DARS overview, IISc Bangalore 18/03/2014

Transcription

1 CH2O Temperatur e Air C2H4 Air DARS overview, IISc Bangalore 18/03/2014

2 Outline Introduction Modeling reactions in CFD CFD to DARS Introduction to DARS DARS capabilities and applications Overview of modules Homogenous reactors Reduction of mechanisms Flame modeling in DARS SRM models DARS 1D models Summary 2

3 Introduction CFD was developed to understand the fluid flow phenomena in various applications CFD provides flow, energy, concentration, and turbulence fields by solving the conservation equations on a discretized domain Although flow characteristics are of major interest in a CFD solution, if applications involve reactions species concentrations are also to be resolved Following slide shows modeling of reactions in a CFD framework 3

4 Modeling reactions in CFD Simulation of systems in which reactions are also involved is one of the major interests in combustion and chemical processing applications Currently, CFD can be used to model reactions, by incorporating detailed kinetic mechanisms or global reactions into it STAR-CCM+ has the capability to solve for reactions and combustion as well On the other hand, it might be time consuming to solve for species concentrations, which is elaborated in the next slide 4

5 Timescales In CFD simulations, flow is typically resolved whose timescales are of the order of milliseconds When reactions are involved, they require very low time step to capture the physics when compared to flow As shown in the picture below chemical time steps range from s to 10s depending upon the chemistry Hence, the overall timestep for the simulation has to be very low(of order of chemical reactions) which would slowdown the simulations greatly 5

6 CFD to DARS Apart from the timescales, if the no. of reactions/species increase in CFD simulations the no. of equations per cell to solve would increase So, if we have to incorporate a detailed mechanism consisting of many reactions into CFD, it would be computationally limiting Lower timescales and higher no. of reactions/species suggest us the idea of studying progress of reactions alone, decoupled from the flow This is where DARS comes into picture 6

7 Introduction to DARS DARS- Digital Analysis of Reaction Systems, a product of CD-adapco, which makes STAR-CCM+ Used to study progress of reactions without taking flow into consideration Standalone tool for simulating chemical reaction systems using detailed kinetic mechanisms It uses transient, 0D and 1D models to study formation of various products as described by the mechanism Can be used as a precursory code to CFD simulations to understand the chemistry alone of the system 7

8 DARS DARS Basic- Standalone tool DARS-CFD-coupled with STAR-CCM+ Emission Libraries-Libraries which can be used in STAR-CCM+ 8

9 DARS assets Ease of setup and tuning One-panel setup for complex fuel chemistries Full GUI support for setup Very few parameters to tune 2D, 3D engine mapping Easy to Use Accurate Combustion and emissions Detailed chemistry mechanisms Methods based on real physics A complete range of models IC Engines Emissions After-treatment Fuels Flames and burners Catalysts and Particulate Filters Fast Based on speedy stochastic reactor models Mechanisms optimized to affordable sizes Library based combustion and emissions Coverage

10 Industries and Applications Automotive SI DICI PPC HCCI Fuel industry Conventional fuels Natural gas Dual fuel Biofuels Synthetic fuels Environment Exhaust manifold Catalytic converters Diesel particulate filters Heavy, energy and chemical industries: Power generation Gas turbines Flames and burners

11 Map of Chemical Simulation Chemical Model Physical Model Detailed Chemistry Reactor Tools Simplified Chemistry CFD Experiments 11

12 DARS Basic Capabilities Essential for engine modeling with detailed kinetics DARS Basic Adds GT-Power or WAVE capabilites Burners Enable for ex. dual fuel applications Homogeneous Theoretical Reactors Engine Reactors SRM Stochastic Reactor Models PSR Piping Coolers DARS 1D Models Aftertreatment DARS-ESM Turbo charging Premixed Flames Counter Flow Flamelet Single Chemical Mechanisms Development Analysis Full powertrain simulation SI (Two- Zones) HCCI DARS- SRM-SI DARS- SRM-HCCI DARS- SRM-DICI Turbulence and gas inhomogeneities Catalytic Converter Diesel Particulate Filter (DPF) Transient! Library Reduction

13 DARS Basic Capabilities (Contd.,) On broader classification DARS contains Reactors(Homogenous and stochastic) used in the chemical industries Mechanism modules majorly used for analyzing and reducing the mechanism Flame models used for general combustion studies Stochastic reactor models(srm) models to account for inhomogeneity 1D models( stochastic) which are used for catalysis and after treatment industries We would briefly touch upon all the modules to give an overview of each one of them and their applicability 13

14 Homogeneous Reactors Reaction mixture is homogenous throughout, in terms of physical quantities such as, concentrations, Temperature etc,. (except plug flow reactor) These contain both open and closed reactors Conservation of mass, species, energy are solved Available modules are: Constant pressure and volume reactors Perfectly stirred and plug flow reactors Rapid compression model: An engine model Various modules are explained in detail in the following slides 14

15 Constant Volume Reactor Closed, stationary and homogeneous Volume is kept constant. Pressure is allowed to increase Used in Calorimetric studies to determine heat of formation of various fuels [ Mass] [ Species] [ Specific Internal energy] 15

16 Constant Pressure Reactor Gas allowed to expand freely in the reactor volume Closed, stationary, homogeneous system Used for ignition delay times, generating PVM table [ Mass] [ Species] [ Specific Enthalpy] 16

17 Perfectly Stirred Reactor Constant pressure, homogeneous flow system Steady-state gas phase combustion [ General Mass] [ General Species] [ General Energy] 17

18 Plug Flow Reactor 1-D model of a tubular reactor No axial mixing (diffusive transport = 0) Perfect radial mixing (diffusive transport = inf) Steady flow With or without surface reactions Heat transfer options: Adiabatic, isothermal, linear [ Mass] [ Momentum] [ Species] [ Energy] 18

19 Rapid Compression Machine Rapid compression machine(rcm) is an engine model, included in the homogenous reactor models of DARS Rapid compression machine: Closed system that represents the time between intake-valve closure and exhaust valve opening in the engine cycle. Equilibrium Model: Can compute adiabatic flame temperatures for gas-phase systems Based on minimization of Gibbs free energy for constant atomic mass fractions Can vary equivalence ratio, temperature, pressure for multiple runs. 19

20 Reduction of kinetic mechanisms Detailed kinetic mechanisms are required to accurately predict the behavior of reacting systems However, the use of these reaction mechanisms for modeling combustors in Computational Fluid Dynamics (CFD) is expensive The reaction mechanism describing oxidation of n-decane, consists of 209 species and 1673 reactions, most of them reversible (Dagaut et al. 2006) Reducing a reaction mechanism to a form having less number of reactions and species The reduced mechanisms can be plugged into CFD simulations 20

21 Degree of Reduction The detailed mechanism can be reduced to any degree of complexity Trade off between accuracy and computational time Following diagram shows levels of reduction and their applications 21

22 Analysis of detailed mechanisms Identification of less important species/reactions by analysis of detailed mechanism Approximations Quasi steady state approximation: If the species is shortlived, it is assumed that the net rate of production of the species is zero Partial equilibria assumption: Fast reactions are taken to be in equilibrium The most common analysis techniques are: Sensitivity analysis Reaction flow analysis Lifetime analysis 22

23 Sensitivity Analysis Sensitivity analysis involves investigation of the change in a quantity of interest due to small changes in the controlling parameters In analyzing kinetic mechanisms, the quantities of interest are generally concentrations of species The highly influential parameter would be the controlling parameter, temperature if reactions are temperature sensitive If rate constant is the controlling parameter, then sensitivity coefficient is defined as 23

24 Analysis of kinetic mechanisms Reaction flow analysis Reaction flow analysis determines the pathway of formation of products from reactants The detailed mechanism is given, into DARS, which would solve the mass fluxes from one species to another Lifetime analysis Lifetime analysis is used for finding species eligible for the Quasi Steady State Assumption (QSSA) It gives the time for which a species is alive Species with lower lifetime and concentration are identified 24

25 Illustration: Reaction flow analysis Arrows denote the reaction pathways Thickness of the lines denotes the mass flux The least significant (thick) pathways can be eliminated from the mechanism for reduction 25

26 Flames: Introduction Flame is a moving combustion zone A self-sustaining propagation of a localized combustion zone The two mechanisms for propagation are Thermal propagation: the mixture is heated by conduction to the point where the rate of reaction is sufficiently rapid to become self-propagating Diffusional propagation: diffusion of active species, such as atoms and radicals, from the reaction zone or the burned gas into the unreacted mixture causes reaction to occur It can vary from laminar to turbulent 26

27 Available flame models Characteristics of Flames in DARS Flames in DARS are one dimensional with a z-axis perpendicular to the flame front Flames are calculated at constant pressure Flames Premixed Counterflow Flamelet Burner stabilized Diffusion Single Freely propagating Back to Back Library Transient 27

28 Premixed Flames In a premixed laminar flame the fuel and oxidant mixture move in the z-direction with the unburned mixture at z - and the burnt mixture at z + The basic equations solved are: Mass conservation Species conservation Energy Conservation Premixed flames can be of two types, burner stabilized and freely propagating, which are discussed in the next few slides 28

29 Burner Stabilized Flames Burner stabilized flames most often used to study chemical kinetics Modeled as one-dimensional, steady-state flames Input: Conditions of the gas at the inlet, burner configuration (inlet gas velocity) 29

30 Freely Propagating Flames Point of reference is a fixed position on the flame Flame speed is thus the velocity of unburned gases moving towards the flame which allows the flame to stay in fixed Input: Conditions of the gas at the inlet Option to include thermal diffusion and radiation available Can calculate temperature profile or read temperature profile 30

31 Counter-flow Flames Counterflow flames are produced in the space between two opposed gas flows Can be either premixed or nonpremixed Non premixed are complicated than the premixed flames as, diffusion is the driving parameter Two types of counterflow flames DARS supports are: Diffusion: Fuel injected on one side, oxidizer on the other Back-to-back: mixture of fuel and oxidizer injected from both sides Gives two premixed flames 31

32 Flamelet In turbulent flows, when chemical time scale is very small compared to the convection/diffusion timescales combustion occurs in thin zones The flame in these thin zones is assumed to be laminar and are called flamelets Features of flamelets Turbulent flame considered to be an ensemble of laminar flamelets Facilitates decoupling of flow and chemistry Conservation equations for species and energy expressed in terms of mixture fraction and scalar dissipation rate 32

33 Flamelet: Models The different models supported by DARS are based on scalar dissipation rates They can either be steady or unsteady The models are : Single Flamelet: Steady state flamelet at user defined scalar dissipation rate Library Steady Flamelet: Runs for a range of scalar dissipation rates until extinction Transient Flamelet model: Solves unsteady equations 33

34 Introduction: Stochastic Reactor Models Drawbacks with homogeneous reactors for engines - Homogeneous composition and temperature» all gas ignites at once» overprediction max pressure, temperature, NOx - Impossibility to account for differences in gas - Turbulence modeling Inhomogeneities exist due to: - Charge stratification - Crevices - Heat transfer to the wall - Injection (DI engine) Stochastic Reactors 34

35 What are Stochastic Reactors? Stochastic Reactor Model: Quasi 0-D model. Homogeneity within the combustion chamber is replaced by statistical homogeneity, with physical quantities described by PDFs In-cylinder conditions such as species concentrations, density, pressure, temperature, cylinder volume, heat release, heat transfer as a function of time can be determined Autoiginition timing and combustion duration also determined 35

36 SRM: General features Gas state (species, enthalpy) is described by PDFs Discretization of gas into virtual particles (SRM-cells) Mixing model (deterministic or stochastic) to model turbulence Stochastic heat transfer Operator splitting technique for solving the system of differential equations An equivalent CFD calculation would take significantly larger resources For example, simulation performed on an In-cylinder engine CFD module consisting around 0.2 million cells takes around 0.5 day (12 hours) of analysis time on 8 processors using 3D CFD, while it takes less than an hour for an SRM run in 1D DARS Basic 36

37 SRM Modeling The mixture is described with Probability Density Functions (PDF): in-cylinder mass is divided into particles representing the discretized PDF Each particle represents a point in the phase space of species mass fraction, and of enthalpy Total heat exchange can also be determined and is defined by Woschni model 37

38 Homogenous vs. Stochastic A comparison between homogeneous and stochastic reactor model for SI engine shows stochastic reactors capture the phenomena better when compared to a homogenous reactor model 38

39 DARS 1D models Catalyst Module A 1D module in which the entire reactor is split into PSRs Stochastic Pipe Module DARS pipe model is a one-dimensional approach based on a series of partially stirred reactors Diesel Particulate Filter DARS DPF is a transient 1D model based on a series of perfectly stirred reactors Stochastic PaSR Model The stochastic PSR is modeled as a single adiabatic, perfectly stirred reactor at constant pressure and with the fixed volume of 1 dm3 39

40 Summary Necessity of DARS as a standalone tool, with an overview of its applicability and capabilities are covered Homogenous models with a detailed elucidation of all the ideal reactors are covered The need for mechanism reduction, along with the various analysis and reduction techniques are discussed A basic introduction of flame modeling and various modules of flames are touched upon The drawbacks of homogenous models, how SRM modules are used to rectify them are discussed 1D modules, which are predominantly used in catalytic reactors and particulate filters are briefed 40