EP1507079A2 - Method for operating an internal combustion engine by adaption of mixture pilot control - Google Patents

Abstract

The internal combustion engine exhaust system contains a pre-catalyser and a main exhaust gas catalyser. The control system may act as a mixture control and may have a circuit block (10) with an input (12) from a sensor downstream of the pre-catalyser and an air mass flow input (14). The circuit block contains a filter (16), a gradient block (18) and a settle-check block (20). An integrator block (22) is connected to the air mass flow input. The settle-check block receives an input from a catalyser model block (40) which has inputs (28,30,42) from the engine. The control system also contains a flash adaptation block (26).

Description

The invention relates to a method for operating an internal combustion engine, in particular Diesel engine or gasoline engine, in particular a motor vehicle, with an exhaust aftertreatment system with at least one pre-catalyst and at least one main catalyst downstream of the primary catalyst, wherein by means of a lambda control from a difference between a lambda desired value and a measured after the pre-catalyst lambda actual value Control intervention for the mixture control is calculated, according to the preamble of claim 1.

The prior art discloses internal combustion engines with exhaust gas systems which have at least one pre-catalyst close to the engine and at least one main catalytic converter arranged downstream of the pre-catalyst. To control the exhaust gas composition is usually upstream of the primary catalytic converter and a lambda sensor downstream of the main catalyst, a further lambda probe or a NO _x sensor arranged with oxygen measuring.

Usually, before the pre-catalyst is a broadband lambda probe and behind the main catalytic converter arranged a step response lambda probe. With a such probe configuration is a mixture control and regulation such possible that a deviation of the Actual mixture composition is detected from a target mixture composition and the detected deviation into a control intervention of a mixture precontrol is converted. The front probe is comparatively close arranged on the internal combustion engine, so that deviations from the Target mixture composition can be quickly detected and corrected.

For fine control, the signal is placed downstream of the main catalyst another lambda probe or NOx sensor with Sauerstoffmeßvorrichtung used. In particular, an accurate calibration of the lambda = 1 Point of the front probe by the signal of the rear probe in operation with Lambda =. 1 The control deviation for achieving a desired lambda value is in the control of the mixture deviation via the front probe is included.

In addition, fuel mixture adaptation methods are known which provide suitable, Detect mostly stationary operating state to this in an adaptation routine to start. Adaptation routines usually consist of slow controllers (I-controller), which the normal internal lambda control on LSU basis (faster Circle) are superimposed. The value of the I component of the higher-level adaptation controller corresponds to the learned systematic mixture pilot control error, this is stored permanently dependent on the operating point and enables the motor control in future driving cycles a more accurate, working point-dependent Mixture pilot control. The regulation is characterized by passing through various Relieves operating points and reduces emissions accordingly. The systematic Error of the mixture precontrol is via an integration of Lambda deviations detected at selected operating conditions and the maps the mixture feedforward control for the amount of fuel are corrected by correction factors or adapted map values adapted. However, this adaptation can only be made slowly, otherwise at the moment of adaptation suddenly two mechanisms, namely the map correction of the mixture precontrol and the lambda control, want to compensate for the same error, resulting in overcorrection and unstable states.

However, these conventional fuel mixture adaptation methods have the Disadvantage that the principle of the superimposed control for stability reasons a enforces relatively slow adaptation controller. This results in long Adaptionseinschwingzeiten. Since an adaptation only in quasi-stationary operating points precise and can be carried out for a sufficiently long time, results from the above Adaption technology either a low adaptation quality or in typical Exhaust test cycles (MVEG or FTP cycle) a rare Ausadaptieren the Mixture pilot control. The described restrictions or disadvantages aggravate even more in vehicle concepts with inherently slower inner Lambda control, such as when laying the Vorkatsonde to a Position behind the precatalyst. Because the correction is only gradually about the integration Deviations are made, for the adaptation relatively long stationary Operating phases of the internal combustion engine needed. Accordingly rare can the Adaptation be performed.

From DE 198 56 367 C1 is a method for purifying the exhaust gas of a Internal combustion engine with a lambda-controlled three-way catalyst and a Trim control known. It is the in the Rohsignalaufbereitung the lambda probe determined with continuous characteristic curve resulting Meßsignalfehler and to the opposite change in the control value of the trim control because the trim control also produces a signal resulting from the raw signal conditioning, lambda sensor-independent error adaptively compensates.

From DE 100 29 633 A1 is a multi-flow exhaust system of a multi-cylinder engine and a method of controlling an air-fuel ratio. The multi-flow exhaust system comprises at least two exhaust lines, in the each one or more cylinders open. Each exhaust system has a separate one Pre-catalyst and one lambda probe downstream of the primary catalyst on. Only one exhaust gas line also has a lambda probe upstream of the primary catalytic converter on. By means of this fully equipped exhaust line a map is determined which with the help of differences between the Nachkat lambda values at defined operating conditions for the individual exhaust strands is modified.

The adaptation of the mixture pilot control is also used to diagnose the fuel supply system used, for example, to leak air sources or faulty Recognize fuel injection valves. In principle, all errors in the fuel and be recognized in the air path.

The invention is based on the object, a method of o.g. Kind regarding a Accuracy in the adaptation of systematic errors of the mixture precontrol to improve as well as to accelerate an adaptation speed.

This object is achieved by a method of o.g. Kind with the in Claim 1 marked features solved. Advantageous embodiments The invention are set forth in the dependent claims.

For this purpose, it is provided according to the invention that at the end of a certain Period of steady state operating condition of the internal combustion engine an I-share of Lambda control as an adaptation value of a mixture precontrol for this operating state stored, immediately for this operating state as an adaptation value used for the mixture precontrol and the I-part of the lambda control on is set to zero.

This has the advantage that a fast and accurate adaptation of, for example achieved by component scattering, systematic Gemischvorsteuerfehler achieved which operates as fast as the lambda control itself without In this case, a stability of the lambda control deteriorates. It will be for the adaptation the mixture precontrol determines and transmits an instantaneous value, whereby the requirements for the engine operating state during the adaptation are significantly lower and the adaptation are carried out accordingly more frequently can.

For example, the adaptation value is an additive and / or multiplicative adaptation value for the mixture pilot control.

To further accelerate the adaptation speed is a lambda value after the precatalyst by means of a catalyst model from an engine speed, a value for the relative air charge of a combustion chamber of the internal combustion engine and a value for the exhaust gas mass flow calculated and therefrom Lambda offset determined.

In a preferred embodiment of the invention for multi-flow exhaust aftertreatment systems is in an exhaust system with two or more exhaust banks with respective primary catalytic converter and respective lambda probe after the pre-catalyst, wherein only one exhaust bank, a lambda probe before the pre-catalyst has, the adaptation value of the mixture precontrol for each exhaust bank separately determined and stored. This has the advantage that for the settlement of Engine specific deviations, such as part tolerances, wear and differences in engine geometry, a procedure available is, which is not a custom engine lambda probe before the pre-catalyst needed.

Conveniently, for the exhaust bank with lambda probe before the pre-catalyst Operating state-dependent adaptation values for the mixture precontrol due to a difference between a lambda setpoint and a pre the lambda actual value measured in the pre-catalytic converter to compensate for all exhaust gas banks interacting, dynamic variables, in particular intake manifold pressure, Engine speed and / or fuel type, performed and on the others Transfer exhaust banks.

For even further approximation of the signal quality to a pre-catalyst measured by exhaust banks without lambda probe before the pre-catalyst real Lambda value becomes from a Lambda actual value before the Vorkatalysator of the exhaust bank with lambda probe before the pre-catalyst, a difference between a Lambda setpoint and a measured before the pre-catalyst lambda actual value the exhaust bank with lambda probe in front of the pre-catalyst and a probe voltage the lambda probe after the Vorkatalysator an exhaust bank without Lambda probe in front of the pre-catalyst for this exhaust bank without lambda probe before the pre-catalyst, a substitute value for the lambda value before the pre-catalyst generated.

In a preferred embodiment of the invention is in an exhaust system with two or more exhaust banks with respective pre-catalyst and respective Lambda probe after the pre-catalyst, with only one exhaust bank a Lambda probe before the pre-catalyst, an additive adaptation value for the Mixture feedforward determined and stored for an exhaust bank and on transmit the other exhaust banks. Appropriately, the additive adaptation value determined on the exhaust bank with the lambda probe in front of the pre-catalyst.

In a preferred embodiment of the invention is in an exhaust system with two or more exhaust banks with respective pre-catalyst and respective Lambda probe after the pre-catalyst, with only one exhaust bank a Lambda probe before the pre-catalyst has a multiplicative adaptation value individually determined and stored for the mixture pilot control for each exhaust gas bank.

Fig. 1

ein schematisches Blockschaltbild einer innere Struktur einer bevorzugten Ausführungsform einer erfindungsgemäßen, modellgestützten Gemischadaption,

Fig. 2

eine schematische Darstellung einer zweiflutigen Abgasnachbehandlungsanlage und

Fig. 3

ein schematisches Blockschaltbild einer bevorzugten Ausführungsform einer erfindungsgemäßen Struktur eines Gemischaufbereitungskonzeptes für die zweiflutige Abgasnachbehandlungsanlage gemäß Fig. 2.

Further features, advantages and advantageous embodiments of the invention will become apparent from the dependent claims, and from the following description of the invention with reference to the accompanying drawings. These show in

Fig. 1

3 is a schematic block diagram of an internal structure of a preferred embodiment of a model-based mixture adaptation according to the invention;

Fig. 2

a schematic representation of a dual-flow exhaust aftertreatment system and

Fig. 3

2 shows a schematic block diagram of a preferred embodiment of a structure according to the invention of a mixture preparation concept for the dual-flow exhaust aftertreatment system according to FIG. 2.

Fig. 1 illustrates an internal structure of a preferred embodiment an inventive adaptation of a mixture precontrol for an internal combustion engine with a precatalyst and downstream of the precatalyst arranged main catalyst. In a block 10 becomes a switch-on condition tested for adaptation. For this purpose, the block 10 receives as input values a probe voltage behind the precatalyst u_Sondehk 12 and a value for the Air mass m_Air 14. The probe voltage behind the pre-catalyst u_Sondehk 12 is fed to a block 16 "Filter". An output of block 16 "Filter" becomes fed to a block 18 "gradient". An output of block 18 "Gradient" becomes a block 20 "Settle-Check" supplied. The value for the air mass m_Air 14 is fed to a block 22 "integrator" and an output of the block 22 "integrator" is also fed to the block 20 "Settle-Check". In block 20 "Settle-Check" is calculated from the input values from the block 18 "gradient" and the Block 22 "integrator" checked whether a current operating state of the internal combustion engine predetermined criteria, for example with respect to a stationary / static Operation satisfied, so that this operating condition are considered quasi-static can and is suitable for adaptation. If this is the case, then the block gives 20 "Settle-Check" a release bit B_adapstart 24 to a block 26 "Flash adaptation" off. The enable bit B_adapstart 24 starts the adaptation in Block 26 "Flash adaptation", wherein an I-part of a lambda controller in a single Transfer step in an operating point-dependent adaptation matrix and then set to zero. In other words, the current value of the I-portion of the lambda controller as assigned to the current operating state Adaptation value stored for a mixture pilot control. This value then becomes used for this operating state as an adaptation value for the mixture precontrol. The enable bit B_adapstart 24 also triggers the reset of the I component in the lambda controller, i. the I component is set to zero. The block 26 "Flash adaptation" also receives as input values a relative air charge of the Combustion chamber rel_Füllung 28, an engine speed n_Motor 30, an intervention of the Lambda controller (factor) f_Regler 32 and a multiplicative intervention of the adaptation the mixture precontrol f_Adapt 34. The block 26 "Flash Adaption" calculated in a block 36 an adaptation factor and stores it in block 38. As an output, the block 26 "Flash Adaption" gives a value for the multiplicative Intervention of the adaptation of the mixture precontrol f_Adapt 34.

Optionally, a block 40 "catalyst model" is additionally provided. This block 40 "catalyst model" receives as input values the engine speed n_motor 30, the relative air charge of the combustion chamber rel_Füllung 28 and an exhaust gas mass flow ms exhaust 42. From this in block 40 "catalyst model" by means of a explicit catalyst model for observer-based lambda offset determination a lambda value calculated after the precatalyst. The output 44 of block 40 "catalyst model" is also added as input value to block 20 "Settle-Check" and the block 26 "Flash Adaption" supplied.

As a result, a fast Lambda adaptation by a fuel quantity neutral Umkopiervorgang achieved in quasi-stationary states, the I component of the lambda controller in a suitable operating state in a single Transfer step in an operating point-dependent adaptation matrix and then set to zero. This is referred to herein as a "flash adaptation". For the actual adaptation becomes only the temporal end of the quasi-stationary state used. By a modeled Nachkatlambda by means of an explicit catalyst model the observer-based lambda offset determination becomes the adaptation speed accelerated. The cat model also allows the use of only briefly stationary operating conditions of the internal combustion engine for Lambda adaptation. By combining a neuronal identifier with a database of typical output lambda values in specific driving situations a lambda offset already at the operating conditions selected for the adaptation with little restriction determinable with good accuracy.

The model-based adaptation strategy according to the invention is more active, responds more frequently faster on changing operating conditions and relieves due to more accurate Pilot control values for the mixture pilot control the lambda control.

The "flash adaptation" according to the invention determines an error in the fuel metering, for example due to component tolerances or aging processes, in an extremely short time and recognizes this as a systematic error (snapshot the lambda deviation). The adaptation value is stored in the adaptation memory copied and by means of the reset of the I-part of the lambda control (up Set zero) is communicated to the lambda control that this error in the metering the fuel already on the adaptation in the mixture precontrol is taken into account and the lambda control accordingly this error is not must and must balance itself. This prevents the control from overshooting during a recording of the correction (adaptation) by the adaptation memory during at the same time the lambda control still tries to compensate for the same error because the lambda control still measures exhaust gas values from before the adaptation. It is particularly advantageous that a moment value (I component of the lambda control) is determined and transmitted, thereby reducing the demands on the engine operating condition, while an adaptation is possible, are considerably lower than at conventional adaptation with the lambda controller superimposed controller with slow Integrator. As a result, the adaptation can be carried out correspondingly more frequently become. Even if the first flash adaptation should be inaccurate, the Accuracy improved in a short time by frequent measurements.

FIG. 2 schematically illustrates a dual exhaust system for an engine 50 having a plurality of cylinders, with corresponding exhaust ports of some cylinders opening into a first exhaust bank 52 and corresponding exhaust ports of the remaining cylinders opening into a second exhaust bank 54. Each exhaust bank 52, 54 has in each case a precatalyst 56 and 58 and in each case one of the precatalyst 56 and 58 downstream lambda probe LSF 60 and 62. The first exhaust bank 52 additionally has a lambda probe LSU 64 in front of the precatalyst 56, whereas such a lambda probe LSU is not provided in front of the precatalyst 58 in the second exhaust bank 54. The two exhaust banks 52 and 54 lead to a common exhaust line 66 together. In the common exhaust line 66, a temperature sensor 68, a main catalytic converter 70 and a NO _x sensor 72 is arranged as seen in the flow direction. The internal combustion engine 50 further includes a fresh air path 74 with throttle 76 and intake manifold pressure sensor 78.

The lambda probe LSU 64 before the precatalyst 56 of the first exhaust bank 52nd Primarily used to compensate for dynamically changing quantities corresponding influence on the mixture precontrol, such as intake manifold pressure, Engine speed, fuel type, etc., which apply to all exhaust banks 52, 54 equally. Therefore, it is sufficient these influences and corrections the dynamic quantities only for the first exhaust bank 52 to determine and transfer to the second exhaust bank 54. For this reason, the deleted Lambda probe LSU before the precatalyst 58 of the second exhaust bank 54th For the Compensation or the correction of engine-specific sizes, which corresponding Have influence on the mixture precontrol, the inventive Adaption control according to flash adaptation executed, which also without lambda probe LSU gets by before the pre-catalyst. This allows this adaptation for Both exhaust banks 52, 54 are performed individually.

For the twin-flow exhaust system of FIG. 2 is a continuous Lambda control to the lambda after the pre-catalysts 56, 58 performed, i.e. There is a lambda measurement after the pre-catalysts 56, 58 by means of Lambda sensors LSF 60 and 62. This is a beginning of the lambda control already enabled with operational LSF probe 60, 62 and it does not have to a later reached, predetermined Hinterkattemperatur at sensor 68 waited become. The fast lambda adaptation is done by the one described above Flash adaptation separately for each exhaust bank. Selected parameters for dynamic improved fuel quantity pilot control are only for the first Exhaust bank 52 determined and mirrored on the second exhaust bank 54. To one Lambda replacement signal for the second exhaust bank 54 to produce symmetrical usable signal components of the first exhaust bank 52 is used. The Signal quality thereby approaches a real measured lambda value.

A structure overview of the mixture preparation for the twin-flow exhaust system 2 is shown schematically in Fig. 3. To the clear Representation is only for the branch of the second exhaust bank 54 a mixture coordination shown. A block 80 represents a lambda control function a regulator variant LR_Bank_1 82 for the first exhaust bank 52 and a controller variant LR_Bank_2 84 for the second exhaust bank 54. The lambda control function 80 receives as input values lambda setpoint lambda_soll 86 and one the first exhaust bank 52 before the pre-catalyst 56 measured lambda actual value Lambda_ist_b1 88. Selected fuel contributions from the controller variant LR_Bank_1 82 for the first exhaust bank 52 become an additional function MIRR_B1_B2 supplied in a block 90. This block 90 reflects these fuel contributions from the regulation of the first exhaust bank 52 LR_Bank_1 82 on the Regulation of the second exhaust bank 54 LR_Bank_2 84, as indicated by arrow 92. The lambda control function 80 then outputs a control factor Regelfaktor_b2 94 for the second exhaust bank 54 to a mixture coordination% GKO_B2 96 for the second Exhaust bank 54 off. This acts on the second exhaust bank 54 assigned Engine part Motor_B2 98 and accordingly to the pre-catalyst 58 of the second exhaust bank 54. A probe voltage 112 of the lambda probe LSF 62nd after the pre-catalyst 58 of the second exhaust bank 54 (LSF_2) becomes an adaptation function fed in a block 100. This adaptation function 100 includes an adaptation variant for the first exhaust bank 52 ADAP_Bank_1 102 as well an adaptation variant for the second exhaust bank 54 ADAP_Bank_2 104. Selected Fuel contributions from the adaptation variant for the first exhaust bank 52 ADAP bank_1 102 are supplied to function supplement MIRR_B1_B2 in block 90. This block 90 reflects these fuel contributions from the adaptation variant for the first exhaust bank 52 ADAP_Bank_1 102 on the adaptation variant for the second exhaust bank 54 ADAP_Bank_2 104, as indicated by arrow 106. The adaptation function 100 then gives an adaptation intervention Adaptionseingriff_b2 108 for the second exhaust bank 54 to the mixture coordination% GKO_B2 96 for the second Exhaust bank 54 off, with corresponding effects on Motor_B2 98 and the Pre-catalyst 58 of the second exhaust bank 54.

In addition, in a block 110, a function for generating a lambda offset value is provided before the pre-catalyst 58 of the second exhaust bank 54 for the second Exhaust bank 54 and provided for probe voltage correction. This block 110 receives as input values, the probe voltage 112 of the lambda probe LSF 62 of second exhaust bank 54, a lambda difference Lambda_differenz_b1 114 of the first Exhaust bank 52 and the lambda actual value Lambda_ist_b1 88 before the pre-catalyst 56 of the first exhaust bank 52. The block 110 then outputs a calculated Lambda actual lambda-is-b2 116 before the pre-catalyst 58 of the second exhaust bank 54 off.

With the concept explained with reference to FIG. 3, symmetry properties become the two exhaust banks 52, 54 exploited for mixture precontrol and is the novel flash adaptation used in block 100.

The mixture adaptation is divided into several parts. So there is ever an adaptation value for an additive error, a multiplicative error and possibly still for a temperature-dependent error. In a single-flow exhaust system is with a lambda probe u.a. determines the mixture deviation for all cylinders. in the The case of a two or more exhaust system is often the sensor executed several times, i. each exhaust bank points in the direction of flow a pre-catalyst, a lambda probe in front of the pre-catalyst and a lambda probe after the precatalyst. Such a concept exists for everyone Adaptation value of the mixture adaptation corresponding to the number of exhaust banks Number of factors. It has proven to be an empirical value that additive error of the mixture precontrol, for example caused by Leakage air in the intake manifold, acting on all exhaust banks alike and that multiplicative Error of mixture precontrol mainly to tolerances or Inaccuracies in the fuel path are due, i. act separately on each exhaust bank. In addition, it is assumed that the air flow of the one common Air path evenly distributed on all cylinders.

Under these conditions, the invention proposes the additive Adaptation value for the mixture precontrol, which contains the additive error of the Mixture feedforward corrected, to measure or calculate only on an exhaust bank and to reflect on the other exhaust banks. Preferably, on the one Exhaust bank, the additive adaptation value for the mixture precontrol is formed, Which immediately after engine leakage a lambda probe before the pre-catalyst having. This simplifies the adaptation of the mixture precontrol for mehrflutige exhaust aftertreatment systems, since only the multiplicative adaptation value for the mixture precontrol, which is the multiplicative error of the mixture precontrol corrected, must be determined separately for each exhaust bank. Exhaust gas tank specific deviations on the exhaust bank / exhaust gas banks, in particular for the multiplicative adaptation range, are over a continuous Lambda control after precatalyst based on a binary or continuous Lambda signal balanced. This can be in the concept with mehrflutiger Exhaust after-treatment system on the lambda probes before the pre-catalyst be dispensed with in all exhaust banks except for an exhaust bank.

Claims (10)

Method for operating an internal combustion engine, in particular a diesel engine or gasoline engine, in particular of a motor vehicle, with an exhaust aftertreatment system having at least one primary catalytic converter and at least one main catalytic converter downstream of the primary catalytic converter, wherein by means of a lambda control a difference between a desired lambda value and a lambda value measured after the primary catalytic converter Actual value a control intervention for the mixture control is calculated,
characterized in that stored at the end of a stationary over a certain period operating condition of the internal combustion engine, an I-part of the lambda control as adaptation value of a mixture precontrol for this operating condition, immediately used for this operating state as an adaptation value for the mixture precontrol and the I-part of the lambda control set to zero becomes. Method according to Claim 1, characterized in that the adaptation value is an additive and / or multiplicative adaptation value for the mixture precontrol. Method according to claim 1 or 2, characterized in that a lambda value after the precatalyst is calculated by means of a catalyst model and from this a lambda offset is determined. Method according to Claim 3, characterized in that an engine speed, a value for the relative air charge of a combustion chamber of the internal combustion engine and a value for the exhaust gas mass flow are supplied to the catalyst model as an input value. Method according to at least one of the preceding claims, characterized in that, in the case of an exhaust system with two or more exhaust banks with respective primary catalytic converter and respective lambda probe downstream of the pre-catalyst, wherein only one exhaust bank has a lambda probe upstream of the pre-catalyst, the adaptation value of the mixture pilot control for each exhaust bank is determined separately and is stored. A method according to claim 5, characterized in that for the exhaust bank with lambda probe upstream of the precatalyst operating state-dependent adaptation values for the mixture precontrol due to a difference between a lambda desired value and a measured before the pre-catalyst lambda actual value to compensate for all exhaust banks jointly influencing dynamic variables , in particular intake manifold pressure, engine speed and / or fuel type, carried out and transmitted to the other exhaust banks. A method according to claim 5 or 6, characterized in that from an actual lambda value before the pre-catalyst of the exhaust bank with lambda probe before the pre-catalyst, a difference between a lambda desired value and a measured before the pre-catalyst lambda actual value of the exhaust bank with lambda probe before the pre-catalyst and a probe voltage of the lambda probe after the pre-catalyst of an exhaust bank without lambda probe before the pre-catalyst for this exhaust bank without lambda probe before the pre-catalyst a replacement value for the lambda value is generated before the pre-catalyst. A method according to claim 1, characterized in that in an exhaust system with two or more exhaust banks with respective pre-catalyst and respective lambda probe after the pre-catalyst, wherein only one exhaust bank has a lambda probe before the pre-catalyst, determines an additive adaptation value for the mixture precontrol for an exhaust bank and stored and transferred to the other exhaust banks. A method according to claim 8, characterized in that the additive adaptation value is determined on the exhaust bank with the lambda probe before the pre-catalyst. A method according to claim 1, 8 or 9, characterized in that in an exhaust system with two or more exhaust banks with respective primary catalytic converter and respective lambda probe downstream of the primary catalytic converter, wherein only one exhaust bank has a lambda probe upstream of the primary catalytic converter, a multiplicative adaptation value for the mixture pilot control for each Exhaust bank is individually determined and stored.

Images