DE10244125A1 - Method for evaluating the time behavior of a NOx sensor - Google Patents

Abstract

A method is described for evaluating the time behavior of a NOx sensor arranged in the exhaust tract of an internal combustion engine, which emits a measurement signal in which a change in concentration in the exhaust gas of the internal combustion engine is used, in which the average rate of change of the change in concentration is greater than an expected maximum rate of change of the measurement signal, and during the change in the measurement signal caused by the change in concentration at a first and a second point in time a first or second value of the measurement signal is recorded and from the first and the second value and the time interval between the first and the second A time constant of the change in the measurement signal is determined.

Description

The invention relates to a Procedure for evaluating the time behavior of a person in the exhaust system Internal combustion engine arranged gas sensor, which is a measurement signal emits at which a concentration change in the exhaust gas of the internal combustion engine is used.

The exhaust gas from internal combustion engines contains various pollutants, such as hydrocarbons, carbon monoxide and nitrogen oxides (NOx). The pollutant concentration in the exhaust gas depends heavily from operating parameters of the internal combustion engine, especially from Air / fuel ratio of the mixture with which the internal combustion engine is operated. Usually are aftertreatment systems in order to improve the exhaust gas quality Use which have one or more catalysts, which convert the pollutants. Operation of the internal combustion engine is then monitored by suitable sensors that sense the exhaust gas and using recourse controlled on the sensors. Typical sensors are e.g. Lambda probes for the Determination of the air / fuel ratio.

The fuel consumption of Otto internal combustion engines To reduce further, internal combustion engines come with lean combustion more frequently for use. With these internal combustion engines, special measures must be taken be taken to set emission limits to comply with NOx. Therefore come in such internal combustion engines reinforced NOx sensors are used to determine the concentration of nitrogen oxides in the Capture exhaust gas.

To ensure compliance with the required exhaust emission limit values via the ensure the entire useful life of an internal combustion engine, increasingly self-diagnosis (on-board diagnosis = OBD) of the entire exhaust gas for the treatment system. Such an OBD must be functional monitor exhaust-relevant components; also the functionality the gas sensors used must be checked. This review requires in particular the dynamic behavior of the sensors used regularly evaluate that a slowdown in the response of the sensor determined and inadequate Dynamic behavior of a faulty sensor can be diagnosed.

Depending on the location of the sensor, this can sometimes be problematic. For example, the exhaust gas downstream of a catalytic converter normally does not have large fluctuations in substance concentration, so that it is difficult to detect a slow response of a sensor that is located downstream of a catalytic converter in the exhaust gas tract of an internal combustion engine. From the DE 198 28 929 A1 In this regard, it is known to utilize a regeneration phase of a NOx storage catalytic converter in which the internal combustion engine is operated with a substoichiometric mixture in order to detect a rate of change in the measurement signal of a measurement sensor. If the rate of change remains below a minimum value, a faulty sensor is diagnosed.

This method has the disadvantage that the excitation of the NOx signal through a change the NOx concentration on the sensor must be known or determined Conditions are sufficient got to. However, it is during operation of an internal combustion engine difficult to always be able to comply with these framework conditions that either only a diagnosis of reduced quality can be carried out, or the frequency of performing diagnosis is severely restricted is.

The object of the invention is therefore to further develop a method for evaluating time behavior, that a more accurate diagnosis is possible is.

This object is achieved with a Method of the type mentioned solved in which the average rate of change the change in concentration is bigger as an expected maximum rate of change of the measurement signal, and during the change caused by the change in concentration of the measurement signal at a first and a second point in time first or second value of the measurement signal is acquired and from the first and the second value and the time interval between the first and second time determines a time constant of the change in the measurement signal becomes.

The method according to the invention thus determines no longer the rate of increase or mean change at an exactly predetermined one Exposure of the sensor to a change in substance concentration, but evaluates the sensor signal in such a way that the Time constant is determined. The change request the substance concentration is loosened to the extent that now only a certain minimum steepness of change is required. This Steepness must be greater than the expected maximum slope of the measurement signal. For the sensor the change in concentration arises then as a step function.

The absolute values, i.e. the level, on which the change are moved for the determination of the time constant is irrelevant. Determining the Time constant only presupposes that in known time Distance at least two values of the sensor signal can be sampled.

The calculation of the time constant from these two values, taking into account the time interval between the values, expediently uses model assumptions regarding the time behavior of the sensor. In most cases, a so-called PT1 characteristic will be assumed. In this case, the excitation, which is brought about by changing the size to be measured by an amplitude A, describes the time profile of the signal S using the following equation

In this equation, T denotes the time constant of the sensor and S0 the initial value of the sensor signal with the change. Such a time behavior of a PT1 element enables in a simple manner a rising or falling edge, i.e. with a decrease or enlarging the concentration to be measured, which is the condition according to the invention Fulfills, after simple connections to calculate the time constant T.

In the event of a decrease in concentration receives one with a signal curve according to equation (1) the time constant T by the time interval between the measured values the difference between the logarithms of the measured values is divided. Consequently can easily calculate the time constant T from the measured values and the known distance can be calculated. Of course, instead of Difference of the logarithms of the logarithm of the quotient from the Both measured values are used, since it is known that the Logarithm of a quotient of two values is synonymous with the difference of the logarithmic values.

The calculation of the time constant can be computationally gentle especially with regard to the performance recourse on a suitable map in which, starting from a predetermined time interval with which the two values of the measurement signal be won for corresponding pairs of values the time constant is stored. Conveniently, such a map between changes with increasing and decreasing substance concentration.

It has, however, surprisingly emphasized that even with relatively simple processes mathematical calculation of the time constant is possible.

With such an approach Calculation possible could not be expected a priori, since with the amplitude A and the initial value S0 and the time constant T three unknowns in the equation mentioned are given, so that two measurements are not actually one successful resolution can expect the equation according to the sought time constant T.

However, the invention makes Take advantage of the knowledge that equation (1) is simplified in special cases. For "switching on" a substance concentration, i.e. a rising edge, the start value S0 can be set to zero become. something similar applies to the "turn off"; then the amplitude A is zero. Equation (1) is then simplified accordingly.

The signal from the gas sensor is closed known at any time t. In contrast, the amplitude A is unknown, the starting value SO and the time constant T. You also have knowledge the time interval between the two times of the measurement. The time constant T that can be calculated in this way now provides a direct one Statement about the dynamic properties of the signal and thus the gas sensor. By cleverly linking the amplitude A of the excitation, i.e. the change in substance concentration, for the Calculation not needed, which a big independence in determining the time constant T.

The method according to the invention for evaluation the dynamics of a gas sensor is therefore independent of the amplitude A Concentration change, whereby the process can be used universally is. This property is particularly valuable for the diagnosis of gas sensors, which are located in an exhaust system downstream of a catalytic converter, since there are usually no quantitative changes in substance concentration are known or permanently adjustable.

The procedure is for linear Lambda probes are just as suitable as for diagnosing NOx sensors. For the method according to the invention enough it, the concentration change to know qualitatively. In this regard, there are various changes in concentration in question.

In an internal combustion engine in the exhaust tract of which a NOx storage catalytic converter is connected upstream of the NOx sensor, such NOx storage catalytic converters are generally operated discontinuously. In a storage phase, they adsorb NOx compounds from the exhaust gas, which are operated with an over-stoichiometric air / fuel mixture when the internal combustion engine is operating. The NOx storage catalytic converter, which can only be absorbed to a limited extent, is emptied in so-called regeneration phases in which the internal combustion engine is operated temporarily with a substoichiometric air / fuel mixture. The carbon monoxides and unburned hydrocarbons contained in the exhaust gas serve the NOx storage tank Talysator as a reducing agent. During such regeneration phases, NH3 also forms within the catalyst, but this is only an intermediate product and does not leave the catalyst. This only occurs when the regeneration phase has ended and there is no more NOx in the storage catalytic converter.

Then NH3 beats just like the exhaust gas components of the substoichiometric Air / fuel mixture to the catalyst outlet through. Therefore the internal combustion engine is now switched back to stoichiometric operation, which means that there is NOx in the exhaust gas again and the requirements for the emission of NH3 at the catalyst outlet are no longer present.

Since NOx sensors are generally cross-sensitive against NH3, they register the occurring reduction in the NH3 concentration, so that the desired one Signal change results, which is evaluated to determine the time constant.

For the inventive method it is therefore possible utilize a decrease in an NH3 concentration that arises if the internal combustion engine from a regeneration operation with substoichiometric Mixture in a company with stoichiometric Mixture is switched.

Another way to change the concentration you want is to decrease the The focus is to utilize a fuel cut-off at the Activate internal combustion engine. With overrun cut-off, the combustion chambers no more fuel is supplied to the internal combustion engine, but instead just air. The air is drawn into the engine Exhaust system pumped. Because air has no or only negligibly little Contains concentrations of NOx or NH3, a drastic Decrease in NOx or NH3 in the exhaust gas. This decrease in substance concentration then serves to diagnose the sensor dynamics.

wobei T die Zeitkonstante und d der zeitliche Abstand ist und Z durch folgende Gleichung gegeben ist, in der y1 den ersten, y2 den zweiten und y3 den dritten Wert bezeichnen,

wobei nur Werte für Z verwendet werden, die größer 0 und kleiner 1 sind.Of course, a change in concentration with an increase in concentration can also be evaluated for diagnosis. In this case, if you want to carry out a calculation of the time constant on an analytical basis, it is also expedient to assume a PT1 time behavior of the NOx sensor. In this case, it is preferred that a third value of the measurement signal is acquired at a third point in time, the first and second as well as second and third points in time each being at the same time interval from one another, using an increase in concentration as a change and the time constant according to the following formula is calculated:

where T is the time constant and d is the time interval and Z is given by the following equation, in which y1 denotes the first value, y2 the second and y3 the third value,

where only values for Z that are greater than 0 and less than 1 are used.

wobei T die Zeitkonstante, d der zeitliche Abstand, y1 der erste, y2 der zweite und y3 der dritte Wert ist und wobei der zeitliche Abstand d kleiner als ein Erwartungswert für die Zeitkonstante T ist.In a computing-saving approach, by approximating the e-function in equation (1) by means of difference equations, the time constant can also be determined when the concentration rises, by detecting a third value of the measurement signal at a third point in time, the first and second as well as second and third point in time are each at the same time interval, an increase in concentration is used as a change and the time constant is calculated using the following formula:

where T is the time constant, d the time interval, y1 the first, y2 the second and y3 the third value, and wherein the time interval d is less than an expected value for the time constant T.

As NOx sensors, as usual, in can generally have a strong cross sensitivity to NH3 an NH3 concentration change can also be advantageously used. A increasing NH3 concentration can e.g. generated by the internal combustion engine at short notice with a substoichiometric air / fuel mixture is operated. Such an operation is usually always there before when a full load enrichment is turned on, so it such a full-load enrichment phase for diagnosis is preferred to use. something similar applies to the transition an internal combustion engine into a catalyst protecting Operation or for the rinsing of a catalytic converter after overrun fuel cut-off.

Of course, an increasing change in the NOx concentration can also be utilized by switching the internal combustion engine from stoichiometric to superstoichiometric operation. However, such a measure usually only leads to an increasing NOx concentration at a NOx sensor. if no NOx storage catalytic converter is connected upstream of this, since such a catalytic converter would otherwise absorb NOx and thus no change in concentration would pass through to the NOx sensor.

Basically, the required changes in concentration can be achieved through targeted intervention in the operation of an internal combustion engine after initiating a Diagnostic function are forced. Alternatively, it is possible to wait until a corresponding one during normal operation of the internal combustion engine Concentration change event occurs, e.g. with sudden Power requirement that is a full load requirement with switching in substoichiometric Operation results in prolonged Downhill descent with overrun cut-off etc.

Another advantage of the method according to the invention lies in the possibility the calculated result, namely to check the time constant T, or over several calculation processes to average. A check or Recalculation and averaging immediately after the first is available Result, i.e. within the same change in concentration.

For example, the second value a first determination of the time constant T as the first value for one second provision can be used. The repetition of the determination the time constant and the averaging of the calculated time constant values elevated the accuracy of the result and effectively protects against misdiagnosis. The method according to the invention is then particularly robust.

The invention is described below Reference to the drawings will be explained in more detail by way of example. In the drawings show:

1 a schematic representation of an internal combustion engine 1 and

2 Time course of a substance concentration in the exhaust gas and a signal from a NOx sensor.

In the 1 illustrated internal combustion engine 1 has an exhaust tract 2 in which an exhaust after-treatment system that has OBD capabilities is arranged. The OBD function is controlled by a control unit 3 taken over, that also for the normal control of the operation of the internal combustion engine 1 responsible is. The control unit 3 controls an injection system, among other things 6 to that of the internal combustion engine 1 Allocated fuel reads the values of the probe 7 and the sensor 8th and executes the OBD.

In the exhaust system 2 the internal combustion engine is a three-way pre-catalyst 4 and a NOx storage catalyst 5 arranged. A single catalyst that shows both properties is also possible. A pre-cat lambda probe is located upstream of these two catalysts 7 and downstream of it a NOx sensor 8th ,

The control unit checks the dynamic behavior of the NOx sensor 8th as follows:
The check assumes that when the NOx sensor is excited by a change in the measured exhaust gas concentration, the sensor signal approximately satisfies a PT1 characteristic as shown in equation (1) above.

2 shows in a curve 9 the course of the excitation, the change in an NH3 concentration being plotted here, for example. Curve 10 shows the measurement signal S of the NOx sensor which satisfies the PT1 characteristic 8th ,

The increasing NH3 concentration is generated by briefly switching the internal combustion engine into substoichiometric operation in a regeneration phase. NH3 is generated in such a regeneration phase in the NOx storage catalytic converter 5 as an intermediate product, so long not at the outlet of the NOx storage catalytic converter 5 emerges as stored in this NOx. Is the NOx storage catalytic converter 5 emptied of NOx, NH3 penetrates to the exit and reaches the NOx sensor with a rapidly increasing concentration 8th , This is the rising edge at time t0 of the concentration of NH3 as it curves 9 shows.

The control unit 3 recognizes the end of the NOx regeneration phase in a manner known to the person skilled in the art and switches the operation of the internal combustion engine 1 back to substoichiometric air / fuel mixture. This means that there is no longer any NH3 in the NOx storage catalytic converter 5 , whereby the NOx sensor at time t1 8th has to detect a strongly decreasing concentration of NH3.

The signal S of the NOx sensor 8th that the control unit 3 is added, reflects this suggestion and follows the curve 10 , At time t0, the concentration increases from a start value S (t0) to a maximum value S (t1), which is reached at time t1. The curve falls due to the then falling concentration of NH3, which jumps back from its maximum value NH3 (t1) to the initial value NH3 (t0) 10 the measurement signal of the NOx sensor 8th also off again.

The control unit 3 now detects the signal S at two points in time after the concentration change has been initiated. This takes place at points in time t2 and t3, so that values S (t2) and S (t3) are present. The times t2 and t3 are spaced in time by a distance d. The time constant is then calculated from these values using the following equation: T = d / ln (S (t2) / S (t3)). (2)

The time constant T is now through a simple calculation process can be calculated.

Equation (2) is based on a suitable transformation and dissolution of equation (1). Naturally Instead of this, a suitable numerical method for evaluating equation (1) can also be used based on an analytical solution of equation (2) for the times t2 and t3 based calculation of the time constant T.

In an alternative approach, three values S (t4), S (t5), S (t6) are recorded at equidistant times t4, t5, t6 with an increasing change in concentration, and the time constant then results from the following equation:

The time interval d = t6 - t5 = t5 - t4 becomes chosen small against the expected time constant, so that applies d <T. In particular, a value for d used, the 10d <T enough.

wobei

ist. Hierbei werden natürlich Werte mit Z ≤ 0 oder Z ≥ 1 verworfen.If this condition for the time interval d cannot be met, the following equation enables the time constant to be calculated:

in which

is. Of course, values with Z ≤ 0 or Z ≥ 1 are rejected.

Claims (12)

Procedure for evaluating the time behavior of an in the exhaust system ( 2 ) an internal combustion engine ( 1 ) arranged gas sensor ( 8th ), which emits a measurement signal in which a concentration change in the exhaust gas of the internal combustion engine is utilized, characterized in that - an average rate of change of the concentration change is greater than an expected maximum rate of change of the measurement signal, - during that due to the change in concentration caused change in the measurement signal at a first and a second time, a first or second value of the measurement signal is detected and - a time constant of the change in the measurement signal is determined from the first and the second value and the time interval between the first and second time. A method according to claim 1, characterized in that as a change a decrease in concentration used and the temporal Distance by the difference of the logarithms of the first and the second Value of the measurement signal is divided. Method according to one of the above claims in an internal combustion engine ( 1 ), in the exhaust tract ( 2 ) the gas sensor ( 8th ) a NOx storage catalytic converter ( 5 ) upstream, characterized in that a reduction in an NH3 concentration is used by the internal combustion engine ( 1 ) is switched from a regeneration operation with substoichiometric mixture to an operation with stoichiometric mixture. Method according to one of claims 1 or 2, characterized in that that a decrease in concentration is used by a Thrust cutout is activated. A method according to claim 1, characterized in that a third value of the measurement signal is recorded at a third point in time, the first and second as well as second and third points in time each being at the same time interval from one another, using an increase in concentration as a change and the time constant after the following Formula is calculated:

where T is the time constant, d the time interval, y1 the first, y2 the second and y3 the third value, and wherein the time interval d is less than an expected value for the time constant T. A method according to claim 1, characterized in that a third value of the measurement signal is recorded at a third point in time, the first and second as well as second and third points in time each being at the same time interval from one another, using an increase in concentration as a change and the time constant after the following Formula is calculated:

where T is the time constant and d is the time interval and Z is given by the following equation, in which y1 denotes the first value, y2 the second and y3 the third value

where only values for Z that are greater than 0 and less than 1 are used. A method according to claim 1, 5 or 6, characterized in that an increase in an NH3 concentration is used as a change in which the internal combustion engine ( 1 ) is switched in operation with substoichiometric mixture. A method according to claim 7, characterized in that a full load enrichment is switched on to the internal combustion engine ( 1 ) to switch to operation with substoichiometric mixture. Method according to Claim 1, 5 or 6, characterized in that an increase in a NOx concentration is used as the change by the internal combustion engine ( 1 ) is switched from stoichiometric to superstoichiometric operation. Method according to one of the above claims, characterized in that for the evaluation an already in normal operation of the internal combustion engine ( 1 ) generated change in concentration is used. Method according to one of the above claims, characterized characterized that the time constant (T) is determined repeatedly and is averaged. A method according to claim 10, characterized in that the repeated determination within the same concentration change carried out becomes.