Exhaust gas purifying apparatus for an internal-combustion engine

Abstract

An exhaust gas purifying apparatus for an internal-combustion engine, including a three-way catalyst, nitrogen oxide purifier that is disposed downstream of the three-way catalyst, an upstream oxygen sensor that is disposed between the three-way catalyst and the nitrogen oxide purifier, and a downstream oxygen sensor that is disposed downstream of the nitrogen oxide purifier. An abnormality determination criterion is set in accordance with the exhaust gas amount of the internal-combustion engine wherein the criterion is equivalent to a time period until the output of the downstream oxygen density sensor is read since the output of the upstream oxygen density sensor is changed to rich after the air-fuel ratio of the exhaust gas was over from lean to rich. Any abnormality of the nitrogen oxide purifier is determined when the output of the downstream oxygen density sensor has been changed to rich before the time period represented by the criterion elapses.

Description

TECHNICAL FIELD

The present invention relates to an exhaust gas purifying apparatus for an internal-combustion engine and, more specifically, to an exhaust gas purifying apparatus for an internal-combustion engine that can accurately measure NOx trapping capability of a nitrogen oxide (NOx) catalyst.

BACKGROUND ART

Conventionally, there is a known technique for decreasing NOx in the exhaust gas by a NOx purifying device that is located in the exhaust system of the internal-combustion engine and incorporates a NOx trapping agent. Japanese Patent Application Unexamined Publication (Kokai) No. H10-299460 discloses a technique for determining deterioration of the NOx purifying device based on delay time of outputs of oxygen density sensors disposed upstream and downstream respectively of the NOx purifying device when the fuel control air-fuel ratio of the engine is set richer than a stoichiometric air-fuel ratio after lean-burn operation is made for a given time period. More specifically, in the conventional technique, deterioration of the NOx purifying device was determined based on a fact that the delay from the time the output of the upstream oxygen density sensor changed to rich to the time the output of the downstream oxygen density sensor changes to rich is relatively long when the NOx trapping capability of the NOx purifying device is high whereas such delay time becomes shorter as the NOx trapping capability degrades.

However, in such conventional techniques, there exists a problem that deterioration of the NOx purifying device cannot be accurately determined because the time period from the time the upstream oxygen density sensor changes to a rich state to the time the downstream oxygen density sensor detects a rich state varies as the air fuel ratio of the gas flowing into the NOx purifier varies at different speeds. Such variation is due to unstable air-fuel ratio of the exhaust gas flowing into the NOx purifying device, which may be caused by the influence of such factors as 1) variation of deterioration of the three-way catalyst disposed upstream of the NOx purifier, 2) variation of activation degree of the three-way catalyst disposed upstream of the NOx purifier and 3) variation of sulfur constituent contained in the fuel.

Thus, it is an objective of the present invention to provide an exhaust gas purifying apparatus that can accurately measure a NOx trapping capability of the NOx purifying device.

SUMMARY OF THE INVENTION

According to one aspect of the present invention, an exhaust gas purifying apparatus for an internal-combustion engine is provided. The exhaust gas purifying apparatus comprises a three-way catalyst that is disposed in an exhaust system of the internal-combustion engine and a nitrogen oxide purifier that is disposed downstream of the three-way catalyst for purifying or cleaning nitrogen oxide that is contained in the exhaust gas when an air fuel ratio of the exhaust gas of the engine is lean. The purifying apparatus further comprises an upstream oxygen sensor that is disposed between the three-way catalyst and the nitrogen oxide purifier, and a downstream oxygen sensor that is disposed downstream of the nitrogen oxide purifier. The purifying apparatus also comprises a correcting means for correcting a criterion for determining abnormality of the nitrogen oxide purifier. The criterion is used for determining abnormality of the NOx purifier based on a change in the output of the downstream oxygen density sensor when the air-fuel ratio of the exhaust gas is changed from lean to rich. Correction of the criterion is made according to the degree of change in the output of the upstream oxygen density sensor.

According to the invention, means is provided for correcting the criterion for determining abnormality of the nitrogen oxide purifier. The criterion is the delay time that the output of the downstream oxygen sensor changes from lean to rich after the output of the upstream oxygen sensor changes from lean to rich. The criterion is corrected in relation to the speed the output of the upstream oxygen sensor changes from lean to rich or the time the upstream oxygen sensor takes to change from lean to rich. This way, the influence of the upstream three-way catalyst on the oxygen density sensors can be excluded, and deterioration of the nitrogen oxide purifier can be accurately determined. In one aspect of the invention, correction of the criterion may include correction of reversal delay time of the downstream oxygen density sensor (that is, the delay time for the downstream oxygen density sensor to change from lean to rich after the upstream oxygen density sensor changes from lean to rich). Correction of the criterion may also include correction of thresholds for determination.

According to another aspect of the present invention, in the exhaust gas purifying apparatus for the engine, the abnormality determination criterion is defined in terms of a delay time from the time the output of the upstream oxygen density sensor reverses as the air fuel ratio is changed from lean to rich to the time the output of the downstream oxygen density sensor reverses. The nitrogen oxide purifier is determined abnormal when the output of the downstream oxygen density sensor reverses before the time period reaches the criterion.

According to one aspect of the invention, the influence of the upstream three-way catalyst upon the oxygen sensors can be eliminated. In one embodiment of the invention, the criterion is modified based on a cumulative intake air amount value GSLFFIN that is an air amount taken-in from the time the upstream oxygen sensor output SVO 2 exceeds a first upstream reference value SVO 2 LNCS, which indicates that SVO 2 started to rise, till the time SVO 2 reaches a second upstream reference value SVO 2 SLF, which SVO 2 would reach within a reasonable time if SOx density is low. When the engine revolution is stable, GSLFFIN corresponds to the time the upstream oxygen sensor took to reverse from a lean state to a rich state. In other words, GSLFFIN corresponds to the speed the upstream oxygen sensor responds to changing air fuel ratio.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a diagram showing an overall arrangement of an internal-combustion engine and its controller according to one embodiment of the present invention.

FIG. 2 is a flowchart showing a process for calculating a target air-fuel ratio coefficient KCMD.

FIG. 3 is a flowchart showing a main routine for a deterioration determination process upon the NOx purifying device.

FIG. 4 is a flowchart showing an execution condition satisfaction determination process.

FIG. 5 is a flowchart, continued from FIG. 4 , of the execution condition satisfaction determination process.

FIG. 6 is a flowchart showing a SOx density determination process.

FIG. 7 is a flowchart showing a deterioration determination pre-processing.

FIG. 8 is a flowchart showing an intake air amount accumulation process.

FIG. 9 is a flowchart showing a deterioration determination process.

FIG. 10 is a flowchart, continued from FIG. 9 , of the deterioration determination process.

FIG. 11 is a table to be used in the deterioration determination process.

FIG. 12 is a flowchart showing a SOx removal process.

FIG. 13 is a flowchart showing a process for determining the lean-burn operation prohibition.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

FIG. 1 shows an overall arrangement of an internal-combustion engine (hereinafter referred to as an “engine”) and its controller including a failure diagnostic device according to one embodiment of the present invention. A throttle valve 3 is disposed in the route of an air intake pipe 2 connected to an engine 1 . The throttle valve 3 is connected to a throttle valve opening degree (THA) sensor 4 . An electric signal that represents an opening degree of the throttle valve 3 is sent from the sensor 4 to an electronic control unit (hereinafter referred to as “ECU”) 5 . The structure of the ECU 5 will be described hereinafter.

A fuel injection valve 6 is provided, for each cylinder, between the engine 1 and the throttle valve 3 slightly upstream of the air intake valve (not shown) of the engine 1 . An absolute air-intake-pipe internal pressure (PBA) sensor 8 and an intake air temperature (TA) sensor 9 are connected to the air intake pipe 2 , so as to detect an absolute pressure and an intake air temperature respectively to provide them to the ECU 5 in the form of electric signals. An engine water temperature (TW) sensor 10 , which is mounted on the main body of the engine 1 , comprises a thermistor and the like. The sensor 10 detects an engine water temperature (cooling water temperature) TW and sends a corresponding electric signal to the ECU 5 .

An engine revolution (NE) sensor 11 and a cylinder identification (CYL) sensor 12 are provided in the peripheries of the camshaft or the crankshaft (not shown) of the engine 1 . The engine revolution sensor 11 outputs a TDC signal pulse at every top dead center point (TDC) when each cylinder of the engine 1 begins its intake stroke. The cylinder identification sensor 12 outputs a cylinder identification signal pulse at a predetermined crank angle for a specific cylinder. Those signal pulses are transmitted to the ECU 5 .

A three-way catalyst 14 and a NOx purifier 15 or a NOx cleaner are disposed in an exhaust pipe 13 . The three-way catalyst 14 is positioned upstream of the NOx purifier 15 . The three-way catalyst has a function of accumulating O 2 contained in the exhaust gas in an exhaust lean condition in which the air-fuel ratio of the mixture to be supplied to the engine 1 is leaner than a stoichiometric air-fuel ratio and the density of the O 2 in the exhaust gas is comparatively high. In contrast, the catalyst oxidizes the HC and the CO contained in the exhaust gas using thus accumulated O 2 in an exhaust rich condition in which the air-fuel ratio of the mixture to be supplied to the engine 1 is richer than the stoichiometric air-fuel ratio, the density of the O 2 contained in the exhaust gas being low and the density of the HC, CO constituents contained in the exhaust gas being high.

The NOx purifier or NOx cleaner 15 incorporates a NOx trapping agent for trapping NOx and a catalyst for promoting oxidization and reduction. The NOx trapping agent traps the NOx in the exhaust lean condition in which the air-fuel ratio of the mixture to be supplied to the engine 1 is leaner than the stoichiometric air-fuel ratio. On the other hand, around the stoichiometric air-fuel ratio or in the exhaust rich condition in which the air-fuel ratio is richer than the stoichiometric air-fuel ratio, the trapped NOx is reduced by HC and CO and discharged in the form of nitrogen gas, while concurrently the HC and CO are oxidized and discharged in the form of steam and carbon dioxide.

When the trapping of NOx continues to reach an trapping capability limit of the NOx trapping agent, i.e., when the maximum NOx trapping volume is reached, no further NOx can be trapped. In such a case, the air-fuel ratio must be set richer in order to reduce and discharge the NOx. This operation is called a reduction-enrichment operation.

A linear oxygen density sensor (hereinafter referred to as an “LAF sensor”) 17 is disposed upstream of the three-way catalyst 14 . The LAF sensor 17 sends to the ECU 5 an electric signal that is substantially proportional to the oxygen density (air-fuel ratio) of the exhaust gas.

A binary type oxygen density sensor (hereinafter referred to as “O 2 sensor”) 18 is disposed between the three-way catalyst 14 and the NOx purifying device 15 and another binary type O 2 sensor 19 is disposed downstream of the NOx purifying device 15 . Signals detected by these sensors are transmitted to the ECU 5 .

The O 2 sensors 18 , 19 have such characteristic that their outputs switch in a binary manner around the stoichiometric air-fuel ratio. That is, the output takes a high level on the rich side and takes a low level on the lean side. In the following description, the O 2 sensor 18 and the O 2 sensor 19 will be referred to as the “upstream O 2 sensor” 18 and the “downstream O 2 sensor” 19 respectively.

The engine 1 has a valve timing switch mechanism 20 that can alternately set the valve timings for the air intake valve and the air exhaust valve at two levels, one being a quick valve timing that is appropriate for a rapidly rotating region of the engine, and the other being a slow valve timing that is appropriate for a slowly rotating region. The switching of the valve timing includes the switching of the distance lifted by the valves. Further, when the slow valve timing is selected, one of the two valves is halted in order to ensure stable combustion, even when the air-fuel ratio is leaner than the stoichiometric air-fuel ratio.

Additionally, an atmospheric pressure sensor 21 for detecting the atmospheric pressure (PA) is connected to the ECU 5 . The detected signal is supplied to the ECU 5 .

The ECU 5 includes a ROM for storing programs and data, and a RAM for providing an operational workspace to store/retrieve programs and data required at the runtime. It also includes a CPU for executing programs and an input interface for processing input signals from various sensors and a drive circuit for sending control signals to such engine sections as the fuel injection valve 6 . All outputs from the various sensors are received by the input interface and are processed in accordance with the programs stored in the ROM. With such a hardware structure, functional blocks in FIG. 1 represent ECU 5 .

The ECU 5 includes functional blocks of an operating conditions detector 22 , an abnormality determination criterion correcting means 24 , abnormality determining means 26 , an air-fuel ratio setting means 27 and a fuel injection controller 28 .

The operating conditions detector 22 detects various engine operating conditions based on the above-described various engine parameter signals. The abnormality determination criterion correcting means 24 corrects an abnormality determination criterion for the NOx purifying device 15 based on a change in the outputs of the downstream O 2 sensor 19 correspondingly to a degree of change in the outputs of the upstream O 2 sensor 18 when the air-fuel ratio of the exhaust gas is turned over from lean to rich. The abnormality determining means 26 determines that the NOx purifying device 15 is abnormal when the output of the downstream O 2 sensor 19 has been changed to rich before the time period corresponding to the abnormality determination criterion elapses.

The air-fuel ratio setting means 27 sets a target air-fuel ratio in accordance with operating conditions detected by the operating conditions detector 22 . The fuel injection controller 28 calculates the fuel injection time TOUT of the fuel injection valve 6 which is opened in synchronization with the TDC signal pulse according to the following equation (1), so as to control the fuel injection valve 6 .

TOUT=TIM×KCMD×KLAF×KPA×K 1 + K 2   (1)

In the equation (1), TIM represents a base fuel amount, or, more specifically, a base fuel injection time of the fuel injection valve 6 , which is to be determined through searching a TI map which is set based on the engine rotational speed NE and the absolute air-intake-pipe internal pressure PBA. The TI map is set so that the air-fuel ratio of the mixture to be supplied to the engine may become almost equal to the stoichiometric air-fuel ratio under the operating condition corresponding to the engine rotational speed NE and the absolute air-intake-pipe internal pressure PBA. In other words, the base fuel amount TIM is almost in proportion to the intake air amount per unit time of the engine (mass flow rate).

KCMD represents a target air-fuel ratio coefficient, which is set in accordance with such engine operating parameters as engine rotational speed NE, throttle valve opening degree THA and engine water temperature TW. The target air-fuel coefficient KCMD is proportional to the fuel-air ratio F/A which is the reciprocal of air-fuel ratio A/F, and has a value of 1.0 at the stoichiometric air-fuel ratio. Accordingly, the coefficient KCMD is also called a target equivalent ratio. Besides, the target air-fuel coefficient KCMD is set to a predetermined enrichment value KCMDRR or KCMDRM for enriching the air-fuel ratio when the reduction enrichment or the deterioration determination for the NOx purifying device 15 is performed as described below.

KLAF represents an air-fuel ratio correction coefficient that is calculated under the STR control so that a detected equivalent ratio KACT, which is obtained from a detected value provided by the LAF sensor 17 , matches the target equivalent ratio KCMD when execution conditions for the feedback control are satisfied.

KPA represents an atmospheric pressure correction coefficient to be set in accordance with the atmospheric pressure PA. It is set to be 1.0 (an uncorrected value) when the atmospheric pressure PA is almost equal to 101.3 kPa. The value of PA is set larger than 1.0 in accordance with the decrease of the atmospheric pressure PA, so that the fuel supply amount may be corrected so as to be increased. Thus, the atmospheric pressure correction coefficient KPA is set so as to increase in accordance with the decrease of the atmospheric pressure PA, and the fuel supply amount is corrected so as to be increased in accordance with the decrease of the atmospheric pressure PA.

K 1 and K 2 represent another correction coefficient and a correction variable that are obtained in accordance with various engine parameter signals. They are determined to be certain predetermined values with which various characteristics such as the fuel characteristics and engine acceleration characteristics depending on the engine operating conditions are optimized.

FIG. 2 is a flowchart showing a process for calculating the target air-fuel ratio coefficient KCMD to be applied to the above-referenced equation (1). The ECU 5 performs this process at a constant time interval.

In step S 31 , when a SOx removal enrichment flag FSRR is set to 1, it indicates that an enrichment of the air fuel ratio is performed for removing the SO 2 accumulated in the three-way catalyst 14 . When FSRR is set to 1, the target air-fuel ratio coefficient KCMD is set to a predetermined value KCMDSF (for example, 1.03) for the SOx removal enrichment in step S 49 .

When FSRR is set to zero, it is determined in step S 32 whether or not the lean operation is underway, in other words, whether or not a stored value KCMDB of the target air-fuel ratio coefficient KCMD to be stored in step S 41 (to be described hereinafter) during the regular control is less than 1.0. When KCMDB is equal to or larger than 1.0, which indicates that the lean operation is not underway, the process proceeds to step S 37 , in which a reduction enrichment flag FRSPOK is set to zero (if it is set to 1, it indicates a reduction-enrichment is being performed). Then, in step S 38 , count-down timers tmRR and tmRM, which will be referred to in steps S 44 , S 47 (to be described later), are started after their initial values are set to a reduction enrichment time TRR and TRM respectively (for example, 5 to 10 seconds).

Next, in step S 39 , it is determined whether or not an enrichment continuation flag FRSPEXT is set to zero. This flag is set to 1 by a deterioration determination process of FIG. 9 (to be described later) so as to indicate that the enrichment of the air-fuel ratio should be continued even after the deterioration determination of the NOx purifying device 15 would have been completed. When FRSPEXT=1, the process proceeds to step S 46 , in which the enrichment of the air-fuel ratio is continued.

When FRSPEXT=0, the operation is performed under the regular control and the target air-fuel ratio coefficient KCMD is set in accordance with the engine operating conditions in step S 40 . The target air-fuel ratio coefficient KCMD is basically calculated in accordance with the engine rotational speed NE and the absolute air-intake-pipe internal pressure PBA. KCMD may be changed to another value depending on different operating conditions, such as the conditions where the engine water temperature TW is low and the conditions where the engine is operated in a heavy load condition. Next, in step S 41 , the target air-fuel ratio coefficient KCMD calculated in step S 40 is stored as a stored value KCMDB and the process exits here. In such engine operating condition where the lean operation is allowed, the target air-fuel ratio coefficient KCMD is set to a value smaller than 1.0.

When KCMDB<1.0 in step S 32 , which indicates that the lean operation is underway, an increment value ADDNOx is determined in accordance with the engine rotational speed NE and the absolute air-intake-pipe internal pressure PBA in step S 33 . The increment value ADDNOx, which is a parameter corresponding to the amount of NOx which is exhausted per unit time during the lean operation, is set such that it increases in accordance with the increase of the engine rotational speed NE and the increase of the absolute air-intake-pipe internal pressure PBA.

In step S 34 , a NOx amount counter CRSP is incremented by the increment value ADDNOx as shown in the following equation (2), so as to obtain a count value which is equivalent to the NOx exhaust amount, that is, the NOx amount trapped by the NOx trapping agent.

CRSP=CRSP+ADDNOx   (2)

Next, in step S 35 , it is determined whether or not an execution condition flag FMCNDF 105 is set to 1. The execution condition flag FMCNDF 105 is set to 1 when the conditions for executing the deterioration determination of the NOx purifying device 15 are satisfied, as will be shown in FIG. 4 and FIG. 5 . Usually, because FMCNDF 105 =0, the process proceeds to step S 36 , in which it is determined whether or not the value of the NOx amount counter CRSP has exceeded an allowance value CNOxREF. When the value of the NOx amount counter CRSP has not exceeded the allowance value CNOxREF, the process proceeds to step S 37 , in which the operation is controlled as usual unless the enrichment continuation flag FRSPEXT is set to 1. The allowance value CNOxREF is set to a value corresponding to, for example, a NOx amount that is slightly smaller than the maximum NOx absorption capability of the NOx trapping agent.

When CRSP>CNOxREF in step S 36 , the reduction enrichment flag FRSPOK is set to 1 in step S 42 , and then the target air-fuel ratio coefficient KCMD is set to a predetermined enrichment value KCMDRR corresponding to about 14.0 of the air-fuel ratio, so as to perform a reduction enrichment in step S 43 . Then, in step S 44 , it is determined whether or not the value of the timer tmRR is zero. While tmRR>0, this process exits. When tmRR=0 in step S 44 , the value of the reduction enrichment flag FRSPOK is set to zero and the value of the NOx amount counter CRSP is also reset to zero in step S 45 . Accordingly, from the next process cycle, the answer in step S 36 becomes “NO”, so that the operation is performed under the usual control.

On the other hand, when the conditions for the deterioration determination are satisfied (that is, when FMCNDF 105 =1 in step S 35 ), the process proceeds from step S 35 to step S 46 , in which the target air-fuel ratio coefficient KCMD is set to a predetermined deterioration determination enrichment value KCMDRM (<KCMDRR) corresponding to a slightly leaner value than a value equivalent to about 14.0 of the air-fuel ratio, so that the deterioration determination may be performed. The reason why the degree of enrichment is set smaller than when the regular reduction enrichment is performed is that the enrichment execution time may be shortened and wrong determination may easily occur at the deterioration determination time for the NOx purifying device 15 if the degree of enrichment is larger. Thus, the accuracy of the deterioration determination could be improved by means of setting the degree of enrichment smaller and prolonging the enrichment execution time. Besides, because of such smaller enrichment degree, the outputs of the O 2 sensors 18 and 19 become sensitive to the SOx, so that the determination accuracy under the high SOx density condition could be improved.

In step S 47 , it is determined whether or not the value of the timer tmRM is zero. While tmRM>0, the process exits here. When tmRM=0, the value of the NOx amount counter CRSP is reset to zero in step S 48 .

According to the processing of FIG. 2 , the reduction enrichment is usually carried out intermittently (S 43 and S 44 ) under such operating condition where the lean engine operation is possible, so that the NOx that has been trapped by the NOx trapping agent of the NOx purifying device 15 can be reduced properly. Also, when the conditions for the deterioration determination for the NOx purifying device 15 are satisfied, the degree of enrichment is set smaller than the reduction enrichment and the deterioration determination is performed over a longer time period than the reduction enrichment (S 46 , S 47 ). Besides, SOx removal enrichment is carried out when the SOx removal is performed (S 31 and S 49 ). Additionally, when the enrichment continuation flag FRSPEXT is set to 1 in step S 174 of FIG. 10 (to be described later), the target air-fuel ratio coefficient KCMD is maintained at the predetermined enrichment value KCMDRM even after the deterioration determination of the NOx purifying device 15 has been completed, so that the air-fuel ratio enrichment may be continued.

FIG. 3 is a flowchart of a main routine for a deterioration determination process of the NOx purifying device 15 . The ECU 5 in synchronization performs this process with the occurrence of the TDC signal pulses. In this process, the deterioration of the NOx purifying device 15 is determined by measuring the NOx trapping capability of the NOx trapping agent based on the output of the downstream O 2 sensor 19 .

In step S 51 , the absolute air-intake-pipe internal pressure PBA is corrected according to the following equation (3):

PBAV=PBA×KPA   (3)

In the equation (3), KPA represents an atmospheric pressure correction coefficient to be decided depending on the output of the atmospheric pressure sensor PA, and PBAV represents an absolute air-intake-pipe internal pressure after correction with the atmospheric pressure (which will be hereinafter referred to as simply “corrected absolute pressure).

In step S 52 , it is determined whether or not the corrected absolute pressure PBAV exceeds a maximum value ( 37 FF” in hexadecimal). When it is smaller than the maximum value, the process proceeds to step S 54 . When it exceeds the maximum value, the maximum value FF is set on the corrected absolute pressure PBAV in step S 53 , and the process proceeds to step S 54 . The corrected absolute pressure PBAV which has been obtained here may be used in some subsequent processes including an intake air amount accumulation process.

In step S 54 , an execution condition determination process to be described later with reference to FIG. 4 and FIG. 5 is performed. In this process, the execution condition flag FMCNDF 105 is set to 1 when the conditions for executing the deterioration determination for the NOx purifying device 15 are satisfied. In step S 55 , it is determined whether or not the execution condition flag FMCNDF 105 is set to 1. When FMCNDF 105 =0 indicating the execution conditions are not satisfied, the process proceeds to step S 56 , in which a deterioration determination pre-processing completion flag FLVLNCEND and a counter CGALNCV, which are set in a deterioration determination pre-processing to be described with reference to FIG. 7 , are set to zero. Subsequently, in step S 57 and step S 58 , a SO 2 density determination completion flag FSLFEND, a first reference-exceeding flag FSVO 2 EXPL and a second reference-exceeding flag FSVO 2 EXPH are all set to zero, and this process exits here.

The SOx density determination completion flag FSLFEND is set to 1 when the SOx density determination process shown in FIG. 6 is completed. The first reference-exceeding flag FSVO 2 EXPL is set to 1 when the upstream O 2 sensor output SVO 2 reaches an upstream reference value SVO 2 LNC (for example, 0.3 volts indicating that SVO 2 started to rise) in step S 113 , FIG. 6 . The second reference-exceeding flag FSVO 2 EXPH is set to 1 when the upstream O 2 sensor output SVO 2 exceeds a second upstream reference value SVO 2 SLF (for example, 0.8 volts, which SVO 2 would exceed if SOx density is low). Thus, FSVO 2 EXPH=1 indicates that SOx density is low.

When the execution condition flag FMCNDF 105 =1 in step S 55 , which indicates that the execution conditions of the deterioration determination for the NOx purifying device 15 are satisfied, it is determined in step S 59 whether or not a downstream sensor determination result waiting flag FTO 2 WAIT has been set to 1 in the deterioration determination process to be described with reference to FIG. 10 . Initially, because FTO 2 WAIT=0, the process proceeds to step S 60 , in which an SOx density determination process shown in FIG. 6 is performed, and then, in step S 61 , it is determined whether or not the first reference exceeding flag FSVO 2 EXPL is set to 1. When FSVO 2 EXPL=1, a deterioration determination process is carried out in step S 63 and the NOx purifying device deterioration determination process exits. When FTO 2 WAIT=1 in step S 59 , which indicates that a failure determination for the downstream O 2 sensor 19 is being waited, the deterioration determination process is carried out immediately in step S 63 . When FSVO 2 EXPL=0 in step S 61 , the flag FDONEF 105 is set to 1 in step S 64 , and this process exits.

FIG. 4 and FIG. 5 are a flowchart of the execution condition satisfaction determination process carried out in step S 54 of FIG. 3 . In this process, in order to stably determine the deterioration of the NOx purifying device 15 and secure the frequencies of the various monitors, the execution possibility of the deterioration determination for the NOx purifying device is decided considering various parameters.

In step S 71 , it is determined whether or not a deterioration determination instruction flag FGOF 105 is set to 1. Because it is sufficient to perform the deterioration determination of the NOx purifying device 15 at a rate of about once one operation period (a period from the engine start to the stop), the deterioration determination instruction flag FGOF 105 is set to 1 at the moment when the state of the engine operation has become stable after the engine is started. It should be noted that the deterioration determination is not permitted when any other monitoring is underway because such monitoring may influence the result of the deterioration determination. When the deterioration determination instruction flag FGOF 105 =1, it is determined in step S 72 whether or not a deterioration determination completion flag FENDF 105 has been set to 1 in step S 179 , FIG. 10 .

When determination in step S 71 is NO indicating that the deterioration determination is not permitted, or when the answer in step S 72 is YES indicating that the deterioration determination has been completed, the deterioration determination completion flag FENDF 105 is reset to zero in step S 73 , and a deterioration determination pre-condition satisfaction flag FLNCMWT is set to zero in step S 86 . This flag is to be set to 1 to indicate the conditions for the deterioration determination are satisfied.

When determination in step S 72 is NO, it is determined in step S 74 whether or not a STR feedback execution flag FSTRFB is set to 1. FSTRFB=1 indicates that the STR feedback control by a STR (Self Tuning Regulator) is underway. The STR will be described later. This STR feedback control is to calculate the air-fuel ratio correction coefficient KLAF according to the equation (1). In another embodiment, such calculation may be performed with a PID feedback control with proportion terms and/or integral terms.

When determination in step S 74 is YES, it is determined in step S 75 whether or not a lean-burn prohibition flag FKBSMJ is set to 1. In order to prohibit the lean-burn operation, the lean-burn prohibition flag FKBSMJ is set to 1 by a lean-burn prohibition determination process. The lean-burn prohibition determination process is carried out under the fuel injection control as well as in parallel to the NOx purifying device determination process shown in FIG. 3 . So, the lean-burn prohibition flag FKBSMJ may be referred to at any time.

When FKBSMJ=0 in step S 75 , it indicates that the lean-burn operation is permitted. Next, in step S 76 , it is determined whether or not the target air-fuel ratio KBSM is equal to or smaller than a predetermined value KBSLBLNC (for example, 20). When KBSM is equal to or smaller than KBSMLNC, it indicates the lean-burn operation is underway. Subsequently in step S 77 , it is determined whether or not the engine rotational speed NE exceeds a map value NELNC. This determination is performed so as not to perform the deterioration determination when the engine rotational speed is lower than a predetermined value.

When determinations in step S 74 , S 76 and S 77 are NO, or when the answer in step S 75 is YES, it is determined that the conditions for the deterioration determination are not satisfied, so that the deterioration determination pre-condition satisfaction flag FLNCMWT is set to zero in step S 86 .

When the answer in step S 77 is YES, it is determined in step S 78 whether or not a deterioration determination execution condition flag FMCNDF 105 is set to 1. Initially, because FMCNDF 105 =0, a lower threshold value PBLNCL is set to a value which is gained through searching a PBLNCLN table based on the engine rotational speed NE in step S 79 and then an upper threshold value PBLNCH is set to a value which is gained through searching a PBLNCLHN table based on the engine rotational speed NE in step S 80 .

When FMCNDF 105 =1 in step S 78 , the lower threshold value PBLNCL is set to a value that is gained through searching a PBLNCSN which value is smaller than the PBLNCLN table based on the engine rotational speed NE in step S 81 and then the upper threshold value PBLNCH is set to a value which is gained through searching a PBLNCSHN table which value is smaller than the PBLNCLHN table based on the engine rotational speed NE in step S 82 . Steps S 79 through S 82 are to set a region for determining the load of the engine 1 in accordance with the absolute air-intake-pipe internal pressure PBA.

In step S 83 , it is determined whether or not the absolute air-intake-pipe internal pressure PBA is larger than the lower threshold value PBLNCL. When PBA>PBLNCL, it is determined in step S 84 whether or not the absolute air-intake-pipe internal pressure PBA is smaller than the upper threshold value PBLNCH. When the answer in step S 83 or S 84 is NO, in other words, when the absolute air-intake-pipe internal pressure PBA is smaller than the lower threshold value PBLNCL or larger than the upper threshold value PBLNCH, the pre-condition satisfaction flag FLNCMWT is set to zero in step S 86 .

When both answers in step S 83 and S 84 are YES, in other words, when PBLNCL<PBA<PBLNCH, it is determined in step S 85 whether or not the reduction enrichment execution flag FRSPOK is set to 1. When FRSPOK=1, the deterioration determination is not performed because the reduction enrichment is being performed, and the process proceeds to step S 86 . When FRSPOK=0, a deterioration determination pre-condition satisfaction flag FLNCMWT is set to 1 in step S 87 .

Next, in step S 88 , it is determined whether or not the value of the NOx amount counter CRSP exceeds a deterioration determination permission value CLNCMACT. When CRSP does not exceed CLNCMACT, the process proceeds to step S 91 in FIG. 5 , in which the enrichment continuation flag FRSPEXT is set to 1, and then, in step S 92 , a downstream O 2 sensor failure determination condition flag FMCDF 103 B is set to zero. When the downstream O 2 sensor failure determination condition flag FMCDF 103 B is set to 1, it indicates that the conditions for executing a failure determination process (not shown) for the downstream sensor 19 are satisfied.

When CRSP>CLNCMACT in step S 88 , it is determined that the amount of the NOx trapped by the NOx trapping agent is large enough to perform the deterioration determination of the NOx purifying device 15 . Accordingly, the downstream O 2 sensor failure determination condition flag FMCDF 103 B is set to 1 in step S 89 , and then, in step S 90 , it is determined whether or not the upstream O 2 sensor determination flag FOK 63 is set to 1.

When determination in step S 90 is YES, it is determined in step S 93 whether or not an execution condition flag FMCNDF 105 has already been set to 1. Initially, because FMCNDF 105 =0, the process proceeds to step S 94 , in which it is determined whether or not the downstream O 2 sensor output LVO 2 is equal to or smaller than a first downstream reference value LVO 2 LNCM (for example, 0.3V). This step is to confirm that the downstream O 2 sensor output LVO 2 before the execution of the deterioration determination enrichment indicates an exhaust lean condition. When FMCNDF 105 =1 in step S 93 , the above-described determination steps are not performed and the process proceeds directly to step S 97 .

When LVO 2 is equal to or smaller than LVO 2 LNCM in step S 94 indicating that the downstream O 2 sensor output LVO 2 indicates the exhaust lean condition, it is determined in step S 95 whether or not an absolute difference value |SVO 2 -LVO 2 | between the upstream O 2 sensor output SVO 2 and the downstream O 2 sensor output LVO 2 is equal to or smaller than a predetermined value DSLVO 2 LN. This step is to confirm that both upstream O 2 sensor output SVO 2 and the downstream sensor output LVO 2 are in a lean condition and further that their difference is very small. When the answer is YES, the process proceeds to step S 102 .

When the answer in step S 95 is NO, a purge cut flag FLNCPG is set to zero in step S 97 and a countdown timer TLNCPG is started after it is set to a predetermined time TMLNCPG (for example, two seconds) in step S 98 . Subsequently, a maximum value parameter SVMAXLNC is set to zero in step S 99 , a flag FSVMAXLNC is set to zero in step S 100 , an execution condition flag FMCNDF 105 is set to zero in step S 101 , and this process exits.

The purge cut flag FLNCPG when it is set to 1 indicates that purging of evaporated fuel in the fuel tank to the intake pipe 2 should be prohibited. The maximum value parameter SVMAXLNC is a parameter representing a maximum value of the upstream O 2 sensor output SVO 2 before the upstream O 2 sensor output SVO 2 reaches a second upstream reference value SVO 2 SLF (for example, 0.8V).

When the answer of step S 95 is YES, the purge cut flag FLNCPG is set to 1 in step S 102 . This is to forcibly cut the purging of evaporated fuel because the density of the purging is uncertain and accordingly a wrong detection may easily happen. Next, in step S 103 , it is determined whether or not the value of the timer TLNCPG which has been started in step S 98 is zero. While TLNCPG>0, the process proceeds to step S 99 . The timer TLNCPG is used for the purpose of waiting for a given time period for the influence of the purging to disappear after the purging is cut.

When the value of the timer TLNCPG becomes zero in step S 103 , the process proceeds to step S 104 , in which it is determined whether or not the upstream O 2 sensor output SVO 2 is less than a third upstream reference value SVLNCMC (for example, 0.7V). When SVO 2 is not less than SVLNCMC, the process proceeds to step S 109 , in which the deterioration determination execution condition flag FMCNDF 105 is set to 1.

When the upstream O 2 sensor output SVO 2 is smaller than the third upstream reference value SVLNCMC, it is determined in step S 105 whether or not the upstream O 2 sensor output SVO 2 exceeds the maximum value parameter SVMAXLNC. Because the maximum value parameter SVMAXLNC is initialized to zero in step S 99 , the answer in step S 105 is YES at first. So, the maximum value parameter SVMAXLNC is set to the current value of the O 2 sensor output SVO 2 in step S 108 , and then the execution condition flag FMCNDF 105 is set to 1 in step S 109 .

When the upstream O 2 sensor output SVO 2 increases monotonously, the answer in step S 105 always becomes YES. However, the output sometimes may decrease temporarily. In such a case, the answer of step S 105 becomes NO and then, in step S 106 , a difference DSV between the maximum value parameter SVMAXLNC and the O 2 sensor output SVO 2 is calculated according to the following equation (4):

DSV=SVMAXLNC−SVO 2   (4)

Then, it is determined in step S 107 whether or not the difference DSV is more than a predetermined value DSVLNCMC. When the answer is NO indicating that the difference is not so significant, the execution condition flag FMCNDF 105 is set to 1 in step S 109 .

When the difference DSV exceeds the predetermined value DSVLNCMC, it is considered that the air-fuel ratio has temporarily become in an exhaust lean condition due to the engine acceleration or other events. If the deterioration determination is continued in such case, there is a possibility of occurrence of wrong determination. Accordingly, the execution condition is regarded to be unsatisfactory and the deterioration determination is stopped. Thus, the flag FSVMAXLNC is set to zero in step S 100 , the execution condition flag FMCNDF 105 is set to zero in step S 101 , and this process exits.

According to the deterioration determination execution condition determination process shown in FIG. 4 and FIG. 5 , the conditions for the deterioration determination execution of the NOx purifying device 15 are basically satisfied when the pre-condition satisfaction flag FLNCMWT is set to 1. However, if the predetermined time has not elapsed since the evaporated fuel purge was prohibited, the execution condition is determined to be unsatisfied in step S 103 . Besides, under such conditions where the upstream O 2 sensor output SVO 2 is smaller than the third upstream reference value SVLNCMC, when the temporary decrease amount (DSV) becomes larger than the predetermined value DSVLNCMC (namely, when the answer in step S 107 is YES), the execution condition is determined to be unsatisfied.

FIG. 6 is a flowchart of the SOx density determination process in step S 60 of FIG. 3 . When high-density-sulfur containing fuel is being used, the three-way catalyst 14 may be influenced by the SOx. In this case, the downstream sensor 19 may not turn to the rich side completely. For that reason, there may be a wrong estimation for the trapped amount of NOx by the NOx purifying device 15 , which may eventually result in a wrong deterioration determination. Therefore, according to the process flow shown in FIG. 6 , SOx density determination is performed so as to determine whether high-density-sulfur containing fuel is used or not.

In step S 111 , it is determined whether or not the SOx density determination completion flag FSLFEND has been set to 1. When the answer is YES, the SOx density determination process is not performed at this time. Initially, because FSLFEND=0, the process proceeds to step S 112 , in which it is determined whether or not the upstream SO 2 sensor output SVO 2 is equal to or larger than a first upstream reference value SVO 2 LNCS (for example, 0.3 volts indicating that SVO 2 started to rise). Initially, because SVO 2 <SVO 2 LNCS, the process skips step S 113 to proceed to step S 114 , in which it is determined whether or not a first reference exceeding flag FSVO 2 EXPL has been set to 1. This flag is set to 1 in step S 113 . Accordingly, initially, this answer is NO. So, GSLFFIN is set to zero in step S 115 and a parameter GSLFTWCH is set to zero in step S 116 . The process proceeds to step S 120 .

In step S 120 , it is determined whether or not the upstream O 2 sensor output SVO 2 has exceeded a fourth upstream reference value SVO 2 SLFT. Initially, since this answer is NO, a parameter GSLFTWCH is set to the cumulative intake air amount value GSLFFIN that is calculated in step S 119 .

Then, in step S 122 , it is determined whether or not the upstream O 2 sensor output SVO 2 has exceeded a second upstream reference value SVO 2 SLF (for example, 0.8V, which SVO 2 would exceed if SOx density is low). Initially, since this answer is NO, the process skips step S 123 to step S 124 .

In step S 124 , it is determined whether or not the first cumulative intake air amount value GSLFFIN is equal to or larger than a determination threshold value GSLFFINT. Initially, since this answer is NO, it is determined in step S 125 whether or not a flag FSVO 2 EXPH to be set in step S 123 has been set to 1. Initially, because this answer is NO, this process exits here.

When the upstream O 2 sensor output SVO 2 has reached the first upstream reference value SVO 2 LNCS, the first reference exceeding flag is set to 1 in step S 113 and the process proceeds to step S 117 via step 114 . In step S 117 , it is determined whether or not a flag FSVO 2 EXPH to be set to 1 in step S 123 has been set to 1. Initially, since this answer is NO, the process proceeds to step S 118 , in which a table KGSLFPBN is searched so as to obtain a value corresponding to the corrected intake air absolute pressure PBAV calculated in step S 51 of FIG. 3 . The obtained value is set to a correction coefficient KGSLFPB.

Then, the value GSLFFIN is calculated according to the following equation (5):

GSLFFIN=GSLFFIN×TIM×KPA×KGSLFPPB   (5)

In the equation (5), GSLFFIN on the right portion is a previously calculated value and TIM and KPA respectively represent the base fuel amount and the atmospheric pressure correction coefficient that are used in the equation (1). Since TIM is the base fuel amount, that is, a fuel amount to be set so that the air-fuel ratio becomes equal to the stoichiometric air-fuel ratio in accordance with the operating conditions (the engine rotational speed NE and the absolute air-intake-pipe internal pressure PBA), it is a parameter which is proportional to the intake air amount (that is, the exhaust gas amount) per unit time of the engine 1 . The first cumulative intake air amount value GSLFFIN, which is obtained according to the equation (5), is a value corresponding to a cumulative value of the exhaust gas amount which has been flowing into the NOx purifying device 15 from the time the upstream O 2 sensor output SVO 2 crosses the first upstream reference value SVO 2 LNCS till the time it reaches SVO 2 SLF.

Because the air-fuel ratio is maintained at the predetermined enrichment value (KCMDRM), which is richer than the stoichiometric air-fuel ratio during the deterioration determination process, the first cumulative intake air amount value GSLFFIN is in proportion to the cumulative amount of the reduction constituents (HC, CO) contained in the exhaust gas. Besides, the value GSLFFIN is in proportion to the elapse time since the accumulation started as long as the engine operation condition is almost constant. The same is true with respect to other cumulative intake air amount values which will be described later.

When the upstream O 2 sensor output SVO 2 is between the first upstream reference value SVO 2 LNCS and a fourth upstream reference value SVO 2 SLFT (for example, 0.7V), the process proceeds from step S 120 to step S 122 via step S 121 . When the upstream O 2 sensor output SVO 2 exceeds SVO 2 SLFT, the process jumps from step S 120 to step S 122 . When SVO 2 exceeds the second upstream reference value SVO 2 SLF, which as described heretofore SVO 2 would exceed if SOx density is low, the process proceeds from step S 122 to step S 123 , and the second reference exceeding flag FSVO 2 EXPH is set to 1. As described heretofore, this flag indicates that SOx density is low.

When the value GSLFFIN is smaller than a determination threshold value GSLFFINT, the process proceeds from step S 124 to step S 125 . At this time, because flag FSVO 2 EXPH is 1, the process proceeds to step S 126 , in which the SOx density determination end flag FSLFEND is set to 1, and the process exits here.

Following will further describe the aforementioned SOx density determination. When the SOx density is high, the output SVO 2 will not exceed SVO 2 SLF by the influence of the SOx even if enough time elapses. When a saturated value of the upstream O 2 sensor output SVO 2 does not reach the reference value within a given time by the influence of the SOx, the fuel is determined to be high-density sulfur containing fuel. In other words, when the upstream O 2 sensor output SVO 2 does not exceed the second upstream reference value SVO 2 SLF when the first cumulative intake air amount value GSLFFIN reaches the determination threshold value GSLFFINT, it is determined that SOx density is high around the upstream O 2 sensor 18 . When the SOx density is high, the time for the upstream O 2 sensor output SVO 2 to reach the value SVO 2 SLF gets longer in some cases. In other cases the upstream O 2 sensor output SVO 2 does not reach the value SVO 2 SLF at all. The process shown in FIG. 6 can determine the SOx influence in either case.

High-density-sulfur containing fuel specifically means that SOx density in the exhaust gas becomes about 600 PPM and more. When such fuel is used, the O 2 sensor output may be influenced by the SOx.

Besides, when the three-way catalyst 14 is deteriorated, the SOx density becomes higher at the downstream side. When the NOx purifying device 15 is disposed downstream of the three-way catalyst 14 as in the embodiment, the O 2 sensor output may change due to SOx to lower the accuracy of determining deterioration of the NOx purifying device 15 . Thus, when the SOx density is high, prohibiting the deterioration-determining process of the NOx purifying device 15 will enhance accuracy of determining deterioration.

Besides, the tendency of lowering of the saturation output of the O 2 sensor is prominent when enrichment degree of the air-fuel ratio is smaller. Therefore, in this embodiment, the target air-fuel ratio coefficient KCMD during the deterioration determination process is set to a predetermined deterioration determination enrichment value KCMDRM corresponding to an air-fuel ratio that is slightly richer (for example, about air-fuel ratio 14.3) than the stoichiometric air-fuel ratio.

Referring now to FIGS. 7 , 9 and 10 , the deterioration determination process of the NOx purifying device 15 will be described. This process is to detect deterioration of the characteristics of the NOx purifying device 15 based on the lean output maintenance time of the downstream O 2 sensor 19 (or exhaust gas amount) when the air-fuel ratio is changed from lean to rich.

FIG. 7 is a flowchart of a deterioration determination pre-process in step S 62 of FIG. 3 . In step S 131 , it is determined whether or not the enrichment continuation flag FRSPEXT is set to 1. Initially, FRSPEXT=0, so the process proceeds to step S 132 , in which it is determined whether or not a deterioration determination pre-processing completion flag FLVLNCEND has been set to 1. The flag FLVLNCEND is set to 1 in step S 139 . Initially, it is zero. So, the process proceeds to step S 133 , in which an intake air amount accumulation process shown in FIG. 8 is performed. When FRSPEXT=1 in step S 131 or when FLVLNCEND=1 in step S 132 , the process exits here.

In step S 141 of FIG. 8 , it is determined whether or not the upstream O 2 sensor output SVO 2 is equal to or smaller than a fifth upstream reference value SVO 2 LNH (for example, 0.6V). When SVO 2 is equal to or smaller than SVO 2 LNH, a second cumulative intake air amount value GALNCS is set to zero in step S 142 , and the process exits here.

When the upstream O 2 sensor output SVO 2 exceeds the upstream reference value SVO 2 LNH, a KNACPBN table is searched based on the corrected absolute pressure PBAV in step S 144 , so as to calculate an air-intake-pipe internal pressure correction coefficient KNACPB.

Next, in step S 147 , a second cumulative intake air amount value GALNCS is calculated according to the following equation (6):

GALNCS=GALNCS+TIM×KPA×KNACPBS   (6)

In the equation (6), GALNCS in the right member represents the previously calculated value and TIM and KPA respectively represent the base fuel amount and the atmospheric pressure correction coefficient. In other words, in the equation (6), the corrected fuel injection amounts are accumulated so as to obtain the intake air amount.

The second cumulative intake air amount value GALNCS, which is obtained according to the equation (6), is a value corresponding to a cumulative value of the exhaust gas amount which has been flowing into the NOx purifying device 15 since the time when the upstream O 2 sensor output SVO 2 exceeded the upstream reference value SVO 2 LNH.

Referring back to FIG. 7 , in step S 134 , a GALNCVN table is searched based on the count value of the counter CGALNCV, so as to calculate a threshold value GALNCV. This table is set so that GALNCVN increases as the count of the counter CGALNCV increases. Next, in step S 135 , it is determined whether or not the second cumulative intake air amount value GALNCS calculated in step S 133 is equal to or larger than the threshold GALNCV. When GALNCS<GALNCV, the process exits here. When GALNCS is equal to or larger than GALNCV, the downstream O 2 sensor output LVO 2 is stored in the buffer LVGALNC that corresponds to the count value of the counter CGALNCV (S 136 ). Thirty buffers LVGALNC in total are provided.

In step S 137 , the counter CGALNCV is incremented by 1. In step S 138 , it is determined whether the count value has reached 30 or not. When it has not reached 30 yet, the process exits here. Thus, the downstream O 2 sensor output LVO 2 is stored in the buffers LVGALNC repeatedly until the counter reaches 30. When the counter reaches 30 (S 138 ), a flag FLVLNCEND is set to 1 (S 139 ).

Following the deterioration determination pre-processing, the deterioration determination process shown in FIG. 9 and FIG. 10 is performed. In step S 151 , it is determined whether or not the enrichment continuation flag FRSPEXT has been set to 1. This flag is set to 1 in step S 174 . Initially, because FRSPEXT=0, the process proceeds to step S 152 , in which it is determined whether or not the downstream O 2 sensor determination result waiting flag FTO 2 WAIT is set to 1. This flag is set to 1 in step S 173 . Initially, because FRSPEXT=0, the process proceeds to step S 153 .

In step S 53 , it is determined whether or not the pre-processing completion flag FLVLNCEND has been set to 1. When FLVLNCEND=0, downstream O 2 sensor output LVO 2 has not completely been stored in the LVGALNC buffers yet. The process exits here. When FLVLNCEND=1 (S 153 ), which indicates that the pre-processing has been completed, it is determined in step S 154 whether or not the SOx density determination flag FSLFEND has been set to 1. When FSLFEND=0, which indicates that the deterioration determination has not been completed yet, the process exits here without performing deterioration determination. When FSLFEND=1 in step S 154 , the process proceeds to step S 155 .

In step S 155 , a NLVGAHN table ( FIG. 11 ) is searched to obtain an upper reference value NLVGAH based on the value GSLFFIN obtained in the SOx density determination process described with reference to FIG. 6 . Next, in step S 156 , the NLVGALN table ( FIG. 11 ) is searched to obtain a lower reference value NLVGAL based on the value GSLFFIN. Then, in step S 157 , the LVO 2 value stored in the LVGALN buffers is retrieved based on the searched NLVGAH, and the retrieved LVO 2 is set as a first value to be examined LVGALNCH. In step S 158 , the value stored in the LVGALN buffers is retrieved based on the searched NLVGAL, and the retrieved value is set as a second value to be examined LVGALNCL. As can be seen from the table shown in FIG. 11 , the value LVGALNCH is a downstream O 2 sensor output LVO 2 having delay relative to the one retrieved from the value LVGALNCL (longer time has elapsed since FMCNDF 105 was set to 1).

Steps S 155 through S 158 decide based on GSLFFIN, which data is to be used out of the downstream O 2 sensor outputs LVO 2 that have been stored in the buffers in certain timing in the deterioration determination pre-process.

Both NLVGAL and NLVGAH tables are arranged as shown in FIG. 11 so as to first determine influence by the SOx or influence of deterioration of the three-way catalyst by experiment and to retrieve data (LVO 2 output) of such timing that can avoid these influence, thereby preventing a wrong detection that may otherwise be caused by the influence of SOx. Thus, the influence on the downstream O 2 sensor 19 by the SOx can be avoided and accordingly accuracy of deterioration determination of the NOx purifying device 15 will improve. In another embodiment, instead of using the buffers, the intake air amount may be accumulated from the reversal of the upstream O 2 sensor output SVO 2 to the reversal of the downstream O 2 sensor output LVO 2 . The accumulated intake air amount may be corrected using tables similar to the one shown in FIG. 11 , or may be compared with a determination threshold value retrieved form the tables shown in FIG. 11 .

In step S 159 , it is determined whether or not the first LVO 2 value LVGALNCH is equal to or smaller than a reference value LVO 2 LNH to determine if LVO 2 is in a lean state. If YES, it is determined that the NOx purifying device 15 is in a good shape and accumulating sufficient amount of NOx. So, in step S 160 , a temporary determination flag FKOKF 105 is set to 1 and a determination pending flag FGRAYF 105 is set to zero, and the process proceeds to step S 171 .

When LVGALNCH>LVO 2 LNH in step S 159 , the process proceeds to step S 161 , in which it is determined whether or not the second LVO 2 value LVGALNCL is equal to or smaller than the reference value LVO 2 LNH indicating that the LVO 2 is in a lean state. If it is YES, it means that a confusing condition is being observed because in step S 159 LVO 2 was determined to be in a rich state. Thus, deterioration determination of the NOx purifying device 15 cannot be readily done, the process proceeds to step S 167 , in which the determination pending flag FGRAYF 105 is set to 1. The flag FGRAYF 105 is to be set to 1 in order to indicate that the deterioration of the NOx purifying device cannot be determined at this moment.

When the answer in step 161 is NO, indicating that the downstream O 2 sensor 19 is in a rich. In this case, it is determined in step S 162 whether or not a SOx removal completion flag FSRMOVEND has been set to 1. This flag is to be set in the SOx removal process which will be later described with reference to FIG. 12 .

When FSRMOVEND=1 indicating that the SOx removal has been completed, it is determined in step S 163 whether or not the exhaust amount parameter GSLFTWCH, which is to be set in step S 121 of FIG. 8 , is equal to or larger than a reference value GSLFJUD. When the answer is NO, that is, GSLFTWCH<GSLFTJUD indicating that the NOx purifying device 15 may be deteriorated, the temporary determination flag FKOKF 105 is set to zero and the determination pending flag FGRAYF 105 is set to zero in step S 164 , and the process proceeds to step S 171 of FIG. 10 .

When FSRMOVEND=0 indicating that the SOx removal has not been completed yet, it is determined in step S 165 whether or not the value GSLFFIN is equal to or larger than the reference value GSLFJUD. When this answer is NO, the process proceeds to step S 163 . When GSLFFIN is equal to or larger than GSLFJUD indicating that there exists some influence of the SOx, the high density flag FSLF is set to 1 in step S 166 , the determination pending flag FGRAYF 105 is set to 1 in step S 167 and the process proceeds to step S 179 of FIG. 10 .

When the answer in step S 163 is YES, that is, GSLFTWCH is equal to or larger than GSLFJUD, the process proceeds to step S 167 as well. This is because a normal deterioration determination is difficult to be carried out when high-density-sulfur containing fuel is being used even if the SOx removal process is performed.

In step S 171 of FIG. 10 , it is determined whether or not a downstream O 2 sensor failure flag FFSDF 103 has been set to 1. The flag FFSDF 103 is to be set to 1 when it is determined that the downstream O 2 sensor 19 is in failure. When FFSDF 103 =1, which indicates that the downstream O 2 sensor 19 is in failure, the process proceeds to step S 179 , in which a deterioration determination completion flag FENDF 105 is set to 1 and a downstream O 2 sensor determination result waiting flag FTO 2 WAIT is set to zero, and the deterioration determination process exits here.

When FFSDF 103 =0 in step S 171 , which indicates that the downstream O 2 sensor 19 is not in failure, it is determined in step S 172 whether or not an O 2 sensor OK flag FKOKF 103 has been set to 1. The O 2 sensor OK flag FKOKF 103 is to be set to 1 when the downstream 2 sensor is determined to be normal. When FKOKF 103 =0 indicating that the downstream 2 sensor is not normal, in order to continue the air-fuel ratio enrichment for performing the failure determination on the downstream O 2 sensor 19 , the determination result waiting flag FTO 2 WAIT is set to 1 in step S 173 and the enrichment continuation flag FRSPEXT is set to 1 in step S 174 , and the process exits here.

When the answer in step S 151 or in step S 152 is YES in the subsequent routine cycles, the process proceeds to step S 171 .

When FKOKF 103 =1 in step S 172 , which indicates that the downstream O 2 sensor 19 is determined to be normal, the SOx removal completion flag FSRMOVEND is set to zero in S 175 and, in step S 176 , it is determined whether or not the temporary determination flag FKOKF 105 has been set to 1. When the temporary determination flag FKOKF 105 has been set to 1, which indicates that the NOx purifying device 15 is normal, the normality flag FOKF 105 is set to 1, the failure flag FFSDF 105 is set to zero, and a deterioration determination done flag FDONEF 105 is set to 1 in step S 177 , and the process proceeds to step S 179 . It is sufficient that deterioration determination for the NOx purifying device 15 is done once in one driving cycle.

When the temporary determination flag FKOKF 105 is set to zero in step S 176 , which indicates that the NOx purifying device 15 is in failure, the normality flag FOKF 105 is set to zero, the failure flag FFSDF 105 is set to 1 and the deterioration determination done flag FDONEF 105 is set to 1 in step S 178 , and the process proceeds to step S 179 .

In step S 179 , the deterioration determination end flag FENDF 105 is set to 1, the determination result waiting flag FTO 2 WAIT is reset to zero, and the process exits here.

According to the process of FIG. 9 and FIG. 10 , deterioration of the characters of the NOx purifying device 15 is determined based on the lean output maintenance period of the downstream O 2 sensor 19 (exhaust gas amount) during the enrichment determination process.

The values LVGALNCH and LVGALNCL which have been retrieved from the outputs LVO 2 of the downstream O 2 sensor 19 that have been buffered in the deterioration determination pre-process of FIG. 7 based on the cumulative intake air amount (or exhaust gas amount) are provided as the downstream sensor output for use with deterioration determination. The LVO 2 values, which are stored in the LVGALN buffers, are retrieved by referring to the table (as shown in FIG. 11 ) that is pre-established based on experiment or simulation. In such a way, the downstream O 2 sensor output can be retrieved after an appropriate time has elapsed (that is, under nominal influence of SOx) since output of the upstream O 2 sensor reversed to the rich side.

When the first determination value LVGALNCH is equal to or smaller than the reference value (that is, when the answer in step S 159 is YES), it is determined that the NOx is substantially accumulated in the NOx purifying device 15 as the downstream sensor output is still lean although sufficiently long time has elapsed since output of the upstream O 2 sensor reversed to the rich side. Accordingly, the temporary determination flag FKOKF 105 , which temporarily determines that the NOx purifying device 15 is normal, is set to 1 in step S 160 .

Secondly, when LVGALNCH>LVO 2 LNH and when the second determination value LVGALNCL exceeds the reference value LVO 2 LNH (that is, when the answer in step S 161 is NO), it is determined that the NOx is not sufficiently accumulated in the NOx purifying device 15 as output of the downstream O 2 sensor 19 reversed to the rich side within a short period since output of the upstream O 2 sensor reversed to the rich side. However, even in this case, when the SOx removal process has not been performed yet (that is, FSRMOVEND=1 in step S 162 ) and when the cumulative intake air amount is equal to or larger than the predetermined value GSLFJUD, the high density determination flag FSLF is set to 1 in step S 166 , and furthermore the determination pending flag FGRAYF 105 is set to 1 in step S 167 , so that the current deterioration determination is suspended. If the deterioration determination is performed based on the output of the downstream O 2 sensor 19 when the SOx removal is not performed, accuracy of deterioration determination will decrease. Therefore, in this case, the SOx removal process is carried out once, which is to be described hereafter.

However, when the SOx removal process has already been performed or when the cumulative intake air amount GSLFFIN is smaller than the predetermined value GSLFJUD, and when GSLFTWCH<GSLFTJUD, the temporary determination flag FKOKF 105 is set to zero because it is possible that NOx purifying device 15 is deteriorated.

When LVGALNCH>LVO 2 LNH and LVGALNCL>LVO 2 LNH, in other words, when the downstream O 2 sensor output is lean after a short time and is rich after a long time, the determination depending flag FGRAY 105 is set to 1 in step S 167 so as to suspend the deterioration determination because there is a high possibility of wrong determination of deterioration of NOx purifying device 15 .

Thus, it is possible to accurately determine deterioration of NOx purifying device 15 because the downstream O 2 sensor output can be referred to at appropriate time utilizing the LVGALN buffer and the table of FIG. 11 with respect to influence of the sulfur constituents in the exhaust gas flowing into NOx purifying device 15 and deterioration state of the three-way catalyst.

The deterioration determination will be suspended if the downstream O 2 sensor 19 is determined to be in failure (that is, FFSDF 103 =1 in step S 171 ) after the value of FKOKF 105 is set. This is because the deterioration determination cannot be properly performed if the downstream O 2 sensor 19 is in failure. When it is not determined that the downstream O 2 sensor 19 is normal (namely, FKOKF 103 =0 in step S 172 ), the enrichment extension flag FRSPEXT is set to 1, and air-fuel ratio enrichment is extended in order to determine failure of the downstream O 2 sensor 19 . This process is performed for the following reasons. Movement of the output of downstream O 2 sensor 19 may not be quick because the downstream O 2 sensor 19 is located downstream of the three-way catalyst 14 and the NOx purifying device 15 . Thus, when the air-fuel ratio is set richer at the time of determining deterioration of the NOx purifying device, it is not possible to determine failure of the downstream O 2 sensor 19 even if its output does not turn rich in a short time. Accordingly, it is required to extend air-fuel enrichment long enough to determine failure. It is determined that the downstream O 2 sensor 19 has failed only when its output does not reverse during such an extended time period.

FIG. 12 shows a flow of a SOx removal process. The ECU 5 performs this process at a predetermined time interval (for example, every 100 millisecond). When it is determined that the influence of the SOx is high in the deterioration determination process for the NOx purifying device (in other words, the high density flag FSLF has been set to 1 in step S 166 ), accuracy of the deterioration determination is improved by performing SOx removal process once after deterioration determination is suspended. It should be noted that lean operation is prohibited during the SOx removal process (in other words, when FSLF=1 and FSRMOVEND=0) as to be described with reference to FIG. 13 .

In step S 181 , it is determined whether or not the high-density flag FSLF has been set to 1. When FSLF=0, a first predetermined value CTSADDS is set to a first countdown counter CSADINT in step S 183 , a second predetermined value CTSDECS is set to a second countdown counter CSDEINT in step S 184 , a third predetermined value CTSMOVS (for example, 6000) is set on a SOx trapping amount counter CSRMOV in S 185 , and then the process exits here. The value CTSRMOVS is set to a value corresponding to a time period within which all SOx can be removed even when trapped SOx amount in the NOx purifying device 15 has reached its maximum amount (saturated state).

When FSLF=1 in step S 181 , which indicates that the SOx density is high, it is determined in step S 182 whether or not the SOx removal end flag FSRMOVEND is set to 1. The answer is YES when the SOx removal process has been completed. The process proceeds to step S 183 . When the answer is NO, it is determined in step S 186 whether or not an estimated temperature TCTLNCH of the NOx purifying device 15 is equal to or larger than a predetermined temperature TCTSRMV. TCTLNCH is calculated by searching a temperature map to be created based on such engine operating conditions as the engine rotational speed NE and the engine load (absolute air-intake-pipe internal pressure PBA). This calculation process is not shown. Alternatively, it is possible to provide a temperature sensor for detecting the temperature of the NOx purifying device and use such detected temperature instead of TCTLNCH.

When TCTLNCH<TCTSRMV in step S 186 , it is determined in step S 188 whether or not the value of the first countdown counter CSADINT is equal to or smaller than zero. Initially, because CSADINT>0, the first countdown counter CSADINT is decremented by 1 in step S 189 , the second predetermined value CTSDECS is set to the second count-down counter CSDEINT in step S 193 , and the process proceeds to step S 200 . In the subsequent cycles, when the value of the counter CSADINT becomes zero, the process proceeds from step S 188 to step S 190 , in which sulfur poisoning amount table DCTSRMPN is searched based on the estimated temperature TCTLNCH to obtain sulfur poisoning amount DCTSRMP. The table DCTSRMPN is set such that sulfur poisoning amount increases as the estimated temperature rises. Next, in step S 191 , a value according to the following equation (7) is set on sulfur poisoning amount counter CSRMOV.

CSRMOV=CSRMOV+DCTSRMP   (7)

In the equation (7), CSRMOV represents the previously calculated value. Then, in step S 192 , the first predetermined value CTSADDS is set to the first countdown counter CSADINT, and the process proceeds to step S 193 .

When TCTLNCH is equal to or larger than TCTSRMV in step S 186 , it is determined in step S 187 whether or not the detected equivalent ratio KACT is equal to and more than a predetermined equivalent ratio KACTSRM (for example, 1.03). The detected equivalent ratio KACT is to be obtained by converting the output of the LAF sensor 17 to a coefficient. When KACT<KACTSRM, the process proceeds to step S 188 . When KACT is equal to or larger than KACTSRM, the process proceeds to step S 194 , in which it is determined whether or not the second countdown counter CSDEINT is equal to or smaller than zero. Initially, because CSDEINT>0, the process proceeds to step S 195 , in which the counter CSDEINT is decremented by 1, and in step S 199 the first predetermined value CTSADDS is set to the first countdown counter CSADINT, and then the process proceeds to step S 200 . In the subsequent cycles, when the value of the second countdown counter CSDEINT becomes zero, the process proceeds from step S 194 to step S 196 , in which a sulfur removal amount table DCTSRMMN is searched based on the estimated temperature TCTLNCH to obtain a sulfur removal amount DCTSRMM. The sulfur removal amount table DCTSRMMN is set such that the sulfur removal amount increases as the estimated temperature rises. Next, in step S 197 , a value according to the following equation (8) is set to sulfur poisoning amount counter CSRMOV.

CSRMOV=CSRMOV−DCTSRMM   (8)

In the equation (8), CSRMOV represents previously calculated value. Then, in step S 198 , the second predetermined value CTSDECS is set to the second countdown counter CSDEINT, and the process proceeds to step S 199 .

In step S 200 , it is determined whether or not the value of the sulfur poisoning amount counter CRSMOV is equal to or smaller than a predetermined value CTSRMOVS. When CRSMOV is equal to or smaller than CTSRMOVS, the process proceeds to step S 202 . When CRSMOV>CTSRMOVS, the sulfur poisoning amount counter CRSMOV is set to a predetermined value CTSRMOVS, and the process proceeds to step S 202 . This process sets an upper limit for sulfur adhesion amount.

In step S 202 , it is determined whether or not the value of the sulfur poisoning amount counter CSRMOV is equal to or smaller than zero. When CRSMOV>0, the process exits. When the value of CRSMOV becomes zero, which indicates that the SOx removal process has been completed, the sulfur poisoning amount counter CRSMOV is set to zero in step S 203 , the SOx removal end flag FSRMOVEND is set to 1 in step S 204 , the high density flag FSLF is set to zero again in step S 205 , the SOx removal enrichment flag FSRR is set to zero in step S 206 , and the process exits.

According to the process of FIG. 12 , when the estimated temperature TCTLNCH exceeds the predetermined temperature TCTSRMV and when the detected equivalent ratio KACT exceeds the predetermined equivalent ratio KACTSRM (that is, when the air-fuel ratio is rich), the process for removing the SOx in the three-way catalyst 14 is performed. The amount of SOx, which has accumulated in the three-way catalyst 14 , is estimated through the sulfur poisoning amount counter CSRMOV. When the value of CSRMOV becomes zero, which indicates that the accumulated SOx has been removed, the SOx removal flag FSRMOVEND is set to 1.

When the estimated temperature TCTLNCH is below the predetermined temperature TCTSRM or when the detected equivalent ratio KACT is smaller than the predetermined equivalent ratio KACTSRM, SOx is not removed but rather it may accumulate in the three-way catalyst 14 . Accordingly, the sulfur poisoning amount counter CSRMOV is incremented every time the value of the second countdown counter CSADINT becomes zero. The accumulation speed and the removal speed of SOx are different depending on the temperature of the three-way catalyst 14 (in other words, SOx accumulates less and is easier to remove as the temperature rises). Accordingly, the sulfur poisoning amount counter CSRMOV is incremented or decremented by amount that is obtained by searching the table. The first countdown counter CSADINT and the second countdown counter CSDEINT are provided as thinning-out counters to compensate for the difference between accumulation speed and removal speed of SOx.

By executing the SOx removal process for the three-way catalyst 14 when the SOx density is determined to be high, it is possible to prevent a wrong determination which may regard a variation in the delay time between the output of the upstream O 2 sensor 18 and the output of the downstream O 2 sensor 19 as a time deterioration of the NOx purifying device 15 .

FIG. 13 is a flowchart of a process for prohibiting the lean operation, which sets the air-fuel ratio leaner than the stoichiometric air-fuel ratio, while the SOx removal process is being carried out. The ECU 5 in synchronization performs this process with occurrence of the TDC signal pulse. This process raises exhaust gas temperature and produces a rich operation so that removal of the SOx becomes easy.

In step S 211 , it is determined whether or not the high-density flag FSLF is set to 1. When FSLF=1 indicating that the SOx density is high, the SOx removal enrichment flag FSRR is set to 1 in step S 212 , and the lean-burn operation prohibiting flag FKBSMJ is set to 1 in step S 222 , so that the lean operation is prohibited.

When the high-density flag FSLF is set to zero, it is determined whether the lean-burn operation may be permitted or not according to the following steps. In step S 213 , it is determined whether or not the engine load PBGA exceeds a value PBKBS which is determined (in a process not shown herein) in accordance with operating conditions including the engine rotational speed. When PBGA>PBKBS, the process proceeds to step S 222 to prohibit the lean-burn operation. When PBGA is equal to or smaller than PBKBS, it is determined in step S 214 whether or not the engine rotational speed NE exceeds a predetermined value NKBSL (for example, 1000 rpm). When the answer is YES, it is further determined in step S 215 whether or not the SOx removal end flag FSRMOVEND is set to 1. When FSRMOVEND=0, the process proceeds to step S 217 . When FSRMOVEND=1 indicating that the SOx removal process has been completed, it is determined in step S 216 whether or not the engine rotational speed NE exceeds a value NKBSSRL (for example, 2000 rpm) which is set slightly larger than NKBSL. When the answer is YES, it is further determined in step S 217 whether or not the engine rotational speed NE is smaller than a predetermined value NKBSH (for example, 3000 rpm). When NE<NKBSH, the process proceeds to step S 218 .

When the engine rotational speed NE is not larger than NKBSL or NKBSSRL in steps S 214 and S 216 , or when the engine rotational speed NE is in a higher rotation region than NKBSH in step S 217 , the process proceeds to step S 222 to prohibit the lean-burn operation.

In step S 218 , it is determined whether or not a gear position (a gear position conversion value of the CVT car) is equal to or higher than the third gear position. When it is equal to or lower than the second gear position, the lean-burn operation is prohibited.

When the answer in step S 218 is YES, that is, when the gear is in a higher position, it is determined in step S 219 whether or not a vehicle speed VP is equal to or larger than a predetermined value VNGRL (for example, 30 km/h). If the answer is YES, it is determined in step S 220 whether not the SOx removal end flag FSRMOVEND is set to 1. When FSRMOVEND=0, the process proceeds to step S 223 . When FSRMOVEND=1 indicating that the SOx removal process has been completed, it is determined in step S 221 whether or not the vehicle speed VP is equal to or larger than a value VNGRSR (for example, 40 km/h) which is slightly larger than VNGRL. If this answer is YES, the lean-burn prohibition flag FKBSMJ is set to zero, so as to permit the lean operation.

When the answer in step S 219 or S 221 is NO, the process proceeds to step S 222 , in which the lean-burn operation is prohibited.

When the high-density-sulfur containing fuel is used (FSLF=1), the lean operation is not permitted in step S 222 so as to perform the SOx removal process. When the high-density flag FSLF is set to zero, the lean operation may be permitted depending on the operating conditions of the engine 1 . Specifically, the lean operation is permitted in step S 223 only when the engine load PBGA is low, the engine rotational speed NE is in a low rotation region, the gear is equal to or higher than the third position and the vehicle speed VP is in an intermediate speed region. Besides, when the SOx removal process has been completed, reference values in terms of the engine rotational speed NE and the vehicle speed VP are reset to higher values so as to determine whether to permit the lean operation. Thus, because the sulfur adhesion to the three-way catalyst 14 relates with the exhaust gas temperature, the sulfur adhesion can be reduced by prohibiting lean-burn operation when the engine temperature is low and the exhaust gas temperature is low or when the engine rotational speed is low. As a result, the influence by the sulfur is excluded, and accuracy of the deterioration determination can be improved.

Although the present invention was described above with respect to specific embodiments, such embodiments are not intended to limit the scope of the invention but various variations and alternatives should be regarded to be included in the scope of the invention.

According to the invention, which provides means for correcting the abnormality determination criterion for the nitrogen oxide purifier based on the change in the outputs of the downstream O 2 sensor correspondingly to the change in the outputs of the upstream O 2 sensor, it is possible to establish an appropriate decision criterion upon the variation of changes in the air-fuel ratio (in the upstream O 2 sensor) on the upstream side of the NOx purifying device. As a result, the influence of the upstream three-way catalyst on the O 2 sensors can be excluded, so that the deterioration of the NOx purifying device can be accurately determined.

Claims

1. An exhaust gas purifying apparatus for an internal-combustion engine, comprising:a three-way catalyst disposed in an exhaust system of the engine; a nitrogen oxide purifier disposed downstream of the three-way catalyst to clean nitrogen oxide contained in the exhaust gas when air fuel ratio of the exhaust gas is lean; an upstream oxygen sensor disposed between the three-way catalyst and the nitrogen oxide purifier; a downstream oxygen sensor disposed downstream of the nitrogen oxide purifier; determining means for determining abnormality of the nitrogen oxide purifier based on change in the output of the downstream oxygen sensor when air-fuel ratio is altered from lean to rich; and correcting means for correcting criterion for determining abnormality of the nitrogen oxide purifier by said determining means in relation to a manner the output of the upstream oxygen sensor changes, wherein said criterion is defined in terms of the delay time from reversal of the output of the upstream oxygen sensor to reversal of the output of the downstream oxygen sensor when air-fuel ratio is turned from lean to rich, the determining means determining abnormality of the nitrogen oxide purifier when the output of the downstream oxygen sensor reverses before the delay time as defined by the criterion elapses, and wherein the correcting means correct the criterion based on the speed that the output of the upstream oxygen sensor changes when air-fuel ratio is turned from lean to rich.

2. The exhaust gas purifying apparatus according to claim 1, wherein said speed is represented by intake air volume taken-in from the time that the output of the upstream oxygen sensor started to rise till the time said output reaches a predetermined value.

3. An electronic control unit for an internal-combustion engine having a three-way catalyst disposed in an exhaust system of the engine, a nitrogen oxide purifier disposed downstream of the three-way catalyst to clean nitrogen oxide contained in the exhaust gas when air fuel ratio of the exhaust gas is lean, an upstream oxygen sensor disposed between the three-way catalyst and the nitrogen oxide purifier, and a downstream oxygen sensor disposed downstream of the nitrogen oxide purifier, the electronic control unit being programmed to;determine abnormality of the nitrogen oxide purifier based on change in the output of the downstream oxygen sensor when air-fuel ratio is altered from lean to rich; and correct a criterion for determining abnormality of the nitrogen oxide purifier by said determining means in relation to a manner the output of the upstream oxygen density sensor changes, wherein said criterion is defined in terms of the delay time from reversal of the output of the upstream oxygen sensor to reversal of the output of the downstream oxygen sensor when air-fuel ratio is turned from lean to rich, the electronic control unit further programmed to determine abnormality of the nitrogen oxide purifier when the output of the downstream oxygen sensor reverses before the delay time as defined by the criterion elapses, and wherein the criterion is corrected based on the speed that the output of the upstream oxygen sensor changes when air-fuel ratio is turned from lean to rich.

4. The electronic control unit according to claim 3, wherein said speed is represented by intake air volume taken-in from the time that the output of the upstream oxygen sensor started to rise till the time said output reaches a predetermined value.

5. A Method for diagnosing an exhaust gas purifying apparatus for an internal-combustion engine having a three-way catalyst disposed in an exhaust system of the engine, a nitrogen oxide purifier disposed downstream of the three-way catalyst to clean nitrogen oxide contained in the exhaust gas when air fuel ratio of the exhaust gas is lean, an upstream oxygen sensor disposed between the three-way catalyst and the nitrogen oxide purifier, and a downstream oxygen sensor disposed downstream of the nitrogen oxide purifier, comprising;determining abnormality of the nitrogen oxide purifier based on change in the output of the downstream oxygen sensor when air-fuel ratio is altered from lean to rich; and correcting a criterion for determining abnormality of the nitrogen oxide purifier by said determining means in relation to a manner the output of the upstream oxygen density sensor changes, wherein said criterion is defined in terms of the delay time from reversal of the output of the upstream oxygen sensor to reversal of the output of the downstream oxygen sensor when air-fuel ratio is turned from lean to rich, and the nitrogen oxide purifier is determined abnormal when the output of the downstream oxygen sensor reverses before the delay time as defined by the criterion elapses, and wherein the criterion is corrected based on the speed that the output of the upstream oxygen sensor changes when air-fuel ratio is turned from lean to rich.

6. The method according to claim 5, wherein said speed is represented by intake air volume taken-in from the time that the output of the upstream oxygen sensor started to rise till the time said output reaches a predetermined value.