CN102817684A - Diagnostic method and diagnostic module for filter of nitrogen oxide sensor of exhaust system - Google Patents

Abstract

The invention discloses a diagnostic method for a filter (5) of nitrogen oxide sensor (4) of an exhaust system (1) having an SCR catalytic converter (3), including: considering a working pattern (11) of the filter (5); caculating (12) an NOx estimated value of a downstream of the catalytic converter (3); building a linear NOx model (13) based on the NOx estimated value; comparing (15) the NOx estimated value and the liner NOx model; and stoping using (18, 19) the filter (5) when a comparsion differential value is below a threshold.

Description

The diagnostic method and the diagnostic module that are used for the filter of vent systems NOx sensor

Technical field

The present invention relates to be used for a kind of diagnostic method and a kind of diagnostic module of the filter of vent systems NOx sensor, in particular for a kind of vent systems of the explosive motor of motor vehicle with SCR catalyst converter.

Background technique

System with SCR (SCR) catalyst converter is applicable to well removes the NOx effulent, especially when diesel engine exhaust.Ammonia (NH feeds intake in the SCR catalyst converter in active SCR system ₃), this ammonia be adsorbed on the catalyzer and with come self-purging NO and NO ₂React.NH ₃Typically directly do not feed intake, but in most of the cases feed intake with a kind of form of urea liquid, this solution partly is converted into NH after injecting ₃In passive/passive SCR system, not initiatively at the upper reaches of this catalyst converter/there is the seedbed to inject NH ₃Or urea.The substitute is, from another upstream component generation NH of this catalyst converter ₃, the LNT (rare NOx catcher) when motor is in rich combustion working state for example.

As the NH that feeds intake or in catalyst converter, store ₃When inadequate, perhaps when temperature is not suitable for the conversion fully of NOx, can't transforms whole Nox and cause Nox to reveal through catalyst converter.If yet the NH that feeds intake ₃Too much and its be stored in the catalyst converter, desorption possibly appear so.Usually NH appears in (for example after engine loading raises) after being rapidly heated ₃Desorption.But at NH ₃Feeding intake when excessive, this also possibly appear under the stationary temperature.

In order accurately to control the process in the SCR catalyst converter, about downstream or catalyst converter after NOx and NH ₃The information of concentration is significant.For this purpose, there is two types sensor, NH ₃Sensor and NOx sensor.NH ₃Sensor only is used for measuring N H ₃Concentration or amount, and the NOx sensor is to NOx and NH ₃All be sensitive.This causes difficulty in measurement, can't confirm simply that promptly this sensor is just detected to be NOx value or NH ₃Value.

A kind of filter in parallel application DE 10 2,011 077 246.4, put down in writing and filter method confirmed by the NOx sensor to signal be NOx measured value or NH ₃Measured value.For this reason, set up two different model or modeling method and/or it is mated.A model description conversion ratio of NOx in catalyst converter, and second model description NH ₃Leakage behavior through catalyst converter.

The degree of accuracy of this filter or efficient depend on the NOx level at the catalyst converter upper reaches or the characteristic of signal.In order can between the NH3 value in NOx value and the downstream NOx signal, to distinguish, the NOx signal at the upper reaches of this catalyst converter must have a kind of known dynamic characteristic.

In catalyst converter, take place under the situation of Nox leakage, the NOx signal will dynamically be followed the tracks of input NOx signal, and NH ₃Value is not typically done like this.Therefore, can't successfully confirm more constant NOx signal and the NOx/NH that is filtered at the upper reaches of catalyst converter (for example in the constant mode of going) ₃Signal should not be used for a kind of feedback control.

Summary of the invention

Basic purpose of the present invention is, improves the working method of the vent systems with catalyst converter.

This purpose is that the characteristic through claim 1 and 10 realizes.Define the favourable improvement project of the present invention in each dependent claims.

According to a first aspect of the invention, a kind of diagnostic method of filter of the vent systems NOx sensor that is used to have the SCR catalyst converter may further comprise the steps:

This filter of-operation comprises the NOx estimated value that calculates the catalyst converter downstream;

-based on this NOx estimated value, set up a LINEAR N Ox model;

-NOx estimated value and LINEAR N Ox model are compared;

-when the difference of this comparison is under the threshold value, this filter of stopping using.

This method makes the filter that is used for the NOx sensor be able to stablize, and has confirmed when credible the decision-making of this filter is.An implementation of this filter does, continuously two different model mated and its accuracy is compared.A model is represented the dynamic model of a simplification of NOx conversion ratio.Another model is a kind of simple linear time algorithm, and this algorithm can be described NH well in the time window that limits ₃The slip behavior.In some cases, this filter can not be at NOx and NH ₃Between distinguish.This typically occurs under the situation of quiescent operation for example, when the NOx signal does not have enough dynamic characteristics.In this case, the NOx in catalyst converter downstream describes through a linear time algorithm equally.Through confirming of said condition, can avoid this filter to carry out wrong supposition or the decision-making that makes the mistake.For example, if the absolute difference between these two comparison signals is lower than a threshold value, this filter of can stopping using, or abandon or do not use its decision-making.

In order to compare, can between a LINEAR N Ox estimated value of this NOx estimated value and this LINEAR N Ox model, carry out Error Calculation.The execution of this Error Calculation is simple, and the rapid reliable result is provided.

When difference relatively surpasses a threshold value, can this filter be opened.This threshold value can be corresponding with that threshold value that is used to stop using, and perhaps possibly adopt two different threshold values.Open or launch and be meant, use the decision-making of this filter.When inactive, the still operation simultaneously in a known way of this filter on the backstage, but its result can not be used.Alternatively, can filter be broken off connection, and when opening, reclose.

When comparing with threshold value, this filter can keep starting state in the time-domain of a qualification.This mode of operation can be called as transient state and start.What in transient state starts, supposed is: if do not move to as yet outside the boundary of qualification having made a decision-making and NOx sensor signal not long ago; " old " filter decision-making so; That is to say before the time-domain of when beginning or qualification, is still effective.Usually, the NOx in catalyst converter downstream and NH ₃Can time changing, even can't make a policy according to current data like this, the decision-making of being made not long ago still possibly be effective.It is not do that the decision-making, the filter that make new advances are still unlatching as yet and use its result to feed back or further handle that transient state starts what supposed.Transient state is enabled in after the preset time section, or when above or will be deactivated when being lower than a predetermined NOx sensor signal boundary.

Under the situation that filter is stopped using, can introduce the variation of the NOx value in catalyst converter downstream.So in this case; When filter can not be made reliable decision-making; Can set a sign (Flag) or information for (from motor) NOx disturbance, perhaps it started reliably so that activate the basic algorithm of this filter again.When filter was stopped using, it did not provide any signal, no matter is NOx or NH ₃In this case, to the feeding intake or inject and be under the situation that does not have feedback, to carry out of catalyst converter, just in the regulating loop (Regelkreis) of an opening, carry out.As long as the NOx sensor signal still is in the boundary known, that allow, it is acceptable that Here it is.Yet, when crossing this boundary, possibly it is desirable for and confirm by the strong hand or calculating NOx sensor signal.This can be enough through disturbance pointedly or change NOx ratio in the exhaust, that leave from motor and realize, for example through completely or partially connecting and/or breaking off exhaust-gas-recirculation valve (EGR).This situation can for example occur in long-term quiescent operation mode, for example through a speed adjusting device (speed regulator).In case obtained this working method and therefore this NOx signal obtained a kind of known dynamic characteristic, then this filter just begins proper functioning automatically.

For the NOx conversion ratio is carried out modeling, can adopt a dynamic model of NOx conversion ratio, preferably can be the dynamic model of a simplification, so that reduce calculation consumption and data processing afterwards.

These steps or calculate to put to a plurality of sampling times and implement, it has sampling time of being preferably one second.The quantity of sampling time point is crossed the young pathbreaker can not obtain consistent model, and excessive quantity will unnecessarily improve calculation consumption.

In order to calculate, these steps can be employed in the value of the sampled point of recalling on the time extraly, preferably within 20 seconds time range.The reliability that this has improved this calculating and has confirmed, its measured value directly records.In order to compare, can after with the mean absolute difference between two signals as criterion.

According to a second aspect of the invention; A kind of diagnostic module of filter of the vent systems NOx sensor that is used to have the SCR catalyst converter, it comprises that one is used for the signal input part of filter signal, a computing unit and a signal output part that is used to export this diagnostic method result that is used to carry out above-mentioned diagnostic method.Advantage mentioned above is effectively same with modification.

Description of drawings

To be explained in more detail the present invention by accompanying drawing below, wherein:

Fig. 1 shows the schematic block diagram of a vent systems, and this vent systems has according to diagnostic module of the present invention.

Fig. 2 shows the block diagram that is used for a kind of diagnostic method of filter according to the present invention.

Fig. 3 and 4 is respectively the chart according to the application of diagnostic method of the present invention.

These diagrammatic sketch only are used to explain the present invention but not limit it.These diagrammatic sketch need not be pro rata with each discrete part.The same identical or similar part of reference mark representative.

Embodiment

Fig. 1 shows the part of an exhaust after treatment system 1 that is used for internal-combustion engine 2 (the for example DENG of motor vehicle).This exhaust after treatment system or vent systems 1 comprise a catalyst converter 3 that is placed in motor 2 downstream, for example a SCR (SCR) catalyst converter.

The arranged downstream of catalyst converter 3 has a NOx sensor 4, is used for measuring the concentration of exhaust NOx or the amount of NOx.NOx sensor 4 does not need directly to be placed in after the catalyst converter 3, and it can also be in more downstream part.

The measured value of NOx sensor 4 is sent to the filter or the NH of this NOx sensor 4 ₃Filter 5.Filter 5 is handled this measured value and is made it can recognize that by NOx sensor 4 measured values be NOx value or NH ₃Value.The corresponding signal of filter 5 outputs is for example exported to an engine controller or exhaust aftertreatment controller.

In the downstream of catalyst converter 3, except the measured value that is provided by NOx sensor 4, filter 5 also needs the amount of NOx concentration or NOx, and described concentration or amount are produced and on the direction of catalyst converter 3, left by motor 2.Have an extra NOx sensor 6 in the downstream of catalyst converter 3, or confirm the NOx concentration in catalyst converter 3 downstream by a model 6, this model has confirmed that the NOx that motor 2 is produced measures and/or concentration.

Between catalyst converter 3 downstream, optional NOx sensors 6 and catalyst converter 3, dispose choose wantonly equally, be used for urea liquid or NH ₃A decanting point 7.NOx sensor 6 or this modeling effect are to be arranged in the downstream of decanting point 7, with the cross-sensitivity (Querempfindlichkeit) of avoiding NH3.

The output terminal 5b of filter 5 links to each other with a diagnostic module 8.Diagnostic module 8 can be an independently unit, and this unit has one and is used for the signal input part 8a of the signal of filter 5, a computing unit and a signal output part 8b who is used to export this methods and results who is used to carry out diagnostic method.Diagnostic module 8 also can be the constituent element of a controller, for example the controller of engine controller or exhaust aftertreatment controller or filter 5.It can be realized with the form of software program or software module.Signal output part 8b can realize in software equally.Result through signal output part 8b output can be an independently value or checking among the result of filter 5, and/or is the filter 5 (freigegebene) that discharged or the result of release not.

Now the working method of diagnostic module 8 and this diagnostic method is described by Fig. 2.NOx sensor 4 beginnings of this diagnostic method after catalyst converter 3.

The process that is schematically shown among Fig. 2 has phase I or the step 10 of setting up a LINEAR N Ox model.In first step or substep 11, move this filter 5, wherein in second step 11, the estimation NOx value NOxPostEstV in catalyst converter 3 downstream is confirmed as a result of filter 5.

In order to carry out the NOx modeling, adopted 0DCSTR (continuously stirred tank reactor (CSTR)) model with dynamics first progression at this.

$\mathrm{NOxPostEstV}=\frac{\mathrm{NOxPreV}}{1+\frac{k_R}{\mathrm{ExhFlowV}}}$

These signals can be scalar (surveying or sampling) or vector (time window with intelligence operation element).For time window, suppose reactive rate coefficient (k _R) be constant.Optimum reactive rate coefficient can reduce method (Least-Squares-Abstiegsverfahren) according to least square by analytical method and confirm:

$\mathrm{\&epsiv;}=(\left[\begin{array}{c}\frac{1}{\mathrm{ExhFlo}{w}_1}\\ \frac{1}{\mathrm{<figure-callout\; id="2"\; label="ExhFlo\; w"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">ExhFlo</figure-callout>}{w}_{}{}_2}\\ ...\\ \frac{1}{\mathrm{ExhFlo}{w}_N}\end{array}\right]k_R-\left[\begin{array}{c}\frac{\mathrm{NOxPr}{e}_1}{\mathrm{NOxPos}{t}_1}\\ \frac{\mathrm{<figure-callout\; id="2"\; label="NOxPr\; e"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPr</figure-callout>}{e}_{}{}_2}{\mathrm{<figure-callout\; id="2"\; label="NOxPos\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPos</figure-callout>}{t}_{}{}_2}\\ ....\\ \frac{\mathrm{NOxPr}{e}_N}{\mathrm{NOxPos}{t}_N}\end{array}\right]+1)=({\mathrm{\&theta;k}}_R-y)$

Wherein

$\mathrm{\&theta;}=\left[\begin{array}{c}\frac{1}{\mathrm{ExhFlo}{w}_1}\\ \frac{1}{\mathrm{<figure-callout\; id="2"\; label="ExhFlo\; w"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">ExhFlo</figure-callout>}{w}_{}{}_2}\\ ...\\ \frac{1}{\mathrm{ExhFol}{w}_N}\end{array}\right],$ $y=1-\left[\begin{array}{c}\frac{\mathrm{NOxPr}{e}_1}{\mathrm{NOxPos}{t}_1}\\ \frac{\mathrm{<figure-callout\; id="2"\; label="NOxPr\; e"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPr</figure-callout>}{e}_{}{}_2}{\mathrm{<figure-callout\; id="2"\; label="NOxPos\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPos</figure-callout>}{t}_{}{}_2}\\ ...\\ \frac{\mathrm{NOxPr}{e}_N}{\mathrm{NOxPos}{t}_N}\end{array}\right]$

Optimum criterion min (ε ^Tε)

Make:

k _R＝(θ ^Tθ) ^-1θ ^Ty

In following step 13, NoxPostEstV makes up LINEAR N Ox model based on this NOx estimated value.A result of this LINEAR N Ox model is that a LINEAR N Ox estimated value NoxLinPostEstV in catalyst converter 3 downstream is available in step 14.

The linear time algorithm (NoxPostEstV-N vector (for example 20 sampled values)) that is used for based on the LINEAR N Ox model of NOx model signals calculates as follows:

NO _xLinPostEstV＝an+b

Wherein a and b are suitable constants, and n is to the number of samples (time horizon line (Zeithorizont)-1) from 0.

Counting criterion through least square minimizes and calculates a and b:

$\mathrm{\&epsiv;}=(\left[\begin{array}{cc}0& 1\\ 1& 1\\ 2& 1\\ ...& \\ N-1& 1\end{array}\right]\left[\begin{array}{c}a\\ b\end{array}\right]-\left[\begin{array}{c}\mathrm{NOxPostEs}{t}_1\\ \mathrm{<figure-callout\; id="2"\; label="NOxPostEs\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPostEs</figure-callout>}{t}_{}{}_2\\ \mathrm{<figure-callout\; id="3"\; label="NOxPostEs\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPostEs</figure-callout>}{t}_{}{}_3\\ ....\\ \mathrm{NOxPostEs}{t}_N\end{array}\right])=(\mathrm{\&theta;}\left[\begin{array}{c}a\\ b\end{array}\right]-y)$

Wherein

$\mathrm{\&theta;}=\left[\begin{array}{cc}0& 1\\ 1& 1\\ 2& 1\\ ...& \\ N-1& 1\end{array}\right],$ $y=\left[\begin{array}{c}\mathrm{NOxPostEs}{t}_1\\ \mathrm{<figure-callout\; id="2"\; label="NOxPostEs\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPostEs</figure-callout>}{t}_{}{}_2\\ \mathrm{<figure-callout\; id="3"\; label="NOxPostEs\; t"\; filenames="HDA00001736777000011.png"\; state="\{\{state\}\}">NOxPostEs</figure-callout>}{t}_{}{}_3\\ ...\\ \mathrm{NOxPostEs}{t}_N\end{array}\right]$

Optimum criterion min (ε ^Tε)

Make:

$\left[\begin{array}{c}a\\ b\end{array}\right]={\left({\mathrm{\&theta;}}^T\mathrm{\&theta;}\right)}^{-1}{\mathrm{\&theta;}}^{T}y$

In second stage 15, NOx estimated value and LINEAR N Ox model or LINEAR N Ox estimated value are compared mutually.

This is through in step 16, between NoxLinPostEstV and NoxPostEstV, carrying out Error Calculation ε according to following formula _NOxLinObtained:

${\mathrm{\&epsiv;}}_{\mathrm{NOxLin}}=\frac{1}{N}\sqrt{{(\mathrm{NOxLinPostV}-\mathrm{NOxPostEstV})}^T(\mathrm{NOxLinPostV}-\mathrm{NOxPostEstV})}$

Error is more little, and the linearity of NOx estimation is good more, and that also just be difficult to definite these sensor 4 measurements thus more is NOx or NH ₃

In stage 17, confirm or confirmed that the NOx signal of sensor 4 is NOx or NH ₃Value.For this reason, in step 18 with error signal _NOxLinWith a predefined boundary or a threshold epsilon _MinCompare.Work as error signal _NOxLinLess than this boundary, then set a sign or signal, so that the decision-making of filter 5 can not implement or consider, perhaps in other words filter 5 is stopped using.Because each sampling time point is calculated a new error signal, so error signal maybe be also considered in decision-making hereto _NOxLinPrevious or preceding several values.

If error signal _NOxLinGreater than this boundary, then move or start filter 5 according to step 19.If non-this situation then can under so-called transient state starting state, make this filter 5 in a time-domain known, that limit in advance, still keep starting according to step 20.What in transient state is opened, supposed is: if made a decision-making at short notice and the NOx sensor signal does not move to outside the boundary of qualification as yet; " old " filter decision-making so; That is to say before the time-domain of when beginning or qualification, be still effectively.

Work as long as transient state starts, just start filter 5, that is to say and consider its result according to step 19.After a predetermined time period or when a NOx sensor signal surpasses or is lower than preestablished limit, transient state is started releasing and jumps to step 21.

In step 21, filter 5 is stopped using, that is to say its disconnection or do not consider its result.

When filter 5 was stopped using, it did not provide any signal, no matter is NOx or NH ₃In this case, to the feeding intake or inject and be under the situation that does not have feedback, to carry out of catalyst converter 3, just in the regulating loop of an opening, carry out.As long as the NOx sensor signal still is in the limit known, that allow, it is acceptable that Here it is.

Be in outside the limit known, that allow (this can be the deviation of forward or negative sense) if in step 22, confirmed the NOx sensor signal, then in step 23, can confirm or calculate the NOx sensor signal by the strong hand.This can or change NOx ratio in the exhaust, that leave from motor through disturbance pointedly realize, for example through completely or partially connecting and/or breaking off exhaust-gas-recirculation valve (EGR).

This method is accomplished in diagnostic module 8.By this method, or exported an effective NOx or NH ₃Value, or exported a signal, this signal provides the reliability of filter 5.Diagnostic module 8 is connected in signal chains between the controller (not shown) of NOx sensor 4, filter 5 and further processing.Corresponding ground, diagnostic module 8 can also be integrated in controller or NOx sensor 4 and the filter 5.

Fig. 3 and 4 shows a plurality of charts, and these charts have shown the working method of this method or diagnostic module 8.What wherein shown is that through this diagnostic method or this diagnostic module 8, working method vent systems 1 or filter 5 becomes sane.

Fig. 3 shows the circulation of going dynamically, wherein can be well to NOx and NH ₃Value is separated.The ε of LINEAR N Ox model _NOxLinError is still higher most of the time.The signal " number filter _ launch " that expression filter 5 starts (filter_enabled), starts most of the time.In two charts below Fig. 3, demonstrated the NOx and the NH in catalyst converter 3 downstream that obtained ₃Value and the actual value that obtains through discrete surveying instrument or sensor, and can see that these values meet well each other.That is to say that it is reliable carrying out work by this method.

Fig. 4 has shown a kind of static more working method, and wherein this diagnostic method has been confirmed: can't make reliable decision-making and filter 5 is inactive.Therefore, on second time section of this chart, signal " filter _ launch " is that do not open or lower.So as stated, the feedback of having stopped using is started up to filter 5 once more.

What can confirm is, though the error ε of LINEAR N Ox model _NOxLinLower, but filter 5 is to keep starting in about 260 seconds of beginning.This has caused the following fact, in any case promptly the NOx sensor measurement low-down value and therefore all the result of filter 5 being calibrated.

This method or diagnostic module 8 can be applied in each has the vent systems of catalyst converter, especially (passive/passive or active/active formula) SCR catalyst converter.

Claims (10)

1. the diagnostic method of the filter (5) of the NOx sensor (4) of a vent systems (1) that is used to have SCR catalyst converter (3), this method has following steps:

Move (11) said filter (5), comprise the NOx estimated value in the said catalyst converter of calculating (12) (3) downstream;

Set up LINEAR N Ox model (13) based on said NOx estimated value;

Said NOx estimated value and said LINEAR N Ox model are compared (15);

When the difference of said comparison (15) is under the threshold value, (18,19) the said filter (5) of stopping using.

2. diagnostic method according to claim 1 wherein in order to carry out said comparison (15), carries out Error Calculation (16) between a LINEAR N Ox estimated value of said NOx estimated value and said LINEAR N Ox model.

3. diagnostic method according to claim 1 and 2 wherein starts said filter (5) when the difference of said comparison (15) surpasses threshold value.

4. according to any one described diagnostic method in the claim 1 to 3, wherein surpass under the situation of said threshold value, make said filter (5) keep starting in the section at the fixed time at comparative result.

5. according to any one described diagnostic method in the claim 1 to 4, wherein under the situation of said filter (5) of stopping using, introduce a variation (23) to the NOx value in said catalyst converter downstream.

6. according to any one described diagnostic method in the claim 1 to 5, wherein, adopted a dynamic model of NOx conversion ratio in order to calculate (12) said NOx estimated value.

7. according to any one described diagnostic method in the claim 1 to 6, wherein, adopted a kind of linear time algorithm for said LINEAR N Ox model is carried out modeling (13).

8. according to any one described diagnostic method in the claim 1 to 7, wherein these steps are implemented at the sampling time point, and the sampling time is preferably a second.

9. diagnostic method according to claim 8 wherein for these steps, has adopted the value of a plurality of sampled points of recalling in time, preferably within 20 seconds time range.

10. the diagnostic module of the filter (5) of the NOx sensor (4) of a vent systems (1) that is used to have SCR catalyst converter (3), said diagnostic module have a signal input part (8a) that is used for filter (5) signal, one and are used to carry out the signal output part (8b) that computing unit and according to the described diagnostic method of claim 1 to 9 are used to export said diagnostic method result.

Images