To model and analyse the after-treatment system for an SI engine system using GT-Power.
Objective : To model and analyse the aftertreatment system for an SI engine system using Gt-Power.
Description :
-SI engine aftertreatment model :
This example demonstrates modeling an SI engine combined with a detailed TWC in a secondary aftertreatment circuit. The original base model SI_4cyl_EGRController was combined with the aftertreatment example model TWC_Ramanathan_and_Sharma, which uses a published reaction mechanism for a passenger car TWC. A secondary circuit is used to solve the reactions using the quasi-steady (QS) flow solver, which is able to use a large master time step so that it does not restrict the computation time. The back pressure from the explicit engine circuit is passed to the QS circuit and actuated as the downstream pressure boundary, and the outlet gas temperature from the QS circuit is passed back to the explicit engine circuit and actuated as the TWC wall temperature so that the exothermic heat release is captured for acoustics. A resonator and muffler could then be added downstream of the TWC. The various monitors show the TWC conversion efficiency, the exhaust lambda, the oxygen storage, the temperature upstream and downstream of the TWC, and the deltaT across the TWC.
There are three cases in Case Setup for three air to fuel ratios for rich, stoich, and lean conditions. The engine speed is set to 3000 rpm, and the throttle angle is set to 45° for a part load cruising situation. The burned fuel fraction in the SIWiebe combustion model is set to 0.995 to generate unburned liquid and vapor indolene fuel as the hydrocarbons, which are then converted to C3H6 and C3H8 by a control system for the TWC model.
-CI engine aftertreatment model :
This example demonstrates an integrated model with detailed engine and detailed aftertreatment system running one steady-state operating point. The detailed engine model comes from the Diesel_WGController example in the Engine Performance directory. The detailed aftertreatment system is a combination of the DOC_Sampara_and_Bissett and SCR_Fe_Zeolite example models along with a DPF setup for passive regeneration only, and same urea injection system used in the EngineState_Vehicle_Aftertreatment_NEDC example model.
The model is set to shut-off at steady-state and has several Convergence RLTs set under Run Setup > ConvergenceRLT to establish convergence. The global wall temperature solver is set to "transient" because the QS Advanced Adaptive solver used in the DOC and SCR catalysts requires it.
The exhaust system is duplicated such that one copy uses the explicit flow solver and the second copy uses the quasi-steady (QS) flow solver with larger, 0.1 second time step. The explicit and QS flow circuit settings are controlled under Home > Run Setup > FlowControl. The explicit solver time step is automatically controlled, but the QS circuit time step is user-imposed. Notice that when the "Improved Solution Sequence for Multi-Circuit Models" is activated in Run Setup > TimeControl, the QS circuit can use a "continuous" flag in Run Setup > FlowControl. A QS time step of 0.1 seconds or less is recommended.
In the "Aftertreatment_Inlet" subassembly, the cycle-averaged mass flow, temperature, and species concentrations are transferred from the explicit circuit to the QS circuit. When transferring species concentrations, conversion of unburned fuel must be considered. The only hydrocarbons emitted from the engine cylinders will be unburned liquid and vapor fuel. Catalytic components are usually characterized by smaller hydrocarbons. Therefore the unburned fuel should be split into representative hydrocarbons for which the aftertreatment components were calibrated. The fuel converter is conceptualized by writing and balancing a stoichiometric equation. Excess H2O is used to help balance the conversion, and O2 is often required as a product, although its generation will not affect the overall lambda that the aftertreatment system sees.
If a total HC measurement is known, it should be entered in the EngCylinder Emissions Map object attribute using a C1 or C3 basis, depending on how the value was measured. In that case, the fuel conversion can be done directly in the Emissions Map object by defining the smaller hydrocarbons and assigning values that are some fraction of the total hydrocarbons.
Emissions map is used in the cylinders of this model to provide an arbitrary amount of soot for the DPF. This model is using the predictive NOx model in the combustion object, which will predict NO only based on temperature and composition in the cylinder. CO will be predicted from chemical equilibrium in the cylinder. In general if accuracy is needed it is best to use emissions maps for all the emission species.
-Integrated Diesel Engine (Map-based), Vehicle, and Detailed Aftertreatment Model :
This example demonstrates an integrated model with Engine, Vehicle, Controls, and Aftertreatment system all in one model running an 1180 sec NEDC. Computation time for this model should be on the order of 10 times faster than RT on a modern computer.
This type of model can be used to quickly analyze and study the effects of changes to transmission gearing, shift strategy, vehicle mass, aftertreatment device sizing, PGM loading, and sequential location of aftertreatment devices upon the emissions and fuel economy of the vehicle. It can be used to test and validate different drive cycles, and for long distance simulations.
Components of the model:
1) Engine - EngineState is a map-based engine with maps of air flow, fuel flow, and emissions vs. engine speed and load in BMEP. Since it is map-based computation speed is very fast. This particular EngineState is based on a 2.0L turbo-charged diesel engine. Alternatively a detailed engine model, a Fast Running engine Model (FRM), or a Mean Value (MV) engine model could be the engine source, given in order of decreasing level of detail, design flexibility, and computation time.
2) Vehicle - Front-wheel drive passenger car with 5-speed transmission.
3) Driver Controls - Providing the accelerator and brake position transient input for the drive cycle.
4) Aftertreatment System - DOC+DPF+SCR with aqueous urea injection. The DOC reduces CO and hydrocarbons, and converts NO to NO2. The NO2 is then used to passively regenerate the uncatalyzed DPF at relatively low temperatures (400-450 K). The SCR is used to reduce the NO and NO2 that reaches it. It does this by reacting NH3 with the NO and NO2, and of course the NH3 comes from the aqueous urea injection upstream, which breaks down into NH3 and CO2.
Important output from the model:
The plot output has been reduced to only show the run time monitors and two pre-processed emissions maps.
1) Fuel Consumption RLT monitor - shows the final value of Average Fuel Consumption RLT in units of L/100 km.
2) Cumulative Emissions shows the cumulative CO, HC, and NOx mass for the drive cycle.
3) DOC Monitor - shows DOC conversion efficiency for CO, HC, and NO.
4) DPF Monitor - shows filter substrate temperature and collected soot mass.
5) SCR Monitor - shows NH3, NO, and NO2 species concentrations entering and exiting the SCR.
6) Vehicle Monitor - shows the target and actual vehicle speed, and the accelerator pedal position.
7) Pre-processed emissions maps for NO and soot - Typically emissions map data will come in the form of scattered data points, either from a set of test bench results, or from a GT-POWER engine model. The output emissions are a function of engine speed and load (BMEP, Brk Tq, IMEP, Ind Tq). Based upon the upper and lower boundary curves, the speed and load operating domain is then established. Since 2D interpolation on scattered data points is expensive in terms of CPU time, a normalized orthogonal map is built through several pre-processing steps. The normalized load represents the % of max load at the given speed. The intermediate steps to generate the map are shown in GT-POST.
Simulation of Detailed SI engine model :
Engine Specifications :
EGR Controller :
Aftertreament model (3- way catalytic convertor) :
Case setup :
Consists of three cases, for rich, stoichiometric and lean A/F mixture .
Post-processing :
Case 1 : A/F ratio - 14 (Rich)
Conversion Efficiency :
Case 2 : A/F ratio - 14.5 ( Stoichiometric)
Conversion Efficiency :
Case 2 : A/F ratio - 15 (Lean)
Conversion Efficiency :
As the mixture becomes leaner, more Nox is formed due to more unburnt air, hence Nox value is more and conversion efficiency is less.
As the mixture becomes richer, HC and CO values are more due to incomplete combustion, and hence their conversion efficiency is less as compared to the leaner case.