US20070256407A1

US20070256407A1 - Reformer temperature control with leading temperature estimation

Info

Publication number: US20070256407A1
Application number: US11/418,938
Authority: US
Inventors: Johannes Reuter
Original assignee: Eaton Corp
Current assignee: Eaton Corp
Priority date: 2006-05-05
Filing date: 2006-05-05
Publication date: 2007-11-08
Also published as: WO2007129203A1

Abstract

One concept relates to a method of controlling fuel reforming within an internal combustion engine exhaust line. Fuel injections are controlled using a predicted temperature, the predicted temperature being a temperature that would occur at some point in the future if predetermined assumptions are met. Preferably, the prediction is made using a model that includes terms for hydrocarbon storage and subsequent reaction within the reformer. The method improves reformer temperature control, particularly over periods during which the fuel supply to the reformer is pulsed. The scope of the invention also includes methods wherein a temperature is not specifically predicted, provided the control method takes into account hydrocarbon storage and subsequent reaction.

Description

FIELD OF THE INVENTION

The present invention relates to pollution control systems and methods for diesel and lean burn gasoline engines.

BACKGROUND

NO_xemissions from diesel engines are an environmental problem. Several countries, including the United States, have long had regulations pending that will limit NO_xemissions from trucks and other diesel-powered vehicles. Manufacturers and researchers have put considerable effort toward meeting those regulations.
In gasoline powered vehicles that use stoichiometric fuel-air mixtures, three-way catalysts have been shown to control NO_xemissions. In diesel-powered vehicles, which use compression ignition, the exhaust is generally too oxygen-rich for three-way catalysts to be effective.
Several solutions have been proposed for controlling NOx emissions from diesel-powered vehicles. One set of approaches focuses on the engine. Techniques such as exhaust gas recirculation and partially homogenizing fuel-air mixtures are helpful, but these techniques alone will not eliminate NOx emissions. Another set of approaches remove NOx from the vehicle exhaust. These include the use of lean-burn NO_xcatalysts, selective catalytic reduction (SCR), and lean NO_xtraps (LNTs).
Lean-burn NOx catalysts promote the reduction of NO_xunder oxygen-rich conditions. Reduction of NOx in an oxidizing atmosphere is difficult. It has proven challenging to find a lean-burn NO_xcatalyst that has the required activity, durability, and operating temperature range. Lean-burn NO_xcatalysts also tend to be hydrothermally unstable. A noticeable loss of activity occurs after relatively little use. Lean-burn NOx catalysts typically employ a zeolite wash coat, which is thought to provide a reducing microenvironment. The introduction of a reductant, such as diesel fuel, into the exhaust is generally required and introduces a fuel economy penalty of 3% or more. Currently, peak NOx conversion efficiencies for lean-burn NOx catalysts are unacceptably low.
SCR generally refers to selective catalytic reduction of NOx by ammonia. The reaction takes place even in an oxidizing environment. The NOx can be temporarily stored in an absorbent or ammonia can be fed continuously into the exhaust. SCR can achieve high levels of NOx reduction, but there is a disadvantage in the lack of infrastructure for distributing ammonia or a suitable precursor. Another concern relates to the possible release of ammonia into the environment.
LNTs are devices that adsorb NOx under lean exhaust conditions and reduce and release the adsorbed NOx under rich condition. A LNT generally includes a NOx absorbent and a catalyst. The absorbent is typically an alkaline earth oxide absorbent, such as BaCO₃and the catalyst is typically a precious metal, such as Pt or Ru. In lean exhaust, the catalyst speeds oxidizing reactions that lead to NOx adsorption. In a reducing environment, the catalyst activates reactions by which adsorbed NOx is reduced and desorbed. In a typical operating protocol, a reducing environment will be created within the exhaust from time-to-time to regenerate (denitrate) the LNT.
A LNT can produce ammonia during denitration. Accordingly, it has been proposed to combine a LNT and an ammonia SCR catalyst into one system. Ammonia produced by the LNT during regeneration is captured by the SCR catalyst for subsequent use in reducing NOx, thereby improving conversion efficiency over a stand-alone LNT with no increase in fuel penalty or precious metal usage. U.S. Pat. No. 6,732,507 describes such a system. U.S. Pat. Pub. No. 2004/0076565 describes such systems wherein both components are contained within a single shell or disbursed over one substrate. WO 2004/090296 describes such a system wherein there is an inline reformer upstream of the LNT and the SCR catalyst.
Creating a reducing environment for LNT regeneration involves eliminating most of the oxygen from the exhaust and providing a reducing agent. Except where the engine can be run stoichiometric or rich, a portion of the reductant reacts within the exhaust to consume oxygen. The amount of oxygen to be removed by reaction with reductant can be reduced in various ways. If the engine is equipped with an intake air throttle, the throttle can be used. The transmission gear ratio can be changed to shift the engine to an operating point that produces equal power but contains less oxygen. However, at least in the case of a diesel engine, it is generally necessary to eliminate some of the oxygen in the exhaust by combustion or reforming reactions with reductant that is injected into the exhaust.
Reductant can be injected into the exhaust by the engine or a separate fuel injection device. For example, the engine can inject extra fuel into the exhaust within one or more cylinders prior to expelling the exhaust. Alternatively, or in addition, reductant can be injected into the exhaust downstream of the engine.
The reactions between reductant and oxygen can take place in the LNT, although it is generally preferred for the reactions to occur in a catalyst upstream of the LNT, whereby the heat of reaction does not cause large temperature increase within the LNT at every regeneration.
In addition to accumulating NOx, LNTs accumulate SOx. SOx is the combustion product of sulfur present in ordinarily fuel. Even with reduced sulfur fuels, the amount of SOx produced by combustion is significant. SOx adsorbs more strongly than NOx and necessitates a more stringent, though less frequent, regeneration. Desulfation requires elevated temperatures as well as a reducing atmosphere. The temperature of the exhaust can be elevated by engine measures, particularly in the case of a lean-burn gasoline engine, however, at least in the case of a diesel engine, it is often necessary to provide additional heat. Typically, this heat is provided through the same types of reactions as used to remove excess oxygen from the exhaust. The temperature of the LNT is generally controlled during desulfation, as the temperatures required for desulfation are generally close to those at which the LNT catalyst undergoes thermal deactivation.
U.S. Pat. No. 6,832,473 describes a system wherein the reductant is reformate produced outside the exhaust stream and injected into the exhaust as needed. During desulfations, the reformate is injected upstream of an oxidation catalyst. Heat generated by combustion of the reformate over the oxidation catalyst is carried by the exhaust to the LNT and raises the LNT to desulfations temperatures.
U.S. Pat. Pub. No. 2003/0101713 describes an exhaust treatment system with a fuel reformer placed in the exhaust line upstream of a LNT. The reformer includes both oxidation and reforming catalysts. The reformer both removes excess oxygen and converts the diesel fuel reductant into more reactive reformate. For desulfations, heat produced by the reformer is used to raise the LNT to desulfations temperatures.
U.S. Pat. Pub. No. 2003/0101713 describes a case in which endothermic reactions dominate and the reformer tends to cool when hydrocarbons are injected at a rate that produces a desired concentration of reformate. Between pulses that produce reformate, fuel is injected at a reduced rate whereby exothermic reactions dominate and the reformer heats.
In spite of advances, there continues to be a long felt need for an affordable and reliable exhaust treatment system that is durable, has a manageable operating cost (including fuel penalty), and is practical for reducing NOx emissions from diesel engines to a satisfactory extent in the sense of meeting U.S. Environmental Protection Agency (EPA) regulations effective in 2010 and other such regulations.

SUMMARY

One of the inventor's concepts relates to a method of controlling fuel reforming within an internal combustion engine exhaust line. A reformer is supplied with fuel injected into the exhaust upstream of the reformer. The fuel injections are controlled using a predicted temperature that is a temperature that would occur at some point in the future if predetermined assumptions are met. In general, the predicted temperature is based in part on a temperature measurement. In a preferred embodiment, the prediction is made using a model that includes terms for hydrocarbon storage and subsequent reaction within the reformer. The method improves reformer temperature control, particularly over periods where the fuel supply to the reformer is pulsed.
A further concept relates to a method of controlling a temperature of a fuel reformer. The method comprises using a model to predict a temperature associated with the reformer and using the predicted temperature in a temperature control algorithm. According to the concept, the temperature prediction is made taking into account the effects of hydrocarbon storage and subsequent reaction, which can result in heating of the reformer following the termination of fuel injection.
A closely related concept is a method of controlling the temperature of a fuel reformer comprising predicting a future reformer temperature. The predicted future temperature is used in a feedback control loop. The predictions take into account the effects of hydrocarbon storage and subsequent reaction.
A further concept relates to a method of reforming within an internal combustion engine exhaust line. Hydrocarbons are injected into the exhaust line upstream of a reformer. The amount of hydrocarbon adsorbed in the reformer is estimated and the reformer temperature is controlled based in part on that estimate.
The primary purpose of this summary has been to present certain of the inventors' concepts in a simplified form to facilitate understanding of the more detailed description that follows. This summary is not a comprehensive description of every one of the inventor's concepts or every combination of the inventor's concepts that can be considered “invention”. Other concepts of the inventor will be conveyed to one of ordinary skill in the art by the following detailed description together with the drawings. The specifics disclosed herein may be generalized, narrowed, and combined in various ways with the ultimate statement of what the inventor claim as his invention being reserved for the claims that follow.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic illustration of an exemplary exhaust treatment system to which the inventor's concepts can be applied.
FIG. 2 is a schematic illustration of another exemplary exhaust treatment system to which the inventor's concepts can be applied.
FIG. 3 is a flow chart of a computational procedure conceived by the inventor.
FIG. 4 is a schematic illustration of control architecture in which some of the inventor's concepts can be applied.
FIG. 5 is a flow chart of a desulfation method in which some of the inventor's concepts can be applied.

DETAILED DESCRIPTION

FIG. 1 provides a schematic illustration of an exemplary power generation system 5 in which various concepts of the inventor can be implemented. The system 5 comprises an engine 9, a transmission 8, and an exhaust aftertreatment system 7. The exhaust aftertreatment system 7 includes a controller 10, a fuel injector 11, a lean NOx catalyst 15, a reformer 12, a lean NOx-trap (LNT) 13, an ammonia-SCR catalyst 14, a diesel particulate filter (DPF) 16, and a clean-up catalyst 17. The controller 10 receives data from several sources; including temperature sensors 20 and 21 and NOx sensors 22 and 23. The controller 10 may be an engine control unit (ECU) that also controls the transmission 8 and the exhaust aftertreatment system 7 or may include several control units that collectively perform these functions.
The exhaust from the engine 9 generally contains products of lean combustion including NOx, particulates, and some oxygen (typically 5-15%). The lean-NOx catalyst 15 removes a portion of the NOx from the exhaust using reductants, typically hydrocarbons that form part of the exhaust or hydrocarbon that have been stored by the lean-NOx catalyst 15. The DPF 16 removes particulates. During lean operation (a lean phase), the LNT 13 adsorbs a second portion of the NOx. The ammonia-SCR catalyst 14 may have ammonia stored from a previous regeneration of the LNT 13 (a rich phase). If the ammonia-SCR catalyst 14 contains stored ammonia, it removes a third portion of the NOx from the lean exhaust. The clean-up catalyst 17 may serve to oxidize CO and unburned hydrocarbons.
FIG. 2 provides another exemplary system 25 in which various concepts of the inventor can be implemented. The system 25 contains many of the same components as the system 5, although it does not include the lean NOx catalyst 15 or the cleanup oxidation catalyst 17. Another difference is that in the system 25 the DPF 16 is placed between the reformer 12 and the LNT 13. In this configuration, the DPF 16 may serve to protect the LNT 13 from high temperatures during denitrations by providing a thermal buffer between the reformer 12 and the LNT 13. Reducing the number and/or magnitude of temperature excursions experienced by the LNT 13 may extend its life.
From time-to-time, the LNT 13 must be regenerated to remove accumulated NOx (denitrated). Denitration may involve first heating the reformer 12 to an operational temperature by injecting fuel at a sub-stoichiometric rate with respect to the oxygen in the exhaust whereby the injected fuel reacts in the reformer 12 in an excess of oxygen. An operational temperature for the reformer 12 depends on the reformer design. Once the reformer 12 is sufficiently heated, denitration may proceed by injecting fuel at a super-stoichiometric rate whereby the reformer 12 consumes most of the oxygen in the exhaust while producing reformate. Reformate thus produced reduces NOx adsorbed in the LNT 13. Some of this NOx is reduced to NH₃, most of which is captured by the ammonia-SCR catalyst 14 and used to reduce NOx during a subsequent lean phase. The clean-up catalyst 17 oxidizes unused reductants and unadsorbed NH₃using stored oxygen or residual oxygen remaining in the exhaust during the rich phases. During regeneration, the lean-NOx catalyst 15 may store reductant for later use.
From time-to-time, the LNT 13 must also be regenerated to remove accumulated sulfur compounds (desulfated). Desulfation may involve heating the reformer 12 to an operational temperature, heating the LNT 13 to a desulfating temperature, and providing the heated LNT 13 with a reducing atmosphere. Desulfating temperatures vary, but are typically in the range from about 500 to about 800° C., more typically in the range from about 650 to about 750° C. Below a minimum temperature, desulfation is very slow. Above a maximum temperature, the LNT 13 may be damaged.
During these operations, the temperature of the reformer 12 is affected by several factors. These factors may include, for example, exothermic reactions by which oxygen is consumed, endothermic reactions by which reformate is produced, convective heat transfer into the reformer 12 by the exhaust feeding the reformer 12, convective heat transfer out of the reformer 12 by exhaust exiting the reformer 12, and the thermal mass of the reformer 12.
In certain modes of operation, the balance of these factors has a tendency to overheat the reformer 12. For example, during desulfation only a small amount of reformate production may be desired and the heat released by exothermic reactions that remove excess oxygen from the exhaust may be far in excess of the heat taken up by endothermic reforming reactions and the heat taken up by the exhaust passing through the reformer 12. Also, during periods of high exhaust oxygen concentration, the characteristics of the reformer 12 may be such that the reformer 12 cannot be operated efficiently, or not at all, at the high fueling rates required for auto-thermal reforming. In either of these cases where the reformer 12 has a tendency to overheat, periods of fuel injection may need to be limited. In between periods of fuel injection, the reformer 12 will cool. Once the reformer 12 has cooled to a sufficient degree, fuel injection can be resumed. This results in pulsed operation.
When the reformer 12 is on the verge of overheating, it may be possible to cool the reformer 12 by increasing the fuel injection rate. Increasing the fuel injection rate will sometimes cool the reformer 12 by increasing the ratio of endothermic to exothermic reactions. Such an operation is within the scope of the inventor's concepts; however, in some situations increasing the fuel injection rate may be undesirable or ineffective. For example, increasing the fuel injection rate may be undesirable if it produces more reformate than can be effectively used by the exhaust aftertreatment system; increasing the fuel injection rate may be undesirable if it would push the reformer 12 into a regime where it does not operate effectively; and increasing the fuel injection rate may be undesirable if it ultimately increases the likelihood of overheating due to the hydrocarbon adsorption phenomena discussed herein. Therefore, it is generally preferred that fuel injection be discontinued when the reformer 12 is almost at a point where it will overheat.
In order to determine when overheating is imminent, the inventor contemplates using a model that takes into account hydrocarbon storage. A preferred model comprises a thermal model, which is a model based in part on an energy conservation equation. A thermal model can be zero, one, two, or three-dimensional, although a zero-dimensional lumped parameter model will generally suffice.
A lumped parameter model generally includes at least a term for heat convection into the model system, a term for heat convection out of the model system, a term for heat taken up by the reformer 12, and a term for heat generated by chemical reaction. Heat losses to the surrounding can also be considered, but generally have a small effect.
Preferably, the model tracks hydrocarbon storage within the reformer 12 and eventual reaction of a portion of that stored hydrocarbon in the reformer 12. In one embodiment, hydrocarbon storage takes place when the reformer 12 is supplied with a rich feed and reaction of previously stored hydrocarbon takes place when the reformer 12 is supplied with a lean feed. By modeling hydrocarbon storage during rich phases and subsequent reaction of a portion of that hydrocarbon during lean phases, the model predicts availability within the reformer 12 of a portion of the injected fuel in periods between temporally adjacent fuel pulses. Temporally adjacent pulses are two periods of continuous fuel injection separated in time by one period during which no fuel is injected.
The heat convection rate into the reformer 12 is the production of the exhaust specific heat, the exhaust temperature, and the exhaust mass flow rate. The exhaust mass flow rate can be measured or estimated, for example using an intake air flow rate measurements, an engine fuel flow rate measurement, or simply with data available from the engine control unit (ECU). The temperature of the exhaust entering the reformer 12 can be measured or determined from the operating point of the engine 9, for example.
The heat convection rate out of the reformer 12 depends on the temperature of the exhaust leaving the reformer 12. That temperature can be measured. Where the temperature of the reformer 12 is measured, the reformer exhaust gas temperature can be approximated as equal to the reformer temperature.
The chemical reactions in the fuel reformer 12 can be modeled as a combination of the three following reactions:
0.684 CH_1.85+O₂→0.684 CO₂+0.632 H₂O (1)
0.316 CH_1.85+0.316 H₂O→0.316 CO+0.608 H₂ (2)
0.316 CO+0.316 H₂O→0.316 CO₂+0.316 H₂ (3)
wherein CH_1.85represents an exemplary reductant, such as a diesel fuel, with a 1.85 ratio between carbon and hydrogen. Reaction (1) is exothermic complete combustion by which oxygen is consumed. Reaction (2) is endothermic steam reforming, which results in reformate production. Reaction (3) is the water-gas shift reaction.
In a preferred embodiment, Reaction (1) is assumed to proceed at a rate and to an extent dependent on temperature, oxygen availability when fuel is in excess, fuel availability when oxygen is in excess, but independently of Reaction (2). The oxygen availability depends at least on the oxygen flow rate into the reformer 12. If the reformer 12 has a significant oxygen storage capacity, then it may be desirable to includes terms for the rates of oxygen adsorption and desorption. The oxygen concentration of the exhaust flowing into the reformer 12 can also be measured or estimated using data available from the ECU. While a significant temperature gradient generally exists within the reformer 12, the purposes of the invention can generally be achieved using a single average temperature for the reformer 12.
In the preferred embodiment, Reaction (2) is considered to proceed to an extent dependent on the availability of fuel after the effect of Reaction (1). Reaction (3) is considered third, and has the least impact on a thermal model. Reaction kinetics, adsorption rates, and desorption rates depend on reactor geometry and composition and are best determined experimentally for a particular system.
FIG. 3 is a flow chart of an exemplary computational procedure 50 for modeling the temperature of the reformer 12. This procedure is implemented in finite time steps. Two instances of the procedure may be used at one time. The first instance may track the actual process. The second instance may be used to look forward in time and make predictions to determine how much the temperature could increase if fuel injection were to cease and excess oxygen were to become available.
At the start of procedure 50, certain values are initialized. These typically include the amount of stored hydrocarbon and the reformer temperature. After initialization, the process begins with operation 51.
Operation 51 accounts for Reaction (1). The reaction is treated as proceeding to the fullest extent of the available reagents, the extent therefore being determined by the limiting reagent, whether that be fuel or oxygen. An amount of heat is added to the reformer 12 based on the extent of Reaction (1) and the size of the time step. The amounts of fuel and oxygen for purposes of the following steps are reduced in accordance with the extent of Reaction (1).
Operation 52 accounts for Reaction (2). Reaction (2) is assumed not to proceed at all if there is excess oxygen and all the fuel is considered to have been consumed by Reaction (1). Where there is fuel remaining after Reaction (1), then Reaction (2) proceeds to some extent. Unlike Reaction (1), the inventor has found that Reaction (2) should not be assumed to proceed to the extent of available reagents. Rather, Reaction (2) is preferably modeled with a limited efficiency. Typical efficiencies may be from about 0.35 to about 0.7 based on the stoichiometry, with values in the range from about 0.45 to about 0.55 having been used in the inventor's work. It is recognized, however, that the best values to use will depend on the particular reformer to which the model is applied and the system in which the reformer is used. Moreover, the efficiency depends on factors including, without limitation, the reactor temperature and the exhaust flow rate, although the inventor does not consider it generally necessary to take these factors into account. Heat is removed from the reformer 12 in accordance with the extent of Reaction (2).
Operation 53 accounts for Reaction (3). In the preferred embodiment, Reaction (3) is assumed to proceed to equilibrium based on the exhaust composition following accounting for Reactions (1) and (2). Heat is added to the reformer 12 in accordance with the extent of Reaction (3).
Operation 54 determines whether there is excess oxygen following Reactions (1)-(3). In general, there will be excess oxygen if the reformer is supplied with fuel at below the stoichiometric rate with respect to the exhaust and there will not be excess oxygen if the fuel is supplied at a stoichiometric or higher rate. If there is excess oxygen, the process 50 proceeds with Operation 55. If there is not, the process 50 proceeds with Operation 58.
Taking the case of excess oxygen, Operation 55 accounts for the release of stored fuel. The release rate may be assumed to be a stoichiometric rate in proportion to the excess oxygen as long as stored fuel is available. Other assumptions may also be used, although the stoichiometric rate assumption is preferred.
Operation 56 accounts for the heat released by reaction of the released fuel. The extent of reaction may be assumed to be stoichiometric with respect to the amount of excess oxygen. The term release is used in a broad sense: the fuel may react without physically moving from its stored location. Operation 57 is a place-holder to account for slip of released fuel. In the inventor's preferred embodiment, there is no fuel slip under conditions of excess oxygen.
Taking the case of excess fuel, Operation 58 determines the amount of the excess fuel that is stored. In one model, the fuel storage amount is a fraction of the fuel that is in liquid form. For example, it may be assumed that about 90% of the excess fuel is in liquid form and that about 45% of this liquid fuel becomes stored on the surfaces of the reformer 12. It should be understood that the inventor's concepts have a largely empirical basis and are independent of the actual mechanism of fuel storage. The actual mechanism may be, for example, physical absorption or chemical adsorption. Whatever the actual mechanism, the inventor has found it can be sufficient to assume that the fuel storage rate is a fixed fraction of excess fuel flow rate. The remaining portion of the excess fuel that is not stored in the reformer 12 is considered to slip from the reformer 12 and is tracked in Operation 59 for use in the management of downstream devices.
Operation 60 adds or removes heat from the reformer 12 based on convective heat transfer: the net heat added to or taken up by the exhaust passing through the reformer 12. Operation 61 advances a clock in preparation for the next iteration of the process 50. If the process 50 is being applied to determine a peak predicted temperature of the reformer 12, the iterations may cease when the amount of stored fuel in the reformer 12 is reduced to zero or when the temperature of the reformer 12 begins to decline.
FIG. 4 provides a schematic of an exemplary control architecture 100 that can be used to control both the temperature of the reformer 12 and the temperature of the LNT 13. The control architecture 100 includes inner and outer loop controls and uses a model of the reformer 12 that tracks hydrocarbon storage and subsequent release.
The LNT temperature controller 102 is activated by a desulfation scheduler/controller 101 that applies any appropriate criteria to determine when to initiate a desulfation process. The LNT temperature controller 102 considers a LNT temperature provided by a state estimator 103. It is preferred to use an observer or state estimator to determine the LNT temperature, because the LNT temperature responds comparatively slowly to controllable parameters. If some form of prediction is not used, there is a risk of the LNT temperature exceeding an intended limit. An extrapolation based on the current measured temperature, its rate of change, and an estimate of the temperature measurement delay is generally sufficient. However, a model of the LNT 13 can be used. Such a model preferably takes into account reactions of hydrocarbons slipping from the reformer 12. These hydrocarbons can react with residual oxygen in the exhaust or with oxygen stored in the LNT 13. When there is no oxygen in the exhaust, some hydrocarbons may become stored in the LNT 13 and subsequently react when oxygen becomes available. These processes can be modeled as they are for the reformer 12.
The output of the LNT temperature controller 102 is instructions for the reformer controller 106. The instructions may simply be instructions for the reformer 12 to switch between active and inactive modes. During the active mode, the reformer 12 is heated to a temperature suitable for reformate production and controlled to produce reformate subject to not overheating the reformer 12. During an inactive mode, the reformer 12 is generally “off”, meaning there is no reductant injection and the reformer 12 is allowed to cool freely.
When the reformer 12 is to be active, the reformer controller 106 regulates the reformer temperature at least by issuing commands to the injection controller 107. The injection influences the state of the reformer 12, which is illustrated by block 108 in the exemplary control architecture. A state includes all properties of the reformer, including its temperature, the composition of the exhaust entering it, and the composition of the exhaust leaving it. The temperature portion of the reformer state 108 is estimated by the reformer temperature estimator 105, and used to provide feedback for the reformer temperature controller 106. Accordingly, steps 105-108 comprise the inner loop of the control process 100.
The reformer state 108 influences the LNT state 109. The temperature portion of the LNT state 109 is temperature estimated by the LNT temperature state estimator 103 to provide a temperature estimate that is used by the LNT temperature control 102 in an outer control loop.
FIG. 5 illustrates a desulfation control process 200 consistent with the control architecture 100. The process 200 begins with operation 201, determining whether desulfation is required. The determination may be made in any suitable fashion. For example, desulfation may be scheduled periodically, e.g., after every 30 hours of operation. Alternatively, the need for desulfation can be determined based on system performance, e.g., based on the activity of the LNT 13 following an extensive denitration or based on the frequency with which denitration is required having increased to an excessive degree.
The desulfation process begins with operation 202, warming the reformer 12. The reformer 12 can be heated in any suitable fashion. In this example, the reformer 12 is heated by injecting fuel at a rate that keeps the exhaust at or below a stoichiometric fuel to oxygen ratio. Substantially all the fuel thereby combusts in the reformer 12 to produce heat with essentially no reformate production.
The LNT 13 heats while the reformer 12 is heating, however, after the reformer 12 is fully heated, the LNT 13 may still require further heating. If necessary, at or below stoichiometric operation may be extended to adequately heat the LNT 13. In one example, the LNT 13 is heated to a temperature of at least about 450° C. prior to commencing rich operation.
Once the warm-up phase is complete, operation 203 begins. The fuel injection rate at this stage may be controlled to give a targeted reformate production rate. Where the controller 10 can throttle the engine air intake or select the transmission gear ratio, these control parameters can be selected to facilitate the efficient production and/or usage of the reformate.
Operation 204 determines whether the reformer 12 is overheating. Preferably, this determination is made on the basis of a predicted temperature wherein a predicted temperature is a temperature that will or could occur at a future time. In other words, a predicted temperature is a temperature that would occur at some point in the future if predetermined assumptions are met. Predetermined means that the assumptions are made first and the temperature predicted second, based on the assumptions. As used herein, the term prediction does not include an estimate of a current temperature based on past information, which will be referred to herein as an estimate to avoid confusion. Also, the term “predicted future temperature” may be used to explicitly distinguish an estimate of a current temperature. The main purpose of using a prediction herein is to account for the effect of hydrocarbon storage and subsequent reaction. The prediction is therefore preferably made on the basis of a model that takes into account this phenomenon.
A prediction of the type described herein is typically made using a measured temperature and a predicted or possible temperature increase. A possible temperature increase could be made on the basis of the assumption that the fuel dosing will stop in the next instant and thereafter an excess of oxygen will become available for combustion of stored hydrocarbons. The model may look ahead over some finite interval of time to determine the value at which the temperature will peak.
Some of the inventor's concepts can be implemented without attempting to accurately predict a temperature. For example, a temperature at which the reformer 12 is determined to be on the verge of overheating can be set as a function that decreases with increasing hydrocarbon storage amount. As a more specific example, it can be assumed that the reformer 12 will heat after fuel cut-off by an amount that is proportional the amount of stored fuel. The amount of heating can be used as the amount by which the limit temperature is reduced.
Another approach contemplated by the inventor is to make an effective prediction through the mechanics by which a temperature estimate is formed. For example, one method of forming a temperature estimate is Kalman filtering. In Kalman filtering, a temperature estimate is made on the basis of a blended average of a measured temperature and a model-based estimate of the current temperature based on a past system state. The Kalman filter estimate can be converted to a prediction by using artificial values to form the model-based estimate whereby the model-based estimate is no longer intended to accurately estimate a current temperature. For example, the Kalman filter can be given accelerated dynamics. Accelerating the dynamics typically involves reducing a term reflecting the heat capacity of the reformer. The model prediction may also depart from an approximation of actual conditions by assuming the presence of excess oxygen not thought to be present under current conditions.
When the reformer 12 is on the verge of overheating, operation 205 shuts off the fuel injection. In operation 206, the process 200 waits while the reformer 12 cools. The length of the waiting period can be determined in any suitable fashion. In one example, operation 206 lasts until the reformer 12 has cooled to a target temperature. In another example, there is a fixed period between each fuel pulse. In a further example, the length of the period is selected dynamically by the controller as part of a process of optimizing the amount of reformate production per unit fuel injected.
Operation 207 determines whether the LNT 13 is getting too hot. A temperature prediction is preferably used in making this determination as the LNT 13 may heat considerably following the termination of fuel injection. If the LNT is getting too hot, operation 208 terminates the fuel injection. Terminating the fuel injection may comprise issuing instructions to the inner loop control. If the LNT is not getting too hot, the process continues with Operation 210, which checks whether desulfation is complete. Fuel injection continues in Operation 203 if fuel injection is not complete and terminates in Operation 211 if desulfation is complete.
Operation 209 is another waiting operation. In one example, this comprises waiting until the LNT 13 has cooled to a target temperature. Preferably, however, there is a fixed period between phases of active fuel injection on the longer time scale determined by the outer loop controls.
Following operation 209, the reformer 12 is heated again in operation 202. If the reformer 12 is of the type that must be heated to operate effectively, heating is generally necessary following a period of no fuel injection during which the LNT 13 is allowed to cool. The periods of no fuel injection to cool the reformer 12 are generally shorter and are normally selected to avoid having to reheat the reformer 12 to a temperature suitable for producing reformate. After the longer periods of no fuel injection to cool the LNT 13, the reformer 12 is generally too cool to effectively produce reformate without a heating period. Such a heating period generally comprises fuel injection at a sub-stoichiometric rate with respect to the exhaust oxygen content.
While the engine 9 is preferably a compression ignition diesel engine, the various concepts of the invention are applicable to power generation systems with lean-burn gasoline engines or any other type of engine that produces an oxygen rich, NOx-containing exhaust. For purposes of the present disclosure, NOx consists of NO and NO₂.
The transmission 8 can be any suitable type of automatic transmission. The transmission 8 can be a conventional transmission such as a counter-shaft type mechanical transmission, but is preferably a CVT. A CVT can provide a much larger selection of operating points than a conventional transmission and generally also provides a broader range of torque multipliers. In general, a CVT will also avoid or minimize interruptions in power transmission during shifting. Examples of CVT systems include hydrostatic transmissions; rolling contact traction drives; overrunning clutch designs; electrics; multispeed gear boxes with slipping clutches; and V-belt traction drives. A CVT may involve power splitting and may also include a multi-step transmission.
A preferred CVT provides a wide range of torque multiplication ratios, reduces the need for shifting in comparison to a conventional transmission, and subjects the CVT to only a fraction of the peak torque levels produced by the engine. This can be achieved using a step-down gear set to reduce the torque passing through the CVT. Torque from the CVT passes through a step-up gear set that restores the torque. The CVT is further protected by splitting the torque from the engine, and recombining the torque in a planetary gear set. The planetary gear set mixes or combines a direct torque element transmitted from the engine through a stepped automatic transmission with a torque element from a CVT, such as a band-type CVT. The combination provides an overall CVT in which only a portion of the torque passes through the band-type CVT.
The fuel injector 11 can be of any suitable type. Preferably, it provides the fuel in an atomized or vaporized spray. The fuel may be injected at the pressure provided by a fuel pump for the engine 9. Preferably, however, the fuel passes through a pressure intensifier operating on hydraulic principles to at least double the fuel pressure from that provided by the fuel pump to provide the fuel at a pressure of at least about 4 bar.
The lean-NOx catalyst 15 can be either an HC-SCR catalyst, a CO-SCR catalyst, or a H₂-SCR catalyst. Examples of HC-SCR catalysts include transitional metals loaded on refractory oxides or exchanged into zeolites. Examples of transition metals include copper, chromium, iron, cobalt, nickel, cadmium, silver, gold, iridium, platinum and manganese, and mixtures thereof. Exemplary of refractory oxides include alumina, zirconia, silica-alumina, and titania. Useful zeolites include ZSM-5, Y zeolites, Mordenite, and Ferrerite. Preferred zeolites have Si:Al ratios greater than about 5, optionally greater than about 20. Specific examples of zeolite-based HC-SCR catalysts include Cu-ZSM-5, Fe-ZSM-5, and Co-ZSM-5. A CeO₂coating may reduce water and SO₂deactivation of these catalysts. Cu/ZSM-5 is effective in the temperature range from about 300 to about 450° C. Specific examples of refractory oxide-based catalysts include alumina-supported silver. Two or more catalysts can be combined to extend the effective temperature window.
Where a hydrocarbon-storing function is desired, zeolites can be effective. U.S. Pat. No. 6,202,407 describes HC-SCR catalysts that have a hydrocarbon storing function. The catalysts are amphoteric metal oxides. The metal oxides are amphoteric in the sense of showing reactivity with both acids and bases. Specific examples include gamma-alumina, Ga₂O₃, and ZrO₂. Precious metals are optional. Where precious metals are used, the less expensive precious metals such as Cu, Ni, or Sn can be used instead of Pt, Pd, or Rh.
In the present disclosure, the term hydrocarbon is inclusive of all species consisting essentially of hydrogen and carbon atoms, however, a HC-SCR catalyst does not need to show activity with respect to every hydrocarbon molecule. For example, some HC-SCR catalysts will be better adapted to utilizing short-chain hydrocarbons and HC-SCR catalysts in general are not expected to show substantial activity with respect to CH₄.
Examples of CO-SCR catalysts include precious metals on refractory oxide supports. Specific examples include Rh on a CeO₂—ZrO₂support and Cu and/or Fe ZrO₂support.
Examples of H₂-SCR catalysts also include precious metals on refractory oxide supports. Specific examples include Pt supported on mixed LaMnO₃, CeO₂, and MnO₂, Pt supported on mixed ZiO₂and TiO₂, Ru supported on MgO, and Ru supported on Al₂O₃.
The lean-NOx catalyst 15 can be positioned differently from illustrated in FIG. 1. In one embodiment, the lean NOx catalyst 15 is upstream of the fuel injector 11. In another embodiment the lean NOx catalyst is downstream of the reformer 12, whereby the lean NOx catalyst 15 can use reformer products as reductants. In a further embodiment, the lean NOx catalyst 15 is well downstream of the LNT 13, whereby the lean NOx catalyst 15 can be protected from high temperatures associated with desulfating the LNT 13.
A fuel reformer is a device that converts heavier fuels into lighter compounds without fully combusting the fuel. A fuel reformer can be a catalytic reformer or a plasma reformer. Preferably, the reformer 12 is a partial oxidation catalytic reformer. A partial oxidation catalytic reformer comprises a reformer catalyst. Examples of reformer catalysts include precious metals, such as Pt, Pd, or Ru, and oxides of Al, Mg, and Ni, the later group being typically combined with one or more of CaO, K₂O, and a rare earth metal such as Ce to increase activity. A reformer is preferably small in size as compared to an oxidation catalyst or a three-way catalyst designed to perform its primary functions at temperatures below 500° C. A partial oxidation catalytic reformer is generally operative at temperatures from about 600 to about 1100° C. A preferred reformer has a low thermal mass and a low catalyst loading as compared to a device designed to produce reformate at exhaust gas temperatures.
The NOx absorber-catalyst 13 can comprise any suitable NOx-adsorbing material. Examples of NOx adsorbing materials include oxides, carbonates, and hydroxides of alkaline earth metals such as Mg, Ca, Sr, and Be or alkali metals such as K or Ce. Further examples of NOx-adsorbing materials include molecular sieves, such as zeolites, alumina, silica, and activated carbon. Still further examples include metal phosphates, such as phosphates of titanium and zirconium. Generally, the NOx-adsorbing material is an alkaline earth oxide. The absorbent is typically combined with a binder and either formed into a self-supporting structure or applied as a coating over an inert substrate.
The LNT 13 also comprises a catalyst for the reduction of NOx in a reducing environment. The catalyst can be, for example, one or more precious metals, such as Au, Ag, and Cu, group VIII metals, such as Pt, Pd, Ru, Ni, and Co, Cr, Mo, or K. A typical catalyst includes Pt and Rh, although it may be desirable to reduce or eliminate the Rh to favor the production of NH₃over N₂. Precious metal catalysts also facilitate the absorbent function of alkaline earth oxide absorbers.
Absorbents and catalysts according to the present invention are generally adapted for use in vehicle exhaust systems. Vehicle exhaust systems create restriction on weight, dimensions, and durability. For example, a NOx absorbent bed for a vehicle exhaust systems must be reasonably resistant to degradation under the vibrations encountered during vehicle operation.
An absorbent bed or catalyst brick can have any suitable structure. Examples of suitable structures may include monoliths, packed beds, and layered screening. A packed bed is preferably formed into a cohesive mass by sintering the particles or adhering them with a binder. When the bed has an absorbent function, preferably any thick walls, large particles, or thick coatings have a macro-porous structure facilitating access to micro-pores where adsorption occurs. A macro-porous structure can be developed by forming the walls, particles, or coatings from small particles of adsorbant sintered together or held together with a binder.
The ammonia-SCR catalyst 14 is a catalyst effective to catalyze reactions between NOx and NH₃to reduce NOx to N₂in lean exhaust. Examples of SCR catalysts include oxides of metals such as Cu, Zn, V, Cr, Al, Ti, Mn, Co, Fe, Ni, Pd, Pt, Rh, Rd, Mo, W, and Ce, zeolites, such as ZSM-5 or ZSM-11, substituted with metal ions such as cations of Cu, Co, Ag, Zn, or Pt, and activated carbon. Preferably, the ammonia-SCR catalyst 14 is designed to tolerate temperatures required to desulfate the LNT 13.
The particulate filter 16 can have any suitable structure. Examples of suitable structures include monolithic wall flow filters, which are typically made from ceramics, especially cordierite or SiC, blocks of ceramic foams, monolith-like structures of porous sintered metals or metal-foams, and wound, knit, or braided structures of temperature resistant fibers, such as ceramic or metallic fibers. Typical pore sizes for the filter elements are about 10 μm or less. Optionally, one or more of the reformer 12, the LNT 13, the lean-NOx catalyst 15, or the ammonia-SCR catalyst 14 is integrated as a coating or within the structure of the DPF 16.
The DPF 16 is regenerated to remove accumulated soot. The DPF 16 can be of the type that is regenerated continuously or intermittently. For intermittent regeneration, the DPF 16 is heated, using a reformer 12 for example. The DPF 16 is heated to a temperature at which accumulated soot combusts with O₂. This temperature can be lowered by providing the DPF 16 with a suitable catalyst. After the DPF 16 is heated, soot is combusted in an oxygen rich environment.
For continuous regeneration, the DPF 16 may be provided with a catalyst that promotes combustion of soot by both NO₂and O₂. Examples of catalysts that promote the oxidation of soot by both NO₂and O₂include oxides of Ce, Zr, La, Y, and Nd. To completely eliminate the need for intermittent regeneration, it may be necessary to provide an additional oxidation catalyst to promote the oxidation of NO to NO₂and thereby provide sufficient NO₂to combust soot as quickly as it accumulates. Where regeneration is continuous, the DPF 16 is suitably placed upstream of the reformer 12. Where the DPF 16 is not continuously regenerated, it is generally positioned as illustrated downstream of the reformer 12. An advantage of the position illustrated in FIG. 2 is that the DPF 16 buffers the temperature between the reformer 12 and the LNT 13.
The clean-up catalyst 17 is preferably functional to oxidize unburned hydrocarbons from the engine 9, unused reductants, and any H₂S released from the NOx absorber-catalyst 13 and not oxidized by the ammonia-SCR catalyst 15. Any suitable oxidation catalyst can be used. A typical oxidation catalyst is a precious metal, such as platinum. To allow the clean-up catalyst 17 to function under rich conditions, the catalyst may include an oxygen-storing component, such as ceria. Removal of H₂S, where required, may be facilitated by one or more additional components such as NiO, Fe₂O₃, MnO₂, CoO, and CrO₂.
The invention as delineated by the following claims has been shown and/or described in terms of certain concepts, components, and features. While a particular component or feature may have been disclosed herein with respect to only one of several concepts or examples or in both broad and narrow terms, the components or features in their broad or narrow conceptions may be combined with one or more other components or features in their broad or narrow conceptions wherein such a combination would be recognized as logical by one of ordinary skill in the art. Also, this one specification may describe more than one invention and the following claims do not necessarily encompass every concept, aspect, embodiment, or example described herein.

Claims

1. A method of fuel reforming within an internal combustion engine exhaust line, comprising:

injecting fuel into the exhaust line upstream of a fuel reformer;

measuring a temperature within the exhaust line;

predicting a temperature based in part on the measured temperature; and

controlling the fuel injection using the predicted temperature;

wherein the predicted temperature is a temperature that would occur at some point in the future if predetermined assumptions are met.

2. The method of claim 1, wherein controlling the fuel injection using the predicted temperature comprises temporarily discontinuing the fuel injection if the predicted temperature meets or exceeds a critical value.

3. The method of claim 1, wherein injecting fuel into the exhaust line upstream of the fuel reformer comprises:

injecting fuel at rate that produces a sub-stoichiometric concentration of fuel in the exhaust line to heat the fuel reformer to a temperature suitable for producing reformate; and

subsequently injecting fuel at a higher rate to produce a super-stoichiometric concentration of fuel in the exhaust line in order to produce reformate.

4. The method of claim 1, wherein the temperature prediction is based in part on a model.

5. The method of claim 4, wherein the fuel is injected in pulses and the model predicts the availability within the reformer of a portion of the injected fuel in periods between temporally adjacent fuel pulses.

6. The method of claim 4, wherein the model takes into account fuel storage within the fuel reformer and subsequent reaction of the stored fuel.

7. The method of claim 4, wherein the model predicts the reformer will heat in periods following the termination of fuel injection due to reactions of previously injected fuel within the reformer.

8. The method of claim 4, wherein the model predicts that some of the injected fuel will absorb within the fuel reformer without immediately reacting, but will subsequently react.

9. An exhaust treatment system comprising a controller, wherein the controller implements the method of claim 1.

10. A vehicle comprising the exhaust treatment system, of claim 9.

11. A method of controlling the temperature of a fuel reformer, comprising:

using a model to predict a temperature associated with the reformer; and

using the predicted temperature in a temperature control algorithm;

wherein the temperature prediction takes into account hydrocarbon storage and subsequent reaction.

12. The method of claim 11, wherein the fuel reformer is configured within an exhaust line upstream of a lean NOx trap.

13. The method of claim 12, wherein fuel is injected into the exhaust line in pulses during a regeneration of the lean NOx trap.

14. The method of claim 11, wherein the model predicts the reformer temperature using dynamics that are faster than the actual dynamics are expected to be.

15. A method of controlling the temperature of a fuel reformer, comprising:

predicting future reformer temperatures; and

using the predicted future reformer temperatures in a feedback control loop;

wherein the predictions take into account hydrocarbon storage and subsequent availability for reaction.

16. The method of claim 15, wherein the fuel reformer is configured within an exhaust line upstream of a lean NOx trap.

17. The method of claim 15, wherein fuel is injected into the exhaust line in pulses during a regeneration of the lean NOx trap.

18. A method of reforming within an internal combustion engine exhaust line, comprising:

injecting hydrocarbons into the exhaust line upstream of a reformer;

estimating hydrocarbon storage by the reformer;

controlling the reformer temperature based in part on the hydrocarbon storage estimate.

19. The method of claim 18, wherein the fuel reformer is configured within an exhaust line upstream of a lean NOx trap.

20. The method of claim 18, wherein fuel is injected into the exhaust line in pulses during a regeneration of the lean NOx trap.

21. The method of claim 18, wherein the hydrocarbon storage estimate is used in a thermal model of the reformer.

22. The method of claim 21, wherein the model is applied with accelerated dynamics.

23. The method of claim 18, controlling the reformer temperature based in part on the hydrocarbon storage estimate comprises temporarily terminating the hydrocarbon injection if the estimated amount of hydrocarbon stored is too high.

24. The method of claim 18, wherein injecting hydrocarbons into the exhaust line upstream of a reformer comprises:

injecting hydrocarbons at rate that produces a sub-stoichiometric concentration of hydrocarbons in the exhaust line to heat the fuel reformer to a temperature suitable for producing reformate; and

subsequently injecting hydrocarbons at a higher rate to produce a super-stoichiometric concentration of fuel in the exhaust line in order to produce reformate.