Patent Application
20140114619
The present disclosure generally relates to a system for treating exhaust gases. More particularly, methodologies for optimizing a burner arrangement to increase an exhaust gas temperature as well as optimizing a reagent injector location are discussed.
To reduce the quantity of NOX and particulate matter emitted to the atmosphere during internal combustion engine operation, a number of exhaust aftertreatment devices have been developed. A need for exhaust aftertreatment systems particularly arises when diesel combustion processes are implemented. Typical aftertreatment systems for diesel engine exhaust may include one or more of a diesel particulate filter (DPF), a selective catalytic reduction (SCR) system, a hydrocarbon (HC) injector, and a diesel oxidation catalyst (DOC).
During engine operation, the DPF traps soot emitted by the engine and reduces the emission of particulate matter (PM). Over time, the DPF becomes loaded and begins to clog. Periodically, regeneration or oxidation of the trapped soot in the DPF is required for proper operation. To regenerate the DPF, relatively high exhaust temperatures in combination with an ample amount of oxygen in the exhaust stream are needed to oxidize the soot trapped in the filter.
The DOC is typically used to generate heat useful for regenerating the soot loaded DPF. When hydrocarbons (HC) are sprayed over the DOC at or above a specific light-off temperature, the HC will oxidize. This reaction is highly exothermic and the exhaust gases are heated during light-off. The heated exhaust gases are used to regenerate the DPF.
Under many engine operating conditions, however, the exhaust gas is not hot enough to achieve a DOC light-off temperature of approximately 300° C. As such, DPF regeneration does not passively occur. Furthermore, NOx absorbers and selective catalytic reduction systems typically require a minimum exhaust temperature to properly operate.
A burner may be provided to heat the exhaust stream upstream of the various aftertreatment devices. Known burners have successfully increased the exhaust temperature of relatively small displacement internal combustion engines for automotive use. However, other applications including diesel locomotives, stationary power plants, marine vessels and others may be equipped with relatively large diesel compression engines. The exhaust mass flow rate from the larger engines may be more than ten times the maximum flow rate typically provided to the burner. While it may be possible to increase the size of the burner to account for the increased exhaust mass flow rate, the cost, weight and packaging concerns associated with this solution may be unacceptable. Therefore, a need may exist in the art for a method to optimize an arrangement to increase the temperature of the exhaust output from a large diesel engine while minimally affecting the cost, weight, size and performance of the exhaust system.
Additional challenges exist relating to dispersion of injected reagent in relatively large exhaust treatment systems. In particular, some systems include multiple parallel flow paths containing exhaust to be treated. Injected reagent may flow into the closest inlet of least restriction. Some of the parallel paths may receive a greater quantity of reagent than required while other paths may receive less than intended. As such, it may be desirable to provide a methodology to optimize injector location in high volume exhaust systems.
To meet EPA Tier IV large diesel engine emission targets, intensive development efforts have been taken to achieve NOx reduction and Particulate Matter (PM) reduction targets. With respect to NOx reduction, liquid urea is typically used as the reagent to react with NOx via SCR catalyst. Regarding PM reduction, additional heat is required to raise exhaust temperature to reach a Diesel Particulate Filter (DPF) active/passive regeneration performance window. Typically the heat may be generated by external diesel burners which allow diesel liquid droplets to react directly with oxygen in the exhaust gas. Alternatively the heat can be generated by catalytic burners which enable diesel vapor to react with oxygen via a Diesel Oxidation Catalyst (DOC) mostly through surface reactions. The latest technology trend is to combine both mechanisms together so that (1) small-scale burners will enhance exhaust temperature to DOC activation temperature; (2) DOC with HC dosing will raise exhaust temperature to 650° C. to achieve active soot reduction.
From the scope of system level design, given the added reagents (HC and urea) and heat to exhaust gases, the need arises to achieve even distributions of exhaust gas velocity, good mixing of burner heat with exhaust gas and good mixing of reagents released from HC or urea injectors with exhaust gas, necessitating optimizations of multiphase heat and mass transport phenomena. To meet these multifaceted heat and mass transport targets, geometrical optimizations of subsystems are deemed important as well as the requirements of subcomponents such as burners and injectors. In this disclosure, a turbo-out manifold box is selected as the subsystem under study for design optimizations. Located between turbo-out ports and an exhaust aftertreatment system, the manifold is subject to burner heat injections and diesel liquid injections at multiple locations. The manifold therefore serves the mechanism to distribute exhaust gas, burner heat and diesel vapor. Downstream, the exhaust aftertreatment system incorporates multiple identical flow paths, each encompassing a DOC, a DPF, urea injectors, a Selective Catalytic Reduction Device (SCR), and an Ammonia Slip Catalyst (ASC) with complete PM and NOx reduction functionalities. Downstream of the exhaust aftertreatment system, a common stack outlet is created to be open to the environment. This disclosure discusses general design considerations, performance metrics and methodology for the development of turbo-out manifold with system targets in scope. Computational Fluid Dynamic modeling has been used as the main tool in performing design iterations.
In large-scale diesel engines such as locomotive, marine, stationary engines, the aftertreatment system development generally dictates modularity, simplicity and manufacturability for subsystems and components because the scalability and versatility have the benefits of being cost-effective and maintenance-friendly for many diverse systems. The system layouts may vary, but they must meet the comprehensive performance targets of back pressure, reductions of NOX, PM, CO and HC, system durability, cost, and maintenance.
This section provides a general summary of the disclosure, and is not a comprehensive disclosure of its full scope or all of its features.
A method of determining an optimized position for a burner in an exhaust aftertreatment system includes estimating temperature distributions across faces of exhaust treatment devices positioned within parallel paths based on an initial burner position upstream of the parallel paths. A temperature distribution across the faces of the exhaust treatment device is again estimated based on a changed burner position. A difference between the estimates is determined. The changing, estimating and determining steps are repeated to correlate the burner position with a temperature variance across the faces. An optimized burner position is determined based on minimizing the temperature variance across the faces of the exhaust treatment devices.
A method of determining an optimized position for a reagent injector in an exhaust aftertreatment system having multiple parallel exhaust gas flow paths includes estimating injected reagent distributions across faces of exhaust treatment devices positioned within the parallel paths based on an initial reagent injector position upstream of the parallel paths. The reagent injector position is changed. Another estimate of injected reagent distribution across the faces of the exhaust treatment devices is made based on the changed reagent injector position. A difference between the estimates is determined. An optimized reagent injector position is determined based on minimizing the reagent distribution variance across the faces of the exhaust treatment devices.
Further areas of applicability will become apparent from the description provided herein. The description and specific examples in this summary are intended for purposes of illustration only and are not intended to limit the scope of the present disclosure.
The drawings described herein are for illustrative purposes only of selected embodiments and not all possible implementations, and are not intended to limit the scope of the present disclosure.
Corresponding reference numerals indicate corresponding parts throughout the several views of the drawings.
Example embodiments will now be described more fully with reference to the accompanying drawings.
Each aftertreatment assembly 30 includes a coupler 32 and a tri-pipe 33. Tri-pipe 33 includes a singular inlet 34 and three outlets identified as 35a, 35b and 35c. Downstream from each of outlets 35a, 35b, 36c is an aftertreatment subassembly identified with one of reference numerals 36a, 36b or 36c. The aftertreatment subassemblies are substantially identical to one another. Accordingly, only aftertreatment subassembly 36a will be described in detail.
Aftertreatment subassembly 36a includes a diesel oxidation catalyst 37a, a diesel particulate filter 38a, a urea injector 40a, a urea mixer 42a, a selective catalytic reduction element 44a and an ASC 46a. An outlet 54 allows the treated exhaust to escape to atmosphere. To properly operate diesel oxidation catalysts 37a, 37b and 37c, at least a first burner 60 is coupled in communication with turbo exhaust manifold 16 to heat the exhaust gas passing through turbo exhaust manifold 16 and raise the temperature of each diesel oxidation catalyst 37a, 37b, 37c to approximately 285° C. to allow active HC dosing for active regeneration of each DPF 38a, 38b, 38c. Additional burners are contemplated and two or more burners are likely to be required. An air pump 64 may provide fresh air to the burners. A supply of fuel is also provided to the burners. Through computer modeling, it has been determined that two burners 60 and 62 may be required to properly increase the temperature of each of the twelve diesel oxidation catalysts implemented within the particular exhaust aftertreatment system 10 modeled. By properly positioning only two burners, as described hereinafter, cost and weight of the aftertreatment system may be minimized. The optimized position of burners 60 and 62 provides even distribution of burner heat to each aftertreatment assembly 30 as well as each aftertreatment assembly 36a, 36b and 36c. Computational Fluid Dynamic modeling estimates a temperature uniformity of more than 93% across all aftertreatment subsystems 36a, 36b and 36c.
Reagent injectors 66, 68 are positioned to inject a reagent, such as a hydrocarbon, into exhaust manifold 16. Computational Fluid Dynamic modeling analysis indicates that only two injectors need be provided as long as the injector positions are optimized. It should be appreciated that more or fewer burners or injectors may be required for an exhaust treatment system configured differently than modeled system 10.
As mentioned,
Given the multitude of its functionalities, the design objectives of turbo exhaust manifold are (1) to distribute the exhaust mass flow evenly to multiple legs, following similar design principles to that of gasoline engine exhaust manifold; (2) to receive and mix heat from multiple burners mounted on the manifold to facilitate active regeneration downstream by raising low exhaust temperatures to DOC light-off temperature such as 285° C.; (3) to receive and mix diesel fuel spray from fuel injectors mounted on the manifold by raising exhaust temperature (equal or higher than 285° C.) to DPF active regeneration temperature of 650° C. To achieve these objectives, it is necessary to formulate multifaceted mixing performance targets in species, temperature and velocity with corresponding solution schemes.
Four performance metrics are derived in Table 1 to evaluate the turbo-box performance. The performance metrics 1 and 2 ensure that each leg will receive equal share of exhaust mass flow with well-distributed exhaust velocity at the inlet of the DOC. If absent of added heat and reagents, as seen in gasoline engine applications, both metrics are sufficient for exhaust manifold design. The performance metric 3 is used to quantify the effect of mixing between burner heat and exhaust flow. The performance metric 4 is used to evaluate the mixing between liquid diesel spray and exhaust gas. Both 3 and 4 are means to satisfy the functionality of PM active regeneration.
With respect to PM reduction for diesel engines, proper thermal management of burner heat includes consideration of multiple factors: (1) burner size and power output; (2) the number of burners; (3) the exit temperature of burner; (4) the locations of burners. As shown in
Burners 60, 62 are able to supply a range of consistent hot gas flow rate at fixed temperature of 650° C., the normal burner operating temperature. The burner stream may be in parallel to the bypass stream as both streams are diverted from the main stream at the turbo-out. The purpose of burner 60, 62 is to raise exhaust temperature to a DOC activation temperature of 285° C., not to raise exhaust temperature to 650° C. The latter scenario may require an in-line burner configuration with higher required external heat flow rate and hence larger burner heat capacities. This configuration may increase system backpressure from the burner in a passive mode. Therefore, if the engine is at low idle condition below 285° C., burner heat is provided to reach DOC activation temperature 285° C. The exhaust gas needs to meet performance metric 3 to achieve temperature distribution criteria. If the engine temperature is beyond DOC activation temperature, indicating no additional heat is required, performance metric 3 is not needed.
Burner activation may not be required if active regenerations are not needed for low idle conditions, as seen in some on-highway heavy-duty truck applications. However, low idle regeneration may become crucial for off-highway heavy duty, stop-and-go refuse trucks or switch locomotive applications where PM have to be regenerated periodically during idle conditions. In these scenarios, the burner may be either standalone or combined with a HC dosing mechanism.
Once the exhaust gas temperature satisfies the DOC activation temperature, a hydrocarbon may be dosed on DOC 37a thereby generating additional heat causing the exhaust temperature to reach 650° C.
Note that the DPF regeneration temperature is assumed to be 650° C. In contrast, catalyzed DPF may enable passive regeneration at lower temperature via NO2 soot oxidization. In the latter scenario, the need of active regeneration via HC dosing or burners may be less demanded, but active regeneration may still provide failsafe benefits as a means to mitigate long term deterioration of catalyst and accumulated urea deposit. Occasional active regenerations may alleviate major risks to system robust performance.
The process for turbo manifold development is shown in
Because of the large dimension of the whole system, validation of the manifold performance poses a considerable challenge. Small scale tests are possible but may not represent the workings of the full geometry. One reason is that mixing, in essence, cannot be downscaled along with other flow characteristics. In contrast, CFD methodology could be a viable approach for this purpose since it has been established and benchmarked to some degrees in multiple test scenarios on spray modeling validations. Application experience indicates that CFD velocity simulations have been correlated relatively well with test results. With respect to spray modeling, it has been observed that system design directions provided by spray modeling correlate with the directions from test observations. Though still subject to limitations, the CFD modeling approach provides fast-turnaround design cycles and cost-effective means to address the complicated mixing and transport phenomena.
With the CFD spray modeling approaches, the system design has been improved stepwise with the final system being able to achieve pre-set performance targets of mass flow, temperature and fuel vapor distribution.
The CFD computational domain is meshed using Ansys ICEMCFD and consists of hybrid elements with 3 million cells. The substrates are modeled using hexahedral elements, manifold and other subsystems are captured using tetrahedral and 2 layer prism elements. Ansys Fluent is used as solver. For burner inlet study, continuity, momentum and energy term equations are solved sequentially. For injector location studies, spray model and species transport are added. In model setup, a steady state pressure based solver (Segregated) and 2 equation realizable k-e turbulence models with standard wall functions is used. As simplifications, radiation and heat transfer through the wall thickness are not modeled. The appropriate equations and model setup are selected for the corresponding analysis. Spray modeling is conducted using ‘Discrete Phase Model’ with quasitransient approach where the fluid flow is steady and particle trajectory unsteady.
The droplets are calculated to account for interaction with the flow field at each iteration step in order to enable two-way couplings. Spherical drag law is assumed for droplets drag parameters and droplet breakup as wave model with default constants is applied. The HC droplets are defined via solid cone spray pattern and diameter distribution is applied via ‘Rosin-Rammler’ fit derived from spray characterization of the injector. DPM wall boundary conditions are treated as wall-jet for the geometry upstream of substrates; for the substrates and the downstream components, the droplet transport is treated as “trap”, meaning the droplet will have a localized effect after hitting walls.
Returning to
This disclosure relates to two sensitivity studies, burner inlet optimization and injector location optimization. Burner inlet studies and injector studies are conducted separately by using the same CFD model geometry. In the burner sensitivity study, three parameters are selected for design optimization, the pressure loss, mass flow non-uniform flow out from turbo split and temperature distribution respectively. The flow split across each leg shows how well the current manifold design distributes the mass flow across the two legs. Different burner flow locations are formulated with the goal to reach optimized burner locations according to performance criteria of temperature distributions across DOC. Once the design direction has been identified, further fine-tuning is conducted for the final burner locations. In the HC injector sensitivity study, hydrocarbon droplets are injected at different locations to determine an optimized position of the injectors according to the performance criteria of even species distributions across the DOC faces. It should be appreciated that the optimization process described may also apply to other emissions control devices such as exhaust flow modifiers or mixers without departing from the scope of the disclosure.
For the full system analysis, a mass flow rate of 3000 Kg/hr and an inlet temperature of 150° C. are applied at the manifold box inlet. At the burner inlet, a total mass flow of 1286 Kg/hr with temperature of 650° C. is applied. Since the current model is symmetric, half the flow rate of 1500 Kg/hr is applied at the turbo outlet. For the cases with 2 or 4 burners, mass flow of each burner is calculated based on the total required burner flow rate divided by the number of burners. In the initial design of experiments, 6 burner options are considered. The initial locations selected are presented in
The rationale for burner location selection is based on packaging space and performance factors. Given the turbo flow impingement effect, ideas of Top-2 and Top-4 are conceived by mounting burner directly opposite to turbo-out flow direction, resembling the concept of mounting PCV valve nearby the strongest turbulence in intake manifold in gasoline engine applications. More design options are created to mount burners in the back and front of the turbo-out manifold.
The pressure drop, flow split and flow distribution are calculated for all 8 cases and presented in the following section. All the burner inlet designs are meshed in a different volume, so based on the design iterations each individual volume can be activated or deactivated. Due to the system symmetry, only two inlets are modeled for a system with 4 burners. Similarly, for the cases with 2 burners, only one is modeled in the half-model.
The total mass flow split and the system pressure loss from are presented in
It is evident that burners play important roles in affecting exhaust mass flow split; typically the duration will last 30 minutes for active regeneration. It is also worth mentioning that during passive soot collection period (without burner heat) the full exhaust system has demonstrated even flow split among multiple legs or aftertreatment assemblies 30. It is understandably so as long as a stable pressure difference is maintained between manifold and exhaust stack-out.
It should be noted that the mass flow rate percentage of only aftertreatment assembly 30 is used in this paper because two aftertreatment assemblies 30 are modeled. If more parallel paths are present in a different system, standard deviations of mass flow rate across all aftertreatment assemblies are recommended, as defined in performance metric 1.
The main objective of introducing burners into the system is to raise the exhaust gas temperature for DOC lightoff especially at idle or low notch conditions. These burner inlets are placed at various locations in order to achieve a uniform temperature not only across each DOC face but also across legs. The static temperature distribution across DOC front faces for all the designs are presented in
From design 1 results, the majority of the flow from the manifold inlet is directed towards leg 1 and hence leg 1 is subject to much higher thermal impact than leg 2. Design 2 results show similar flow patterns that lead to higher temperature at leg 2. From design 3, the burner flow facing the DOC inlet has better distribution compared to previous two designs. In design 4, one inlet at the top resulted in similar trend of first two designs, that is, the outer leg experiences higher temperature. From design 5 and 6, one burner at the top back or bottom back result in good distribution across both legs. Design 5 has slightly higher temperature at leg 1 whereas design 6 has slightly higher temperature at leg 2. It is therefore deduced that an inlet location in between these two designs would result in improved distribution. Comparisons of these design options indicate that design 3, as an intuitive option of facing burners directly to inlet of legs, fails to deliver the expected results. This phenomenon stresses the dominance of turbulence and swirl effect over laminar flow patterns. Based on the existing results from the flow split calculation and the temperature distribution, two more additional designs were created by offsetting the back inlets. The results of Design 7 and Design 8 are shown in
The velocity uniformity is defined as,
(6) and the temperature/enthalpy uniformity are defined by
(7), where {dot over (m)} is the mass flow rate and Q is enthalpy flux.
A similar CFD geometry is used for a sensitivity study of injector locations. The injector location optimization is similar to burner studies in that the locations of injector affect the diesel mixing performance. Based on the available packaging space, the injector locations are selected and spray modeling is conducted to predict the droplet trajectories in discrete phases. Via evaporation of droplets, the resulting species distribution across the DOC faces helps identify the optimized location for the specific manifold design.
The run condition is selected such that mass flow rate is 18000 Kg/hr with inlet temperature of 320° C. Steady state particle tracking in discrete phase model is applied. Hydrocarbon property of C10H22 with a total mass flow rate of 177 Kg/hr is applied with droplets being released into the exhaust stream via discrete parcels. Rosin Rammler diameter distribution for the spray modeling is determined from the PDPA tests of the injector. The measured spray angle of 67 degree and injector velocity of 20 m/sec are applied as the initial properties. The tentative injector locations are shown in
The mass fraction of hydrocarbon (HC) vapor evaporated from the droplets is the main factor of interest. The uniform distribution of species across the substrate faces is expected to indicate the mixing performance (Explicit in Metric 4). Contours of vapor concentrations are shown in
Option 1 result shows that most of the droplets tend to get into leg 2 due to turbulence effect, leading to a low dosing rate for DOC at leg 1. In Option 2, even though the injector is mounted on the inlet of leg 2, the incoming flow takes away the particles after impinging on the top wall and carries them to the leg 1. In both cases, the distributions across legs are significantly different but the species distribution at a specific DOC may be uniform, indicating that uniformity indices should be used across all six DOCs. To address the misdistribution of HC, a tentative idea was proposed that each injector flow rate may vary to mitigate flow induced misdistribution patterns. However, this idea may not be feasible because of added complexity and unnecessary burden in injector controls. Generally, hardware optimization is the preferred way to optimize system performance; varying dosing rate alone should only be used as the last resort.
From vapor contours in Option 3, it was observed that the individual injector for each leg reaches desired dosing rate, but vapor distributions are poor for individual DOC due to short mixing length. For leg 2, the top DOC receives higher concentration than the bottom one; for leg 3, the trend is opposite. Such difference in distribution is a direct result of interaction between turbulent flow patterns and HC droplet transport trajectory prior to reaching the DOC faces. To better understand the design implications, the pros and cons of two mounting options are described in Table 4.
Understanding of tradeoffs allows one to select a location that enables droplets to mix well with the flow before they enters the leg. One way is to mount the injectors directly on top of the turbo outlet pipe upstream of the manifold.
As evident in Option 1, the resulting diesel mass flow rates may vary across legs. One way to circumvent this limitation is to use more injectors to allow better dosing allocations. In Option 4, four injectors are mounted on top of turbo-out box. The species distribution for this design is shown in
Similar to temperature uniformity, the species or vapor uniformity can be calculated for these design options to quantify the performance of injector location optimizations. The results are presented in
A method of determining an optimized position for a reagent injector in an exhaust aftertreatment system having multiple parallel exhaust gas flow paths is accomplished in a substantially similar manner to that previously described in relation to optimizing the position for the burner as depicted at
This disclosure provides general design principles and methodologies for a representative manifold geometry with the objective to achieve multifaceted design targets for diesel turbo-out manifold. Such principles and methodologies can be applied to optimize similar large diesel turbo manifold geometry.
Overall aftertreatment system layout, the design rules, and performance metrics are provided to evaluate a turbo manifold subsystem. Given the tradeoffs between tests and simulations, CFD thermal and spray modeling are adopted as the main tool due to its cost-effectiveness and rapid development cycle. It is employed to optimize burner mounting locations and injector mounting locations. Equipped with such tools, a representative manifold is provided to explain the concept of design optimization. The specific manifold design was to evenly distribute burner heat, diesel vapor and exhaust gas through catalytic subsystems which are composed of multiple identical legs with complete functionality to reduce HC, CO, PM and NOx. Given this specific manifold, design optimizations on burner and injector locations have been performed.
Successes have been achieved from burner and HC layout optimizations. In the burner layout study, flow split per leg and temperature distribution are predicted for progressive designs, resulting in the optimized design of burner mounting locations. Among the 8 designs, Design 5 (2 burners on the back of the manifold) yields the best temperature distribution across the DOC face. In injector layout study, the fuel mass flow rate per leg and the mass fraction of evaporated HC species distribution are analyzed. Among 4 design options, Option 4 (4 injectors mounted on top of the manifold box) resulted in the best design in diesel vapor distribution and vapor mass flow rate per leg.
The foregoing description of the embodiments has been provided for purposes of illustration and description. It is not intended to be exhaustive or to limit the disclosure. Individual elements or features of a particular embodiment are generally not limited to that particular embodiment, but, where applicable, are interchangeable and can be used in a selected embodiment, even if not specifically shown or described. The same may also be varied in many ways. Such variations are not to be regarded as a departure from the disclosure, and all such modifications are intended to be included within the scope of the disclosure.
