Indian Institute of Science Bangalore
etd@IISc Electronic Theses and Dissertations at Indian Institute of ScienceNot a member yet
6204 research outputs found
Sort by
Experimental and Numerical Studies on High-pressure Trapped Vortex Combustion
No full textRecent climate change issues demand stringent emission standards for natural gas-fired power generation applications. The Trapped Vortex Combustor (TVC) is a relatively new concept where the flame is stabilized using a physical cavity with direct injection of reactants. This configuration offers superior performance and emission benefits over conventional gas turbine combustors. The present thesis involves design and development of a unique high-pressure combustor with optical access to enable visualization and detailed experiments. Experimental studies on this high-pressure combustor are complemented by Reynolds Averaged Navier-Stokes (RANS) numerical simulations. The first part of the work involves numerical analysis of the TVC to understand the flame stabilization methods in the cavity and main flow of the combustor. The cavity flame can be stabilized in the upper vortex, lower vortex, or between the vortices based on the jet momentum flux ratio. It is observed from the simulations that the flame stabilization and pollutant emissions are related to the jet momentum flux ratio and the height of the main duct. These parameters regulate the penetration of the cavity flow into the main flow and entrainment of the main flow into the cavity. Insight from these numerical simulations is used in the design of the experimental facility. The second part of the thesis involves experimental studies at atmospheric pressure which bring out the effect of main flow velocity, jet momentum flux ratio, and cavity/main flow equivalence ratios on the static stability of the flame, pattern factor, combustion efficiency, and pollutant emissions. Efficient operation of the TVC requires higher values of jet momentum flux ratio, but the blockage of the main flow is the drawback. The cavity equivalence ratio is observed to be limited by the lean and rich blow-out limits. It is further observed that the fueling of the main flow is mandatory to achieve low pollutant emissions by preventing the quenching of the cavity flow reactants. The optimum control parameters derived from this phase of the study are used to investigate the performance of the TVC at high pressures in the third part of the thesis. An increase in the combustor efficiency and NOx emission is observed with a noticeable increase in combustor noise. It is observed that control of cavity equivalence ratio and decrease in jet momentum flux ratio are crucial for reducing the combustor noise at high pressure. A detailed experimental investigation is performed at 5 bar (145 kW) to understand the effect of fuel stratification on the performance of the combustor. The variation of the stratification ratio by fixing the flow parameters and overall equivalence ratio results in two modes of TVC operation, lean premixed (LP) and rich-burn quick-mix lean-burn (RQL). An optimum operating condition is identified for each operating mode of the TVC by comparing pattern factor, combustion efficiency, NOx emission and combustor noise level. In the last part of the thesis, the dynamic stability of the combustion process is studied by analyzing the pressure time-series data and high-speed chemiluminescence images of the OH* radical. Large amplitude pressure oscillations (2% of mean combustor pressure) are observed when the cavity equivalence ratio transitions from the LP to the RQL regime. The OH* images exhibit oscillations near the cavity bottom for cavity equivalence ratio near 1.4 leading to limit cycle oscillations (LCO). A Spectral Proper Orthogonal Decomposition (SPOD) analysis is conducted to understand the dominant mechanism of the oscillations for optimum LP and RQL operations and the LCO condition. The shear layer oscillation on the top of the cavity is observed to be the dominant mechanism in all conditions except for the LCO. The high-pressure TVC is optimized to operate in both LP and RQL regimes. The LP operation offers ultra-low NOx emission (3 ppm), whereas the RQL operation leads to relatively lower combustor noise. Overall, the results can be used to design and choose optimum conditions for the TVC operating on natural gas fuel.Siemens Technology and Services Private Limite
New tools to constrain EFTs and CFTs
No full textIn this thesis, we develop new methods for the S matrix bootstrap in the context of 2-2 scattering amplitudes and four-point correlators in conformal field theories (CFTs).
For 2-2 scattering in quantum field theories, we consider manifestly three-channel crossing symmetric dispersion relation (CSDR), unlike the two-channel symmetric fixed-t dispersion relation. We give simple derivations of certain known positivity conditions for effective field theories, including the null constraints, which lead to two-sided bounds and derive a general set of new nonperturbative inequalities. We derive the analyticity domain of the CSDR analogously to the Lehmann-Martin ellipse. We present a fascinating correspondence between an area of mathematics called geometric function theory (GFT) and the scattering amplitudes focusing on the case with O(N) global symmetry. We obtain two-sided bounds on Wilson coefficients of physical Pion amplitudes via positivity and GFT.
Then we consider Bell correlations in light-by-light (LbyL) scattering at low energies. The known contributions in the Standard Model (SM) lead to Bell violation at all scattering angles except for a small transverse region, leading to a fine-tuning problem. Incorporating a light axion/axion-like particle (ALP) removes this problem and constrains the axion-coupling--axion-mass parameter space.
In the second part of the thesis, we consider CSDR for Mellin amplitudes of scalar four-point correlators in conformal field theories. This allows us to rigorously set up the nonperturbative Polyakov bootstrap for the conformal field theories in Mellin space, fixing the contact term ambiguities in previous work. Our framework allows us to connect with the conceptually rich picture of the Polyakov blocks being identified with Witten diagrams in anti-de Sitter space. We also give two-sided bounds for Wilson coefficients for effective field theories in anti-de Sitter space. The derivation of the Polyakov bootstrap allows rigorous epsilon expansion solely from bootstrap principles
Average Reward Actor-Critic with Deterministic Policy Search
No full textThe average reward criterion is relatively less studied as most existing works in the Reinforcement Learning literature consider the discounted reward criterion. There are few recent works that present on-policy average reward actor-critic algorithms, but average reward off-policy actor-critic is relatively less explored. In this work, we present both on-policy and off-policy deterministic policy gradient theorems for the average reward performance criterion. Using these theorems, we also present an Average Reward Off-Policy Deep Deterministic Policy Gradient (ARO-DDPG) Algorithm. We first show asymptotic convergence analysis using the ODE-based method. Subsequently, we provide a finite time analysis of the resulting stochastic approximation scheme with linear function approximator and obtain an -optimal stationary policy with a sample complexity of . We compare the average reward performance of our proposed ARO-DDPG algorithm and observe better empirical performance compared to state-of-the-art on-policy average reward actor-critic algorithms over MuJoCo-based environments
Elucidating the role of SNARE Syntaxin 1A in regulating lysosome and melanosome function
No full textSNAREs are transmembrane proteins that localize to specific membranes and facilitate the process of membrane/vesicle fusion. SNAREs mediate the fusion by binding with their respective cognate SNAREs. SNAREs mediate the fusion by binding with their respective cognate SNAREs. The typical structure of a SNARE contains a SNARE motif which consists of typically 60-70 amino acids and interacts with other cognate SNAREs; the N-terminal regulatory; and the transmembrane domain (TM) at the C-terminus. Exceptionally, only some SNAREs may have a TM domain or an unstructured N-terminal region (Fasshauer, 2003). Further, SNAREs are classified into Q- and R- SNAREs. The Q-SNARE family comprises several members of syntaxins (STX), including STX1A and STX1B, which were extensively described in the literature. STX1A or 1B SNAREs are believed to function only in neuronal cells. However, further studies have shown that they are present in other cells as well (Bennett, Calakos, & Scheller, 1992) (Bennett et al., 1993). Studies have also shown that STX1A localizes to endosomes (Brandhorst et al., 2006). Proteomics analysis of a Drosophila S2 cell line revealed that only one isoform, STX1 is present in the cell line, and it localizes at phagosomes (Stuart et al., 2007). However, the precise function of endosomal STX1 or at phagosomes is not yet well-understood (Dingjan et al., 2018). Although STX1A is known to function majorly at the plasma membrane (PM), its intracellular localization or function remains poorly understood. In this study, we have attempted to study the intracellular function of STX1A using HeLa cells/melanocytes as a model.
To elucidate the role of SNARE STX1A in regulating lysosome and melanosome function, we have divided the work into three objectives. We have studied the localization, its functional regulation in HeLa cells and targeting mechanisms of STX1A to the intracellular organelles (Objective 1), screening and identification of cognate SNAREs of STX1A (Objective 2), and its role in melanosome biogenesis (Objective 3). These studies demonstrate that STX1A regulates lysosome exocytosis and melanosome biogenesis in mammalian cells
Understanding the role of Cysteine desulfurases in Mycobacterium tuberculosis physiology and pathogenesis
No full textThe Mycobacterium tuberculosis (Mtb) persistence inside the human host relies on successful adaptation to host-induced stresses such as reactive oxygen species (ROS), reactive nitrogen species (RNS), iron starvation, low pH, and hypoxia. Iron-sulfur (Fe-S) clusters, which are the most archaic protein prosthetic groups, are sensitive targets for ROS and RNS. Iron starvation also adversely affects the biogenesis of Fe-S clusters. Mtb harbors more than 50 Fe-S cluster proteins associated with cellular metabolism, gene regulation, drug resistance, and persistence. Therefore, the knowledge of biogenesis and repair of Fe-S clusters is critical for understanding the basis of persistence for this human pathogen. Two multiprotein assembly systems, SUF and ISC, coordinate Fe-S clusters in most prokaryotes. Mtb encodes a complete SUF system, crucial for both in vitro and in vivo growth. However, the ISC system is extensively truncated to a single gene iscS coding for a cysteine desulfurase, whose function remains uncertain in Mtb.
Mtb IscS was shown to be involved in maintaining the activity of the Fe-S cluster containing enzymes succinate dehydrogenase and aconitase. Furthermore, Mtb lacking IscS exhibits sensitivity to oxidative stress. However, important questions remain unanswered: (1) What are the mechanisms by which IscS contributes to oxidative stress resistance? and (2) what is the consequence of an IscS loss in the backdrop of a fully functional SUF system on the persistence and virulence of Mtb. In the present study, we systemically characterize the contribution of IscS in maintaining redox balance, respiration, and metabolism of Mtb. We performed RNA-sequencing to assess if IscS loss affects the transcriptome of Mtb and examined the consequence of IscS loss on the pathogen’s fitness in response to vitro stresses, anti-TB drugs, and macrophages. Lastly, we investigated the survival and persistence of the IscS mutant in mice. A brief summary of the findings is given below:
Using a redox biosensor, XF Flux analyzer, and mass-spectrometry, we find that Mtb-iscS mutant cannot maintain redox balance, central carbon metabolism, and oxidative phosphorylation. Mtb-iscS mutant exhibited slow aerobic growth and reduced survival in response to oxidative stress, antimicrobials, and hypoxia. A transcriptome analysis of Mtb-iscS mutant reveals diminished expression of the DOS dormancy regulon and an overlap with the transcriptomes of the Fe-S cluster-containing transcription factors such as WhiB3, WhiB1, and SufR. In contrast to the SUF system, IscS deficiency did not reduce the survival of Mtb under nitrosative stress or in activated macrophages. Surprisingly, unlike wild-type Mtb, Mtb-iscS mutant could not enter a stable, persistent state of infection, continued to replicate in mice, and showed hypervirulence. We found that the suf operon was specifically overexpressed in Mtb-iscS mutant derived from murine lungs, and reducing SUF expression abrogated hypervirulence.
In summary, our observations show that IscS is required to maintain redox balance, bioenergetics, antibiotic susceptibility, ROS resistance, and macrophage survival. In the context of infection, deletion of IscS led to hypervirulence in mice, a phenotype linked to uncontrolled induction of the Suf system. Our data indicate that Mtb employs the IscS and Suf systems to attain an intermediate degree of virulence critical for persistence
A Learnable Distillation Approach For Model-agnostic Explainability With Multimodal Applications
No full textDeep neural networks are the most widely used examples of sophisticated mapping functions from feature space to class labels. In the recent years, several high impact decisions in domains such as finance, healthcare, law and autonomous driving, are made with deep models. In these tasks, the model decisions lack interpretability, and pose difficulties in making the models accountable. Hence, there is a strong demand for developing explainable approaches which can elicit how the deep neural architecture, despite the astounding performance improvements observed in all fields, including computer vision, natural language processing, generates the output decisions. The current frameworks for explainability of deep models are based on gradients (eg. GradCAM, guided-gradCAM, Integrated gradients etc) or based on locally linear assumptions (eg. LIME). Some of these approaches require the knowledge of the deep model architecture, which may be restrictive in many applications. Further, most of the prior works in the literature highlight the results on a set of small number of examples to illustrate the performance of these XAI methods, often lacking statistical evaluation.
This thesis proposes a new approach for explainability based on mask estimation approaches, called the Distillation Approach for Model-agnostic Explainability (DAME). The DAME is a saliency-based explainability model that is post-hoc, model-agnostic (applicable to any black box architecture), and requires only query access to black box.
The DAME is a student-teacher modeling approach, where the teacher model is the original model for which the explainability is sought, while the student model is the mask estimation model. The input sample is augmented with various data augmentation techniques to produce numerous samples in the immediate vicinity of the input.
Using these samples, the mask estimation model is learnt to generate the saliency map of the input sample for predicting the labels. A distillation loss is used to train the DAME model, and the student model tries to locally approximate the original model. Once the DAME model is trained, the DAME generates a region of the input (either in space or in time domain for images and audio samples, respectively) that best explains the model predictions.
We also propose an evaluation framework, for both image and audio tasks, where the XAI models are evaluated in a statistical framework on a set of held-out of examples with the Intersection-over-Union (IoU) metric. We have validated the DAME model for vision, audio and biomedical tasks. Firstly, we deploy the DAME for explaining a ResNet-50 classifier pre-trained on ImageNet dataset for the object recognition task. Secondly, we explain the predictions made by ResNet-50 classifier fine-tuned on Environmental Sound Classification (ESC-10) dataset for the audio event classification task. Finally, we validate the DAME model on the COVID-19 classification task using cough audio recordings. In these tasks, the DAME model is shown to outperform existing benchmarks for explainable modeling.
The thesis concludes with a discussion on the limitations of the DAME approach along with the potential future directions
Statistics of Conductance Fluctuations in 2D systems
No full textStudy of 2D materials is interesting due to the various phase transitions observed
in these materials. I shall present the results obtained in the study of phase transi tions in three different 2D material systems – monolayer graphene in integer quan tum Hall (IQH) regime, topological surface states (TSS) of 3D topological insulator
(TI) BiSbTeSe2 (BSTS), and a thin film of rare-earth nickelate with composition
(La0.2Pr0.2Nd0.2Sm0.2Eu0.2)NiO3 ([LPNSE]NO).
First, I shall talk about the conductance fluctuations in IQH transition in mono layer graphene. The results shall provide the first experimental evidence for the the oretical prediction that observables such as conductance fluctuations are multifractal
in an Anderson localization-delocalization transition [1].
Next, I shall discuss the measured universal conductance fluctuations (UCF) aris ing from the TSS of BSTS and show that the UCF is multifractal at low temperatures
thus proving the theoretical prediction that in the presence of electron-electron inter actions, the TSS of a 3D TI exists at Anderson criticality.
Finally, I shall discuss the phase transitions in [LPNSE]NO thin films probed
through resistance fluctuations spectroscopy measurements in the temperature range
80 K to 240 K. The results establish the co-existence of two phase transitions (metal insulator and magnetic Néel transitions) in this sample in the temperature window un der consideration and demonstrate the existence of an electronically phase-separated
regime [2]
Methanol Synthesis from Simulated Bio-syngas: Experimental and Modeling Studies
No full textMethanol is increasingly being considered as an alternative fuel due to its potential to reduce environmental pollution and its ability to be used directly in internal combustion engines without any engine modification. This study focuses on using biomass as a source for methanol synthesis. The use of biomass for methanol synthesis has two main advantages: it addresses the air pollution caused by biomass incineration, and it generates value-added chemicals from waste biomass. In this work, the suitability and profitability of producing methanol from biomass-derived syngas (bio-syngas) were investigated. Methanol is typically produced in industries by catalytic conversion of syngas obtained from the steam methane reforming process (SMR). However, syngas obtained from biomass gasification is a relatively newer route since this process generates syngas with lower H2 and higher CO2 content.
The hydrogen content in biomass is only 5-7%, in comparison to the carbon value of 48-52%, making the choice of gasifying agent and gasifier design crucial as it determines the H2 percentage and tar content in the syngas. Thus, the simulated syngas composition obtained from oxy-steam downdraft gasification was chosen as an input for studying its effect on the final methanol yield. Thermodynamic analysis of methanol generation from both bio-syngas and SMR showed that methanol yield is sensitive to temperature, pressure, and stoichiometric number (S) in both cases. The optimized methanol yield was achieved at 61.87% for SMR-based syngas and 39.54% for bio-syngas at 483 K and 5 MPa, respectively.
Before developing an Aspen Plus® model for biomass to methanol (B2M) conversion, a kinetic-based downdraft gasification model was developed in Aspen Plus® software. The model included tar kinetics and considered the downdraft gasification process in four separate zones with major reaction kinetics. The model was validated with literature data for different feedstocks and three different gasifying agents: air, oxygen, and oxy-steam. And as for the methanol generation system, limited literature was available for methanol production via bio-syngas. This thesis includes experiments with simulated bio-syngas composition for methanol production. Methanol synthesis experiments were performed in a high-pressure reactor using commercial Cu/ZnO/Al2O3 catalysts. The experimental values were used for the validation of the Aspen Plus® methanol kinetic model. The methanol yield values were optimized for the parameters like temperature, pressure and the S for the methanol reactor setup. These optimized values of the parameter were then considered for the B2M process optimization as well.
Surrogate models were created using multi-variable analysis to predict and optimize the methanol yield value for the entire B2M process. The model predicts that the maximum achievable methanol yield was 37.77% for bio-syngas. This can be achieved at a gasification condition where the Equivalence Ratio (ER), temperature, Steam to Biomass Ratio (SBR) values were 0.2, 1173 K, and 4, respectively.
Finally, a techno-economic analysis of B2M process was done to assess the feasibility of this new alternative route in comparison to the existing natural gas reforming process for methanol synthesis. The techno-economic studies show that biomass to methanol technology can be developed in a country like India where surplus biomass is available. This process becomes economically viable at a methanol selling price of Rs 28 per litre or 0.3 Euro per litre and above a plant capacity of 2000 Tonnes per day of methanol production
Activity induced phase separation and the emergence of liquid crystal phases in chiral and achiral systems, and development of an efficient method to compute the entropy of various liquid crystal phases
No full textThe phase behaviour of shape-anisotropic particles is an emerging field of research that
gives rise to various liquid crystal phases. In this thesis, we explore various equilibrium and
non-equilibrium properties of shape-anisotropic particles by modelling them as soft repulsive
spherocylinders (SRSs) and soft helical rods.
In the first part, we introduce the two-temperature model to study the phase behaviour of
scalar active SRS and soft helical rods. Most realisations of activity are vectorial in nature due
to the force of self-propulsion. Recent studies have shown that many physical and biological
processes, like phase separation in colloidal systems, chromatin organisation in the nucleus,
are operated by the unequal sharing of energy by the constituents of the system. Such systems
are classified as scalar active systems. In the simplest case, these systems can be modelled by
connecting half the particles with a thermostat of higher temperature (labelled ‘hot’/‘active’)
while maintaining temperature of the rest constant (labelled ‘cold’/‘passive’) at a lower value.
The relative temperature difference between the two constituents of the system is a measure
of activity. This model is known as two-temperature model that has been found to capture
many essential properties of scalar activity. Starting from a homogeneous isotropic phase at
a definite temperature, we show that this model leads to phase separation into hot and cold
regions and induces liquid-crystal ordering of the cold particles while hot particles remain in
the isotropic phase. In particular, we find that activity drives the cold particles through a phase
transition to a more ordered state and the hot particles to a state of less order compared to
the initial equilibrium state. Hence, the phase boundary of the isotropic-nematic transition
shifts towards lower densities for cold particles and higher densities for the hot particles with
respect to its equilibrium location. Remarkably, we find liquid crystalline phases for the aspect
ratios [length(L)/diameter(D)] as low as L/D = 2, 3 which do not satisfy the minimum
shape-anisotropy criteria that Onsager’s theory demands in equilibrium. Similar model we
have employed in a system of soft helical particles of various intrinsic chiralities and found different liquid crystal ordering in these cases as well. The following nonequilibrium features
emerge from our study: an enhancement of the temperature of half of the particles gives rise
to LC ordering in the remaining half of the particles at any density. The hot and cold domains
should not be viewed as bulk equilibrium phases with non-equilibrium behaviour only at the interfaces.
By calculating the stress anisotropy and heat current, we find that the non-equilibrium
behaviour is not restricted to the hot-cold interfaces but pervades the system as a whole, driving
various ordering transitions in the cold zone. Thus, our study unravels various aspects of
non-equilibrium scalar active rods in the framework of the two-temperature model.
In the second part, we discuss the Two-phase thermodynamic (2PT) model for computing
entropy, free energy, and other thermodynamic properties of various liquid crystal phases in
equilibrium. In the 2PT method, the density of state (DoS) of the LC phases is decomposed
into vibrational (solid) and diffusive (gas) components. The thermodynamic quantities are
then calculated using harmonic oscillator approximations for the solid component, hard sphere
approximations for the gas component, and the rigid rotor approximation for the rotational
mode. In the 2PT method, the entropy of a definite state point is calculated from a single
MD trajectory, which makes it advantageous for systems for which the analytical form of the
equation of state is unknown (such as SRS). Our method can be used to calculate entropy and
other thermodynamic quantities of different liquid crystal phases formed by the SRS system.Inspire fellowshi
Hydrodynamic stability of shear flows : Jet in crossflow, particle laden planar and swirling jets
No full textJet in crossflow is a flow scenario that involves momentum exchange of a round jet injected into a crossflow. This flow configuration has industrial importance because of its application in gas turbine combustors which requires mixing of hot combustion products with cold air bled from the compressor to achieve the required temperature profile at the exit of the combustion chamber. Baseflow model proposed by Kelly and Alves (2008) is considered to study the near field of the jet with a thin shear layer and a weak crossflow. Linear stability calculations for the direct injection jet (leading order free jet) shows that the axi-symmetric mode m=0 is more unstable than m=1. The effect of addition of a weak crossflow is studied. The crossflow to jet velocity ratio is small, the eigenspectra for weak crossflow cases are different in that, at small wavenumbers the zero crossflow eigenmodes has a sightly higher growth rate compared to the weak crossflow cases. We find that addition of weak crossflow results in the most amplified mode that is predominantly axi-symmetric (m=0) while higher helical modes both even and odd are present but with smaller magnitude. As the velocity ratio is increased, the contribution of axi-symmetric modes to the most unstable mode decreases and helical disturbances (both lower and higher helical modes) shows contribution which possibly explains why the smoke visualizations of Fric and Roshko showed the near field of the jet (in presence of crossflow) are dominated by distorted vortex rings.
Spatiotemporal analysis of a constant density zero crossflow jet at Re=800 is found to be convectively unstable.
Particle laden flows are commonly seen in many industrial applications such as fluidized beds in process industry, air laden with abrasive particles in abrasive machining and particle laden plumes in chemical industries. In the present work, we perform local analysis of a particle laden planar jet in the dilute suspension regime. Unladen parallel planar jets have been extensive studied using normal modes and is shown to have two unstable modes namely sinuous and varicose modes. Sinuous modes are found to be more unstable compared to the varicose modes. In the present study, we investigate the effect of particles on the stability of planar jets. Addition of particles at low Stokes numbers (St) (fine particles) results in higher growth rates than that of the unladen jet. In the intermediate Stokes number regime, addition of particles have a stabilizing effect on both the sinuous and the varicose modes. Interestingly for St~10, the unstable varicose mode is completely damped. Increasing the Stokes number by increasing the particle size, both sinuous and varicose modes show increasing growth rates, while increasing density ratio has a stabilizing effect on the flow. For non uniform particle loading, additional modes apart from the sinuous and varicose modes are observed. These modes suggests occurrence of compositional instability with an increased particle accumulation in the shear layer that is an order of magnitude higher compared to that of the sinuous and varicose modes.
Linear stability of annular swirling jet laden with particles in a swirl-stabilized combustor is considered. Eigenspectra of the particle laden jet for low, intermediate and large Stokes numbers with uniform particle concentration shows three unstable modes (centre, sinuous and varicose modes) and a new set of neutrally stable modes which are absent in the unladen jet eigenspectrum. As Stokes number is increased, the growth rates of the centre and shear layer modes reduces compared to that of the unladen swirling jet. Magnitude of the velocity eigenmodes peaks in the vortex core and reduces radially outward. Variation in the particle concentration occurs mostly in the vortex core and almost none in the shear layer indicating that particles migrate from the centreline to the vortex core. But as the swirl number is increased further, it is seen that although the fluid velocities peak in the vortex core, the particle velocities peak in the shear layer. The increase in swirl number results in the particle concentration magnitude peaking in the shear layer and not in the vortex core. The effect of increase in backflow parameter is to increase the growth rate of the centre mode. Sinuous and varicose modes also have reduced growth rates at intermediate Stokes numbers. As the backflow parameter is increased, the growth rate of sinuous and varicose modes increases. The increase in baseflow azimuthal vorticity leads to increase in the net generation rate of vorticity fluctuations from rearrangement of base flow vorticity by velocity disturbances, unsteady vortex stretching due to base flow velocity gradients.
We look at the effect of non uniformity of the particle concentration in the base flow. It is found that when the peak of the concentration profile lies within the vortex core, the centre modes are stable. Sinuous and varicose modes remain unstable for regardless of where the peak is located, although the growth rates are smaller than the unladen flow. Ring modes remain stable as in the case of unladen flow and uniform concentration. Recent studies on the unladen flow by Manoharan and weakly non linear calculations by Manoharan et al on a turbulent single phase annular swirling jet indicate that m=1 mode causes precessing vortex core (PVC) oscillations. The radial, azimuthal velocities and particle concentration fields for the present calculation at St=1, suggests that the effect of addition of particles is to reduce the growth rate of the fluctuations, thus reducing the PVC oscillations