SE534479C2 - Estimation of a deviation for at least one model variable of a catalyst model - Google Patents

Description

10 534 475 Nitrogen oxides NOX include both nitrogen monoxide NO and nitrogen dioxide NO2 throughout this document. For the desired reaction in the catalyst 10, there is a stoichiometric ratio between the ammonia formed and the nitrogen oxides NUX in the exhaust gases to be purified. An ammonia molecule is needed for each molecule of nitrogen oxides NOX. For the catalyst process to be efficient, therefore, the concentration of ammonia, and thus also the fl fate of dosed urea, must correspond to and follow the concentration of nitrogen oxides NO, in the exhaust gases you want to purify.

The concentration of nitrogen oxides NOX in the exhaust gases 120 before the catalyst and also the operating conditions, such as the temperature, vary over time. These variations may be due, for example, to how the vehicle is driven and to the ambient temperature. For example, the ability of the catalyst to accumulate ammonia varies with temperature, so that the ability to accumulate ammonia is greater at low temperature than it is at high temperature.

The dosage of urea must be balanced with these variations to provide the desired purification of the exhaust gases. For example, more urea should be dosed if the temperature drops and less urea should be dosed if the temperature rises. The dosage of urea must therefore be adjusted to provide an efficient stoichiometric catalyst process, and to counteract the presence of nitrogen oxides NOX and / or ammonia in the exhaust gases 130 flowing from the exhaust pipe or catalyst.

In order to calculate the required fl fate of dosed urea so that fl the fate of dosed urea is adapted to the concentration of nitrogen oxides NOX in exhaust gases 120 upstream from the catalyst 110, ie the exhaust gases the system 100 receives from the engine, and to the required amount of accumulated ammonia a model of the catalyst l 10 is used. The catalyst model includes conditions for the catalyst, such as, for example, temperatures, and the accumulated amount of ammonia in the catalyst. A description of such a catalyst model is given below.

The concentration of nitrogen oxides NO X upstream of the catalyst 110 can be determined by using a NO X sensing sensor 122, which is positioned so that it is brought into contact with the exhaust gases 120 before passing the catalyst 10. The concentration of nitrogen oxides NOX in the exhaust gases 120 from the engine can also be determined. is determined by using a model, which describes the concentration of nitrogen oxides NOK in each drift point. 534 479 For the calculation of the required fl fate dosed urea, the exhaust fl fate from the engine can also be taken into account. The exhaust gas fate can be measured, but can also be calculated based on operating parameters such as speed, charge pressure, intake temperature and injected fuel quantity.

Figure 1 schematically shows a sensor 122, which may comprise a temperature sensor and a NOX sensor, where the sensor 122 is arranged so as to be brought into contact with the exhaust gases from the engine 120. The schematically illustrated sensor 122 can here sense the concentration of nitrogen oxides NO, and / or temperature and / or the fate of the exhaust gases 120. In Figure 1, the sensor 122 is shown schematically as located upstream of the metering nozzle 121.

However, the temperature sensor included in the sensor 122 may be located either upstream or downstream of the metering nozzle 121. However, the NO X sensor contained in the sensor 122 should be located upstream of the metering nozzle so as not to be affected by the metering. Sensor 122 provides one or more signals corresponding to the concentration of nitrogen oxides NO X and / or temperature and / or the flow of the exhaust gases 120 to a control unit 160. Sensor 122 may be located both before and after the metering nozzle 121. The control unit 160 is also arranged to receive a signal 132 from a NOX sensor 131, which is positioned so as to be brought into contact with exhaust gases 130 downstream of the catalyst 1 10, i.e. with exhaust gases which have passed the catalyst 1 10, and is intended to shield nitrogen oxides NOX, that is to say both nitrogen monoxide NO and nitrogen dioxide NO,, as well as ammonia NHa, in these exhaust gases 130.

The control unit 160 controls an actuator 140, which communicates with a urea tank 150 and with the dosing nozzle. The actuator 140 controls the flow of urea from the urea tank 150 to the metering nozzle 121 according to the control signals the controller 160 provides to the actuator 140. The controller 160 utilizes the catalyst model to determine how the actuator is to be controlled.

A model-based regulation of the urea dosage for an SCR catalyst generally enables efficient utilization of the catalyst's capacity for NOX reduction. Crucial to a performance in the control (regulation) of urea dosing is, however, the accuracy of the model.

The accuracy of the model depends on the accuracy of one or fl your model variables of the model, which correspond to input signals to the model, and then in particular input signals dry concentration 534 479 of NOX in the exhaust gases l20 from engine and fl fate of dosed urea, condition in the catalyst, and used properties for catalyst.

However, the accuracy of the model variables is not always satisfactory. There is a relative deviation of one or more model variables in the catalyst model, which means that the catalyst model does not correspond to reality. These deviations can vary with time and can also be different sizes depending on the size of the model variable. For example, a deviation of the insigrrals for both NOX before catalyst and the fate of dosed urea gives rise to a time-varying error in modeled concentrations of nitrogen oxides NOX and ammonia NH 3 after catalyst.

Brief Description of the Invention An object of the present invention is to provide a method and a system for estimating a deviation for at least one model variable of a catalyst model, which completely or partially solve the above-mentioned problems.

This object is achieved by a method for estimating a deviation for at least one model variable of a catalyst model according to the characterizing part of claim 1.

The object is also achieved by a system for estimating a deviation for at least one model variable of a catalyst model according to the characterizing part of claim 20.

The object is also achieved by a computer program and a computer program product which implement the method according to the invention.

The model variables for which deviations are estimated correspond to one or more of the input signals to the catalyst, the state of the catalyst and the properties of the catalyst. The invention produces one or two deviation parameters, each of which describes a relationship between a fair value of any of these inputs to the catalyst, state of the catalyst and properties of the catalyst and a value of a corresponding model variable. These one or two deviation parameters can then be used to correct for errors in input signals to, or properties of, the catalyst model. These deviation parameters can, according to the present invention, be calculated very computationally efficiently. 534 479 NOX sensor 131 after the catalyst is used here to estimate deviations related to model variables, i.e. to estimate deviation parameters, whereby to these related input signals to and properties for the catalyst model can be corrected, so that better accuracy of the model is achieved. This also results in a better performance in the regulation of urea dosing.

The estimated deviation for a model variable can thus, for example, be determined for a state, while the correction is made on an input signal or a property that gives rise to the deviation in the state. An example of this is the case of estimating deviation for the catalyst temperature condition, which is then used to correct the input signal for exhaust temperature before the catalyst. Here, the input signal that relates to the estimated deviation for the condition is corrected.

By utilizing the present invention, the input signals of the catalyst model can be adjusted so that they correspond well with reality. This means that the regulation of urea can be done very precisely, ie that urea consumption can be reduced while the emission requirements are met effectively.

Since the exhaust gas purification becomes very efficient when the present invention is utilized, the emission requirements can be met with a smaller catalyst. A smaller catalyst also provides less back pressure for the engine, which results in reduced fuel consumption.

Thus, the present invention can be used in model-based control of the urea dosage of an SCR catalyst. The present invention can also be used for other types of catalysts. For example, it can be used for applications in which a model is used to determine how much nitrogen monoxide NO is oxidized to nitrogen dioxide NO 2 over an oxidation catalyst. For this application, the method is used to estimate deviations in the concentration of nitrogen oxides NO, upstream of the oxidation catalyst and to estimate deviations for the activity in the oxidation catalyst. According to an embodiment of the present invention, the value of the deviation parameter c NO is determined, which describes the deviation of the input signal for the concentration of nitrogen oxides NOX upstream from the catalyst.

According to an embodiment of the present invention, the value of the deviation parameter cr is determined, which describes the deviation of the state temperature in the catalyst. This deviation is related to a deviation of the input signal for temperature upstream of the catalyst.

According to an embodiment of the present invention, the value of the deviation parameter ck, which describes the deviation of a property relating to activity in the catalyst, or the deviation of the input signal relating to the fate of the catalyst is determined.

According to an embodiment of the present invention, the value of the deviation parameter is determined, which describes the deviation for a condition relating to the degree of coverage of ammonia.

According to an embodiment of the present invention, the value of the deviation parameter cum is determined, which describes the deviation of the input signal for the fate of dosed urea.

By utilizing these embodiments of the invention, values for deviation parameters are thus obtained, which can then be used to correct the respective input signal to, or property of, the catalyst model. That is, the deviation parameters are used to correct the input signals and properties of the catalyst to which they relate. The development of these deviation parameters thus enables the control for dosing of urea to be made efficient and precise.

The deviation parameters can also be used for misdiagnosis of the system. For example, a fault in a sensor upstream of the catalyst can be diagnosed if large errors in the input signal for the concentration of nitrogen oxides NO X upstream of the catalyst are detected.

BRIEF DESCRIPTION OF THE DRAWINGS The invention will be further elucidated below with reference to the accompanying drawings, in which: 534 479 Figure 1 shows a schematic sketch of a catalyst system, Figure 2 shows a flow diagram of the method according to the present invention, Figure 3 schematically shows a control unit.

Description of Preferred Embodiments The method of the present invention involves estimating errors of model variables of the catalyst model, such as inputs to the model, state of the model, and properties of the model. The estimate is based, among other things, on the sensor signal provided by the NO X sensor 131. The NOx sensor 131 can give either separate signals for nitrogen oxides INOY and ammonia YNH], respectively, or a common signal for these imm, (i.e. the NO 1 sensor 131 is a cross-sensitive NOX sensor).

The cross-sensitive NO 1 sensor 131 has approximately the same sensitivity to nitrogen monoxide NO (non-sensitive sensitivity with a value of 1) as for ammonia NH 3 (non-sensitive sensitivity with a value of 1), but has a slightly lower sensitivity to nitrogen dioxide NO 2 (standardized sensitivity with a value of 0.7). However, if it is an efficient catalyst process, most nitrogen monoxide NO has normally been left in the exhaust gases. However, with an estimation method, a separate signal for the concentration of nitrogen oxides may have a separate signal for the concentration of ammonia obtained from the sensor signal Ixenx ".

The following is a description of a catalyst model. As will be appreciated by one skilled in the art, catalyst models can be defined in a variety of ways, of which this described model is an example. A description of variables and indices used in the description of the catalyst model and also the following equations is presented at the end of the description.

The catalyst model is formulated as a tank series model where active material is divided into layers. The catalyst model includes the following reactions: Urea + H 2 O -> 2NH, + CO 2 (eq. 1) S + NH 3 -> S-NH 3 (eq. 2) S-NH, -> S + NH 3 (eq. 3) 10 534 479 s 4s _ NH, + 4 / vo + o, _ »4s + sH2o + m, (eq. 4) 4s _ NH; + 502 s 4s + sH, o + 4 / vo (eq. 5) The model also includes urea decomposition (first reaction) from the metering nozzle 121 up to the catalyst 110. The model uses the simplification that the isocyanic acid (HNCO) formed during the decomposition of urea considered equivalent to ammonia.

The urea decomposition is treated as a homogeneous reaction with the reaction rate: fnak = km 'Gram' Yufeaßp (ÛkV-Ö) Other reactions are catalytic. The reaction rate for adsorption in tank k and layer n is expressed as: fc_a.k.n = ka; 'Crank' YNH ,, I <, n '(1 _ gm) (eq- 7) The reaction rate for desorption in tank k and layer n is expressed as: fc, d, k, n = kan' Ûm (CkV- 8) The reaction rate for NO X - reduction in tank k and layer n is expressed as: fc, r, k, n = km 'Except' Y ~ o ,, k, n 'Ûm (ekV- 9) The reaction rate for ammonia oxidation in tank k and layer n expressed as: fail / nn = biscuit 'Crow' Ûk fl (eq- m) The rate constant for both homogeneous and catalytic reaction is determined from Arrhenui's equation according to: _ EH k M = ka] -e RT ”(eq. ll) 534 479 The total the gas concentration is determined from the general gas laws according to: c - p *: ork - R_Ts, k (eq. 12) The total pressure Pk can be added to the pressure e fi er catalyst 110 which is equal to the atmospheric pressure if the catalyst is mounted last in the exhaust system. For a more accurate value, the pressure drop from the current tank to the end of the catalyst can be calculated from the laminar flow funnel in a channel and added to the pressure of the catalyst. If the catalyst is not the last in the system, the pressure can be measured or calculated from the pressure drop across the unit or units that are located in the catalyst.

For the catalyst, the following material balance regarding the gas fl fate can be set for substance i: F '(Yi, «- 1, o _ Yfymo)„ rim:' (Yf.v <, o _ Yfym l * z VL; 'fm / m' Vx = Û (ekV- 13) I The following material balance can be set up for the gas phase in layer n for substance i (homogeneous reactions neglectedyri, k, n-1 '(Yi, k, n-1 “Yf, k, n)“ room '(Yi, / <, n' Y /, k, n + 1) + 21 / 1,1 'fajkn' Wu = Û (ekV- 14) I For the innermost layer, the second term is deleted according to: ff, k, ~ - 1 '(Yi, k, ~ -1 r Yr, k, ~) + 2% /' fe, 1, k, ~ 'Wmv = Û (ekV- 15) 1 The material balances for the gas fl fate and for the gas phase in the active layers form an equation system from which the mole fractions for the different substances in the gas och and the gas phase in the active layers are determined.Mass transport coefficient for transport between the gas och and the first layer is determined according to: 534 479 1 O Åc / Lk 'Umm 11x10 1 pñlâlêxl (eq. k The '1 k Mass transport coefficient for transport between the layers is determined according to: L ”= (eb / _ 17) AX n + Axn + 1 The film transport coefficient is determined according to: h - D» kw., = S d' - * (ek v. 18) For Sherwood's numbers, Sh, asymptotic values for current channel geometry can be used. The ordinary diffusivity is determined according to: T 1.75 Dilk: Dæfj ref Knudsendi fl The excessivity is determined according to: d * 8 - R - T DKJ-Ik = ïp '(eq.

The effective diffusivity is determined according to: De fl ihk = 1 fo 1 51 :: + DKLK The time-dependent material balance for adsorbed ammonia gives the time derivative for degree of coverage according to: 534 479 11 døk fl 1 = * 'ZVS-NH j'fcjkn'W / d: / vcj = - '(eq. 22) The coverage ratio at each time point is then integrated based on the time derivative. The temperature in the gas stream through the catalyst 110 is determined by solving a heat balance which takes into account the heat transferred to the solid material and the heat of reaction for homogeneous reactions.

The power from the heat of reaction for homogeneous reactions is determined according to: oh, = Xrmvk (_ AH) (eq. 23) in The heat balance for the gas fl is then solved by: TgIk = F “czw 'Ta / m * hk' Åatx 'Tu + QM (eq. 24) F-Cplg + hk -Achyk Where the heat transfer coefficient is determined according to: hk _ 'l / ”Jlg (eq. 25) For Nusselt's numbers, Nu, asymptotic values for the current channel geometry can be used. The power from the heat of reaction for catalytic reactions is determined according to: Oc, = ZZrCJJU, wkl ”- (- AH, -) (eq. 26) ni The time-dependent heat balance of the solid in the catalyst gives the time derivative for temperature according to: dT i = fw *. ACM .ggvk k. 7 dt mM-cns (e v 2) -Tsrwøc fi The temperature at each time point is then integrated based on the time derivative. 534 479 12 If the catalyst model, including its input signals, is completely in line with reality, the regulation of the flow of dosed urea becomes relatively accurate and efficient purification is achieved.

However, the accuracy of the catalyst model depends on the accuracy of its input signals. The two input signals that are most important for the accuracy of the catalyst model are the input signal for NOX from the engine, ie for the concentration of NO, upstream from the catalyst 110, and the input signal for the fate of dosed urea. When the model is used to control urea dosing, errors in the input signals will affect the emissions of nitrogen oxides NOX and ammonia NH 3 downstream of catalyst 110. For example, if the input signal for the concentration of NOX upstream of catalyst 1 10 is too high, the control will dose more urea in proportion to NO X, which gives rise to a higher concentration of ammonia NH 3 after the catalyst 110, i.e. downstream of the catalyst 110. A similar deviation, which gives rise to a similar consequence, will occur if the actual fl fate of dosed urea is greater than requested.

The accuracy of the catalyst model also depends on the accuracy of its properties used. For example, due to the aging of the catalyst, a deviation in the property activity may occur, as the activity deteriorates over time.

The present invention utilizes the principle that deviations for one or more model variables of the catalyst model can be estimated by comparing actual concentrations of nitrogen oxides NOX and ammonia NH 3 downstream of the catalyst 110 with modeled respective concentrations. The NO X sensor 131 after the catalyst is thus used here to estimate deviations related to model variables, i.e. to estimate deviation parameters.

The idea is that the model calculates output signals, in the form of, for example, concentrations after the catalyst. How well these outputs correspond to reality depends on the accuracy of a number of model variables, which correspond to input signals to the catalyst, state of the catalyst or properties of the catalyst. It is also possible to make samband am relationships that represent how these model variables deviate from reality, ie 534 479 13 deviation parameters. Since one can establish a relationship between these deviations and the input signals to the respective properties of the catalyst model, one can also estimate errors in the input signals and the properties, respectively, and correct for these, so that better accuracy for the model is achieved.

For example, in order to correct input signals based on the estimate, one can determine how the input signals used (input signals used by the model, which have errors / deviations) relate to the actual input signals (the input signals without errors / deviations). The errors in the modeled concentrations of nitrogen oxides NOX and ammonia NH 3 downstream of the catalyst then relate to the errors / deviations in the input signals for, for example, NO X from the engine, the fate of dosed urea, and ternperitrite and fate. Therefore, it is also possible to correct for the deviations in these input signals.

Figure 2 shows a flow chart for a method according to the present invention. In a first step 201 of the method, an actual signal is compared with an estimated signal. The actual signal here depends on the sensor signal measured by the NO X sensor 131. The estimated signal depends on at least one estimation function and constitutes the catalyst model's equivalent to the actual signal.

In a second step 202 of the method, the comparison is used to determine at least one deviation parameter for each at least one estimation function. Each of these one or more deviation parameters describes a relationship between a fair value and a value for the corresponding model variable.

The value of the model variable thus corresponds to the fair value, which is why the deviation parameter relates the fair value to the value of the model variable. Thus, they describe one or fl your deviation parameters how well the model corresponds to reality.

The value of the model variable can be based on, for example, a measurement or on a separate model, where this separate model can be a different model than the catalyst model which is used according to the present invention. 534 479 14 The invention thus identifies a difference between how the model says that a model variable should look like and how reality says that the corresponding fair value should look like, and relates this difference to input signals to and properties of the catalyst.

Because these one or two deviation parameters can be determined in this way, the errors in the insigrials to and the properties of the catalyst, which relate to these deviation parameters, can be determined and corrected for. When these errors are corrected for, a higher accuracy of the input signals to the respective properties of the catalyst is obtained, which results in a more effective control of the urea dosage. Thus, the method of the present invention provides a more efficient purification process for the catalyst, which reduces the consumption of urea. Fuel consumption can also be reduced because a smaller catalyst, with less back pressure for the engine, can be used to meet the emission requirements.

According to various embodiments of the present invention, different estimation functions are used to determine the deviation parameters. These estimation functions will be described in the following. One of these estimation functions consists of an estimation function for nitrogen oxides f N01 and one consists of an estimation function for ammonia f NH].

The estimation function for nitrogen oxides f N01 depends on a material balance for a modeled reaction in the catalyst model and relates a modeled concentration of nitrogen oxides NO, downstream from the catalyst l10, to the sensor signal YNUX measured by the NO1 sensor. As described above, a separate signal ym for the concentration of nitrogen oxides and a separate signal fw for the concentration of ammonia can be obtained by simultaneous estimation from the sensor signal Ismw ”.

The relationship between the modeled concentration of nitrogen oxides NOX downstream of catalyst 110 and the measured sensor signal ïm is described by one or fl of the deviation parameters c NO,, c,, and ch. Deviation parameter c m describes the deviation for the concentration of nitrogen oxides NOX upstream of the catalyst 110. Deviation parameter c, describes the deviation for the state temperature in the catalyst l 10. Deviation parameter ck, 10 534 479 describes the deviation for the property activity in the catalyst or for the input signal fl fate in the catalyst.

The estimation function for nitrogen oxides can be derived from the material balance of a first order reaction in a tank series. The estimation function then becomes a product between the deviation parameter c NO, for nitrogen oxides NOX upstream from catalyst 110, the mole fraction for nitrogen oxides NOX upstream from catalyst l 10 and a factor for each tank describing how much is left of nitrogen oxides NO, after each tank: f = . _ ______, "_ K 1 dk zs No, UNO, yuo ,, o lll 1 + (krk kw af (e V) - the turnover variable (kr) k represents a first order velocity constant times the residence time and is determined from input and output nitrogen oxides NO X from each tank according to: Y ~ o ,, o _ 1 k: 1 Y (k1) k = y ~ z ° ïffo (eq. 29) _ 1 ke [2, K] Y / vo ,, k, o The deviation parameter for turnover gives the effect of deviation in activity / flow, degree of coverage of ammonia and temperature according to: tt »'~ ° ~' e RR * ° f here (ekvao) fkak = Ckf '9NH ,, k ck, is the deviation parameter for the property activity or the input signal flow, 0 ,, is the deviation parameter for the state coverage of ammonia, cr is the deviation parameter for the state temperature in the catalyst, and EA is the activation energy for reduction of NO, .10 534 479 16 Common to the deviation pair frames is that the value one corresponds to the absence of deviation. deviation parameters except for the degree of coverage of ammonia indicate the value of the deviation parameter by which factor r the value used needs to be multiplied by to correspond to the fair value. For ammonia coverage, 2 minus the deviation parameter indicates the value to which the coverage should be increased to correspond to fair value.

According to one embodiment of the present invention, the estimation function for nitrogen oxides f NOA comprises, a first relationship between a modeled concentration of nitrogen monoxide NO downstream of the catalyst 110 and the sensor signal jmxo, measured by the NO, sensor 131, and a second relationship between a modeled concentration of nitrogen dioxide NO2 downstream of the catalyst 110 and the sensor signal ~ y sensor 'measured by the NOX sensor 131' Furthermore, according to one embodiment of the present invention, the estimation function of ammonia f NH: a modeled concentration of ammonia NH; downstream of the catalyst to the sensor signal L ", measured by the NO 1 sensor 131, and depends on a balance between modeled adsorbed ammonia and modeled ammonia in gas form at an outlet end of the catalyst 110. The outlet end of the catalyst is normally the last tank in the catalyst 1 lO.

The estimation function of ammonia f NH: can be derived from the equilibrium between adsorbed ammonia and ammonia in gas phase in the last tank of the catalyst. The deviation parameters relate the value of the catalyst model for ammonia NH 3 after the catalyst to fair value by deviation in temperature and degree of coverage of ammonia NH 3 according to: frvH, = YNH ,, K '(ekV-31) - ce is the deviation parameter for the state degree of ammonia, - cr is the deviation parameter of the state temperature in catalyst, and 534 479 17 - AH is adsorption enthalpy for adsorption of ammonia NH3.

Thus, according to one embodiment of the present invention, the estimation function of ammonia relates to NH; a modeled concentration of ammonia NH 3 downstream of the catalyst 11 to the sensor signal ÉNH measured by the NO 1 sensor 131; . This is done by using at least one deviation parameter cT for the state temperature in the catalyst, and a deviation parameter c ,, for the state degree of coverage of ammonia.

Furthermore, the estimation för functions for nitrogen oxides f m, respectively for ammonia f ”H3 are used to estimate deviations of one or more model variables of the catalyst model.

The model variables may be catalyst inputs, catalyst states, or catalyst properties, as described below.

According to an embodiment of the present invention, the estimation function for nitrogen oxides f NO! for estimating errors of an input signal corresponding to a concentration of nitrogen oxides NO X upstream of catalyst 110. This is done by estimating the deviation parameter c NÛX for concentration of nitrogen oxides NOX upstream from said catalyst.

According to an embodiment of the present invention, the estimation function for nitrogen oxides fm: is used, wherein the estimation function for nitrogen oxides f NOK comprises a first relation between a modeled concentration of nitrogen monoxide NO downstream of the catalyst 110 and the sensor signal íNo, measured by NO, sensor 131, and a second relation signal. between a modeled concentration of nitrogen dioxide NO 2 downstream of the catalyst 110 and the sensor signal YNOI measured by the NO 1 sensor 131.

By using this estimation function NO! , which includes these first and second relationships, together with two separate deviation parameters for nitrogen monoxide c NO and for nitrogen dioxide cNOZ, an estimate of a deviation for model variables relating to nitrogen monoxide NO concentration upstream of the catalyst and concentration of nitrogen dioxide NO 2 upstream of the catalyst can be made. . Then, errors of two separate inputs, which correspond to a concentration of nitrogen monoxide NO upstream of the catalyst 11 and a concentration of nitrogen dioxide NO 2 upstream of the catalyst 11, can be determined, as they are related to the model variables.

According to an embodiment of the present invention, the estimation function for nitrogen oxides f ML is used for estimating errors of an input signal, which relates to a temperature in the catalyst 110. This is done by determining a deviation parameter c, for the model variable for the condition relating to the temperature in the catalyst 110. . This deviation parameter c, is related to an input signal which affects the temperature in the catalyst, which allows the error of this input signal to be determined, since it is a deviation of this input signal which results in the deviation in the state of temperature in the catalyst.

For example, an error of the exhaust gas temperature input signal upstream of the catalyst from the sensor 122 may give rise to a deviation in the state of the temperature in the catalyst 10. By determining this deviation of the state of the temperature in the catalyst, the error of this input signal can also be determined and corrected for.

According to an embodiment of the present invention, the estimation function for nitrogen oxides f N01 is used for estimating deviation of a property relating to an activity in the catalyst 110 or for estimating deviation of an input signal corresponding to fl fate in the catalyst 1 10. This is done by a deviation parameter ck , for the property activity or the input signal flow in said catalyst 110 is determined.

According to one embodiment of the present invention, any of the deviation parameters c NUX, which describes the deviation of the input signal for concentrating nitrogen oxides NOX upstream of the catalyst, is estimated, cr, which describes the deviation of the state temperature in the catalyst 1, and ck, which describes the deviation of the property activity. or the input signal flow in the catalyst, by a comparison between the sensor signal 2.0 measured by NO 1 sensor 131! , which corresponds to a concentration of nitrogen oxides NO,, and the estimated corresponding signal is made. 534 4? 9 19 This comparison uses the relationship: ym = f ,, 0X (c ,, ..., ck) + g, where (eq. 32) - YNOX is the separate value for nitrogen oxides fi 'from the measured sensor signal, - fNOX (c, ck) is the estimated signal, - cl, ..., ck is at least one deviation parameter, which is one of the deviation parameters c NO! , c, and ch, and - s are a residual.

According to one embodiment of the present invention, the estimation function for ammonia f NH is used: for estimating one of the deviation paranieters cT, which describes the deviation of the state temperature in the catalyst 110, and ck, which describes the deviation of the property activity or the input signal fl fate in the catalyst, by comparison of the sensor signal 2% measured by the NO 1 sensor 131, which corresponds to a concentration of ammonia, and the estimated corresponding signal is performed.

This comparison uses the relation: íw] = fw, (c ,, ..., ck) + g, where (eq. 33) - YNHJ is the separate value for ammonia from the measured sensor signal, - fw: (c ,,. .., ck) is the estimated signal, - c ,, ..., c ,, is at least one deviation parameter, which is one of the deviation parameters cr and ch, and - s is a residual.

According to an embodiment of the present invention, both the estimation function for nitrogen oxides f Nor and the estimation function ßr ammonia f NH! for estimating any of the deviation parameters c M, which describes the deviation of the input signal for concentration of 10 534 479 nitrogen oxides NOX upstream of the catalyst, c, which describes the deviation of the state temperature in the catalyst 110, and ck, which describes the deviation of the property activity. or the input signal flow in the catalyst, by performing a comparison between the sensor signal fsenso measured by the NOX sensor 131, and the estimated corresponding signal.

According to this embodiment, the measured sensor signal corresponds to Ixenm, the sum of a concentration of nitrogen oxides NO, and a concentration of ammonia NH 1, which is the sum normally given as the output signal of the cross-sensitive NO X sensor 13 l. The estimated signal here consists of the model's equivalent to the sum of the concentration of nitrogen oxides NO, downstream of the catalyst 110 and the concentration of ammonia NH; downstream of catalyst 110.

The estimation is performed by using a comparison of the measured sensor signal with the estimated signal according to: ïmßo, = fN0x (c ,, ..., c ,,) + f ,, H3 (c ,, ..., c, <) + s, where (eq. 34) - lens' is said measured sensor signal, which is the sum of a concentration of nitrogen oxides NOX and a concentration of ammonia NII3, - fw, (c ,, .. ,, ck) + fNHa (c ,, ..., ck) is the estimated signal, - c,, ..., ck is at least one deviation parameter, which is one of the deviation parameters cm, c, and ck,, and - s is a residual.

Through this embodiment, deviations in model variables can be estimated directly from the output signal from the cross-sensitive NOX sensor 13 l.

According to an embodiment of the present invention, a deviation estimate is made for a model variable, which is an input signal corresponding to the fate of dosed urea, where a deviation parameter cum, for the fate of dosed urea is estimated and utilized. A problem in estimating deviation in the fate of dosed urea is that there is a dynamic between dosing urea and reduction of nitrogen oxides and ammonia NH 3 after the catalyst. The dynamic effect arises when the flow of dosed urea first gives rise to ammonia NH, which is then adsorbed in the catalyst. This adsorbed ammonia NH, then reduces NO X or at temperature transients gives rise to ammonia NH 3 after the catalyst, when ammonia desorbs.

In order to get around this problem, according to this embodiment, other equivalent measures of the dose of dosed urea are used instead. It is utilized here that the urea that is dosed gives rise to either a reduction of nitrogen oxides NOX or ammonia NH; is a catalyst.

The sum of reduced nitrogen oxides NO; and ammonia NH 3 after catalyst is thus an equivalent measure of the fl fate of dosed urea. An equivalent fl fate of dosed urea according to the catalyst model (modeled fl fate of dosed urea) is also produced and used together with an estimation function for urea f "which relates said modeled 1 fate of dosed urea to said actual fl fate of dosed urea, to obtain a measures of modeled fl fate of dosed urea. These equivalent measures of the fl fate of dosed urea are then used in a comparison to determine an error in the input signal for the fl fate of dosed urea.

Thus, according to the embodiment, the actual signal corresponds to an equivalent actual fl fate of dosed urea Ém, which is determined based on actual exhaust flow, a measured concentration of nitrogen oxides NOI upstream of the catalyst, saint on the sensor signal measured by the NOX sensor l3l, where the measured signal comprises a measured concentration of nitrogen oxides NO, and a measured concentration of ammonia NH, downstream of said catalyst l 10: Fuæa = FfY / vopo 'Y / vo, * YA / HJ »where (ekV-35) _ E / rea is equivalent actual fl fate of dosed urea, - F is exhaust gas fl fate, - y M, _0 is measured concentration of nitrogen oxides NO, upstream of the catalyst, 534 479 22 - YNOX is the separate value for nitrogen oxides from the measured sensor signal, - ïNm is the separate value for ammonia from the measured sensor signal.

The sensor signal measured by the NO 1 sensor can here include actual separate values for nitrogen oxides NO, and for ammonia NH 2. These separate values for nitrogen oxides NO, and for ammonia NH 3 can be determined with separate sensors, i.e. sensor 131 comprises separate sensors for nitrogen oxides NO; and for ammonia NH, These separate values for nitrogen oxides NO: and for ammonia NH 3 can also be determined by simultaneous estimation from a NO 1 sensor which measures the sum of NOI and NH 3, i.e. from a cross-sensitive sensor 131.

The equivalent fl fate of dosed urea according to model E can be expressed as: / rea Puma = F '(Y ~ o., O' Y ~ o, .K + YNH, .K), then? (CkV-36) - Em is an equivalent flow of dosed urea according to model, - F is the exhaust gas flow, - y NOA, is measured concentration of nitrogen oxides NO, upstream of the catalyst, - y Nov K is a modeled value for nitrogen oxides downstream of the catalyst , and - yNHPK is a modeled value for ammonia downstream of the catalyst.

When estimating deviation parameters for urea, the relationship is used: »Euma = fu, e _., (F ,,, e,, c ,, ..., ck) + e, where (eq. 37) - lïuæ, the equivalent is real fl the fate of dosed urea, - fm, (Fm,, c ,, ..., ck) is estimated fl the fate of dosed urea, - c1, ..., c ,, are deviation parameters for urea, and - s is a residual. In the simplest embodiment the estimation function is linear, according to: fuæa (Fuæ ,, c ,, ..., c ,,) = cum -Fma, where (eq. 38) - fam, (Fuma, c ,, ..., c ,,) is estimated fl fate of dosed urea, - cum, is a deviation parameter for fl fate of dosed urea, and - F Urea is an equivalent flow of dosed urea according to model.

Alternatively, for example, a polynomial function may be used, the deviation being described by your deviation parameters.

In order to estimate the error in the fl fate of dosed urea, a representative value is thus used for the actual fl fate of dosed urea, ie the equivalent value for the actual fl fate of dosed urea than,. A representative value for actual flow of dosed urea Fuæa can be determined from actual converted nitrogen oxides NO, (the difference between nitrogen oxides NOX before and after catalyst) plus actual ammonia NH, or catalyst. This value is an accurate measure of the dose of dosed urea, as the dose of dosed urea for the most part either reduces NO, or gives rise to ammonia NHa or the catalyst.

The error in the det fate of dosed urea can then be estimated by comparing the representative (equivalent) value for actual fl fate of dosed urea lïuæ, with the corresponding equivalent modeled value F l / rca from the catalyst model. The comparison is thus made between the actual sum of and the modeled sum of converted nitrogen oxides NO, by catalyst plus ammonia NH, by catalyst. In this way, the impact of the problematic dynamic effect on the fate of dosed urea, which is due to the adsorption and desorption of ammonia in the catalyst, is minimized.

If a linear estimation function is used (according to equation 38), the estimation gives a value for the deviation parameter cuma, which indicates the value the input signal corresponding to the fl fate of dosed urea should be multiplied by the feed to correspond to the fair value. The value one thus means the absence of deviation. According to an embodiment of the present invention, the at least one deviation parameter c1, ..., c1 is determined by analyzing for which value of the at least one deviation parameter c1, ..., c1 as the residual in equations 32, 33, 34, 37 and 38 are minimized. The value of the at least one deviation parameter cl, ..., cl that minimizes the residual s is selected. This way of determining the value of the at least one deviation parameter c1, ..., c1 can be done efficiently and with the use of small calculation capacity, since efficient estimation algorithms can be used.

These efficient estimation algorithms include, according to one embodiment of the invention, utilizing a Kalman filter if equations 32, 33, 34, 37 and 38 are linear. If the equations are non-linear, an extended Kalman filter (English name: Extended Kalman filter) or a value-oriented Kalman filter (English name: Unscented Kalman filter) can be used. The extended Kalman filter is preferably used for smaller nonlinearities for the equations and the expected value Kalman filter is used if the nonlinearities are large.

According to an embodiment of the invention, a recursive least squares method is used. This recursive least squares method can be used if the equations are linear or moderately nonlinear.

According to an embodiment of the present invention, developed deviation parameters c1, ..., c1 are used for correcting at least one input signal in, or at least one property of, the catalyst model, where the at least one input signal or property has given rise to the determined deviation of the model variable. The input signal or properties or properties to be corrected thus have a relation to one or more respective model variables. Once the deviations, i.e. the deviation parameters, for these one or more model variables have been determined, these can be used to correct the respective input signal or property.

Thus, generated values are used for any of the deviation parameters c m, cl, cl,, cl,, cum, to correct the corresponding input signal or property. c m 'describes the deviation of the signal for concentration of nitrogen oxides NO1 upstream of the catalyst. cl describes the deviation of the state temperature in the catalyst, which is related to the input signal for temperature upstream of the catalyst 1010. el, describes the deviation of a property which 534 479 relates to activity in the catalyst, or describes the deviation of the input signal for activity, which relates to flow in the catalyst. c, describes the deviation for a condition related to the degree of coverage of ammonia. cum describes the deviation of the input signal for the fate of dosed UfCä.

By utilizing this embodiment of the invention, values for deviation parameters are thus obtained, which can then be used to correct the input signals and properties they relate to.

For all deviation parameters except for the degree of coverage of ammonia approx., The determined value for each deviation parameter indicates the value by which factor the input signal needs to be multiplied by to correspond to the actual value. For ammonia coverage, 2 minus the value of the deviation parameter indicates approximately the value the input signal for the coverage should be raised to to correspond to fair value.

The degree of coverage of ammonia depends on a number of model variables and is affected by a number of deviations for the model. If all input signals and properties of the catalyst model, through the use of the invention, have been corrected to correspond to reality, the degree of filling of ammonia for the catalyst model will also correspond to reality.

When the deviation parameters describe a linear deviation, the correction according to this embodiment of the invention proceeds so that the input signal, which is defective, is multiplied by the value for each deviation parameter to give a corrected value corresponding to the actual value, which after the multiplication is not defective.

Thus, the corrected value for the concentration of nitrogen oxides NOX is obtained by multiplying the input signal for the concentration of nitrogen oxides NOK by the deviation parameter cN Correspondingly, the corrected value for temperature in the catalyst is obtained by multiplying the input signal for the temperature by the deviation parameter cT, and the corrected value for the flow of dosed urea by multiplying the input signal for the fate of dosed urea by the deviation parameter cum,. The corrected value for degree of coverage of 534 475 26 ammonia is obtained by raising the input signal for degree of coverage of ammonia by two (2) minus the value of the deviation parameter ce. The value one (1) for these deviation parameters thus means the absence of deviation.

According to an embodiment of the present invention, the deviation pair frames can also be used for misdiagnosis of parts of the system 100. For example, if a sensor 100 upstream of the catalyst is broken, it may provide output signals with large errors. This can be easily diagnosed by using the present invention, by detecting large errors in the input signal for concentrating nitrogen oxides NOX upstream of the catalyst.

Correspondingly, for example, errors in the control system for the dosing of urea can be detected if large errors are detected in the input signal for the flow of dosed urea. One skilled in the art realizes that the other deviation parameters described above can similarly be used to diagnose other defects.

The present invention also includes a system 100 for estimating a deviation for at least one model variable of a catalyst model. The system 100 is arranged to base the estimate on the catalyst model described above and on at least one measured sensor signal from at least one sensor 131, which is arranged to be brought into contact with exhaust gases downstream of a catalyst 110. According to the present invention, the system 100 comprises a means which is arranged to compare an actual signal with an estimated signal, where the actual signal depends on the measured sensor signal and where the estimated signal depends on at least one estimation function and constitutes the catalyst model's equivalent to the actual signal.

Furthermore, the system 100 comprises a means, which is arranged to use this comparison for determining at least one deviation parameter for each at least one estimation function, each of the one or more deviation parameters describing a relationship between a value of a model variable and a corresponding fair value. .

The person skilled in the art realizes that a method for estimating deviation of at least one model variable of a catalyst model according to the present invention can also be implemented in a computer program, which when executed in a computer causes the computer to perform the method. The computer program is comprised in a computer-readable medium of a computer program product, said computer-readable medium consisting of a suitable memory, such as, for example: ROM (Read-Only Memory), PROM (Programmable Read-Only Memory), EPROM (Erasable PROM), Flash- memory, EEPROM (Electiically Erasable PROM), a hard disk drive, etc.

Figure 3 schematically shows control unit 160. Control unit 160 comprises a calculation unit 161, which may be constituted by substantially any suitable type of processor or microcomputer, e.g. a Digital Signal Processor (DSP), or an Application Specific Integrated Circuit (ASIC).

The calculation unit 161 is connected to a memory unit 162 arranged in the control unit 160, which provides the calculation unit 161 with the control algorithm and the catalyst model used, and for example the stored program code and / or the stored data the calculation unit 161 needs to be able to perform calculations. The calculation unit 161 is also arranged to store partial or final results of calculations in the memory unit 162.

Furthermore, the control unit 160 is provided with devices 163, 164, 165 for receiving and transmitting input and output signals, respectively. These input and output signals may contain waveforms, pulses, or other attributes, which of the input signals means 164, 165 may be detected as information and may be converted into signals which may be processed by the computing unit 161. These signals are then provided to the computing unit 161.

The output signal transmitting device 163 is arranged to convert signals obtained from the computing unit 161 for creating output signals by e.g. modulate the signals, which can be transmitted to other parts of the system.

Each of the connections to the devices for receiving and transmitting input and output signals, respectively, may be one or two of a cable; a data bus, such as a CAN bus (Controller Area Network bus), a MOST bus (Media Orientated Systems Transport bus), or any other bus configuration; or by a wireless connection. The connections between the sensors 122, 131 and the control unit 160 and between the operating member 140 and the control unit 160 shown in figure 1 can also be constituted by one or more of these cables, buses, or wireless connections.

One skilled in the art will appreciate that the above-mentioned computer may be comprised of the computing unit 161 and that the above-mentioned memory may be constituted by the memory unit 162. 534 479 28 The person skilled in the art will also recognize that the above system may be modified according to the various embodiments of the method. The present invention is not limited to the above-described embodiments of the invention but relates to and encompasses all embodiments within the scope of the appended independent claims. 534 479 29 Description of variables in the catalyst model Variable Description AW, Geometric area, tank k D ,,,, ¿k Effective difisivity, substance i, tank k You, Ordinary diffusivity, substance i, tank k DK, k Knudsen diffusivity, substance i , tank k DM, Reference diffusivity, substance in EA, j Activation energy, reaction j F Mol fl fate, exhaust M, - Molar mass, substance in NC Number of active seats Nu Nusselt's number Pk Total pressure, tank k QCJ. Power, heat of reaction catalytic reactions QM Power, heat of reaction homogeneous reactions R General gas constant S Active seat Sh Sherwoods speech Tæf Reference temperature TM Temperature gas, tank k TM Temperature catalyst, tank k Vk Gas volume in catalyst duct, tank k CM Speci fi k heat, gas cp, Spec cp, Spec , catalyst if, Total gas concentration, tank kd Duct size dp Pore diameter fD Coefficient for porosity and pore winding hk Heat transfer coefficient boundary layer, tank k can Mass coefficient coefficient boundary layer, substance i, tank kk M Speed constant, coefficient, reaction j, tank j, reaction. , Mass catalyst, tank k fal / JU, Reaction rate, catalytic reaction j, tank k, layer no ,,,, ¿k Reaction rate, homogeneous reaction j, tank kt Time y, ¿k ,,, Molar fraction, substance in, tank, k, layer nw ”Mass active material, tank k, layer n F, ¿k ,,, Mass transport coefficient, substance in, tank k, layer n AH in Reaction enthalpy, reaction j Ax ,, Thickness, layer n Å. g Heat conduction a gas 534 479 3 0 v, ¿, - Stoichiometric coefficient, sleeve i, reaction j 0.0, Degree of coverage (dimensionless accumulation), tank k, layer n 534 479 31 Description of index in the catalyst model Index Meaning Number of tanks / last tank Number layer / innermost layer In gas fl fate Adsorption Desorption Ammonia oxidation NOK reduction: fi oqmzzx Urea decomposition

Claims (20)

10 15 20 25 30 534 479 32 Patent claims A method for estimating a deviation for at least one model variable of a catalyst model, said estimation being based on said catalyst model and on at least one measured sensor signal from at least one sensor (131), said at least one sensor (131) being arranged to be brought in contact with exhaust gases downstream of a catalyst (1 10), characterized by: - comparing an actual signal with an estimated signal, wherein said actual signal depends on said measured sensor signal and where said estimated signal depends on at least one estimation function and constitutes the equivalent of the catalyst model to said real signal, wherein said catalyst model is a tank series model, and - utilizing said iron configuration for determining at least one deviation parameters for said at least one estimation function, respectively, each of said at least one deviation parameter describes a relationship between a value for a model variable and a corresponding fair value. The method of claim 1, wherein said model variable corresponds to an input signal, which relates to a concentration of nitrogen oxides NO X upstream of said catalyst (1 10), wherein - one of said at least one deviation parameter is a deviation parameter c NO for concentration of nitrogen oxides NO X upstream of said catalyst (110). The method of claim 2, wherein said model variable corresponds to two separate inputs, which relate to a concentration of nitrogen monoxide NO upstream of said catalyst (1 10) and a concentration of nitrogen dioxide NO 2 upstream of said catalyst (1 10), respectively. The method of claim 1, wherein said model variable corresponds to at least one state, which refers to a temperature in said catalyst, wherein - one of said at least one deviation parameter is a deviation parameter of the state of temperature in said catalyst (1 10). 10 15 20 25 30 534 479 33 The method of claim 1, wherein said model variable corresponds to a property relating to an activity in said catalyst, or an input signal relating to a fl fate in said catalyst, wherein - one of said at least one deviation parameter is a deviation parameter ck , for the property of activity in said catalyst or for the input signal for flow in said catalyst (1 10). A method according to any one of claims 2-5, wherein: - said actual signal is said measured sensor signal, which corresponds to a concentration of nitrogen oxides NOX downstream of said catalyst (1 10), - said estimated signal consists of a modeled concentration of nitrogen oxides NO, downstream of said catalyst (110), and - one of said at least one estimation function is a estimation function for nitrogen oxides f NO, which relates said modeled concentration of nitrogen oxides NO, to said measured sensor signal. The method of claim 6, wherein said comparing said measured sensor signal with said estimated signal utilizes the relationship: ivo, = fNu, (c1 »~ -» c1 <) + 5 v where - jïmh is said measured sensor signal, - fm ( c,, ..., c ,,) is said estimated signal, - c ,, ..., ck is at least one deviation parameter, which describes how well the model corresponds to reality, where 2 l, and - e is a residual. A method according to any one of claims 4-5, wherein: - said real signal is said measured sensor signal, which corresponds to a concentration of ammonia NH, downstream of said catalyst (l 10), and - said estimated signal is a modeled concentration of ammonia NH, downstream of said catalyst (10 10), and - one of said at least one estimation function is an estimation function for ammonia 10 15 20 25 30 534 479 34 f NH:, which relates said modeled concentration of ammonia NH 3 to said measured sensor signal . The method of claim 8, wherein said ironing of said measured sensor signal with said estimated signal utilizes the relationship: YNH, = fNH, (cv - wc / Jl 'E "where - ÉFN", is said measured sensor signal, - fw! ( c ,, ..., c ,,) is said estimated signal, - cl, ..., ck is at least one deviation parameter, which describes how well the model corresponds to reality, where 2 l, and - e is a residual. A method according to any one of claims 2-5, wherein: - said actual signal is said measured sensor signal, which corresponds to a sum of a concentration of nitrogen oxides NO, and a concentration of ammonia NH, - said estimated signal is a sum of a modeled concentration of nitrogen oxides NO, downstream of said catalyst (110) and a modeled concentration of ammonia NH, downstream of said catalyst (110), - one of said at least one estimation function is an estimation function for nitrogen oxides fm, which relates said modeled concentration of nitrogen oxides NO, to said measured sensor signal, and - one of said at least one estimation function is an estimation function for ammonia f W *, which relates said modeled concentration of ammonia NH, to said measured sensor signal. The method of claim 10, wherein said ironing of said measured sensor signal with said estimated signal utilizes the relationship: lean »= f1v0, (ci = ~ - = Ck) + fNH, (c | 1 '~ -1ck) + 5 i there - ïxçutw, is said measured sensor signal, - [vor (c,, ..., ck) + f NH! (cl, ..., ck) is said estimated signal, c, ..., c ,, is at least one deviation parameter, which describes how well the model corresponds to reality, where 2 1, and - e is a residual. The method of claim 1, wherein said model variable corresponds to an input signal relating to a fl fate of dosed urea, wherein - one of said at least one deviation parameter is a deviation parameter cum, for the fl fate of dosed urea. The method of claim 12, wherein - said actual signal corresponds to an actual fl fate of dosed urea, which is determined based on a measured concentration of nitrogen oxides NOX upstream of said catalyst, and on said at least one measured sensor signal from said at least one sensor (131) , which is arranged to be brought into contact with exhaust gases downstream of the catalyst (110), wherein said at least one measured sensor signal comprises a measured concentration of nitrogen oxides NO, and a measured concentration of ammonia NH 3, - said estimated signal consists of a modeled dosed urea, and - one of said at least one estimating function is an estimating function for urea fum, which relates said modeled fl fate of dosed urea to said actual fl fate of dosed urea. The method of claim 13, wherein said comparing said actual fl fate of dosed urea with said estimated signal utilizes the relationship: Em = fu ,, ¿, (F ,,,, _ ,,, c ,, ._., Ck) + ¿, Where - Fuæa is said actual fl fate of dosed urea, - fuæ, (FW,, c ,, ..., ck) is estimated fl fate of dosed urea, - c,, ..., ck is at least one deviation parameter , which describes how well the model corresponds to reality; and - s is a residual. 15. l5. The method according to any one of claims 7, 9, 11 or 14, wherein, for estimating said at least one deviation parameter c1, c2 ..., ck, which minimizes said residual s, a calculation method is used in the group: - a calculation method using a Kalman filter, - a calculation method using an extended Kalman filter, - a calculation method using a value-based Kalman filter, and - a recursive least squares method. Method for correcting at least one input signal in a catalyst model, characterized in that said correction is performed by using estimation of said deviation for at least one model variable of said catalyst model according to any one of claims 1-15, each of said at least one input signal relates to at least one of said at least one model variable. Method for correcting at least one property of a catalyst model, characterized in that said correction is performed by using estimation of said deviation for at least one model variable of said catalyst model according to any one of claims 1-15, wherein each of said at least one property relates to at least one of said at least one model variable. A computer program comprising program code, which when said program code is executed in a computer causes said computer to perform the method according to any of claims 1-17. A computer program product comprising a computer readable medium and a computer program according to claim 18, wherein said computer program is included in said computer readable medium belonging to any of the group comprising: ROM (Read-Only Memory), PROM (Programmable ROM), EPROM (Erasable PROM), Flash memory, EEPROM (Electrically Erasable EPROM) and hard disk drive. A system (100) for estimating a deviation of at least one model variable of a catalyst model, said system being arranged to base said estimate on said catalyst model and on at least one measured sensor signal from at least one sensor (13 l), wherein said at least one sensor (131) is arranged to be brought into contact with exhaust gases downstream of a catalyst (11), characterized by a means arranged to compare a real signal with an estimated signal, said real signal depending on said measured sensor signal and wherein said estimated signal depends on at least one estimation function and constitutes the equivalent of the catalyst model to said actual signal, wherein said catalyst model is a tank series model, and - means arranged to use said comparison to determine at least one deviation parameter for said at least one estimation function, respectively, one of said at least one deviation parameter describes a relationship between a value for a model variable and a corresponding fair value.