predetermined amount of reducing agent dependent on the operating state of the internal combustion engine and/or of the exhaust aftertreatment system. The amount of reducing agent supplied is adapted. For this, in stationary operation of the internal combustion engine, the emission of nitrogen oxides and/or of ammonia is measured by means of sensors and is compared with desired values stored for this stationary operating state. If a deviation is detected, the control of the exhaust aftertreatment system determines a correction value with which the amount of reducing agent supplied is adapted. What is disadvantageous with this system is that the adaptation, owing to the inertia of the system described above, can only take place in a relatively long stationary operating phase at predetermined operating parameters of the internal combustion engine. "Stationary operating conditions" are understood in this case to mean that the operating variables which determine the admetering of reducing agent may not, or may only minimally, change. Such stationary operating phases do not occur over relatively long periods in certain operating modes of the internal combustion engine, e.g. if it is operated in a vehicle and the vehicle is moving in urban traffic. This means that a correction of the amount of reducing agent cannot take place over relatively long periods, and therefore an increased pollution emission of nitrogen oxides and/or ammonia occurs. A further approach for countering the problems described can be inferred from DE 195 36 571 A1. Therein, a method and also an associated device for dosing the amount of a reducing agent introduced into the exhaust stream or exhaust-air stream of combustion systems, in particular combustion engines, with a subsequent catalyst is described. The amount of reducing agent introduced is set, departing from operation-relevant parameters of the combustion system, the exhaust and the catalyst, via (sets of) characteristic curves, the position of the (sets of) characteristic curves being checked and adapted to the current state and also the current operating conditions of the combustion engine, the exhaust and the catalyst. Therefore the characteristic curves or sets of characteristic curves is/are adapted by comparing the actual pollutant concentration determined by means of sensors with stored desired values. With this procedure too, the great inertia of the system, in particular of the sensors, means that checking is only possible when stationary operating conditions of the internal combustion engine exist, so the disadvantages described above for DE 101 00 420 A1 apply here too. It is an object of the invention to avoid the disadvantages of the known methods. This object is achieved by the method according to Claim 1; advantageous embodiments of the method are characterised in the dependent claims. The method according to the invention is applied in conjunction with an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air. As is conventional with such arrangements, the dosing of the reducing agent takes place with the aid of a dosing means associated with the exhaust aftertreatment system and controlled by a control unit. For the reduction of nitrogen oxides, at least one SCR catalyst is arranged downstream of the dosing means in the exhaust stream, as a further part of the exhaust aftertreatment system. The dosing amount of the reducing agent - in systems which are nowadays conventional this is usually aqueous urea solution, but alternatively other reducing agents (ammonia or compounds which split off ammonia, such as urea or ammonium formate in solid or solution form) are conceivable - is determined by the control unit with the aid of a model stored therein which covers all possible operating points of the internal combustion engine or of the exhaust aftertreatment system. A model in this case is to be understood to mean, in the simplest case, a characteristic curve or a set of characteristic curves, but it may also be a plurality of characteristic curves, sets of characteristic curves or alternatively one-parameter or multi-parameter functions which are determined or established with the aid of what is called a reference arrangement and/or by theoretical considerations. The reference arrangement in the present case is an arrangement, typical of a model series, of internal combustion engine and exhaust aftertreatment system which may already be installed in a vehicle. By measurements on the reference arrangement and/or by theoretical considerations, for a plurality of operating points of the arrangement on one hand a dosing amount of the reducing agent is determined and on the other hand a desired emission is determined for these operating points. Each operating point in this case is defined by the size of at least one operating parameter of the reference arrangement. The dosing amounts determined and the associated desired emissions or desired conversions are held in the form of a model in the control units of the corresponding model series, such that a variable which is proportional to the dosing amount and a variable which is proportional to the desired emission or desired conversions can be obtained directly or by interpolation by means of the control unit from the model, for all possible values which the at least one operating parameter may adopt, i.e. for all the operating points which occur. From this model, i.e. the characteristic curves, sets of characteristic curves or functions, the control unit determines the dosing amount, dependent on the at least one operating parameter, evaluated by the control unit, of the internal combustion engine and/or of the exhaust aftertreatment systems. The respective momentary value of the at least one operating parameter in this case determines the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system. In practice, it is frequently a question not of only one, but of a plurality of operating parameters which influence the correct dosing amount of the reducing agent; consequently this multiple dependency has to be taken into account in the model. Since this multiple dependency is however not constant, but is subject in particular to a change over time, i.e. one determined by the number of operating hours of a system, and also to a change due to environmental influences, the dosing amount determined by means of the model has to undergo a correction. For this, the procedure is advantageously such that, during the operation of the internal combustion engine, the control unit determines a deviation by comparison of a variable determined from stored values for the internal combustion engine and/or of the exhaust aftertreatment system by means of the control unit of the internal combustion engine, which variable is proportional to the desired emission or the desired conversions, with a variable detected by measurement by the control unit, which variable is proportional to the actual emission or the actual conversions, and determines a correction value for the dosing amount dependent on this deviation. For subsequent dosing operations, the stored model is then modified with this correction value by the control unit. The modification of the model with the correction value is in each case retained until the control unit, by comparing a variable which is proportional to the desired emission or the desired conversions and is determined from stored values for the internal combustion engine and/or the exhaust aftertreatment system by means of the control unit of the interna! combustion engine, with a variable re-detected by measurement by the control unit which is proportional to the actual emission or the actual conversions, determines a deviation and dependent on this deviation determines a new correction value for the dosing amount. For subsequent dosing operations, the stored model is then modified with this new correction value. For detection of the actual emission value or the actual conversions of an internal combustion engine with subsequent exhaust aftertreatment system, on one hand a certain time is necessary over which the signal of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor is added up or integrated; additionally the operating conditions may in so doing not change if at all possible. This restricted the determination of actual emission values or actual conversions per se to stationary operating conditions. In order nevertheless advantageously to obtain independence from stationary operating conditions for the determination of the correction value, in the determination of the actual emission or the actual conversions according to the method of the invention the procedure is such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor arranged downstream of the SCR catalyst, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges being determined by variables stored in the control unit. The addition or integration may take place for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integration of at least one operating parameter. This operating parameter may for example be the amount of exhaust and/or the amount of fuel and/or the amount of reducing agent and/or the work performed by the internal combustion engine. Furthermore, the operating parameter intended for monitoring the value ranges can also be used for determining the duration of the addition or integration. Of course, determination of the untreated emissions is additionally necessary for determining the desired and actual conversions. The reaching of the amount of emission can in this case be determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows. Due to the above-mentioned allocation of value ranges within which the at least one operating parameter may vary during the adding-up or integration, i.e. by class formation, the frequency at which actual emission values or actual conversions and hence correction values for dosing the reducing agent can be obtained is drastically increased. In this case, by selecting the value ranges within which the at least one operating parameter may vary, the error produced by this class formation can be kept within a negligible order of magnitude. If the at least one operating parameter leaves the current value range, i.e. the class, during the adding-up or integration, two alternative process configurations are yielded. On one hand, it is possible to proceed such that the control unit, upon establishing that the current value range is being left, discards the added-up or integrated sum. On the other hand, upon establishing that the current value range is being left, the added-up or integrated sum can be temporarily stored by the control unit in order, when the control unit establishes a return to the value range previously left, to continue with the adding-up or integration until a pre-settable amount of emission, work produced or sum or integral of a further operating parameter is reached or the pre-settable time t for adding-up or integration has passed. Both alternatives have advantages and disadvantages. In the first case, it is ensured that the adding-up or integration takes place in one go, i.e. there are only seconds between the start and end of the operation. In such short periods, temporal influences or environmental influences do not become apparent, and therefore do not have an effect on the measurement result for the actual emission. On the other hand, however, the frequency at which new actual emission measured values are available is greatly reduced. In the second case, at least theoretically, days or even weeks can pass before a determination of actual emissions is concluded, which means that temporal influences or environmental influences might very well have an effect and falsify the measurement result. This can be countered by time-limiting the validity of temporarily stored values. What is advantageous in the process configuration with temporary storage is that the frequency at which actual emission measured values and hence current correction values are available is greatly increased. In the further course of the process, the control unit uses the sum added-up or integrated over the pre-settable time t or until the pre-settable amount of emission, work produced or sum or integral of a further operating parameter is reached, or a variable which is proportional thereto as actual emission for the comparison with the stored desired emission, and determines a correction value for the dosing amount with the established difference. The current correction values in each case which are valid for a value range of the at least one operating parameter are held in the control unit, so that the control unit can resort thereto, in order to modify the dosing amount from the model with the correction value determined for the respective value range of the at least one operating parameter, if the value of the at least one operating parameter currently determined by the control unit lies in the corresponding value range. The model for determining the dosing amount can thus be adapted efficiently and hence advantageously to changes over time and environment-determined changes. Alternatively, there is the possibility that the control unit, dependent on the current value, determined by the control unit, of the at least one operating parameter, from correction values which were determined for values of the at least one operating parameter directly neighbouring to this current value, determines an intermediate value as correction value by interpolation. In this manner, on one hand the error due to the class formation can be made advantageously smaller, because intermediate values to the stored correction values can be formed for each operating point, and on the other hand value ranges for which no correction value has yet been determined can be advantageously bridged by interpolation. One particularly simple and hence advantageous method of determining the desired emission or the desired conversion from the model consists in that the control unit in the pre-settable time t or until a pre-settable amount of emission or work produced or sum or integral of a further operating parameter is reached, i.e. timewise in parallel with the determination of the actual emission or the actual conversion, adds up or integrates the values for the desired emission or the desired conversion obtained from the model dependent on operating point and the sum thus obtained or a variable which is proportional thereto is used as desired emission or desired conversion for the comparison of the actual emission or the actual conversion with the desired emission or the desired conversion. The significant advantage with this procedure is that, with the exception of the operating parameter for which the correction value is to be determined and which may therefore vary only within the associated value range, i.e. the class, all other operating parameters may assume any desired values, because the influences yielded thereby are eliminated by the adding-up or integration of the actual emissions and the desired emissions. Or, in other words, the same operating points are passed through both in determining the actual emission or the actual conversion and in determining the desired emission or the desired conversion, so that the influences thereby produced are eliminated upon comparison of both values and as a result only the deviation relative to the operating parameter for which the correction value is to be determined is left over. In order to modify the dosing amount determined from the models with the correction value, the control unit forms the sum of dosing amount and correction value if the correction value is stored in the form of a positive or negative corrective dosing amount, or the product of dosing amount and correction value if the correction value is stored as a factor. With regard to the operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system which are evaluated by the control unit, the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the emissions and/or the fuel/air ratio and/or the change in these variables over time are considered. This means for the dosing of the reducing agent that, in the simplest case, i.e. when only one operating parameter, e.g. the exhaust mass stream, is evaluated by the control unit, the model held in the control unit would contain only one characteristic curve. If several operating parameters are evaluated by the control unit, these are one or more sets of characteristic curves or one or more multi-parameter function(s). Owing to the numerous operating parameters which may have an influence on the dosing amount for the reducing agent, it is useful for the control unit to determine and store different correction values and/or different parameters of at least one correction function and/or a plurality of correction functions for different operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system, and for the control unit, dependent on the current operating point of the internal combustion engine and/or of the exhaust aftertreatment system, to determine an operating-point-related correction value from these correction values and/or correction functions. Owing to the multiple dependency addressed above, the importance of the allocation of the operating parameters to classes becomes particularly clear, because only in this way can multiply-dependent correction values be determined at all within an acceptable period. The class formation causes operating-parameter ranges within defined limits in each case to be viewed as constant, so that as a consequence an operating-point range is yielded which can be treated as quasi-stationary operation. The probability that the real operating point at which the internal combustion engine and/or the exhaust aftertreatment system currently is will lie for long enough in such an operating-point range, which is regarded as quasi-stationary, to determine the correction value increases drastically due to the aforementioned class formation. HNCO sensors and/or N20 sensors and/or NOx sensors and/or NH3 sensors and/or lambda sensors can be used as sensors for determining the actual emissions, the untreated emissions and the actual conversions. These sensors are commercially available, and their construction therefore does not require any further explanation. The invention will be explained in greater detail below with reference to examples of embodiment. For the following embodiments, the starting point is an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air. As has already been indicated at the beginning, it is not sufficient for such systems to dose the reducing agent in controlled manner dependent on operating point, rather it is necessary, dependent on changes over time of the exhaust aftertreatment system or of the internal combustion engine, and also dependent on environmental influences, to correct the dosing amount. For this, operating-point-dependent correction values are determined. The method of operation to determine a correction value with respect to the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system, will be indicated below by way of example. In this case it is a prerequisite that the described sequences are carried out as programmed control sequences in a control unit, e.g. a programmable, electronic engine control unit, such as are used in modern internal combustion engines. The control unit is connected via a plurality of sensors to the internal combustion engine and the exhaust aftertreatment system, and determines via these sensors all current values of relevant operating parameters of the internal combustion engine or of the exhaust aftertreatment system, and also the current actual emission or the current actual conversion of the exhaust aftertreatment system. The current values of the relevant operating parameters in this case define the current operating point of the internal combustion engine or of the exhaust aftertreatment system. First of all, continuous monitoring of the relevant operating parameters takes place during the operation of the internal combustion engine. This on one hand serves the purpose of gathering, dependent on operating point, from the model stored in the control unit, the dosing amount of the reducing agent, on the other hand the monitoring of the operating parameters serves to determine, dependent on operating-parameter values, correction values for the amount of reducing agent which is held, dependent on operating point, as a stored value. This determination of the correction values will be gone into in detail below. As already stated, the determination of the correction values takes place in class-related manner. This is to be understood to mean that the values which the relevant operating parameters can adopt are divided up into value ranges or classes. At least two value ranges or classes are defined for each relevant operating parameter. How many classes exist for an operating parameter, or of what nature the classes are, i.e. what initial value and what final value they have, depends on in which value ranges of the corresponding operating parameter a negligible change in the correction value is yielded. The corresponding value ranges were determined with the aid of a reference system. The reference system, as already explained, is an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, which corresponds to the series specification and on which the corresponding measurements are carried out in tests. The value ranges or classes for each of the relevant operating parameters thus determined are stored in the control units of the series, so that they can be resorted to. During operation of a system in accordance with the series specification, and on the assumption that initially no correction values are stored in the control unit, e.g. upon the first starting-up of the internal combustion engine, the dosing takes place using the dosing amounts held in the model, i.e. uncorrected. At the same time, the control unit begins to interrogate the current values of the relevant operating parameters cyclically and to compare them with the value ranges or classes stored in the control unit. Dependent on the result of this comparison, two possibilities are yielded: The current value of the operating parameter remains within the limits of one class; The current value of the operating parameter exceeds the class limit. Both the aforementioned cases will be gone into in greater detail below. A further operation takes place in parallel with the aforementioned operation. In this case, with respect to the current class or classes in which the operating parameter(s) are located, the value supplied by an exhaust sensor or the conversion determined with the aid of this sensor is added up or integrated by the control unit. For the adding-up or integration a time t, an amount of emission, a work produced by the internal combustion engine or the sum or the integral of a further operating parameter can be pre-set in the control unit, after the expiry of which, under the prerequisite that until the expiry thereof, the class or classes has/have not been left, the control unit terminates the adding or integration operation and uses the value thus obtained as equivalent to the actual emission or to the actual conversion of the system for an actual value/desired value comparison. The pre-settable time t should in that case be at least 15 seconds. In the event that a pre-settable amount of emission serves as interruption criterion instead of the time t, the amount of emission, as already mentioned, can be determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows. If the pre-settable amount of emission is given in units of mass by adding-up or integration of emission masses and/or emission mass flows, the pre-settable amount of emission for NOx should be at least 1 mg and/or the pre-settable amount of emission for NH3 at least 0.01 mg and/or the pre-settable amount of emission for N20 at least 0.02 mg and/or the pre-settable amount of emission for HNCO at least 0.01 mg. The value of the desired emission or the desired conversion may be determined in that stored values are resorted to which are analogous to the actual emission values determined by the control unit, but were determined by theoretical considerations and/or by tests with the aid of the reference internal combustion engine in a preceding operation and, linked to the dosing amounts stored in the memory of the control unit, are likewise stored there. Such stored values may, analogously to the adding-up or integration of the value for the actual emission or the actual conversion, likewise be added up or integrated over the pre-settable time t or alternatively until the pre-settable amount of emission, work produced or the sum or the integral of a further operating parameter is reached, so that the value thus obtained is proportional to the desired emission. If the current value of the operating parameter or operating parameters exceeds the class limit, as addressed above as the second possibility, there are likewise again two possible methods of further processing. On one hand, the partial sum already obtained or the partial integral can be discarded, so no, or no new, correction value is formed. On the other hand, there is the possibility of interrupting the adding-up or integration operation if the class limit is exceeded, and to resume it at the interrupted point upon a return to the class. The pre-settable time t or alternatively the pre-settable amount of emission, the pre-settable work produced or a pre-set sum or an integral of a further operating parameter provided for the adding-up or integration operation is thus subdivided into partial times tx or partial amounts of emission, partial work or partial sums or partial integrals of a further operating parameter. Which operating parameters (e.g. oil, exhaust, engine water and external temperature etc.) have to be taken into account in determining the actual emission depends on the respective conditions. In practice, of the theoretically possible parameters listed further above, only a few will be taken into account. If the deviation between the desired and the actual emission or the desired and the actual conversion is determined, a correction value can be determined therefrom and this, linked with the respective operating-parameter class or the operating-parameter classes, can be stored in the memory of the control unit. If there are then correction values present in the control unit, the dosing amount is corrected dependent on these correction values which relate to the operating-parameter classes. This is done by the control unit linking the dosing amount held in the context of the model stored therein, which is fetched from the memory dependent on the current operating point, with the correction value(s). The following general statements can be formulated for the above embodiments: For n operating parameters, hereinafter referred to as influencing parameters E, at least 2n correction values and/or at least one correction function Af with n- parameters have to be determined. The amount of reducing agent mreduang agem actually added at the time f is then yielded from an amount of reducing agent mreduc/ng agent,moaei determined from stored data, in the form of curves, sets of characteristic curves, tables or functions, and at least one correction value K, linked with at least one influencing parameter E. This correction value is dependent on the current value of the influencing parameter Eand the correction value k, determined at the time f and linked with the influencing parameter E, the correction value ku as described above, having been determined in class-related manner. Generally it can be written: (Formula Removed) describing the current time and V a time in the past. In addition to multiplication, of course addition according to (Formula Removed) is also possible. To determine the effect of different values of the influencing parameter, the deviation of desired and actual emissions, or desired and actual conversions for at least two different value ranges for the influencing parameter E and hence at two different times in the past has to be determined. An example of this: first of all, in a preceding operation the operating parameters are classed as described above, the correction values for each class are then determined by integration or addition of the actual values within the classes and the comparison with the desired value. In this case, a certain measuring time t has to be pre-set which is necessary in order to reliably determine a correction value. By way of example, the method of the operating-parameter classes or influencing-parameter classes is illustrated below by the example of the influencing parameter "exhaust mass stream". In the actual value/desired value comparison during the operation, at 100 -1000 kg/h a correction value of 120% at the time t' is determined. at 1001 - 2000 kg/h a correction value of 90% at the time t" is yielded at 2001 - 3000 kg/h a correction value of 130% at the time t'" is yielded. The correction values thus determined are plotted in the form of a characteristic curve and are associated with the exhaust mass stream classes 100 -1000 kg/h, 1000 -2000 kg/h and 2000 - 3000 kg/h: Exhaust mass stream classes [kg/h] 100-1000 1001 -2000 2001 -3000 Correction value [%] 120 90 130 The correction value is thus linked to a class of an influencing variable and no longer to an individual value of an influencing variable. During non-stationary engine operation, the correction value determined can then apply either to the entire breadth of the class, or in each case only to one value of the class, for example lower limit, middle or upper limit, the current correction value then being determined, advantageously by linear interpolation, from the characteristic curve and being used for adapting the controlled dosing amount. In the first case, a constant correction value of 90%, by which the dosing amount determined from the models would be corrected by multiplication with the correction value, would be yielded for a mass stream of between 1001 and 2000 kg/h; from 2001 kg/h upwards a correction would be made with 130%. In the other case, a correction value of 102% would be yielded for an exhaust mass stream of 1800 kg/h with linear interpolation, on the assumption that the correction value is related in each case to the middle of the two closest correction classes. The correction values can, as described above, be determined and stored in the form of relative values or in the form of absolute values, such as changed amounts of reducing agent. The following example is intended to make the difference clear, with the deviations being adopted directly as correction values for simplicity both for the relative and for the absolute correction values. In reality, it is however advisable to limit the maximum values and/or the change in the correction values permitted per check, in order to prevent oscillation of the system. As has already been described further above, the relative correction values are to be determined for the influencing variable exhaust mass stream, the NOx conversion being used as assessment criterion. Exhaust mass stream classes [kg/h] Desired NOx conversion [%] Actual NOx conversion [%] Correction value [%] If absolute values are used, it is advisable when using NOx sensors to use the NOx concentrations downstream of the SCR system as assessment criterion. Exhaust mass stream classes [kg/h] Desired NOx concentration [ppm] Actual NOx concentration [ppm] ■ Correction value [ppm] When determining the amount of reducing agent mreduangagenl(t) at a later time, when using absolute correction values kf care should be taken that the correction values must reproduce reducing-agent dosing amounts. That is to say that in the above example, the concentration value determined from the correction values must be converted with the aid of the current amount of exhaust into a current reducing-agent dosing amount supplement. This can be avoided if it is not exhaust concentrations but already amounts of reducing agent which are stored as correction values. It is possible to proceed analogously to the above examples for determining absolute and relative correction values in determining the correction values for the other influencing parameters such as the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the exhaust mass stream and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or the amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the untreated NOx emission and/or the operating hours and/or the air humidity and/or the atmospheric pressure. If n influencing parameters E1 to En are yielded, then for n correction values the dosing amount actually added can be determined for example by multiplication (Formula Removed) or addition (Formula Removed) of the correction values. The addition of the correction values is mostly used when the correction values are absolute values, and multiplication when the correction values contain relative values. It is also conceivable to draw up a multi-parameter correction function K, in which the influencing parameters E are contained: The individual correction values for the influencing parameters may have been, but do not have to have been, determined at different times t', t", t'"etc. The correction values are frozen, i.e. stored unchanged and used further to correct the stored values and hence the control, until again a return to the respective operating-parameter class is achieved and/or a new integration or adding-up operation is concluded, which means that a renewed check of the emissions can be carried out and new correction values determined. In the preceding examples, the deviation of the desired value from the actual value was adopted one to one as correction value. This is not always successful. In the case of large deviations, large correction values result therefrom, which may lead to oscillation of the system. It is therefore useful to limit the change in the correction values per checking step. This can be achieved by pre-setting a maximum permitted change in correction values per checking step and/or by pre-setting a minimum and/or maximum correction value. One further possibility consists in determining the correction value by multiplying the deviation with a value of between zero and one. In determining several correction values for several operating or influencing parameters, care should be taken that the different influencing parameters have a different degree of influence on the actual emissions or actual conversions. It is therefore useful to realise the influence of the individual influencing-parameter-dependent factors which result in correction of the dosing amount by weighting factors wh wz, w3, ...wn for the individual correction values. This results for example in the following approach: Above all, in determining the dosing amount by means of multiplication and/or drawing up a multi-parameter correction function, it is also conceivable to influence the different influencing parameters already during determination of the individual correction values, so that smaller correction values result for parameters with lesser influence than for parameters with great influence. In addition, it is also conceivable to introduce different maximum permitted changes in correction value per checking step and/or different minimum and/or maximum correction values for different influencing parameters. One further possibility is to determine the different correction values by multiplication with different weighting factors which lie between zero and one and thereby represent the influence of the respective influencing parameter. The weighting factors in the simplest case may be defined as constants. One further possibility is to determine the weighting factors with the aid of a function and/or a set of characteristic curves and/or a characteristic curve which is/are additionally dependent on the size of at least one operating or influencing parameter and/or the deviation between desired and actual emissions/conversions. Since the influence of an influencing parameter can change during the course of operation, e.g. due to ageing of the catalysts, it is additionally conceivable to adapt the weighting factors over the running time. This may be done e.g. by making the weighting factor dependent on the number of changes to the associated correction value and/or the size of the change in the correction value. In addition, it is possible to effect the weighting of the correction values and/or the determination of the correction values dependent on at least one influencing parameter with the aid of at least one neural network. Claims: 1. A method for use in conjunction with an exhaust aftertreatment system for dosing a reducing agent, which splits off ammonia, into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, wherein the dosing of the reducing agent takes place by means of a dosing means controlled by a control unit and associated with the exhaust aftertreatment system into the exhaust stream, at least one SCR catalyst is arranged downstream from the dosing means in the exhaust stream, as a further part of the exhaust aftertreatment system, the dosing amount is determined by the control unit by means of a model stored therein which covers all possible operating points of the internal combustion engine and/or of the exhaust aftertreatment system, dependent on at least one operating parameter of the internal combustion engine and/or of the exhaust aftertreatment system evaluated by the control unit, the respective momentary value of the at least one operating parameter determining the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system, during the operation of the internal combustion engine, the control unit determines a deviation by comparison of a variable, proportional to the desired emission or the desired conversion, determined for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system by means of the control unit of the internal combustion engine from stored values, with a variable detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, and determines a correction value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this correction value, the model thus modified is used by the control unit for dosing until the control unit, by comparison of a variable stored for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system in the control unit of the internal combustion engine which is proportional to the desired emission or the desired conversion with a variable again detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, determines a deviation and determines a new correction value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this new correction value, the determination of the actual emission or the actual conversion takes place such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor and/or of an HNCO sensor and/or of a lambda sensor arranged downstream of the SCR catalyst or the actual conversion determined before the SCR catalyst with the aid of at least one of these sensors and the untreated emission, for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integrating at least one operating parameter, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges being determined by variables stored in the control unit, the control unit, when it establishes that the current value range has been left during the adding-up or integration, discards the added-up or integrated sum, or the control unit, when it establishes that the current value range has been left during the adding-up or integration, temporarily stores the added-up or integrated sum and then when the control unit establishes a return to the value range previously left, continues with the adding-up or integration until the pre-settable time t for adding up or integration has passed or until a pre-settable amount of emission or, by adding-up or integrating at least one operating parameter, a pre-settable value is reached, the control unit uses the sum added up or integrated over the pre-settable time t or the pre-settable amount of emission or pre-settable work, or a variable proportional thereto as actual emission or actual conversion for the comparison with the desired emission or the desired conversion determined from stored values, and determines a correction value for the dosing amount, the control unit, with the correction value determined for the respective value range of the at least one operating parameter, modifies the dosing amount from the model if the value currently determined by the control unit of the at least one operating parameter lies in this value range, or the control unit, dependent on the current value, determined by the control unit, of the at least one operating parameter, from correction values which had been determined for values of the at least one operating parameter directly neighbouring to this current value, determines a correction value by interpolation, and with this correction value modifies the dosing amount from the model. 2. A method according to Claim 1, characterised in that timewise in parallel with the determination of the actual emission or the actual conversion the control unit determines the desired emission or the desired conversion in that for the operating points of the internal combustion engine and/or of the exhaust aftertreatment system which are passed through the ideal emission values stored in each case in the control unit for the operating points, or ideal conversions in the pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integration of at least one operating parameter, are added up or integrated by the control unit, and the control unit uses the sum thus obtained or a variable proportional thereto as desired emission or desired conversion for the comparison of the actual emission or the actual conversion with the desired emission or the desired conversion. 3. A method according to one of Claims 1-3, characterised in that the dosing amount determined from the models is linked by the control unit to the correction value by multiplication or addition. 4. A method according to one of Claims 1-4, characterised in that the at least one operating parameter of the internal combustion engine and/or of the exhaust aftertreatment system which is evaluated by the control unit is the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the emissions and/or the fuel/air ratio and/or the change in these variables over time. 5. A method according to one of the preceding claims, characterised in that the control unit determines and stores different correction values and/or different parameters of at least one correction function and/or a plurality of correction functions for different operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system, and in that the control unit, dependent on the current operating point of the internal combustion engine and/or of the exhaust aftertreatment system, determines an operating-point-related correction value from these correction values and/or correction functions. 6. A method according to one of the preceding claims, characterised in that the pre-settable time t is at least 15 seconds. 7. A method according to one of the preceding claims, characterised in that the pre-settable amount of emission is determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows. 8. A method according to one of the preceding claims, characterised in that the pre-settable amount of emission is stored in units of mass and that the reaching of the pre-settable amount of emission is brought about by adding-up or integration of emission masses and/or emission mass flows, the pre-settable amount of emission for NOx being at least 1 mg and/or the pre-settable amount of emission for NH3 being at least 0.01 mg and/or the pre-settable amount of emission for N20 being at least 0.02 mg and/or the pre-settable amount of emission for HNCO being at least 0.01 mg. 9. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values are identical to the operating parameters, the value ranges of which are checked or which are used for determining the correction value. 10. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values differ from the operating parameters, the value ranges of which are checked or which are used for determining the correction value. 11. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values are the amount of exhaust and/or the amount of fuel and/or the amount of reducing agent and/or the work produced by the internal combustion engine.

Full Text

Method of setting the dosing of the reducing agent in selective catalytic reduction
The invention is directed at a method according to the preamble of Claim 1.
In addition to solids particles, nitrogen oxides are one of the limited exhaust constituents which are produced during combustion processes and the permitted emissions of which are being reduced further and further. Various methods are nowadays used to minimise these exhaust constituents in internal combustion engines operated in motor vehicles. The reduction of the nitrogen oxides takes place mostly with the aid of catalysts. In oxygen-rich exhaust gas a reducing agent is additionally necessary in order to increase the selectivity and the NOx conversions.
These methods have become known under the collective term "SCR methods", SCR standing for "selective catalytic reduction". They have been used for many years in the power station industry and recently also in internal combustion engines. A detailed account of such methods can be gathered from DE 34 28 232 A1. Mixed oxides containing V205, for example in the form V2O5AA/O3/T1O2, can be used as SCR catalysts. Typical V205 contents are in that case between 0.2 and 3%. The use of iron-containing and/or copper-containing zeolites is also conceivable.
In practical application, ammonia or compounds which split off ammonia, such as urea or ammonium formate, in solid or solution form, are used as reducing agents.
Thus the urea decomposes at high temperatures to form isocyanic acid and ammonia.
(Formula Removed)
The isocyanic acid hydrolyses further with water contained in the exhaust to form NH3 and C02.
(Formula Removed)
With complete hydrolysis of one mole of urea, thus two moles of ammonia and one mole of carbon dioxide are produced.
(Formula Removed)
Owing to the hydrolysis of urea, this makes available the same proven reducing agent as in the power station industry, namely ammonia.
One mole of ammonia is in that case necessary for reacting one mole of nitrogen monoxide.
(Formula Removed)
With an ideal catalyst this means that, for a feed ratio of one, all the nitrogen oxides are reduced, i.e. a 100% NOx conversion is achieved, since XNOx applies for the NOx conversion:
(Formula Removed)
If the SCR catalysts are preceded by a platinum-containing NO oxidation catalyst for the formation of N02,
(Formula Removed)
the SCR reaction can be considerably accelerated and the low-temperature activity can be noticeably increased.
(Formula Removed)
However, in the presence of N02 increased nitrous oxide emissions in accordance with the following reaction also have to be expected:
(Formula Removed)
With internal combustion engines operated in vehicles, the reduction of nitrogen oxides with the aid of the SCR method is difficult because changing operating conditions, such as fluctuating exhaust temperatures, amounts of exhaust and untreated NOx emissions, prevail therein, which makes it more difficult to admeter the quantities of reducing agent. Although on one hand the highest possible conversion of nitrogen oxides should be achieved, on the other hand it should be ensured that there is no emission of nitrous oxide, isocyanic acid or unspent ammonia.
Currently, two methods for determining the correct dosing amount of reducing agent are used when admetering the reducing agent for the SCR method in vehicles.
Firstly, this is a pure control without check-back of sensors for determining the actual emissions after the catalyst system. The dosing amount in this case is determined with the aid of models, from data which are stored and/or are determined in a memory of an electronic engine control unit of an internal combustion engine in the form of tables or curves, sets of characteristic curves or functions, and optionally with the aid of sensors for determining the catalyst temperature, the amount of NOx and amount of exhaust. Thus for example the untreated emission of the engine is calculated from the injection quantity, the engine speed, the injection pressure, the fuel/air ratio etc. The possible NOx conversions and the dosing amounts of reducing agent necessary therefor in turn depend on the catalyst temperature, the untreated NOx emission, the amount of exhaust etc. The actual emissions after the system are not detected and thus have no influence on the dosing amount (DE 43 15 278 A1, DE 195 36 571 A1, DE 199 06 344 A1, EP 898 061 A1).
The disadvantage of this method is that faults, defects or environmental influences can scarcely be compensated for owing to the lack of check-back about the actual emissions.
The second possibility of admetering the reducing agent is to construct a conventional closed control loop with the aid of HNCO, N20, NOx and/or NH3 sensors after the system. For this, the actual values currently supplied by the sensors are compared with the desired values and the dosing amount is constantly adapted. However, there is the
problem of permanent regulation in the inertia of the system and the sensors and also of the simultaneously highly dynamic operation of the internal combustion engine in vehicles. Thus for example during acceleration operations or additional loads on engines on exhaust-charged internal combustion engines the NOx emissions may • increase by a factor of 10 within one second. In the case of aspirated engines the increase is even faster owing to the lack of inertia of the exhaust turbocharger. The same applies in the event of loss of load or upon transition into the overrun condition.
The sensors for determining the emissions are not capable of detecting these highly-dynamic operations. This is due firstly to the inertia of the sensors, the typical t90 time of which, i.e. the time at which 90% of the final value is reached, should be settled at 300-500 ms, and secondly to the necessary positioning of the sensors after the catalyst system. Thus merely the gas travel time from emerging from the cylinder head to emerging from the catalyst system, depending on exhaust-gas flow rate and volume of the exhaust system, is 200-2000 ms.
In order to remedy this problem, a generic exhaust aftertreatment system and a method for controlling this exhaust aftertreatment system, in particular for an internal combustion engine, which partially solves this problem, is proposed in DE 101 00 420 A1. The exhaust aftertreatment system, which comprises at least one catalyst, is supplied with a predetermined amount of reducing agent dependent on the operating state of the internal combustion engine and/or of the exhaust aftertreatment system. The amount of reducing agent supplied is adapted. For this, in stationary operation of the internal combustion engine, the emission of nitrogen oxides and/or of ammonia is measured by means of sensors and is compared with desired values stored for this stationary operating state. If a deviation is detected, the control of the exhaust aftertreatment system determines a correction value with which the amount of reducing agent supplied is adapted.
What is disadvantageous with this system is that the adaptation, owing to the inertia of the system described above, can only take place in a relatively long stationary operating phase at predetermined operating parameters of the internal combustion engine. "Stationary operating conditions" are understood in this case to mean that the operating variables which determine the admetering of reducing agent may not, or may only minimally, change. Such stationary operating phases do not occur over relatively long periods in certain operating modes of the internal combustion engine, e.g. if it is
operated in a vehicle and the vehicle is moving in urban traffic. This means that a correction of the amount of reducing agent cannot take place over relatively long periods, and therefore an increased pollution emission of nitrogen oxides and/or ammonia occurs.
A further approach for countering the problems described can be inferred from DE 195 36 571 A1. Therein, a method and also an associated device for dosing the amount of a reducing agent introduced into the exhaust stream or exhaust-air stream of combustion systems, in particular combustion engines, with a subsequent catalyst is described. The amount of reducing agent introduced is set, departing from operation-relevant parameters of the combustion system, the exhaust and the catalyst, via (sets of) characteristic curves, the position of the (sets of) characteristic curves being checked and adapted to the current state and also the current operating conditions of the combustion engine, the exhaust and the catalyst. Therefore the characteristic curves or sets of characteristic curves is/are adapted by comparing the actual pollutant concentration determined by means of sensors with stored desired values.
With this procedure too, the great inertia of the system, in particular of the sensors, means that checking is only possible when stationary operating conditions of the internal combustion engine exist, so the disadvantages described above for DE 101 00 420 A1 apply here too.
It is an object of the invention to avoid the disadvantages of the known methods. This object is achieved by the method according to Claim 1; advantageous embodiments of the method are characterised in the dependent claims.
The method according to the invention is applied in conjunction with an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air. As is conventional with such arrangements, the dosing of the reducing agent takes place with the aid of a dosing means associated with the exhaust aftertreatment system and controlled by a control unit. For the reduction of nitrogen oxides, at least one SCR catalyst is arranged downstream of the dosing means in the exhaust stream, as a further part of the exhaust aftertreatment system.
The dosing amount of the reducing agent - in systems which are nowadays conventional this is usually aqueous urea solution, but alternatively other reducing agents (ammonia or compounds which split off ammonia, such as urea or ammonium formate in solid or solution form) are conceivable - is determined by the control unit with the aid of a model stored therein which covers all possible operating points of the internal combustion engine or of the exhaust aftertreatment system.
A model in this case is to be understood to mean, in the simplest case, a characteristic curve or a set of characteristic curves, but it may also be a plurality of characteristic curves, sets of characteristic curves or alternatively one-parameter or multi-parameter functions which are determined or established with the aid of what is called a reference arrangement and/or by theoretical considerations. The reference arrangement in the present case is an arrangement, typical of a model series, of internal combustion engine and exhaust aftertreatment system which may already be installed in a vehicle. By measurements on the reference arrangement and/or by theoretical considerations, for a plurality of operating points of the arrangement on one hand a dosing amount of the reducing agent is determined and on the other hand a desired emission is determined for these operating points. Each operating point in this case is defined by the size of at least one operating parameter of the reference arrangement. The dosing amounts determined and the associated desired emissions or desired conversions are held in the form of a model in the control units of the corresponding model series, such that a variable which is proportional to the dosing amount and a variable which is proportional to the desired emission or desired conversions can be obtained directly or by interpolation by means of the control unit from the model, for all possible values which the at least one operating parameter may adopt, i.e. for all the operating points which occur.
From this model, i.e. the characteristic curves, sets of characteristic curves or functions, the control unit determines the dosing amount, dependent on the at least one operating parameter, evaluated by the control unit, of the internal combustion engine and/or of the exhaust aftertreatment systems. The respective momentary value of the at least one operating parameter in this case determines the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system. In practice, it is frequently a question not of only one, but of a plurality of operating parameters which influence the correct dosing amount of the reducing agent; consequently this multiple dependency has to be taken into account in the model.
Since this multiple dependency is however not constant, but is subject in particular to a change over time, i.e. one determined by the number of operating hours of a system, and also to a change due to environmental influences, the dosing amount determined by means of the model has to undergo a correction. For this, the procedure is advantageously such that, during the operation of the internal combustion engine, the control unit determines a deviation by comparison of a variable determined from stored values for the internal combustion engine and/or of the exhaust aftertreatment system by means of the control unit of the internal combustion engine, which variable is proportional to the desired emission or the desired conversions, with a variable detected by measurement by the control unit, which variable is proportional to the actual emission or the actual conversions, and determines a correction value for the dosing amount dependent on this deviation. For subsequent dosing operations, the stored model is then modified with this correction value by the control unit.
The modification of the model with the correction value is in each case retained until the control unit, by comparing a variable which is proportional to the desired emission or the desired conversions and is determined from stored values for the internal combustion engine and/or the exhaust aftertreatment system by means of the control unit of the interna! combustion engine, with a variable re-detected by measurement by the control unit which is proportional to the actual emission or the actual conversions, determines a deviation and dependent on this deviation determines a new correction value for the dosing amount. For subsequent dosing operations, the stored model is then modified with this new correction value.
For detection of the actual emission value or the actual conversions of an internal combustion engine with subsequent exhaust aftertreatment system, on one hand a certain time is necessary over which the signal of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor is added up or integrated; additionally the operating conditions may in so doing not change if at all possible. This restricted the determination of actual emission values or actual conversions per se to stationary operating conditions. In order nevertheless advantageously to obtain independence from stationary operating conditions for the determination of the correction value, in the determination of the actual emission or the actual conversions according to the method of the invention the procedure is such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor
arranged downstream of the SCR catalyst, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges being determined by variables stored in the control unit. The addition or integration may take place for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integration of at least one operating parameter. This operating parameter may for example be the amount of exhaust and/or the amount of fuel and/or the amount of reducing agent and/or the work performed by the internal combustion engine. Furthermore, the operating parameter intended for monitoring the value ranges can also be used for determining the duration of the addition or integration. Of course, determination of the untreated emissions is additionally necessary for determining the desired and actual conversions.
The reaching of the amount of emission can in this case be determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows.
Due to the above-mentioned allocation of value ranges within which the at least one operating parameter may vary during the adding-up or integration, i.e. by class formation, the frequency at which actual emission values or actual conversions and hence correction values for dosing the reducing agent can be obtained is drastically increased. In this case, by selecting the value ranges within which the at least one operating parameter may vary, the error produced by this class formation can be kept within a negligible order of magnitude.
If the at least one operating parameter leaves the current value range, i.e. the class, during the adding-up or integration, two alternative process configurations are yielded. On one hand, it is possible to proceed such that the control unit, upon establishing that the current value range is being left, discards the added-up or integrated sum. On the other hand, upon establishing that the current value range is being left, the added-up or integrated sum can be temporarily stored by the control unit in order, when the control unit establishes a return to the value range previously left, to continue with the adding-up or integration until a pre-settable amount of emission, work produced or sum or integral of a further operating parameter is reached or the pre-settable time t for adding-up or integration has passed.
Both alternatives have advantages and disadvantages. In the first case, it is ensured that the adding-up or integration takes place in one go, i.e. there are only seconds between the start and end of the operation. In such short periods, temporal influences or environmental influences do not become apparent, and therefore do not have an effect on the measurement result for the actual emission. On the other hand, however, the frequency at which new actual emission measured values are available is greatly reduced. In the second case, at least theoretically, days or even weeks can pass before a determination of actual emissions is concluded, which means that temporal influences or environmental influences might very well have an effect and falsify the measurement result. This can be countered by time-limiting the validity of temporarily stored values. What is advantageous in the process configuration with temporary storage is that the frequency at which actual emission measured values and hence current correction values are available is greatly increased.
In the further course of the process, the control unit uses the sum added-up or integrated over the pre-settable time t or until the pre-settable amount of emission, work produced or sum or integral of a further operating parameter is reached, or a variable which is proportional thereto as actual emission for the comparison with the stored desired emission, and determines a correction value for the dosing amount with the established difference. The current correction values in each case which are valid for a value range of the at least one operating parameter are held in the control unit, so that the control unit can resort thereto, in order to modify the dosing amount from the model with the correction value determined for the respective value range of the at least one operating parameter, if the value of the at least one operating parameter currently determined by the control unit lies in the corresponding value range.
The model for determining the dosing amount can thus be adapted efficiently and hence advantageously to changes over time and environment-determined changes.
Alternatively, there is the possibility that the control unit, dependent on the current value, determined by the control unit, of the at least one operating parameter, from correction values which were determined for values of the at least one operating parameter directly neighbouring to this current value, determines an intermediate value as correction value by interpolation. In this manner, on one hand the error due to the class formation can be made advantageously smaller, because intermediate values to the stored correction values can be formed for each operating point, and on the other hand value ranges for
which no correction value has yet been determined can be advantageously bridged by interpolation.
One particularly simple and hence advantageous method of determining the desired emission or the desired conversion from the model consists in that the control unit in the pre-settable time t or until a pre-settable amount of emission or work produced or sum or integral of a further operating parameter is reached, i.e. timewise in parallel with the determination of the actual emission or the actual conversion, adds up or integrates the values for the desired emission or the desired conversion obtained from the model dependent on operating point and the sum thus obtained or a variable which is proportional thereto is used as desired emission or desired conversion for the comparison of the actual emission or the actual conversion with the desired emission or the desired conversion.
The significant advantage with this procedure is that, with the exception of the operating parameter for which the correction value is to be determined and which may therefore vary only within the associated value range, i.e. the class, all other operating parameters may assume any desired values, because the influences yielded thereby are eliminated by the adding-up or integration of the actual emissions and the desired emissions. Or, in other words, the same operating points are passed through both in determining the actual emission or the actual conversion and in determining the desired emission or the desired conversion, so that the influences thereby produced are eliminated upon comparison of both values and as a result only the deviation relative to the operating parameter for which the correction value is to be determined is left over.
In order to modify the dosing amount determined from the models with the correction value, the control unit forms the sum of dosing amount and correction value if the correction value is stored in the form of a positive or negative corrective dosing amount, or the product of dosing amount and correction value if the correction value is stored as a factor.
With regard to the operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system which are evaluated by the control unit, the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed
and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the emissions and/or the fuel/air ratio and/or the change in these variables over time are considered.
This means for the dosing of the reducing agent that, in the simplest case, i.e. when only one operating parameter, e.g. the exhaust mass stream, is evaluated by the control unit, the model held in the control unit would contain only one characteristic curve. If several operating parameters are evaluated by the control unit, these are one or more sets of characteristic curves or one or more multi-parameter function(s).
Owing to the numerous operating parameters which may have an influence on the dosing amount for the reducing agent, it is useful for the control unit to determine and store different correction values and/or different parameters of at least one correction function and/or a plurality of correction functions for different operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system, and for the control unit, dependent on the current operating point of the internal combustion engine and/or of the exhaust aftertreatment system, to determine an operating-point-related correction value from these correction values and/or correction functions.
Owing to the multiple dependency addressed above, the importance of the allocation of the operating parameters to classes becomes particularly clear, because only in this way can multiply-dependent correction values be determined at all within an acceptable period. The class formation causes operating-parameter ranges within defined limits in each case to be viewed as constant, so that as a consequence an operating-point range is yielded which can be treated as quasi-stationary operation. The probability that the real operating point at which the internal combustion engine and/or the exhaust aftertreatment system currently is will lie for long enough in such an operating-point range, which is regarded as quasi-stationary, to determine the correction value increases drastically due to the aforementioned class formation.
HNCO sensors and/or N20 sensors and/or NOx sensors and/or NH3 sensors and/or lambda sensors can be used as sensors for determining the actual emissions, the untreated emissions and the actual conversions. These sensors are commercially available, and their construction therefore does not require any further explanation.
The invention will be explained in greater detail below with reference to examples of embodiment.
For the following embodiments, the starting point is an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air.
As has already been indicated at the beginning, it is not sufficient for such systems to dose the reducing agent in controlled manner dependent on operating point, rather it is necessary, dependent on changes over time of the exhaust aftertreatment system or of the internal combustion engine, and also dependent on environmental influences, to correct the dosing amount. For this, operating-point-dependent correction values are determined.
The method of operation to determine a correction value with respect to the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system, will be indicated below by way of example. In this case it is a prerequisite that the described sequences are carried out as programmed control sequences in a control unit, e.g. a programmable, electronic engine control unit, such as are used in modern internal combustion engines. The control unit is connected via a plurality of sensors to the internal combustion engine and the exhaust aftertreatment system, and determines via these sensors all current values of relevant operating parameters of the internal combustion engine or of the exhaust aftertreatment system, and also the current actual emission or the current actual conversion of the exhaust aftertreatment system. The current values of the relevant operating parameters in this case define the current operating point of the internal combustion engine or of the exhaust aftertreatment system.
First of all, continuous monitoring of the relevant operating parameters takes place during the operation of the internal combustion engine. This on one hand serves the purpose of gathering, dependent on operating point, from the model stored in the control unit, the dosing amount of the reducing agent, on the other hand the monitoring of the operating parameters serves to determine, dependent on operating-parameter values, correction values for the amount of reducing agent which is held, dependent on
operating point, as a stored value. This determination of the correction values will be gone into in detail below.
As already stated, the determination of the correction values takes place in class-related manner. This is to be understood to mean that the values which the relevant operating parameters can adopt are divided up into value ranges or classes. At least two value ranges or classes are defined for each relevant operating parameter. How many classes exist for an operating parameter, or of what nature the classes are, i.e. what initial value and what final value they have, depends on in which value ranges of the corresponding operating parameter a negligible change in the correction value is yielded. The corresponding value ranges were determined with the aid of a reference system. The reference system, as already explained, is an exhaust aftertreatment system for dosing a reducing agent which splits off ammonia into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, which corresponds to the series specification and on which the corresponding measurements are carried out in tests. The value ranges or classes for each of the relevant operating parameters thus determined are stored in the control units of the series, so that they can be resorted to.
During operation of a system in accordance with the series specification, and on the assumption that initially no correction values are stored in the control unit, e.g. upon the first starting-up of the internal combustion engine, the dosing takes place using the dosing amounts held in the model, i.e. uncorrected. At the same time, the control unit begins to interrogate the current values of the relevant operating parameters cyclically and to compare them with the value ranges or classes stored in the control unit. Dependent on the result of this comparison, two possibilities are yielded:
The current value of the operating parameter remains within the limits of one
class;
The current value of the operating parameter exceeds the class limit. Both the aforementioned cases will be gone into in greater detail below.
A further operation takes place in parallel with the aforementioned operation. In this case, with respect to the current class or classes in which the operating parameter(s) are located, the value supplied by an exhaust sensor or the conversion determined with the aid of this sensor is added up or integrated by the control unit. For the adding-up or integration a time t, an amount of emission, a work produced by the internal combustion
engine or the sum or the integral of a further operating parameter can be pre-set in the control unit, after the expiry of which, under the prerequisite that until the expiry thereof, the class or classes has/have not been left, the control unit terminates the adding or integration operation and uses the value thus obtained as equivalent to the actual emission or to the actual conversion of the system for an actual value/desired value comparison.
The pre-settable time t should in that case be at least 15 seconds. In the event that a pre-settable amount of emission serves as interruption criterion instead of the time t, the amount of emission, as already mentioned, can be determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows.
If the pre-settable amount of emission is given in units of mass by adding-up or integration of emission masses and/or emission mass flows, the pre-settable amount of emission for NOx should be at least 1 mg and/or the pre-settable amount of emission for NH3 at least 0.01 mg and/or the pre-settable amount of emission for N20 at least 0.02 mg and/or the pre-settable amount of emission for HNCO at least 0.01 mg.
The value of the desired emission or the desired conversion may be determined in that stored values are resorted to which are analogous to the actual emission values determined by the control unit, but were determined by theoretical considerations and/or by tests with the aid of the reference internal combustion engine in a preceding operation and, linked to the dosing amounts stored in the memory of the control unit, are likewise stored there. Such stored values may, analogously to the adding-up or integration of the value for the actual emission or the actual conversion, likewise be added up or integrated over the pre-settable time t or alternatively until the pre-settable amount of emission, work produced or the sum or the integral of a further operating parameter is reached, so that the value thus obtained is proportional to the desired emission.
If the current value of the operating parameter or operating parameters exceeds the class limit, as addressed above as the second possibility, there are likewise again two possible methods of further processing. On one hand, the partial sum already obtained or the partial integral can be discarded, so no, or no new, correction value is formed. On the other hand, there is the possibility of interrupting the adding-up or integration
operation if the class limit is exceeded, and to resume it at the interrupted point upon a return to the class. The pre-settable time t or alternatively the pre-settable amount of emission, the pre-settable work produced or a pre-set sum or an integral of a further operating parameter provided for the adding-up or integration operation is thus subdivided into partial times tx or partial amounts of emission, partial work or partial sums or partial integrals of a further operating parameter.
Which operating parameters (e.g. oil, exhaust, engine water and external temperature etc.) have to be taken into account in determining the actual emission depends on the respective conditions. In practice, of the theoretically possible parameters listed further above, only a few will be taken into account.
If the deviation between the desired and the actual emission or the desired and the actual conversion is determined, a correction value can be determined therefrom and this, linked with the respective operating-parameter class or the operating-parameter classes, can be stored in the memory of the control unit.
If there are then correction values present in the control unit, the dosing amount is corrected dependent on these correction values which relate to the operating-parameter classes. This is done by the control unit linking the dosing amount held in the context of the model stored therein, which is fetched from the memory dependent on the current operating point, with the correction value(s).
The following general statements can be formulated for the above embodiments:
For n operating parameters, hereinafter referred to as influencing parameters E, at least 2n correction values and/or at least one correction function Af with n- parameters have to be determined.
The amount of reducing agent mreduang agem actually added at the time f is then yielded from an amount of reducing agent mreduc/ng agent,moaei determined from stored data, in the form of curves, sets of characteristic curves, tables or functions, and at least one correction value K, linked with at least one influencing parameter E. This correction value is dependent on the current value of the influencing parameter Eand the correction value k, determined at the time f and linked with the influencing parameter E,
the correction value ku as described above, having been determined in class-related manner.
Generally it can be written:
(Formula Removed)
describing the current time and V a time in the past.
In addition to multiplication, of course addition according to
(Formula Removed)
is also possible.
To determine the effect of different values of the influencing parameter, the deviation of desired and actual emissions, or desired and actual conversions for at least two different value ranges for the influencing parameter E and hence at two different times in the past has to be determined.
An example of this: first of all, in a preceding operation the operating parameters are classed as described above, the correction values for each class are then determined by integration or addition of the actual values within the classes and the comparison with the desired value. In this case, a certain measuring time t has to be pre-set which is necessary in order to reliably determine a correction value.
By way of example, the method of the operating-parameter classes or influencing-parameter classes is illustrated below by the example of the influencing parameter "exhaust mass stream".
In the actual value/desired value comparison during the operation,
at 100 -1000 kg/h a correction value of 120% at the time t' is determined.
at 1001 - 2000 kg/h a correction value of 90% at the time t" is yielded
at 2001 - 3000 kg/h a correction value of 130% at the time t'" is yielded.
The correction values thus determined are plotted in the form of a characteristic curve and are associated with the exhaust mass stream classes 100 -1000 kg/h, 1000 -2000 kg/h and 2000 - 3000 kg/h:
Exhaust mass stream classes [kg/h] 100-1000 1001 -2000 2001 -3000
Correction value [%] 120 90 130
The correction value is thus linked to a class of an influencing variable and no longer to an individual value of an influencing variable.
During non-stationary engine operation, the correction value determined can then apply either to the entire breadth of the class, or in each case only to one value of the class, for example lower limit, middle or upper limit, the current correction value then being determined, advantageously by linear interpolation, from the characteristic curve and being used for adapting the controlled dosing amount. In the first case, a constant correction value of 90%, by which the dosing amount determined from the models would be corrected by multiplication with the correction value, would be yielded for a mass stream of between 1001 and 2000 kg/h; from 2001 kg/h upwards a correction would be made with 130%. In the other case, a correction value of 102% would be yielded for an exhaust mass stream of 1800 kg/h with linear interpolation, on the assumption that the correction value is related in each case to the middle of the two closest correction classes.
The correction values can, as described above, be determined and stored in the form of relative values or in the form of absolute values, such as changed amounts of reducing agent. The following example is intended to make the difference clear, with the deviations being adopted directly as correction values for simplicity both for the relative and for the absolute correction values. In reality, it is however advisable to limit the maximum values and/or the change in the correction values permitted per check, in order to prevent oscillation of the system.
As has already been described further above, the relative correction values are to be determined for the influencing variable exhaust mass stream, the NOx conversion being used as assessment criterion.

Exhaust mass stream classes [kg/h] Desired NOx conversion [%] Actual NOx conversion [%] Correction value [%]

If absolute values are used, it is advisable when using NOx sensors to use the NOx concentrations downstream of the SCR system as assessment criterion.

Exhaust mass stream classes [kg/h] Desired NOx concentration [ppm] Actual NOx concentration [ppm] ■ Correction value [ppm]

When determining the amount of reducing agent mreduangagenl(t) at a later time, when using absolute correction values kf care should be taken that the correction values must reproduce reducing-agent dosing amounts. That is to say that in the above example, the concentration value determined from the correction values must be converted with the aid of the current amount of exhaust into a current reducing-agent dosing amount supplement. This can be avoided if it is not exhaust concentrations but already amounts of reducing agent which are stored as correction values.
It is possible to proceed analogously to the above examples for determining absolute and relative correction values in determining the correction values for the other influencing parameters such as the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the exhaust mass stream and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or the amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the untreated NOx emission and/or the operating hours and/or the air humidity and/or the atmospheric pressure.
If n influencing parameters E1 to En are yielded, then for n correction values the dosing amount actually added can be determined for example by multiplication
(Formula Removed)
or addition
(Formula Removed)
of the correction values.
The addition of the correction values is mostly used when the correction values are absolute values, and multiplication when the correction values contain relative values.
It is also conceivable to draw up a multi-parameter correction function K, in which the influencing parameters E are contained:
The individual correction values for the influencing parameters may have been, but do not have to have been, determined at different times t', t", t'"etc.
The correction values are frozen, i.e. stored unchanged and used further to correct the stored values and hence the control, until again a return to the respective operating-parameter class is achieved and/or a new integration or adding-up operation is concluded, which means that a renewed check of the emissions can be carried out and new correction values determined.
In the preceding examples, the deviation of the desired value from the actual value was adopted one to one as correction value. This is not always successful. In the case of large deviations, large correction values result therefrom, which may lead to oscillation of the system. It is therefore useful to limit the change in the correction values per checking step. This can be achieved by pre-setting a maximum permitted change in correction values per checking step and/or by pre-setting a minimum and/or maximum correction value. One further possibility consists in determining the correction value by multiplying the deviation with a value of between zero and one.
In determining several correction values for several operating or influencing parameters, care should be taken that the different influencing parameters have a different degree of influence on the actual emissions or actual conversions. It is therefore useful to realise the influence of the individual influencing-parameter-dependent factors which result in correction of the dosing amount by weighting factors wh wz, w3, ...wn for the individual correction values. This results for example in the following approach:
Above all, in determining the dosing amount by means of multiplication and/or drawing up a multi-parameter correction function, it is also conceivable to influence the different influencing parameters already during determination of the individual correction values, so that smaller correction values result for parameters with lesser influence than for parameters with great influence. In addition, it is also conceivable to introduce different maximum permitted changes in correction value per checking step and/or different minimum and/or maximum correction values for different influencing parameters. One further possibility is to determine the different correction values by multiplication with different weighting factors which lie between zero and one and thereby represent the influence of the respective influencing parameter. The weighting factors in the simplest case may be defined as constants. One further possibility is to determine the weighting factors with the aid of a function and/or a set of characteristic curves and/or a characteristic curve which is/are additionally dependent on the size of at least one operating or influencing parameter and/or the deviation between desired and actual emissions/conversions. Since the influence of an influencing parameter can change during the course of operation, e.g. due to ageing of the catalysts, it is additionally conceivable to adapt the weighting factors over the running time. This may be done e.g. by making the weighting factor dependent on the number of changes to the associated correction value and/or the size of the change in the correction value. In addition, it is possible to effect the weighting of the correction values and/or the determination of the correction values dependent on at least one influencing parameter with the aid of at least one neural network.

Claims:
1. A method for use in conjunction with an exhaust aftertreatment system for dosing a reducing agent, which splits off ammonia, into the exhaust stream of an internal combustion engine, installed in a vehicle, which is operated with excess air, wherein
the dosing of the reducing agent takes place by means of a dosing means controlled by a control unit and associated with the exhaust aftertreatment system into the exhaust stream,
at least one SCR catalyst is arranged downstream from the dosing means in the exhaust stream, as a further part of the exhaust aftertreatment system, the dosing amount is determined by the control unit by means of a model stored therein which covers all possible operating points of the internal combustion engine and/or of the exhaust aftertreatment system, dependent on at least one operating parameter of the internal combustion engine and/or of the exhaust aftertreatment system evaluated by the control unit, the respective momentary value of the at least one operating parameter determining the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system,
during the operation of the internal combustion engine, the control unit determines a deviation by comparison of a variable, proportional to the desired emission or the desired conversion, determined for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system by means of the control unit of the internal combustion engine from stored values, with a variable detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, and determines a correction value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this correction value,
the model thus modified is used by the control unit for dosing until the control unit, by comparison of a variable stored for the respective operating point of the internal combustion engine and/or of the exhaust aftertreatment system in the control unit of the internal combustion engine which is proportional to the desired emission or the desired conversion with a variable again detected by measurement by the control unit which is proportional to the actual emission or the actual conversion, determines a deviation and determines a new correction
value for the dosing amount dependent on this deviation, and for subsequent dosing operations modifies the stored model with this new correction value, the determination of the actual emission or the actual conversion takes place such that the control unit adds up or integrates the measured value of an NOx sensor and/or of an NH3 sensor and/or of an N20 sensor and/or of an HNCO sensor and/or of a lambda sensor arranged downstream of the SCR catalyst or the actual conversion determined before the SCR catalyst with the aid of at least one of these sensors and the untreated emission, for a pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integrating at least one operating parameter, and at the same time monitors whether the at least one operating parameter lies within one value range of at least two value ranges, the value ranges being determined by variables stored in the control unit,
the control unit, when it establishes that the current value range has been left during the adding-up or integration, discards the added-up or integrated sum, or the control unit, when it establishes that the current value range has been left during the adding-up or integration, temporarily stores the added-up or integrated sum and then when the control unit establishes a return to the value range previously left, continues with the adding-up or integration until the pre-settable time t for adding up or integration has passed or until a pre-settable amount of emission or, by adding-up or integrating at least one operating parameter, a pre-settable value is reached,
the control unit uses the sum added up or integrated over the pre-settable time t or the pre-settable amount of emission or pre-settable work, or a variable proportional thereto as actual emission or actual conversion for the comparison with the desired emission or the desired conversion determined from stored values, and determines a correction value for the dosing amount, the control unit, with the correction value determined for the respective value range of the at least one operating parameter, modifies the dosing amount from the model if the value currently determined by the control unit of the at least one operating parameter lies in this value range, or the control unit, dependent on the current value, determined by the control unit, of the at least one operating parameter, from correction values which had been determined for values of the at least one operating parameter directly neighbouring to this current value, determines a correction value by interpolation, and with this correction value modifies the dosing amount from the model.
2. A method according to Claim 1, characterised in that timewise in parallel with the determination of the actual emission or the actual conversion the control unit determines the desired emission or the desired conversion in that for the operating points of the internal combustion engine and/or of the exhaust aftertreatment system which are passed through the ideal emission values stored in each case in the control unit for the operating points, or ideal conversions in the pre-settable time t or until a pre-settable amount of emission is reached or until a pre-settable value is reached by adding-up or integration of at least one operating parameter, are added up or integrated by the control unit, and the control unit uses the sum thus obtained or a variable proportional thereto as desired emission or desired conversion for the comparison of the actual emission or the actual conversion with the desired emission or the desired conversion.
3. A method according to one of Claims 1-3, characterised in that the dosing amount determined from the models is linked by the control unit to the correction value by multiplication or addition.
4. A method according to one of Claims 1-4, characterised in that the at least one operating parameter of the internal combustion engine and/or of the exhaust aftertreatment system which is evaluated by the control unit is the coolant temperature and/or the oil temperature and/or the fuel temperature and/or the fuel-injection pressure and/or the suction-air temperature and/or the charge-air temperature and/or the turbocharger speed and/or the charging pressure and/or the driving speed and/or the engine speed and/or the amount of fuel injected and/or the exhaust temperatures and/or the catalyst temperature and/or amount of reducing agent injected and/or the exhaust-gas recirculation rate and/or the reducing-agent pressure and/or the emissions and/or the fuel/air ratio and/or the change in these variables over time.
5. A method according to one of the preceding claims, characterised in that the control unit determines and stores different correction values and/or different parameters of at least one correction function and/or a plurality of correction functions for different operating parameters of the internal combustion engine and/or of the exhaust aftertreatment system, and in that the control unit, dependent on the current operating point of the internal combustion engine and/or of the exhaust aftertreatment system, determines an operating-point-related correction value from these correction values and/or correction functions.
6. A method according to one of the preceding claims, characterised in that the pre-settable time t is at least 15 seconds.
7. A method according to one of the preceding claims, characterised in that the pre-settable amount of emission is determined by adding-up or integration of concentration values and/or emission masses and/or emission mass flows.
8. A method according to one of the preceding claims, characterised in that the pre-settable amount of emission is stored in units of mass and that the reaching of the pre-settable amount of emission is brought about by adding-up or integration of emission masses and/or emission mass flows, the pre-settable amount of emission for NOx being at least 1 mg and/or the pre-settable amount of emission for NH3 being at least 0.01 mg and/or the pre-settable amount of emission for N20 being at least 0.02 mg and/or the pre-settable amount of emission for HNCO being at least 0.01 mg.
9. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values are identical to the operating parameters, the value ranges of which are checked or which are used for determining the correction value.
10. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values differ from the operating parameters, the value ranges of which are checked or which are used for determining the correction value.
11. A method according to one of the preceding claims, characterised in that the operating parameters used for determining the duration of the integration or adding-up of the desired and actual values are the amount of exhaust and/or the amount of fuel and/or the amount of reducing agent and/or the work produced by the internal combustion engine.

Documents:

http://ipindiaonline.gov.in/patentsearch/GrantedSearch/viewdoc.aspx?id=uthPN8L8lFCs3ihQhC2Gvw==&loc=+mN2fYxnTC4l0fUd8W4CAA==

« Previous Patent

Next Patent »

Patent Number

272498

Indian Patent Application Number

512/DEL/2010

PG Journal Number

15/2016

Publication Date

08-Apr-2016

Grant Date

05-Apr-2016

Date of Filing

05-Mar-2010

Name of Patentee

MAN TRUCK & BUS AG

Applicant Address

DACHAUER STRASSE 667, D-80995 MUNCHEN, GERMANY,

Inventors:

#	Inventor's Name	Inventor's Address
1	DORING, ANDREAS	HEIMERANSTRAßE 2, D-80339 MUNCHEN, GERMANY,
2	WALDE, FLORIAN	BUCHENWEG 1, D-90599 DIETENHOFEN, GERMANY,
3	PHILIPP, JOCHEN	ROTHANGER 13, D-91080 UTTENREUTH, GERMANY,
4	MUNITZK, HENRY	TULPENWEG 16, D-90562 HEROLDSBERG, GERMANY,
5	STEINERT, RALF	WITTELSBACHERSTRASSE 23, D-90475 NURNBERG, GERMANY,

PCT International Classification Number

B01D53/56;

PCT International Application Number

N/A

PCT International Filing date

PCT Conventions:

#	PCT Application Number	Date of Convention	Priority Country
1	10 2009 012 093.9	2009-03-06	Germany