US20070012027A1

US20070012027A1 - NOx control for an internal combustion engine

Info

Publication number: US20070012027A1
Application number: US11/492,938
Authority: US
Inventors: Koon Yang
Original assignee: Individual
Current assignee: Individual
Priority date: 2001-10-08
Filing date: 2006-07-26
Publication date: 2007-01-18
Also published as: US20040261397A1; EP1442201A1; AUPR812301A0; WO2003031782A1

Abstract

An improved NOX control for an internal combustion engine having an exhaust system with a NOX storage catalyst (105) includes a gas composition sensor (110) upstream of the NOX catalyst. Whether the NOX catalyst is storing NOX onto the catalyst or purging NOX from the catalyst is estimated from the signal output from the gas composition sensor. The composition of exhaust gasses input to the catalyst is also estimated using variables like engine load, engine speed, engine temperature and space velocity of the exhaust gases. This enables the NOX flow rate input to the catalyst and the aggregate quantity of NOX stored on the catalyst to be estimated.

Description

FIELD OF INVENTION

This invention relates to internal combustion engines and emissions standards for internal combustion engines and more particularly to calibration, control and operation of direct injection engines and lean burn stratified charge engines so as to comply with certain emissions standards.

BACKGROUND

A majority of fuel injected engines/vehicles currently produced by the major engine and vehicle manufacturers of the world are of the conventional manifold multipoint or port fuel injection (MPI/PFI) type. However, with the ongoing efforts to extract better performance from internal combustion engines and the emerging requirements to meet stringent emission laws, more and more engine manufacturers are investigating and developing direct fuel injection engine technologies. Such direct injection (DI) engine technologies are considered by some to be the next evolutionary step for internal combustion engines and examples of automotive vehicles incorporating DI engines are in fact already available to the consumer in certain automotive markets.
Almost all current design MPI and DI engines include a catalytic converter or exhaust gas after treatment system of some nature located in the exhaust system of the vehicle. The catalytic converter typically acts to convert undesirable exhaust emissions such as hydrocarbons (HC), carbon monoxide (CO), and oxides of nitrogen (NOx) into non-harmful substances such as carbon dioxide, nitrogen, oxygen and water. Increasingly stringent internal combustion engine emissions legislation around the world, such as the proposed US ULEV II & SULEV emissions regulations and European Euro IV regulations, is placing increasing pressure on engine and vehicle manufacturers to reduce engine emissions.
In meeting these stringent emissions standards most MPI and PFI vehicles suffer a fuel consumption penalty even though overall tail pipe emission levels decrease. This increase in fuel consumption arises for various reasons including increased engine hardware requirements that serve to increase the level of parasitic loading on the engine, increased fuel consumption due to the catalyst light-off strategies used and an increase in fuel consumption that arises when the engine is calibrated to produce reduced levels of hydrocarbon emissions and NOx emissions. DI engines promise fuel economy savings and are being looked to as a potential technology to offset increases in fuel consumption that are likely to result from the introduction of low emissions vehicles.
A significant portion of the current activity taking place in respect of DI engine technology development is in relation to lean burn engines. A lean burn engine is able to operate with a lean air fuel ratio. Typically this occurs at lower load operating points. Lean operation allows significant fuel consumption reductions to be realised. Typically, lean burn engines also operate with a stoichiometric air/fuel ratio at higher operating load points.
Lean operation is typically associated with the provision of a stratified fuel-air charge in the combustion chamber whilst stoichiometric operation is typically associated with a homogenous fuel-air charge.
There are however certain challenges to be faced when effecting lean burn engines. In particular, lean operation often results in the formation of NOx emissions. Reduction of NOx emissions by after treatment systems can be difficult to effect. In this regard, conventional three-way catalytic converters (TWC's) have been found to be unsatisfactory for efficiently treating NOx emissions. One present way of addressing this is by incorporating a further Lean NOx catalyst (LNC) catalyst which acts to adsorb NOx gases emitted from the engine until its operating conditions are more favourable for reduction of the trapped NOx gases by the exhaust gas treatment system. The favourable engine condition is typically when the engine is operating with a rich or stoichiometric air/fuel ratio. Accordingly, in current systems incorporating an LNC, it has been found necessary for the engine to temporarily run with a rich air/fuel ratio to promote reduction of NOx stored/trapped on the LNC. However, these rich excursions impose a fuel consumption penalty that detracts from the fuel consumption benefits available from DI engines. Accordingly, it is desirable to reduce the extent of any rich excursion used in regenerating an LNC.
It is therefore an object of the present invention to provide improved NOx control for an internal combustion engine.

SUMMARY OF INVENTION

With this in mind, there is provided in accordance with a first aspect of the present invention a method of operating an internal combustion engine having a first catalyst with NOx storage capability; said method comprising the steps of determining the composition of gasses input to said catalyst to thereby determine an aggregate quantity of nitrogen oxide compounds (NOx) stored on said catalyst.
Preferably said estimate of said aggregate quantity of NOx stored on said catalyst is performed at intermittent time intervals.
Preferably said intermittent time intervals further comprise a background loop of an electronic control unit of said engine.
Preferably said method further comprises the step of estimating a quantity of NOx that has either been stored onto said catalyst or reduced from said catalyst over a recent one of said intermittent time intervals, and summing said quantity of NOx estimated to have been stored or reduced over said recent one of said intermittent time intervals with an aggregate quantity of NOx estimated to have been stored on said catalyst over a previous one of said intermittent time intervals.
Preferably said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating NOx flow to said catalyst.
Preferably said NOx flow is derived from at least engine speed and engine load
Preferably said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating engine temperature.
Preferably said NOx flow and said intermittent time interval is used to determine a total quantity of NOx input to said catalyst during a recent one of said intermittent time intervals.
Preferably said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating the space velocity of said gasses input to said catalyst.
Preferably said space velocity and said total quantity of NOx input to said catalyst are used to determine said quantity of NOx stored on said catalyst.
Preferably said step of estimating NOx reduced from said catalyst further comprises the step of estimating a NOx purge time period based on said aggregate quantity of NOx stored on said catalyst.
Preferably said NOx purge time comprises the time required to purge a pre-determined quantity of NOx from said catalyst under predetermined engine operating conditions.
Preferably said step of estimating NOx reduced from said catalyst comprises the further step of estimating a NOx reduction flow to said catalyst.
Preferably said NOx reduction flow is estimated from the space velocity of gasses input to said catalyst and from the composition of said gasses.
Preferably said step of estimating the quantity of NOx reduced from said catalyst is derived from said NOx reduction flow and said NOx purge time.
Preferably said step of monitoring the composition of gasses into to said catalyst further comprises determining the molar weight ratio of carbon monoxide (CO) and NOx in said gasses.
Preferably said method further comprises the steps of estimating that NOx is stored on said catalyst when said molar weight ratio is less than 3.0 and estimating that NOx is reduced from said catalyst when said molar weight ratio exceeds 3.0.
Preferably said engine further comprises a first gas composition sensor located upstream of said storage catalyst and wherein said step of determining the composition of gasses input to said catalyst further comprises monitoring an output signal of said first gas composition sensor.
Preferably said method further comprises the steps of monitoring said aggregate quantity of NOx, determining if said aggregate quantity exceeds a threshold value and operating said engine so as to reduce NOx stored on said catalyst when said aggregate quantity exceeds said threshold value.
Preferably said engine is operated with a rich air fuel ratio providing exhaust gasses with molar weight ratio of CO and NOx of 3.0 or greater in order to reduce NOx stored on said catalyst.
Preferably said engine is operated with a closed loop air fuel ratio to thereby provide said rich air fuel ratio for reducing NOx stored on said catalyst.
Preferably said engine further comprises a second catalyst upstream of said first catalyst, said second catalyst being a three way catalyst and said first gas composition sensor located intermediate said first and second catalysts.
In accordance with another aspect of the present invention there is provided an internal combustion engine and exhaust system; the engine comprising an exhaust manifold with outlet for egress of raw combustion gasses and an electronic control unit (ECU) for controlling operation of the engine and the exhaust system comprising a three way catalyst (TWC) located separately from a lean NOx catalyst (LNC), the TWC having an inlet port located so as to receive the raw combustion gasses from the outlet of the exhaust manifold and an outlet port for egress of partially treated combustion, the LNC having an inlet port to receive the partially treated combustion gasses and an outlet port for egress of exhaust gasses and an O₂sensor located intermediate the outlet of the TWC and the inlet to the LNC wherein during at least pre-determined TWC operating conditions said ECU monitors the output of said O₂sensor to thereby control engine air fuel ratio so as to control composition of treated exhaust gas products input to the LNC.
Preferably said ECU monitors the output of said O₂sensor to thereby determine whether said LNC is storing NOx or being purged of NOx.
Preferably said ECU monitors the output of a first O₂sensor operatively associated with an inlet to the TWC to thereby provide closed loop air fuel ratio control.
Preferably the closed loop air fuel ratio control is λ≈1 control.
Preferably the closed loop air fuel ratio control perturbates the air fuel ratio and wherein this perturbation is between lean and rich air fuel ratios to thereby provide λ≈1 control.
Preferably said rich perturbations are controlled in response to the output of O₂sensor associated with the output of the TWC so as to control whether said LNC is purging or storing NOx.
According to a further aspect of the present invention there is provided a NOx model for estimating the aggregate quantity of NOx stored on a lean NOx catalyst for use with a lean burn engine and an exhaust system having a first catalyst for treating engine out emissions and a second catalyst wherein said second catalyst is a lean NOx catalyst and an oxygen sensor located intermediate said first catalyst and said second catalyst wherein said model receives as input data at least
m said oxygen sensor wherein said data indicates the composition of exhaust gas input to said lean NOx catalyst whereby said model estimates whether said lean NOx catalyst is storing NOx or reducing NOx.
According to a further aspect of the present invention there is provided an electronic control unit for controlling operation of an engine having an exhaust after treatment system with a NOx storage catalyst; said ECU determining the composition of gasses input to the catalyst to thereby determine an aggregate quantity of nitrogen oxide compounds (NOx) stored on said catalyst.
Preferably said ECU further determines the composition of gasses input to the catalyst further comprises determining the molar weight ratio of carbon monoxide (CO) and NOx in said gasses.
Preferably said ECU estimates that NOx is stored on said catalyst when said ratio is less than 3.0 and estimates that NOx is reduced from said catalyst when said ratio exceeds 3.0.
Preferably said exhaust after treatment system further comprises a gas composition sensor upstream of said catalyst; said ECU receiving signals output by said gas composition sensor which said signals indicate said molar weight ratio.

DESCRIPTION OF THE DRAWINGS

Preferred embodiments will now be described by way of example only and with reference to the accompanying drawings in which:
FIG. 1 a is a schematic representation of a direct injection internal combustion engine utilising a three way catalyst (TWC) and an lean NOx catalyst (LNC).
FIG. 1 b is a pictorial representation of the exhaust system of FIG. 1 a and shows the LNC and the TWC.
FIG. 2 is a graph depicting the characteristics of feed gas into the LNC of FIG. 1 under closed loop stoichiometric air fuel ratio operating conditions.
FIG. 3 is a flow chart representing a method of NOx control.
FIG. 4 is a control diagram for estimating the aggregate quantity of NOx adsorbed by an LNC.

BEST MODE FOR CARRYING OUT THE INVENTION

A preferred embodiment utilises an exhaust system having a three way catalyst (TWC) and a lean NOx catalyst (LNC). Through knowledge of the properties of the TWC and the LNC these embodiments estimate the amount of NOx stored on the LNC and the amount purged when the engine is operated in a NOx purge mode (i.e. an LNC regeneration mode). The engine is controlled in accordance with the amount of NOx estimated to have been stored on the LNC.
The TWC is located upstream of the LNC. The TWC receives raw emissions from the engine. The TWC processes these emissions and outputs emissions that may be referred to as intermediate emissions. The intermediate emissions are processed by the LNC. The TWC may be a close coupled catalyst, i.e. a catalyst that is located in close proximity to the engine's exhaust ports. Gasses treated by the LNC are expelled to atmosphere as exhaust gasses or tail pipe gasses/emissions. The LNC may be an underbody catalyst.
A gas composition sensor, such as an “oxygen” or an “O₂” sensor is associated with an inlet to the LNC. By using this sensor, the composition of gasses input to the LNC can be understood, which allows the level of NOx stored by the LNC to be estimated accurately.
Referring now to FIGS. 1 a and 1 b which depict a direct injection engine with an exhaust after treatment system having a TWC and an LNC. FIG. 1 b is a schematic representation of an internal combustion engine 120 utilising an exhaust system 100 having a TWC canister 115 that houses a TWC160. The TWC 160 is housed internally to the TWC canister 115 (as depicted in FIG. 1 b). The TWC canister 115 is located separately from an LNC canister 105. An LNC 165 is housed internally to the LNC canister 105 (as depicted in FIG. 1 b). The internal combustion engine 120 is an example of a direct injection engine where fuel is injected directly into the combustion chambers 126 of the engine 120 by use of a direct injection fuel system using direct injection fuel injectors 170.
A typical direct injection engine 120 has the capacity to operate with a stratified distribution of fuel and air in the combustion chamber 126. Such a stratified distribution may consist of a first region in the combustion chamber with a stoichiometric or near stoichiometric air fuel ratio and a second region with only a limited amount of fuel if any and therefore a very lean air fuel ratio. Engines capable of supporting such a stratified distribution of fuel are often referred to as stratified charge engines. As only a portion of the combustion chamber 126 in a stratified charge engine has a stoichiometric, or near stoichiometric air fuel ratio, the combustion chamber will often operate with an overall lean air fuel ratio. Accordingly, such direct injection engines are also commonly referred to as ‘lean burn’ engines due to this capacity to operate with combustion chambers 126 having lean air fuel ratios. Such engines may be arranged so that the stoichiometric or near stoichiometric charge of fuel and air is positioned within the combustion chamber 126 so that it may be ignited by an ignition device such as a spark plug.
A stratified charge engine may be operated under certain part load conditions so that the quantity of fuel delivered to the engine is varied relatively independently of throttle position, i.e. independently of the quantity of air induced into a combustion chamber for a combustion event. Where this occurs, the quantity of fuel typically varies with operator demand rather than engine load (i.e. throttle position or air flow into the engine). A stratified charge engine can operate in this manner as any increase or decrease in fuelling level results in a corresponding increase or decrease in the volume of the stratified charge within the combustion chamber. An engine operated in this manner may be said to be operated in a “fuel led” mode. By providing fuel to an engine relatively independently of its throttle position the pumping work of the engine against the throttle can be reduced which results in fuel economy advantages. This is particularly the case where the engine utilises exhaust gas recirculation (EGR). EGR is where exhaust gas is fed back into the combustion chamber in a controlled manner. Due to exhaust back pressure, the pumping work for EGR is lower than for atmospheric air.
The ability to operate with a lean air fuel ratio means that the air/fuel charge in the combustion chamber 126 is oxygen rich compared with a charge distributed homogenously throughout the combustion chamber 126. This results in such direct injection engines having higher quantities of NOx (oxides of Nitrogen) in their combustion gasses than would normally be seen for a homogenous charge engine. N.B. a homogenous charge engine operates with a stoichiometric or near stoichiometric air fuel ratio substantially throughout the volume of its combustion chambers. Typically a multi-point fuel injected engine operates as a homogenous charge engine. The level of NOx produced by a stratified charge engine can be tempered by introducing EGR into the combustion chamber.
Direct injection engines may utilise different mechanisms to generate a stratified mixture of fuel and air in a combustion chamber. This includes spray guided (or jet guided) mechanisms, wall guided mechanisms and air guided (or motion guided) mechanisms. Spray guided mechanisms typically locate a fuel injector in an axial or near axial position relative to the combustion chamber. A spark plug is typically located down stream from a nozzle of the fuel injector so that spray issuing from the fuel injector passes over a spark gap in the spark plug. In this way the spark plug and fuel injector are arranged so that the spray can be ignited as it issues from the fuel injector. Typically the tail end of the spray issuing from the injector is ignited directly by a spark plug without first deflecting off any other surfaces in the combustion chamber. The lead portion of the spray may deflect off the piston prior to ignition of the tail portion. Typically ignition of the tail of the spray occurs after the fuel injector has ceased operation or immediately prior to the fuel injector ceasing operation. An engine arranged in this manner to ignite fuel as it issues from an injector can be referred to as ‘spray guided’.
Engines using gasoline as fuel, and therefore a spark plug to ignite the fuel charge may be referred to as ‘gasoline’ engines or alternatively they may be referred to as a ‘spark ignited’ engines.
LNCs are catalyst that store NOx and reduce this NOx to N₂in the presence of CO. A typical LNC reduction reaction may be conceptualised as: NO_x+CO→N₂+CO₂. Internal combustion engines tend to produce greater quantities of NOx when the air/fuel ratio in the combustion chamber 125 is lean (i.e. there is an excess of air in the combustion chamber) than when operated with a stoichiometric air fuel ratio. The symbol “λ” is commonly used to designate an air fuel ratio normalised by 14.6. 14.6 is selected as the normalisation factor as the stoichiometric air fuel ratio is 14.6 parts air to one part fuel. Hence an air fuel ratio designated by λ=1.0 is a stoichiometric air fuel ratio. λ>1.0 designates a lean air fuel ratio as there is more than 14.6 parts of air to every one part of fuel. Similarly λ<1.0 designates a lean air fuel ratio as there is less than 14.6 parts of air to one part of fuel. A engine operating with a lean air fuel ratio will not necessarily produce excess NOx, temperature and pressure conditions within a combustion chamber operating at λ>1 must be sufficient for excess NOx to be produced.
When the air/fuel ratio is rich (λ<1), oxidation of fuel may not be as complete as compared with combustion of an air/fuel mixture having a stoichiometric ratio (λ=1). This incomplete combustion results in a higher proportion of CO appearing in the combustion gasses of a rich air fuel mixture. This excess of CO enables NOx stored on an LNC to be purged (i.e. released from the catalyst substrate) and then reduced to N₂.
TWCs are catalysts that operate to promote further oxidation of partially combusted engine gasses. Typical TWC oxidation reaction is HC+CO+O₂→H₂O+CO₂. A typical TWC will also promote reduction of NOx under stoichiometric and rich operating conditions, however a TWC commonly will not store or reduce NOx under lean operating conditions. Accordingly the TWC can be represented as facilitating the following reactions under stoichiometric and rich operating conditions: HC+CO+NOx+O₂→H₂O+CO₂+N₂. Under lean operating conditions the TWC can be represented as facilitating the following reactions: HC+CO+O₂→H₂O+CO₂. In other words, under lean operating conditions, the amount of NOx present in engine out exhaust emissions should be the same as the amount present in intermediate emissions that are output from the TWC and that form feed gas for the LNC.
Certain embodiments may utilise an air assisted or dual fluid injection system. Such systems typically use compressed air as a propellant for delivery of a metered quantity of fuel into a combustion chamber of an engine. For example, a dual fluid fuel injector may utilise a holding chamber which is in constant communication with a source of compressed air. Such a holding chamber is at an elevated pressure relative to atmosphere. A predetermined quantity of fuel is metered into the holding chamber each cylinder cycle. This typically occurs when the holding chamber is shut off from the combustion chamber. Once the fuel has been metered, the holding chamber can be opened to the combustion chamber and the pressure differential between the holding chamber and the combustion chamber allows the compressed air to be expelled from the injector into the combustion chamber. The compressed air that flows into the combustion chamber also carries the fuel from the holding chamber into the combustion chamber. The fuel is entrained and atomised by the compressed air. Such dual fluid injection systems are often referred to as low pressure direct injection systems as the pressure of the compressed air is typically less than the pressure used in single fluid direct injection systems, i.e. systems that simply inject gasoline directly into a combustion chamber using a pressure time metering principle. Further detail on dual fluid direct injection fuel systems may be found in the applicants patents U.S. Pat. No. RE 36768 & WO99/28621.
To maintain tail pipe NOx emissions at certain desired levels, present embodiments estimate the quantity of NOx stored on the LNC 165 during engine operation. This estimate is based on firstly estimating the quantity of NOx stored by an LNC under engine operating conditions conducive to the production of NOx gasses and secondly estimating the quantity of NOx reduced to N₂under engine operating conditions conducive to the reduction of NOx. The embodiments monitor exhaust gas composition input to the LNC in order to make these estimates. As such, the embodiments are particularly well suited to lean burn engines that have a tendency to produce higher levels of NOx gasses under lean operating conditions.
The exhaust treatment system 100 has a front and a rear gas composition sensors 135 and 110 respectively, which may be oxygen sensors. The front oxygen sensor 135 is located in the exhaust manifold 120 of the engine adjacent an inlet to the TWC housing 115 and the rear oxygen sensor 110 is located in the intermediate exhaust section 150 adjacent an outlet of the TWC housing 115.
The front sensor 135 may be used to effect closed loop air fuel ratio (AFR) control (commonly referred to as “λ=1.0” control). Such closed loop AFR control monitors the output of the front oxygen sensor 135 to ensure that the air fuel ratio input to the engine's combustion chambers is stoichiometric. Typically closed loop AFR control is used under high load operating conditions with a homogenous distribution of air and fuel of a substantially stoichiometric ratio (λ=1.0) throughout the combustion chamber.
Gas composition sensors such as the front and rear oxygen sensors 135 & 110 respectively, typically provide an output proportional to the amount of oxygen in the exhaust gas. Through knowledge of the quantity of air inlet to an engine's combustion chambers from cycle to cycle, and knowledge of the quantity of fuel combusted from cycle to cycle, it is possible to determine a set point voltage for the oxygen sensor, which set point voltage indicates combustion of stoichiometric mixtures of air and fuel in the engine's combustion chambers. Typically an oxygen sensor voltage of 0.45 volts will indicate that the engine is operating with a stoichiometric mixture of air and fuel (λ=1.0).
The rear oxygen sensor 110 indicates the amount of oxygen in the partially treated exhaust gasses that are expelled from the TWC housing 115 and input to the LNC housing 105. This enables the quantity of CO in the partially treated exhaust gasses to similarly be determined. When sufficient CO is present in the partially treated exhaust gasses, NOx stored on the LNC is purged from the LNC substrate and then reduced. The rear oxygen sensor indicates whether or not there is sufficient CO present in the feed gas to the LNC to purge and reduce NOx stored on the LNC. Additional information on the space velocity of the feed gas for the LNC in combination with the rear oxygen sensor voltage 110 indicates the rate at which NOx is purged and reduced from the LNC.
Present embodiments monitor the rear oxygen sensor 110 in order to determine when the partially treated exhaust gasses will cause purging of the LNC and to estimate the quantity of NOx purged. This enables accurate control of tail pipe NOx emissions as it prevents the LNC from inadvertently reaching saturation level where it can no longer store NOx. The condition where an LNC saturates is often referred to as a “breakthrough” condition.
The LNC housing 105 also locates a first catalyst temperature sensor 185 that monitors the temperature of the LNC 165. The temperature of the LNC 165 is affected by any chemical reactions that it promotes along with the temperature of the intermediate emissions that pass though it.
An electronic control unit (ECU) 190 monitors the front and rear oxygen sensors 135 & 110, and first catalyst temperature sensor 185. The ECU controls operation of the engine to implement various control strategies. The ECU receives signals from various sensors mounted on the vehicle engine. It executes control strategies in response to these signals and transmits control signals to various actuators in response to the computation of the control strategy. The ECU controls operation of the engine through this actuation of components, such as fuel injectors and spark plugs at appropriate points in time for appropriate durations. The ECU operates as a “real time” system with background and foreground processes. A foreground processes is a process, such as actuation of a fuel injector or a spark plug, that is time critical and so must occur at a specified point in time in order to ensure continued operation of the engine. Processes that must occur at least once every cylinder cycle are controlled in a foreground loop. A background process is a process that is not time critical and can occur when the ECU has sufficient processing resources in between foreground processes.
Referring now to FIG. 2 which is a graphical representation of a typical intermediate feed gas composition for the present embodiments (i.e. composition of feed gasses for the LNC) when the engine is operated under closed loop air fuel ratio control about λ=1.0. The graph indicates typical intermediate feed gas composition for air fuel ratios between λ=1.02 and λ=0.98 (i.e. for air fuel ratio perturbations within the air fuel ratio window of between λ=1.02 and λ=0.98). This range of air fuel ratios is indicated by the rear oxygen sensor voltage 110 that forms the horizontal axis of the graph. λ=1.02 is typically indicated by an oxygen sensor voltage of 0.2 v and λ=0.98 is typically indicated by an oxygen sensor voltage of 0.8 v. It can be observed that the level of CO increases with increasing oxygen sensor voltage and that the level of NOx and O₂decreases with increasing oxygen sensor voltage. In other words, the level of CO in the intermediate feed gas increases as the air fuel ratio becomes rich and the level of NOx and O₂increases as the air fuel ratio becomes lean.
A typical oxygen sensor is constructed so as to be highly sensitive to the level of oxygen in engine out emissions when the engine is operated with air fuel ratios in a band between λ=0.98 and λ=1.02. Small changes either side of the stoichiometric ratio cause large swings in the oxygen sensor output voltage.
Under closed loop stoichiometric air fuel ratio control the air fuel ratio is caused to perturbate about λ=1.0 by typically ±0.02. These perturbations cause rapid cyclical changes to the feed gas composition for the TWC. The TWC operates by storing excess oxygen when the air fuel ratio is slightly lean and releasing the stored oxygen when the air fuel ratio is slightly rich. This storage and release of oxygen facilitates oxidation reactions promoted by the TWC. The oxygen sensor output voltage has a substantially binary response to these perturbations in air fuel ratio (i.e. the output of the oxygen sensor has primarily digital output between two voltages in response to the air fuel ratio perturbations). This binary like response is used to control switching of the air fuel ratio from a lean state to a rich state and from a rich state to a lean state at least during the air fuel ratio perturbations.
During closed loop air fuel ratio control about λ=1.0 where perturbation of the air fuel ratio is utilised, the effect of the TWC on the engine out feed gas is to produce an intermediate feed gas whose composition changes at a slower rate than the composition of the engine out gasses. This slower rate is seen as the output voltage of the rear oxygen sensor 110 having a more constant level over time with less dramatic changes in voltage level compared with the output voltage of the front oxygen sensor 335.
When the molar weight ratio of CO/NOx in the intermediate feed gas input to the LNC is greater than 3.0, the LNC efficiently purges stored NOx, reduces this purged NOx and also reduces free NOx in the feed gas. Present embodiments monitor the output voltage of the rear oxygen sensor 110 to determine whether or not the CO level in the intermediate feed gas to the LNC is sufficient to purge NOx from the LNC. When the rear Oxygen sensor voltage is greater than a threshold (e.g. 0.6 volts) then the present embodiments determine that the LNT is purging and reducing NOx and accordingly the quantity of NOx purged and reduced from the LNC is estimated. When the rear oxygen sensor voltage is below a threshold, present embodiments estimate the quantity of NOx that is being stored on the LNC.
Engines typically produce intermediate feed gas with a molar weight ratio of CO/NOx of greater than 3.0 under operating conditions richer than stoichiometric. How rich the air fuel ratio should be to produce a molar weight ratio exceeding 3.0 depends on the level of engine out NOx produced. For low engine out NOx, air fuel ratio with A in the range, λ=09.6-0.98 may be used, however λ=0.80 may be required for engines producing higher levels of engine out NOx.
Forced operation of an engine so as to produce a molar weight ratio of CO/NOx greater than 3.0 typically occurs by operating the engine under closed loop air fuel ratio conditions. Accordingly, closed loop air fuel ratio conditions may be used to effect purging and reduction of some or all of the NOx stored on an LNC. By monitoring the output voltage of the rear oxygen sensor during such closed loop operation, the rich bias of the air fuel ratio perturbations may be varied to ensure that the molar ratio of CO/NOx in the intermediate feed gasses input to the LNC is greater than 3.0. Such monitoring also allows for variations to the air fuel ratios used to purge the LNC under closed loop stoichiometric air fuel ratio control conditions. These variations compensate for TWC performance degradation with age and/or use. For example, as the performance of the TWC degrades it may be necessary to optimise the level of rich bias used during rich biased perturbation.
Referring now to FIG. 3 which is a flow chart detailing a NOx estimate process 300 by which the ECU 190 estimates the aggregate quantity of NOx stored on the LNC 165 during operation of the engine 120.
The NOx estimate process 300 determines whether the composition of the intermediate feed gas to the LNC results in NOx being stored on the LNC 165 or purged from the LNC 165. Using this information it estimates the aggregate quantity of NOx stored on the LNC 165 at any point during operation of the engine. This assists with preventing LNC breakthrough conditions from occurring, as the engine can be controlled to provide operating conditions that promote purging of NOx from the LNC 165 when required. This can also reduce fuel consumption as the LNC purge process is estimated from the feed gas to the LNC. This can reduce the need for reliance excessively rich air fuel ratios to ensure effective purging.
The NOx estimate process may be a background loop of ECU 190 which may be requested at step 305 on an intermittent time basis relative to the foreground processes of the ECU at least. Once triggered, the NOx estimate process 300 proceeds to step 310 where the ECU 190 determines whether or not a request to purge the LNC 165 of some or all of its NOx is current. Such a request may be generated by ECU 190 when the aggregate quantity of NOx estimated to be stored on LNC 165 exceeds a pre-determined threshold. If such a request is not current then the NOx estimate process 300 moves to step 320 where it is determined whether or not the engine 120 is operating in a “power enrichment” mode. Power enrichment modes typically occur at or near wide open throttle (WOT) conditions and as such typically represent homogenous operation of the engine 120. A typical power enrichment mode operates the engine with a slightly rich air fuel ratio and so produces an excess of CO that may reduce NOx stored on the LNC 165 if the molar weight ratio of CO/NOx is sufficient. If at step 320 the engine 120 is operating in a power enrichment mode then the NOx estimate process 300 moves to step 325 otherwise it moves to step 312. At step 312 a NOx storage model is activated and the quantity of NOx stored on the LNC 165 since last calculation is determined and the aggregate quantity of NOx stored on the LNC 165 is updated.
If at step 310 it is determined that a request to purge the LNC 165 of NOx is current then the process moves to step 315 where the ECU 190 checks whether or not the engine has entered into closed loop AFR control. If the engine 120 has entered closed loop AFR control then the NOx estimate process 300 moves to step 325 otherwise the NOx estimate process 300 moves to step 320. As detailed above, step 320 determines whether or not the engine is being operated in a power enrichment mode.
At step 325 the NOx estimate process 300 investigates the output of the rear oxygen sensor 110 in order to determine the quantity of CO in the partially treated exhaust gases. If the quantity of CO is insufficient for NOx stored on the LNC 165 to be purged and reduced, then the NOx estimate process 300 moves to step 312 and the quantity of NOx stored on the LNC 165 is calculated. If the quantity of CO is sufficient for the NOx stored on the LNC to be reduced then the NOx estimate process 300 moves to step 330.
At step 330 a NOx purge model is enabled and the quantity of NOx purged from the LNC 165 since the last calculation of aggregate NOx is determined and the aggregate quantity of NOx stored on the LNC 165 is then updated by subtracting the purged quantity of NOx from the previously estimated aggregate quantity of NOx.
Referring now to FIG. 4, which is a control diagram of an aggregate NOx model 400 for estimating the aggregate quantity of NOx stored on the LNC 165. The aggregate NOx model 400 estimates the aggregate quantity of NOx stored on the LNC 165 when the engine 125 is in operation. To do this, the aggregate NOx model 400 operates as a background loop, referred to as a NOx loop, of the ECU and estimates the quantity of NOx either stored onto or purged from the LNC 165 during the most recent ECU NOx loop. This estimate is then summed with the aggregate quantity of NOx that was calculated to have been stored on the LNC 165 at completion of the previous ECU NOx loop. In this way the aggregate quantity of NOx stored on the LNC 165 is updated from NOx loop to NOx loop. This allows the aggregate quantity of NOx to be to be an estimate based on continuous operation of the engine between engine start up and engine stopping. However the estimate is based on a series of calculations which are performed at intermittent time intervals through out engine operation. The aggregate quantity of NOx is stored as a variable called Residual_NOx_Mass_n. Accordingly, present embodiments calculate the aggregate quantity of NOx stored on the LNC 165 on an intermittent basis as ECU 190 processing permits.
To calculate the aggregate amount of NOx on the LNC 165 the aggregate NOx model 400 calculates the amount of NOx stored or purged during each ECU NOx loop. The aggregate NOx model 400 uses a Stored_NOx_mass _—n 428 variable and a Purge_NOx_mass _—n 450 variable to represent the amount of NOx stored or purged respectively from the LNC 165 during an ECU NOx loop. Calculation of the Stored_NOx_mass _—n 428 variable corresponds with, or is analogous to, the NOx storage model activated at step 312 in FIG. 3. Similarly, calculation of the Purge_NOx_mass _—n 450 variable corresponds with, or is analogous to the NOx purge model activated at step 330 of FIG. 3. Two blocks are identified by broken outline within the NOx model 400. Block A encloses the elements used to estimate the Stored_NOx_Mass _—n 428 variable. Block B encloses the elements used to estimate the Purge_NOx_Mass _—n 450 variable.
To determine whether LNC 165 is storing NOx or purging NOx, the NOx model uses a logical selection criteria 430, which selects between updating either the Stored_NOx_mass _—n 428 variable or the Purge_NOx_mass _—n 450 variable. The selected variable, once updated, is summed with a Residual_NOx_mass _—n−1 454 variable in order to update the Residual_NOx_Mass_—nvariable. The Residual_NOx_Mass_—nvariable indicates the aggregate amount of NOx stored on the LNC 165 at the end of the current ECU NOx loop. The Residual_NOx_mass _—n 456 variable for the current ECU NOx loop becomes the Residual_NOx_mass _—n−1 454 variable for the next ECU NOx loop.
The logical selection criteria 430 makes its selection by preferably referencing at least an output signal 434 of the rear oxygen sensor 110 and a NOx storage constant 432 that details the maximum amount of NOx that the LNC 165 is permitted to store. When the output signal 434 of the rear oxygen sensor 110 indicates sufficient CO in the exhaust gas for purging of the LNC to occur, the model 400 selects the Purge NOx_mass _—n 450 variable. Otherwise it selects the Stored_NOx_mass _—n 428 variable. Alternately, when the Residual_NOx_mass _—n 456 variable is equal to or exceeds the NOx storage constant 432, then the model 400 may request that the engine operate in a manner that will purge NOx from the LNC 165. Typically a closed loop air fuel ratio control mode is selected for controlling the engine during a purge process when the level of NOx stored on the LNC 165 exceeds a pre-determined threshold. The logical selection criteria 430 may also look at other variables such as the temperature of the engine 120 or the temperature of the exhaust gas input to the LNC 165 as these variables may also affect NOx storage and NOx purge rates or require the engine to be forced to operate for a period of time in a mode of operation that will purge NOx from the LNC 165.
The logical selection criteria 430 may also select the Purge_NOx_mass _—n 450 variable when the engine is operated in a power enrichment mode. Such a mode typically utilises an air fuel ratio that is biased toward being rich (i.e. λ=0.82) and is typically utilised under wide open throttle operating conditions. This rich operation will tend to produce an excess of CO resulting in purging and reduction of NOx adsorbed by the LNC165.
Typically, logical selection criteria 430 selects the Stored_NOx_mass, 428 variable for updating when the LNC is estimated to be storing NOx. This typically occurs when the engine is operated with lean or stoichiometric air fuel ratios over an ECU NOx loop and the rear oxygen sensor has output voltage below a threshold value of, for example, 0.6 v.
Assume now that the LNC is storing NOx and that the logical selection criteria 430 has selected the Stored_NOx_mass _—n 428 variable to update the Residual_NOx_mass _—n 456 variable. The aggregate NOx model 400 integrates a NOx flow rate 408 over a period of time represented by an event time 402 variable to provide a first estimate of NOx output from the engine 120. The event time 402 variable represents the amount of time that has elapsed since last the Residual_NOx_mass _—n 456 variable was updated. Accordingly the event time 402 corresponds with elapsed ECU 190 processing time since last commencing a background loop to update the Residual_NOx_mass _—n 456 variable (i.e. event time 402 corresponds with the duration of the ECU NOx loop).
The NOx flow rate 408 variable represents the average flow rate of NOx into the LNC 165 during the event time 402. This NOx flow rate 408 is determined with reference to an engine speed 404 variable and an engine load 406 variable. The engine speed 404 and the engine load 406 may be determined from the latest value for these variables at commencement of calculating the NOx flow rate 408. An average of these values over the event time 402 may alternately be provided.
The engine speed 404 and engine load 406 variables are used as input parameters to a NOx flow lookup map 408. The NOx flow lookup map 408 may be determined during calibration of the engine (e.g. calibration of an engine family as typically occurs prior to serial or low volume production) and provides data on the level of engine out NOx (i.e. engine out emissions prior to processing by the TWC 160) for a particular engine speed and load. Engine out NOx may be used to compile the NOx flow lookup map 408 as typically any NOx produced by an engine operating with a lean air fuel ratio will pass through the TWC 160 without being stored or reduced. Accordingly it has been found that a reasonable approximation is to assume that all engine out NOx passes through the TWC 160.
The level of engine out NOx produced by the engine 120 however may vary under different engine operating temperatures. For example, the engine 160 may produce more NOx when operating at higher temperatures compared with lower temperatures. Accordingly the aggregate NOx model 400 may compensate NOx flow rate data output from the NOx flow lookup map 408 for various engine temperatures. This compensation may be provided by way of a NOx compensation lookup map 414 that has as its input engine coolant temperature 412, or some other suitable measure of engine temperature. NOx compensation data output from the NOx compensation lookup map 414 is a variable that is multiplied 410 with the NOx flow data output from the NOx flow lookup map 408. The NOx compensation lookup map 414 may be generated during calibration of the engine 120, in which case the level of engine out NOx produced at varying engine speed and load points for different engine temperatures is determined.
As referred to above, the NOx flow data output from the NOx flow lookup map 408 is multiplied 410 with the NOx compensation data from the NOx compensation lookup map 414. The resultant compensated NOx flow data may then be filtered using a first order filter 416 to produce filtered compensated NOx flow data which may then be integrated 418 over the event time 402. This integration 418 of the filtered and compensated NOx flow data produces data representative of the amount of NOx output from the engine 120 during the event time 402 and accordingly input to the LNC 165.
The first order filter 416 allows the aggregated NOx model 400 to utilise data as to engine speed 440, engine load 406 and coolant temperature 412 upon commencement of calculating an update of the Residual_NOx_mass _—n 456 variable. At steady state operation of the engine, selection of data at commencement of an update calculation is considered to be an accurate approximation of system behaviour throughout the event time 402. Accordingly integration 418 over the event time 402 should introduce minimal errors into the updated Residual_NOx_mass _—n 456 variable. Where a transient occurs during the event time 402, the first order filter 416 operates as an approximation of the response of exhaust gas flows in the exhaust system to the transient. That is, the response of the exhaust gasses output from the engine 120 and the response of gasses within the exhaust system 100 to transient engine behaviour may be modelled as a first order response. Accordingly, applying compensated NOx flow data to a first order filter approximates the response of the system to the engine transients. Changes in exhaust gas flows to transient conditions during the time interval of an ECU NOx loop may be approximated as a first order response.
To determine the proportion the of NOx passing through the LNC that is stored on the LNC during an event time 402, a NOx storage factor is output from a NOx storage factor lookup map 426. The NOx storage factor is preferably a number between 0.0 and 1.0 (i.e. between zero and one) and is multiplied 420 with the amount of NOx output from the engine 120 over an event time 402 as determined by integrator 418. That is, the output of the NOx storage factor lookup map 426 is multiplied 420 with output from the integrator 418 in order to derive a number representing the quantity of NOx adsorbed by the LNC 165 during event time 402. Data resulting from this multiplication 420 forms the Stored_NOx_mass 428 variable.
NOx storage factors which form data elements of the NOx storage factor lookup map 426 may be determined during calibration of the engine 120 and exhaust system 100 by monitoring the amount of NOx input to the LNC 165 and output from the LNC 165 at various exhaust gas flow rates. The ratio of NOx stored by the LNC to NOx input to the LNC is one form of NOx storage factor. Exhaust gas flow rates are typically specified by the term “space velocity” which has units of ‘hr⁻¹’ (per hour). Space velocity is a measure of the quantity of gas that can be processed by a catalyst in an hour. However space velocity is a normalised value allowing the processing capability of catalysts of different volumes be compared.
The space velocity of a catalyst may be calculated according to the following equation:
(1000*m _air/ρ_air)/V _catalyst[hr⁻¹]

- m_air: measured air mass [kg/hr]
- ρ_air: 1.293 kg/m³
- V_catalystCatalyst volume [1]

Space velocity may also be used to specify the quantity of gas that has passed through catalyst 165 during an event time 402. Using space velocity allows the present embodiment to be utilised on different exhaust systems with different catalyst volumes.
Space velocity can be derived from knowledge of air flow through the engine as might be provided by a mass air flow sensor located in an inlet manifold 130 to the engine 120. A space velocity 424 variable is preferably provided from foreground processing by the ECU. This enables space velocity information to be provided on a real time or near real time basis. Alternate embodiments may also account for the quantity of fuel input to engine 120 during event time 402 when calculating space velocity.
Space velocity data specifying the amount of gas that has passed through the catalyst 165 during an event time 402 forms an input the NOx storage factor lookup map 426.
Alternate embodiments may use one or more additional inputs, to the NOx storage factor lookup map 426, that affect adsorption rate of NOx by the LNC 165. Exhaust gas temperature 422 is one such parameter that may be used to compensate for any temperature window within which LNC 165 may operate most efficiently. For example, LNC 165 may have an operating widow between 200° C. and 450° C. It is preferable that the exhaust gas temperature 422 variable is also provided from foreground processing by the ECU.
Accordingly whilst NOx flow rate data output from the NOx flow lookup map 408 can be approximated as being time invariant over a typical event time 402, it is preferable that the NOx adsorption factor be calculated on a real time or near real time basis in an ECU foreground loop.
Assume now that the LNC is purging or about to purge NOx and that the logical selection criteria 430 has accordingly selected to update the Purge_NOx_mass _—n 450 variable. The Purge_NOx_mass _—n 450 variable details the amount of NOx purged from the LNC. It is calculated as a result of the engine entering a power enrichment mode of operation or when the ECU 190 determines that NOx should be purged from the LNC. When the ECU determines that NOx should be purged from the LNC, the ECU determines the period of time for which the engine needs to be operated under current load conditions in order to purge NOx from the LNC. The engine is then operated so that there is sufficient CO in the intermediate feed gas to ensure that NOx is purged and reduced. This calculation may be repeated across several ECU NOx loops until the level of NOx on the LNC drops below a predetermined threshold. This threshold may be different to a threshold at which the ECU determines that NOx should be purged from the LNC.
Where the level of NOx stored on the LNC 165 exceeds a predetermined threshold value the model 400 selects a NOx_Purge_Flow 448 variable from a NOx purge flow look up map and a NOx_Purge_Time 438 variable from a NOx purge time look up map. The NOx_Purge Flow 448 variable is integrated over the NOx_Purge Time 438 to produce the Purge_NOx_mass _—n 450 variable. Where the engine 120 has entered a power enrichment mode then the NOx_Purge_Time 438 variable is not calculated. Instead the NOx_Purge_Flow 448 variable is integrated over the event time 402 of either successive ECU NOx loops for which the engine is in the power enrichment mode or the duration for which the engine is operated in the power enrichment mode.
The NOx_Purge_Time 438 variable is calculated as the amount of time required to purge the NOx stored on the LNC 165 to a predetermined level when the engine is operated with a pre-determined engine load and with a pre-determined air fuel ratio or within a range of air fuel ratios. Conveniently the predetermined engine load is WOT conditions where power enrichment often occurs and where typically the air fuel ratio will be rich, e.g. λ=0.82. Open loop operation of the engine 120 with a pre-determined air fuel ratio may be provided. Alternatively, a wide range oxygen sensor may be used which could allow closed loop air fuel ratio control over a wider range of air fuel ratio than provided by oxygen sensors.
Once the NOx_Purge_Time 438 variable has been calculated, the engine may then be operated at a predetermined air fuel ratio or within a range of pre-determined air fuel ratios for the amount of time indicated by the NOx_Purge_Time 438 variable. At the end of this period the level of NOx stored on the LNC 165 will have been reduced in accordance with the prevailing engine load(s) throughout the duration of the purge process.
To calculate the NOx_Purge_Time 438 variable, the process 400 uses a NOx Purge time lookup map 438 which uses the Residual_NOx_mass_—nvariable 456, calculated during the preceding ECU NOx loop, as its input parameter. This NOx Purge time lookup map 438 is derived during calibration of the engine 120 and exhaust system 100 to determine the amount of time required to purge a pre-determined amount of NOx from the LNC at a pre-determined engine load, such as WOT. Once the NOx purge time has been determined the engine 120 may then be forced to operate with a pre-determined air fuel ratio for a period of time corresponding to the NOx_Purge_Time 438 variable. Alternate embodiments may utilise additional input parameters to the NOx purge time look up map 438, such as current engine load.
The NOx_Purge_Flow 448 variable is output from a NOx Purge Flow lookup map 446. NOx_Purge_Flow 448 variable estimates the rate at which NOx is purged from the LNC 165 during an ECU NOx loop. This rate is affected by the quantity of exhaust gas that is passed through the LNC 165 and the composition of this exhaust gas.
The NOx Purge Flow lookup map 446 has as its input parameters, the space velocity 424 of the exhaust gas, as detailed above and the rear oxygen sensor 110 output. The Space Velocity variable 424 details the quantity of exhaust gas that has passed through the LNC 165. The composition of gas entering the LNC 165 is determined from the rear oxygen sensor 110 output. The NOx Purge Flow lookup map 446 may be determined during calibration of the engine 120 and the exhaust system 100.
Where the engine is operated with a predetermined air fuel ratio or within a predetermined range of air fuel ratios in order to purge the LNC 165 of NOx, it is preferable that the air fuel ratio be controlled so that the output of the rear oxygen sensor 110 indicates a sufficient amount of CO is present in the exhaust gas entering the LNC 165 to effect purging of the LNC 165. By controlling the air fuel ratio in accordance with the output of the rear oxygen sensor 110, estimates of the quantity of NOx purged from the LNC 165 may be calculated without the need to also estimate the effect that the TWC 160 has on the level of CO and HC in the raw engine out exhaust gasses.
The NOx_Purge_Time 438 variable may be calculated using the assumption that the engine is operated at WOT during the purge process. This however may not always be the case and accordingly the level of CO and HC passing over the LNC 165 may vary with actual engine load during the purging. Accordingly the Purge_NOx_mass, 450 variable should reflect that a lower quantity of NOx has been purged from the LNC 165. In calculating the Purge_NOx_mass _—n 450 variable, lower levels of HC and CO that occur for part engine loads are accounted for by the space velocity variable 424 that is an input parameter to the NOx Purge Flow lookup map 446. As detailed above, the space velocity variable 424 is a measure of the quantity of exhaust gas treated by the exhaust after treatment system 100. During a purge process the air fuel ratio is controlled so that it is within a predetermined range. Accordingly, from knowledge of both the air fuel ratio, the space velocity variable 424 during the purge process and emissions data from calibration of the engine 120, the quantity of CO and HC treated by the LNC 165 is known. Accordingly, the calculation of the Purge_NOx_mass _—n 450 variable is self-correcting for different engine loads over the duration of the purge event. However, where the engine 120 is operated at part load during a purge event, the desired level of purging may not take place. This may necessitate operating the engine 120 so as to purge the LNC 165 of NOx for one or several more ECU NOx loops

Claims

1. A method of operating an internal combustion engine having a first catalyst with NOx storage capability; said method involving determination of the composition of exhaust gas products input to the catalyst to thereby estimate an aggregate quantity of nitrogen oxide compounds (NOx) stored on said catalyst, said method including the steps of

estimating a quantity of NOx that has recently either been stored onto said catalyst or reduced from said catalyst, and

summing said quantity of NOx estimated to have recently been stored or reduced with an aggregate quantity of NOx estimated to have previously been stored on said catalyst;

said determination of the composition of gases input to said catalyst also being used during certain engine operating conditions to control the engine air fuel ratio so as to control the composition of exhaust gas products produced by the engine and input to the catalyst.

2. A method as claimed in claim 1 wherein said estimate of said aggregate quantity of NOx stored on said catalyst is performed at intermittent time intervals.

3. A method as claimed in claim 2 wherein said intermittent time intervals further comprise a background loop of an electronic control unit of said engine.

4. A method as claimed in claim 2 further comprising the step of estimating the quantity of NOx that has either been stored onto said catalyst or reduced from said catalyst over a recent one of said intermittent time intervals, and summing said quantity of NOx estimated to have been stored or reduced over said recent one of said intermittent time intervals with an aggregate quantity of NOx estimated to have been stored on said catalyst over a previous one of said intermittent time intervals.

5. A method as claimed in claim 1 wherein said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating NOx flow to said catalyst.

6. A method as claimed in claim 5 wherein said NOx flow is derived from at least engine speed and engine load.

7. A method as claimed in claim 5 wherein said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating engine temperature.

8. A method as claimed in claim 5 wherein said NOx flow and said intermittent time interval is used to determine a total quantity of NOx input to said catalyst during a recent one of said intermittent time intervals.

9. A method as claimed in claim 1 wherein said step of estimating the quantity of NOx stored on said catalyst further comprises the step of estimating the space velocity of said gases input to said catalyst.

10. A method as claimed in claim 9 wherein space velocity and said total quantity of NOx input to said catalyst are used to determine said quantity of NOx stored on said catalyst.

11. A method as claimed in claim 1 wherein said step of estimating NOx reduced from said catalyst further comprises the step of estimating a NOx purge time period based on said aggregate quantity of NOx stored on said catalyst.

12. A method as claimed in claim 11 wherein said NOx purge time comprises the time required to purge a pre-determined quantity of NOx from said catalyst under predetermined engine operating conditions.

13. A method as claimed in claim 11 wherein said step of estimating NOx reduced from said catalyst comprises the further step of estimating a NOx reduction flow to said catalyst.

14. A method as claimed in claim 13 wherein said NOx reduction flow is estimated from the space velocity of gases input to said catalyst and from the composition of said gases.

15. A method as claimed in claim 13 wherein said step of estimating the quantity of NOx reduced from said catalyst is derived from said NOx reduction flow and said NOx purge time.

16. A method as claimed in claim 1 wherein determination of the composition of gases input to said catalyst further comprises determination of the molar weight ratio of carbon monoxide (CO) and NOx in said gases.

17. A method as claimed in claim 16 further comprising the steps of estimating that NOx is stored on said catalyst when said molar weight ratio is less than 3.0 and estimating that NOx is reduced from said catalyst when said molar weight ratio exceeds 3.0.

18. A method as claimed in claim 1 wherein said engine further comprises a first gas composition sensor located upstream of the first catalyst and wherein said determination of the composition of gases input to said catalyst further comprises monitoring an output signal of said first gas composition sensor.

19. A method as claimed in claim 1 further comprising the steps of monitoring said aggregate quantity of NOx, determining if said aggregate quantity exceeds a threshold value and operating said engine so as to reduce NOx stored on said catalyst when said aggregate quantity exceeds said threshold value.

20. A method as claimed in claim 19 wherein said engine is operated with a rich air fuel ratio providing exhaust gases having molar weight ratios for CO and NOx of 3.0 or greater in order to reduce NOx stored on said catalyst.

21. A method as claimed in claim 20 wherein said engine is operated with a closed loop air fuel ratio to thereby provide said rich air fuel ratio for reducing NOx stored on said catalyst.

22. A method as claimed in claim 1 wherein said engine further comprises a second catalyst upstream of said first, said second catalyst being a three way catalyst and said first gas composition sensor located intermediate said first and second catalysts.

23. An internal combustion engine and exhaust system, the engine comprising an exhaust manifold with outlet for egress of raw combustion gases and an electronic control unit (ECU) for controlling operation of the engine and the exhaust system comprising a three way catalyst (TWC) located separately from a lean NOx catalyst (LNC), the TWC having an inlet port located so as to receive the raw combustion gases from the outlet of the exhaust manifold and an outlet port for egress of partially treated combustion gases, the LNC having an inlet port to receive the partially treated combustion gases and an outlet port for egress of exhaust gases, and an O₂sensor located intermediate the outlet of the TWC and the inlet to the LNC wherein during at least pre-determined TWC operating conditions said ECU monitors the output of said O₂sensor to thereby control the engine air fuel ratio so as to control the composition of treated exhaust gas products input to the LNC.

24. An internal combustion engine and exhaust system as claimed in claim 23 wherein said ECU monitors the output of said O₂sensor to thereby determine whether said LNC is storing NOx or being purged of NOx.

25. An internal combustion engine and exhaust system as claimed in claim 23 wherein said ECU monitors the output of a second O₂sensor operatively associated with the inlet to the TWC to thereby provide closed loop air fuel ratio control.

26. An internal combustion engine and exhaust system as claimed in claim 23 wherein the closed loop air fuel ratio control is λ≈1 control.

27. An internal combustion engine and exhaust system as claimed in claim 23 wherein the closed loop air fuel ratio control perturbates the air fuel ratio and wherein this perturbation is between lean and rich air fuel ratios to thereby provide λ=1 control.

28. An internal combustion engine and exhaust system as claimed in claim 27 wherein said rich perturbations are controlled in response to the output of said O₂sensor associated with the output of the TWC so as to control whether said LNC is purging or storing NOx.

29. A NOx model for estimating the aggregate quantity of NOx stored on a lean NOx catalyst for use with a lean burn engine and an exhaust system having a first catalyst for treating engine out emissions and a second catalyst wherein said second catalyst is a lean NOx catalyst and an oxygen sensor located intermediate said first catalyst and said second catalyst wherein said model receives input data at least derived from said oxygen sensor wherein said data indicates the composition of exhaust gas input to said lean NOx catalyst whereby said model estimates whether said lean NOx catalyst is storing NOx or reducing NOx.

30. A NOx model as claimed in claim 29 wherein the input data derived from said oxygen sensor is also used during certain engine operating conditions to control the engine air fuel ratio so as to control the composition of the exhaust gas products produced by the engine and input to the lean NOx catalyst.

31. An electronic control unit for controlling operation of an engine having an exhaust after treatment system with a NOx storage catalyst, said ECU determining the composition of gases input to the catalyst to thereby determine an aggregate quantity of nitrogen oxide compounds (NOx) stored on said catalyst.

32. An electronic control unit as claimed in claim 31 wherein determining the composition of gases input to the catalyst further comprises determining the molar weight ratio of carbon monoxide (CO) and NOx in said gases.

33. An electronic control unit as claimed in claim 32 wherein said ECU estimates that NOx is stored on said catalyst when said ratio is less than 3.0 and estimates that NOx is reduced from said catalyst when said ratio exceeds 3.0.

34. An electronic control unit as claimed in claim 33 wherein said exhaust after treatment system further comprises a gas composition sensor upstream of said catalyst, said ECU receiving signals output by said gas composition sensor which said signals indicate said molar weight ratio.

35. An electronic control unit as claimed in claim 34 wherein the signals output by said gas composition sensor are also used during certain engine operating conditions to control the engine air fuel ratio so as to control the composition of the exhaust gases produced by the engine and input to the NOx storage catalyst.