Krithika Narayanaswamy

Assistant Professor
Computational Chemistry & Combustion group,
Thermodynamics and Combustion Engineering Lab,
Department of Mechanical Engineering,
Indian Institute of Technology Madras,
Chennai - 600036
Email: krithika (at) iitm (dot) ac (dot) in

Research Experience
Faculty page
Scholar page
ResearchGate page
Scholar Profiles @ IITM

My research focuses on development of chemical kinetic models to describe oxidation of fuels. I am interested in predicting global combustion characteristics of conventional and alternative fuels and interpreting these observations based on insights gained from molecular level kinetic descriptions.

If you are interested in being a part of our team, peruse the list of available projects to identify your interests and write to me with a subject line that includes the title of the topic. Please note that I am *not* interested in hiring interns.

Abstracts for topics floated for MS admissions (Jan 2020)

** We have received a grant to carry out research in development of compact kinetic models. Open positions are likely to be adverstised for project assistants soon. Please watch out if you are interested. **

Research highlights


Simulation-driven surrogate formulation.


Time taken by our proposed fast algorithm to compute mulit-component diffusion velocities versus existing method.


Chemical model developed in our earlier works reorganized into the component library framework.


Ignition delays of jet fuel surrogate.


Extinction strain rates of DME-air mixtures measured in a counter-flow diffusion flame burner.


Ignition delay times of methylbutanoate.


Flame speeds of methylmethacrylate at various unburnt fuel-air temperatures


Our research focuses on development of chemical kinetic models to describe oxidation of fuels. I am interested in predicting global combustion characteristics of conventional and alternative fuels and interpreting these observations based on insights gained from molecular level kinetic descriptions.

In the recent times, my team has been studying kinetics of oxygenated fuels, which being renewable and resulting in reduced emissions, are immediately relevant to the country's energy needs. In addition, we are also recently explored ideas to generate compact reaction mechanisms and optimize kinetic schemes. A brief tour of our research activities is provided below. The funding agencies that support our research are gratefully acknowledged.

Glimpse of our Research work

Kinetics of small oxygenates

Fig: Flame speeds of methylmethacrylate at various unburnt fuel-air temperatures; symbols - experiments, lines - simulations
Fig: Extinction strain rates of DME-air mixtures are measured in a counter-flow diffusion flame burner. symbols - experiments (present work), lines - simulations
(Left) In order to analyze the high temperature gas phase oxidation of PMMA, and thereby predict its fire behaviour with less computational effort, a compact kinetic model for the oxidation of its primary decomposition product, MMA, is most essential. A compact reaction model, consisting of 88 species and 1092 reactions is developed to describe the oxidation of MMA and validated against fundamental experimental datasets obtained in premixed flames.

(Right) Blending small amounts of dimethylether (DME) with diesel is found to increase its resistance to extinction. In order to critically evaluate the overall combustion behaviour of DME via numerical simulations, a short kinetic mechanism consisting of 23 species and 88 elementary reactions is proposed. In addition to the extinction data, the short mechanism accurately reproduces the available experimental data in the literature.

Kinetics of biodiesel surrogates

Fig: Effect of pressure on ignition delays of methylbutanoate

Kinetics of conventional fuel surrogates

Fig: Ignition delays of jet fuel surrogate
(Left) A compact reaction scheme is derived for methylbutanoate, which is an important component of biodiesel surrogate. Starting with an existing detailed mechanism, revisions based on recent experimental measurements and theoretical rate constant calculations have been introduced. The revised kinetic model has been validated comprehensively against a wide range of experimental data along with ignition delays at intermediate temperatures measured in an RCM.

(Right) Simulations are performed by representing the real jet fuel as a surrogate containing 30.3% n-dodecane, 21.2% m-xylene, and 48.5% methylcyclohexane (mole %). The kinetics of this surrogate mixture are extracted from a multi-component reaction mechanism using the component library approach.

Surrogate formulation

Fig: Simulation-driven surrogate formulation

Mechanism Optimization

Fig: Framework for mechanism optimization
(Left) Our study suggests that in a computational modeling context, surrogate compositions should be determined not just based on global characteristics, as has been done so far, but also on the characteristics and predictive capabilities of the chemical kinetic mechanism used for the simulations.

(Right) A framework for reaction mechanism optimization is developed and the idea has been implemented as a black box tool. Scope of performing optimization in multiple stages utilizing the underlying kinetic knowledge is also explored.

Mechanism Reduction, data-driven ignition prediction, and more ...

Fig: Time taken by our proposed fast algorithm to compute diffusion velocities versus the total time taken by the iterative biconjugate gradient method.
Fig: Chemical model developed in our earlier works reorganized into the component library framework.

(Left) An accurate, fast, direct and robust algorithm to compute multi-component diffusion velocities has been proposed. To our knowledge, this is the first provably accurate algorithm scaling at a computational complexity of O(N) in finite precision.

(Right) A flexible and evolutive component library framework has been proposed to derive short chemical mechanisms with only the necessary kinetics for the desired surrogate mixture. A script to extract a chemical mechanism for a surrogate mixture, the kinetics of whose individual components are described in this parent chemical mechanism, is available.

View our funding sources

Detailed description of Research work

Kinetics of small oxygenates

  • Compact kinetic model for methyl methacrylate oxidation:

    In these days, synthetic and natural polymeric esters find applications in transport and construction sectors, where fire safety is an important concern. One polymer that is wide used is poly methyl methacrylate (PMMA), which almost completely undergoes thermal decomposition into methyl methacrylate (its monomer) CH2=C(CH3)-C(=O)-O-CH3 (MMA) at ~250-300oC. In order to analyze the high temperature gas phase oxidation of PMMA, and thereby predict its fire behaviour (such as burning rate, temperature of the material and heat fluxes) with less computational effort, a compact kinetic model for the oxidation of its primary decomposition product, MMA, is most essential.

    As a part of this work, a compact reaction model to describe the oxidation of MMA is developed and validated against fundamental experimental datasets obtained in premixed flames, starting with a detailed mechanism proposed by the Lawrence Livermore group.

    Evaluating the model against these data sets point to the need to revise the kinetic model, which is achieved by adopting rate constants of key reactions from recent literature for analogous molecules. The updated compact kinetic model is able to predict the major species in the flat flame as well the burning velocity of MMA satisfactorily. The final `short MMA mechanism' consists of 88 species and 1092 reactions.

    Contributor: Shanmugasundaram D.
    Publication(s): Dakshnamurthy et al., CST 2018

  • Extinction studies of oxygenated fuels:

    Addition to oxygenates to diesel fuel has been found to reduce soot emissions. Nonetheless, the resulting changes in the extinction and auto-ignition characteristics of the fuel mixture have not yet been fully understood. Considering two important oxygenates, dimethyl ether and methanol, we undertook a fundamental analysis, which reveals that blending small amounts of oxygenates with diesel increases its resistance to extinction, which can be of use in applications such as burners, furnaces and pre-vaporizers. We also find that the auto-ignition characteristics are not altered much due to blending. This study suggests that in the presence of oxygenate, a combustion system can be made more stable even at higher strain rates and at the same time operate with reduced emissions.



    In order to critically evaluate the overall combustion behaviour of DME via numerical simulations, an accurate as well as compact kinetic mechanism to describe its oxidation is most essential. A short kinetic mechanism consisting of 23 species and 88 elementary reactions is proposed to describe the oxidation of DME based on the detailed San Diego mechanism. In the model development, a "bottom-up approach" is adopted to incorporate just the necessary kinetics to describe the combustion characteristics with good fidelity. The short mechanism accurately reproduces the available experimental data for ignition delays, laminar flame speeds, and species profiles in flow reactors as well as jet-stirred reactors.

    To assess the validity of this reaction mechanism in non-premixed systems, extinction strain rates of DME-air mixtures, which are not available in the literature, are measured in a counter-flow diffusion flame burner. Details on the experimental set up can be found here.

    The 23-species reaction mechanism is able to predict the experimental data for extinction within the uncertainty limits. This mechanism is further reduced by introducing quasi-steady state assumptions for six intermediate species to finally obtain a 14-step global kinetic scheme. A code is developed in MATLAB to obtain these 14 global steps and their corresponding rate expressions in terms of the elementary reaction rates. The 14-step mechanism performs as good as the 23-species mechanism for all the experimental data sets tested for.

    Extending the bottom-up approach proposed to derive the DME mechanism to other fuels is one of our near term objectives.

    Contributor: Rohit Sanjay Khare
    Publication(s): Khare et al., CNF 2018



Kinetics of biodiesel surrogates

  • Kinetics of biodiesel surrogate components:

    One of our recent interests concerns with developing a kinetic scheme for a small ester, which is a potentially important candidate to represent the longer chain molecules in the real biodiesel fuel, namely methylbutanoate. A compact reaction scheme is derived for methylbutanoate from an existing detailed mechanism, revised based on recent experimental measurements and theoretical rate constant calculations. The revised kinetic model has been validated comprehensively against a wide range of experimental data along with ignition delays at intermediate temperatures measured in an RCM at PTB, Germany.


    Thereafter, a novel approach is used to propose surrogates to represent biodiesel fuel, consisting of methylbutanoate and n-dodecane, and assessed thoroughly (Fig. below). This work serves as our first step towards the development of a compact reaction scheme for a biodiesel surrogate which will be coupled with combustion studies to investigate the use of biodiesels and its blends with diesel in CI engines.


    Fig: Ignition delays in a shock tube and Species profiles in a jet stirred reactor; symbols: experiments for biodiesel, lines: Surrogate (presented at the 10th U.S. National Combustion Meeting, 2017)


    While methylbutanoate is a saturated ester, the real biodiesel also contains a significant degree of unsaturation (depending on the source of the oil), which results in considerable differences in combustion characteristics, such as in ignition delay times. The importance of including an unsaturated molecule in the biodiesel surrogate needs to be investigated.

    A kinetic scheme is being developed for methyl crotonoate, which is a simple unsaturated ester, that is representative of the long chain unsaturated esters in the real biodiesel. To validate the kinetic model, experimental data at intermediate temperatures for ignition delays measured in an RCM and species profiles in a JSR will be utilized, in addition to the experimental data available at high temperatures. Besides ascertaining the importance of including an unsaturated ester constituent in biodiesel surrogates, the reaction scheme proposed and the rate constants used therein will act as a basis to build reaction mechanisms for longer chain unsaturated esters, which are less understood presently.

    Contributors: Aditya D. Lele, Praise Noah Johnson
    Publication(s): Lele et al., CNF 2018

Kinetics of conventional transportation fuel surrogates

  • Kinetics of n-dodecane:

    Coming up soon ...

  • Kinetics of methylcyclohexane:

    Coming up soon ...

  • Kinetics of aromatics:

    Coming up soon ...

Surrogate formulation

  • Simulation driven surrogate formulation

    Coming up soon ...

Mechanism Optimization

  • Reaction mechanism optimization:

    A framework for reaction mechanism optimization has been developed and the idea has been implemented as a black box tool. This tool requires target information and uncertainty of rate constants as inputs. This modular Python script uses FlameMaster (Pitsch, H. FlameMaster v3.3.10) for kinetics simulations and NLopt for minimization of objective function.

    Numerical simulations are decoupled from the optimization by a response surface that approximates the target values as algebraic functions of normalized reaction rate constants. As opposed to the factorial design approach, a sensitivity based method adopted here (Davis et al. 2004). It generates a response surface with fewer computations, but with comparable accuracy. Finally, the minimization is carried out on an objective function indicative of the error between calculated and observed target values. To demonstrate the applicability of this tool, a reaction mechanism for n-heptane is considered.

    Scope of performing optimization in multiple stages utilizing the underlying kinetic knowledge is also explored. The optimization tool developed as a part of this work is made available on github in a black-box fashion for the use of the combustion community.
    Currently, ideas to improve existing methods in order to to achieve reduced computational cost as well ensure consistency in the optimized models are being worked upon.

    Contributors: Vaisakh Vasudevan, Krunal Rajeshkumar Panchal

Fast computation of multi-component diffusion velocities, Data-driven ignition predictions, and more ...

  • SMILES based interface to Thermodynamic Property Estimation Using THERM:

    Fuels, both renewable and non-renewable, are major sources of energy. Reaction schemes describing the oxidation of fuels are increasingly incorporated within computations where the focus is in capturing reactivity and the amounts of emissions. Thermodynamic properties of the species, which include the heat of formation, specific heat capacity and entropy, must be provided as part of such a kinetic scheme. These properties are used in determining the (a) equilibrium constant and thereby, the (b) reverse reaction rate constant using the principle of microscopic reversibility and (c) the heat release term of the energy conservation equation.

    To find the thermodynamic property values of a species, group additivity approach can be used, which is a simple and feasible method. It hinges on the idea that: for any species, the thermodynamic property values can be calculated as the sum of contributions to the property from all groups present in that species. THERM is a software that uses this method to find property values, taking in the group type, number of rotors and symmetry number of the species, as inputs. In this project, a program is developed that can calculate all the inputs needed for THERM, from either molfile or SMILES of a given species. This program is capable of processing any non-cyclic stable or single radical species, made up of carbon, hydrogen and oxygen atoms. The code is available open-source on github.

    Contributor: M. C. Sanjay

  • Component Library approach:

    Coming up soon ...

  • Fast computation of multi-component diffusion velocities:

    Coming up soon ...

  • Data driven prediction of ignition delays:

    Ignition delay has an important role in combustion. Ignition delay is affected by different parameters such as temperature, pressure, Equivalence ratio, etc. The objective is to predict ignition delays of fuels based on the molecular structure and the aforementioned parameters.

    Contributor: Pragnesh Rana

  • Methods for mechanism reduction:

    Reaction mechanisms that describe the oxidation of a fuel are huge with several thousands of individual steps. While using them in computations, compact representation of these schemes are essential. The state-of-the-art graph based reduction methods will be implemented to begin with. Improvements will be suggested to existing methods, their pros and cons will be evaulated, and the perfect order in which they must be applied to achieve best results will be investigated.

    Contributor: Page Kedar Govind

Funding sources

  • Core Research Grant, awarded by SERB, DST, Project No. -- : December 2019--2022
  • New Faculty Seed Grant, IIT Madras, Project No. MEE/15-16/845/NFSG: November 2017--2020
  • InnoINDIGO BiofCFD project awarded by DST: November 2017--2020
  • Indo-Russian project awarded by DST under Grant No. 16-49-02017: October 2016--2019
  • Exploratory Research Project, IIT Madras, May 2016--2017
  • New Faculty Initiation Grant, IIT Madras, Project No. MEE/15-16/845/NFIG: September 2015--2017

Collaborations


Krithika Narayanaswamy

Research Scholars

Krunal Rajeshkumar Panchal
PhD scholar, Research

Background:
B.E. in Mechanical Engineering from Government Engineering College, Modasa, Gujarat (2016)
M. Tech in Mechanical Engineering from Nirma University, Ahmedabad, Gujarat (2018)

Research interests: Optimization methods in combustion kinetics

Email: me18d005@smail.iitm.ac.in

Shanmugasundaram D
PhD scholar, Research

Background: B. Tech and M. Tech in Mechanical Engineering from Pondicherry Engineering College, Puduchery (2014,2016)

Research interests: Chemical kinetic modeling and reactive simulations of fuels relevant to fire research

Co-advisor: Prof. V. Raghavan, IIT Madras

Email: me16d025@smail.iitm.ac.in

Page Kedar Govind
MS scholar, Research

Background: B.E. in Mechanical Engineering from PVG's College of Engineering and Technology, Pune (2017)

Research interests: Developing open source tools to generate compact kinetic models

Email: me18s033@smail.iitm.ac.in

Pragnesh Rajubhai Rana
MS scholar, Research

Background: B.E. in Mechanical Engineering from Gujarat Technological University (2016)

Research interests: Data-driven prediction of ignition delays

Co-advisor: Dr. Sivaram Ambikasaran, IIT Madras

Email: pragneshrana244@gmail.com

Praise Noah Johnson
MS scholar, Research

Background: B.E in Mechanical Engineering from Maharaja Engineering College, Coimbatore (2017)

Research interests: Kinetics of unsaturated ester content in biodiesel fuel

Email: praisenoahjohnson@gmail.com



Alumni

Aditya Dilip Lele
MS 2016--18, Research

Thesis: Development of a skeletal chemical kinetic mechanism for a biodiesel surrogate and its application in CI engine simulations

Co-advisor: Dr. K. Anand, IIT Madras.

Currently: PhD scholar, Penn State.

Rohit Sanjay Khare
MS 2016--18, Research

Thesis: A comprehensively validated compact mechanism for dimethyl ether oxidation: an experimental and computational study

Co-advisor: Prof. V. Raghavan, IIT Madras.

Currently: PhD scholar, Combustion Technology department, Karlsruhe Institute of Technology

M. C. Sanjay
M. Tech 2015--17, Research

Project: SMILES based interface to Thermodynamic Property Estimation Using THERM

Currently: Scientific officer, Nuclear Power Corporation of India Limited (NPCIL)

Srinu Naik
M. Tech 2017--19, Research

Project: Development of the heat flux method for measuring burning velocities

Co-advisor: Dr. Varunkumar S.

Currently:

Vaisakh Vasudevan
M. Tech 2015--17, Research

Project: Development of a framework for optimization of reaction nmechanisms and its application to n-heptane

Currently: Postgraduate engineer, Lennox India Technology Centre, Chennai

Open source codes

Conference proceedings

2018 -- Present

  • Praise Noah Johnson, K. Narayanaswamy, "A combined kinetic model for high temperature oxidation of methyl crotonate, methyl butanoate, and n-dodecane, as components of biodiesel surrogate", 12th Asia-Pacific Conference on Combustion, 2019.

  • Shanmugasundaram D., Raghavan V., K. Narayanaswamy, "A compact kinetic model for high temperature oxidation of MMA using DRGEPSA", 12th Asia-Pacific Conference on Combustion, 2019.

  • A. D. Lele, K. Anand, K. Narayanaswamy, "Development of a chemical kinetic mechanism for biodiesel surrogate", 10th U.S. National Combustion Meeting, 2017.

  • R. Khare, V. Raghavan, K. Narayanaswamy, "A chemical kinetic modeling study of the effects of oxygenated species on soot emissions from diesel engines", Proceedings of the International Conference on Sustainable Energy and Environmental Challenges, 2017.

  • R. Khare, V. Raghavan, K. Narayanaswamy, "Study of auto-ignition and extinction characteristics of diesel blended with oxygenates in laminar opposed non-premixed flames", 10th U.S. National Combustion Meeting, 2017.

  • M. Hunyadi-Gall, G. Mairinger, R. Khare, K. Narayanaswamy, V. Raghavan, K. Seshadri, "The Influence of Stoichiometric Mixture Fraction on Extinction of Laminar, Nonpremixed DME Flames", 10th U.S. National Combustion Meeting, 2017.

Journal publications

*Submitted versions of manuscripts that are still under review and pre-prints of publications are provided to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by the authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's or publisher's copyright. These documents may not be reposted without the explicit permission of the copyright holder.

Book Chapter

  • A. D. Lele, K. Anand, and K. Narayanaswamy. "Surrogates for Biodiesel: Review and Challenges." Biofuels. Springer Singapore, 2017. 177-199.

For a Masters thesis*

  • Alternative Fuel Combustion

    • Renewable and alternative fuels for aviation have advantages of lesser emissions and clean burning ability. Identifying such a suitable surrogate mixture to represent such fuels and arriving at a kinetic scheme to predict the fuel's combustion behavior from a fundamental approach are the objectives of this work. This opens up avenues to study several `drop-in alternatives' to current day aviation fuels. As a part of this work, an approach automatically integrate kinetics of new fuels into existing reaction mechanisms will be explored.

  • Bottom-up approach to kinetic modeling of larger hydrocarbons

    • A bottom-up approach to compact model development has been put forth in one of our recent works to successfully develop a 23 species model for dimethylether. How can this be extended to higher hydrocarbons? A major challenge in adopting this methodology for higher hydrocarbons is the development of the fuel specific sub-mechanism. To begin with, the student will use the approach to develop the kinetics of a small molecule and thereafter investigate how to extend the approach.

  • Development of structure based compact reaction models

For a M.Tech/DD project*

  • Automatic reaction mechanism generation

    • The primary objective is to develop stand alone code to generate mechanisms for any hydrocarbon fuel. This has been done earlier -- but the challenge is building one on our own and how can we do better than the existing ones.

  • Thermodynamic property estimation for cyclic species

    • Estimating thermodynamic property for cyclic species by analyzing a database of known properties. One of our earlier stand-alone codes on estimating thermochemical properties will be extended to consider cyclic alkanes. Check this out under the publications (open source codes) tab.

  • Evaluating the effect of accuracy of kinetic models in realistic simulations

    • Reaction mechanisms that describe the oxidation of a fuel are huge with several thousands of individual steps. While using them in computations, compact representation of these schemes are essential. Do the reduced mechanism when run through complex calculations yield accurate results? To what degree? This will be prime focus of this project.

  • Simulation driven surrogate formulation

    • Can simulations be used to define surrogates (representative mixtures) for real fuels? We have made some progress on this along with a preliminary code development exercise. How can this be made more flexible and meaningful is the challenge.

  • Solver for homogeneous reactors and analysis tools

    • Homogenous reactor configurations are often used to validate kinetic models. The objective is to develop a stand alone in-house solver for simulating these configurations. As the second part of the work, the aim is to develop a GUI based interface to path flux analysis (through which routes does the reaction system evolve the most) commonly used in kinetic model development.

  • Kinetics of iso-alkanes

    • Can the effect of cycloakane representative in real fuel surrogates be imitated by iso-alkanes? The aim is to investigate the common branched alkanes, iso-octane and iso-cetane and explore their role as real fuel surrogates in place of cycloalkanes. As a first task, reaction scheme for iso-alkanes have to be arrived at and integrated with an existing master kinetic model. Getting grips with kinetic model development is an important part of this project.
*Please watch my home page to find out if I am interested in admitting students this cycle.

Alma mater

Vivekananda Vidyalaya, Perambur
DAV Girls, Gopalapuram
Indian Institute of Technology Madras
Stanford University
Cornell University

Useful web-based tools

Web plot digitizer
Coding ground
WriteLaTeX
Detexify
WordClouds
W3Schools
QR code generator

Articles/Opinions

For those starting their graduate studies
Some Important Things Most Students Never Ask About Graduate School
No power points for lectures
Top five communication mistakes

Unclassified

Notes
Musings
HackerRank
Calvin & Hobbes
PhD comics
Quotes

QR code for the group