Virtual ICE Development - Performance and Emissions - Release 2020 R1
Release 2020 R1
Virtual ICE Development - Performance and Emissions
The latest release of ICE performance and emissions portfolio – version 2020 R1 – includes a range of enhancements that bring a variety of new benefits to its users.
- AVL CRUISE™ M now offers comprehensive layout assessment for all kinds of exhaust gas aftertreatment components and systems
- New features and enhancements to simplify and accelerate thermodynamic cycle simulations in CRUISE M without compromising accuracy
- Less time required for model preparation and setup of in-cylinder flow, combustion and emission simulations in AVL FIRE™
- Advanced spark ignition model CADIM and Eulerian Flame Tracking Model E-FTM now coupled to FIRE FGM, allowing closer combustion examination
- The investigation of new fuels is supported by offering a multi-component flash boiling model in FIRE
- Higher accuracy in thermal load simulations due to enhanced FIRE RPI Boiling model
- The effect of wall adhesion is accounted for, improving heat transfer calculations in FIRE and FIRE M
- Sophisticated models in FIRE M allow resolving more details of boiling flows during immersion quenching, and are applicable to both aluminum and steel parts
- Proven exhaust gas aftertreatment capabilities of BOOST™ and FIRE are now applicable to fuel reformers
- Important enhancements have been incorporated into CRUISE M to offer even better support for function development and calibration of engines and aftertreatment systems
Solution: Engine thermodynamics
Engine Parameterization Wizard
The setup of high-accuracy engine thermodynamics models suitable for HiL applications is key to compliance with the typically tight timelines of the engine development process. CRUISE M supports efficient model setup with a focused and automated workflow that considers engine specification sheets, measurement data and standard user-defined names for measurement positions.
The workflow comprises the following steps:
- General Specification: Input of a model type (mean value filling and emptying), fuel type (gasoline, diesel), charging type (naturally aspirated, turbocharged), etc. for defining the basic model layout
- Sensor Channel Names: Specification of channel names in a mapping table consisting of AVL standard names and user-defined names for all sensor and actuator channels as well as for measurement data
- Gas Path Component Geometries: Specification of all required component geometries based on a defined list of inputs. All parameters feature default values when entered onto the page. These values are calculated for the selected engine size and layout based on AVL’s experience. They are then refined by the user
- Measurement Data: Input of data from steady-state measurements (and/or reference simulation models) in a given pre-defined table featuring user-defined channel names
- Input Summary: Feedback on the provided input data and a preview of the next steps in the workflow
- Model Layout Generation: Creation of model topology and the population of all input parameters based on the previous inputs. Creation of additional infrastructure components to monitor the model
- Component Parameterization: Application of Parameterization Wizards for pressure drop and turbocharger (in case of turbocharged engines). Depending on the model layout, wizards are automatically opened, one after the other, for all gas path components. Parameter optimization is automatically performed when the wizard is opened
- Simulation Setup: Creation of simulation cases following the given input measurement data. After that the simulation model is ready to go. When running the simulations, the model results can be immediately compared to the measurement data given at the start, and the final model accuracy can be assessed
VTG and Waste Gate Add-Ons to TC Parameterization Wizard
The Turbocharger Parameterization Wizard is enhanced to handle VTG (Variable Turbine Geometry) and waste-gate turbines. With the new update, the wizard automatically determines the mapping from the ECU signal to VTG or waste-gate position.
For VTG turbines, the wizard parametrization workflow creates a map that compensates for deviations of the baseline turbine correction map. The map is automatically populated with data over the turbine pressure ratio and the ECU signal on the VTG position.
For waste-gate turbines, the wizard workflow automatically generates an external restriction component and steering map. The split between turbine and restriction bypass flow is set with the help of a map that spans the turbine pressure ratio and the ECU duty signal (ECU-driven waste gate) or the compressor pressure ratio (mechanical waste gate).
Figure 1.3.2 1 Automatized parameterization: VTG position mapping and waste gate turbine – flow coefficient map
Gas Path Wizards Upgraded with Automated Parametrization
The efficient parametrization of components to match reference data (i.e. measurements, 3D simulations) is a key and often a time-consuming activity during the setup of thermodynamic engine models. CRUISE M supports these efforts with dedicated parametrization wizards for adjusting the pressure drop and heat transfer behavior of selected gas path components.
In this new version, the existing manual parameter search is completed with automated parameter optimization. Regression methods are used to identify those model parameters that best match the model and the reference data on a global level. Longer running optimizations can be interrupted and the – up to this point – best fit is offered.
The automated parameter search raises the efficiency of a model parameterization by getting the right model parameters quickly and by immediately identifying reference data that is potentially wrong.
Solution: In-Cylinder Flow, Combustion and Emission
Engine Selections – Autodetection of Pre-Chambers
This release features extended functionality of the engine selection tool available in the FIRE Pre-processor to automatically recognize pre-chambers.
Thus, the following additional selections are found automatically:
- BND_Pre_Chamber for temperature BND condition definition
- BND_Spark_plug for spark plugs placed in the pre-chamber
- v_INI_Pre_Chamber to generate a cell selection on the final volume mesh for initialization and solver output
- v_INI_Connecting_Channel_[N] for the transfer channels.
The latter selection results in a cell selection INI_Connecting_Channel_[N] on the final volume mesh. The numbering starts with the channel oriented towards the intake ports. In addition to the selections all relevant feature edges are generated.
Multi-Cycle Run Mode
In FIRE a multi-cycle option has been implemented to simplify the setup of in-cylinder flow simulations that include several engine operating cycles. The new option allows cycle duration (angle or time, according to the run-mode) and the number of cycles to be specified, instead of the end angle or end time. All cycles are calculated in a single simulation run.
All time-dependent input data which are provided in a table in the Solver GUI – such as boundary condition tables – may be provided either for all cycles or for one cycle only. If they are given for one cycle only, the solver will automatically repeat the tables periodically.
Additionally, cycle-convergence criteria may be specified for single cycles so that they can be automatically repeated if the criterion is not met. A limit can be prescribed for the number of such repetitions. Each cycle-convergence criterion can be defined by a scalar value (e.g. an average pressure) and a maximal difference value. Cycles are repeated if the difference of these scalar values between the beginning and the end of the cycle is larger than the prescribed maximum.
Currently, convergence-values can only be defined by formula. In future, it is planned to provide useful values via a pull-down menu.
AVL CADIM Spark Ignition Model Coupled to FGM
The AVL CADIM Spark Ignition Model uses Lagrangian particles to describe the detailed behavior of the spark channel. This model has already been available for FIRE General Gas Phase Reactions (GGPR) as well as for ECFM / ECFM-3Z and the L-FTM and E-FTM combustion models. With this new release, the AVL CADIM has also been coupled to the FGM model, which features tabulated chemistry for premixed combustion.
FGM Coupled with FIRE Eulerian Flame Tracking
The FGM Model and the Eulerian Flame Tracking Model (E-FTM) have already been available in earlier versions of FIRE. The latest update couples the two models in order to improve the calculated flame propagation by applying the E-FTM and also to improve the prediction of heat release and chemical species concentrations by applying the FGM model.
ECFM-3Z Combustion Model in Conjunction with the PANS Method
In combustion modelling ECFM-3Z was adopted to take advantage of the Partially Averaged Navier-Stokes method, which provides two variables separately. These variables are unresolved and modelled resolved kinetic energy. A total kinetic energy for combustion modelling is now used as the sum of these two kinetic energies which make results less dependent on computational meshes.
See also: SAE Paper 2020-01-1107 "The Prospect and Benefits of Using the Partial-Averaged Navier-Stokes Method for Engine Flows"
Solution: Nozzle Flow, Cavitation and Erosion
New Multi-Component Flash Boiling Model
The aim of the multi-component flash-boiling model implemented in the FIRE Eulerian multiphase module is to simulate the flashing phenomenon in injection nozzles operated with blended fuels.
The model solves additional species transport equations by considering each phase as mixture of species. The pressure of each species is chosen as its partial pressures and the activity coefficient is calculated based on the UNIFAC method.
The attached figures show an impressive application of the new multi-component flash boiling model: Flashing of Hexane (50 % C6H14) and iso-Octane (50 % C8H18) in a Spray-G 8-hole injector nozzle from the Engine Combustion Network (ECN).
The first figure shows the molar density fraction (MDF) values of the C8H18 and C6H14 mixture tested in flash boiling conditions. In the second figure, string cavitation, which influences the flow field inside the nozzle holes, can be observed. These highly unsteady vapor structures appear upstream of the injector holes and inside the nozzle sac. In the third figure it can be seen how string vortices develop, which eventually could lead to string cavitation.
See also: SAE Paper 2020-01-0827 “Numerical Investigation and Experimental Comparison of ECN Spray G at Flash Boiling Conditions”
Solution: Thermal Load
Improved RPI Wall Boiling Model
In FIRE 2020 R1 the RPI wall boiling model reliability is enhanced. The energy balance outputs consider evaporation heat fluxes and vapor in/out mass flows, leading to exact energy conservation for converged equations. The robustness of the treatment of heat flux boundary conditions has been increased too. Model sub iterations to converge to the user defined heat flux have been reduced and additional information is provided to the user if unattainable conditions are imposed.
Wall Adhesion Model available in Eulerian Multiphase Module
The wall adhesion of liquids has an essential effect on wall heat transfer. The heat flux between the structure and an ambient gas like pure air is rather poor, whereas there is a significantly better heat exchange with fluids like water or oil. Consequently, there is a need for more accurate predication of the contact between liquid fluids, in particular thin liquid films, with a surrounding structure. The new wall adhesion functionality implemented in FIRE and FIRE M fully meets this demand by offering two suitable models:
- The In Height model: Here, the user has to input a minimum film height and the adhesion pressure and forces are calculated by keeping this minimum
- The Bond Number model: The Bond Number is the ratio between gravitational force and surface tension forces. So, the Bond Number is inversely proportional to film thickness and is also related to surface tension
Additionally, version 2020 R1 offers selection-based multiphase 2D outputs for the new wall adhesion force model. Also, output for average film height, average adhesion force and average adhesion pressure can be activated. Furthermore, provided 3D multiphase outputs are the film height and the adhesion pressure.
New Sophisticated Immersion Quenching Model
An advanced heat and mass transfer model, the "Advanced Quenching Model", has been implemented in FIRE™ M v2020 R1.
The new model enables the detailed and realistic simulation of liquid boiling during immersion quenching. The model can be used to predict all boiling regimes where four regimes – the nucleate, the developed nucleate, the transitional, and the film boiling regime – are distinguished. Model applications include, for example, the immersion quenching of hot metal parts covering different fluid-solid combinations. For steel quenching, the new model provides better results than the Standard Quenching model.
Liquid / vapor water interface when quenching a test specimen, calculated using the new “Advanced Quenching Model”
Solution: Model Based Calibration
The engineering-enhanced gasoline cylinder model of CRUISE M (MOBEO cylinder) has been enhanced in three areas:
- The use of variable valve lift (VVL) settings for the intake and exhaust is improved. A user needs to explicitly activate VVL in order to get data bus channels for setting the actual maximum valve lift
- Residual gas content can be calculated by considering an externally defined exhaust gas temperature. Doing so, the impact of cold start conditions on residual gas can be considered
- Additional combustion parameters (i.e. wall heat loss corrections, indicated power corrections) are offered for manual refinement
The engineering-enhanced diesel cylinder model of CRUISE M (MOBEO cylinder) is enhanced with a refined map of FMEP and by adding several out-sensor channels such as: exhaust mass flow, effective power, temperature at SOI and delivery ratio.
General Species Transport
The engineering-enhanced cylinder models of CRUISE M (MOBEO cylinder) have been extended to handle inputs from “General Species Transport”. General species transport enables the definition of a gas flow network composition in an arbitrary manner. The properties of transported species (i.e. O2, N2, CO, NO, etc.) are calculated using the NIST-JANAF approach together with an internal property database.
The application of general species transport together with the engineering-enhanced cylinders simplifies the setup of models, especially when species and particulate matter from the cylinders are transported to the exhaust gas aftertreatment system. Furthermore, it allows the consideration of phenomena related to species conversion in the gas path e.g. condensation.
The engineering-enhanced cylinder models of CRUISE M (MOBEO cylinder) feature a continuous change in a variety of aspects, such as details on gas exchange, combustion, and emissions. These changes are driven by HiL-based calibration tasks which require models that adequately capture the behavior of the latest combustion systems.
Version compatibility is introduced to better handle these model changes. This compatibility allows the selection of dedicated models from the current or the previous release. Beyond this additional choice, backward compatibility also enables models to be loaded from the previous version of CRUISE M (2019 R2) and to simulate in a manner that is compatible in respect to model parameters and results. This compatibility applies to the cylinder components and the corresponding parameterization wizards.
Solution: Exhaust Gas Aftertreatment
This version of CRUISE M provides comprehensive new functionalities to simulate all kinds of exhaust gas aftertreatment systems:
- Component Library: A model library of the most commonly used components in diesel and gasoline exhaust systems has been created. Dedicated models for DOC, DPF, cDPF, SCRF, SCR, ASC, NSC, pipes, Urea Doser, HC Doser, TWC, GPF and cGPF are available. The components can be freely assembled to form any kind of exhaust line topology
- Modelling depth: The models are transient 1D, 1D+1D or 2D depending on user’s selection. 1D models are typically applied for single layer wash coats, 1D+1D resolutions are applied for double-layer wash coats (i.e. TWC, ASC) and to describe the advection, diffusion, reaction problem in the wall of wall flow filters. Cylindrical, 2D, resolutions are applied to resolve the insulation of multi-layered walls
- Predefined Reaction Modelling: Predefined reaction models are offered. The models are taken from literature and they cover the typical reaction mechanisms given by the component list above. Furthermore, all reaction rates are fully accessible to the user. This access is offered via a dedicated User Coding Interface (UCI) which enables the modification of any part of the reaction rates
- Customized Reaction Modelling: UCI is a tailored interface to describe reaction rate and transfer (heat, mass) models via guided UI input masks and c-code. In a compile step, the complete reaction mechanism is translated into c-code and compiled. One key aspect of this process is that the reaction engineer can specify which reaction parameters shall be made public and thus made available in the CRUISE M model. With that the level of IP protection can be scaled
- Monitoring: The different components allow the definition of arbitrary measurement positions and the creation of corresponding sensor channels. These channels can be connected to any interfacing component including the Monitor element for online monitoring and actuation
- Engine and Exhaust Gas Aftertreatment: The aftertreatment models can be set up and operated stand-alone with given input traces representing engine outlet emissions or any kind of measurement data. The aftertreatment models can also be invoked into the thermodynamic network of an engine model. Here, dedicated co-simulation techniques are applied to combine the computational requirements of the thermodynamic and aftertreatment models
- Real-time Simulation: A real-time capable solver technology enables the simulation of complex exhaust gas lines including chemical kinetics in real-time. The computational effort of individual components is measured and it offers additional support to the user to identify more demanding components and, if needed, to apply model adaptations.
Figure 1.3.7 1 - Comprehensive CRUISE™ M Aftertreatment domain
CRUISE M Aftertreatment provides 100 % compatibility with existing BOOST Aftertreatment models and simulation results. Beyond that, the incorporation of BOOST Aftertreatment into CRUISE M combines the established aftertreatment modelling capabilities with latest user interface technology.