Department of Mechanical Engineering

IIT Madras is a public technical and research university located in Chennai, Tamil Nadu. Founded in 1959, it is recognised as an Institute of National Importance.

IIT Madras has been ranked as the top engineering institute in India for four years in a row by the National Institutional Ranking Framework of the MHRD

It currently offers undergraduate, postgraduate and research degrees across 16 disciplines in Engineering, Sciences, Humanities and Management. About 596 faculty belonging to science and engineering departments and centres of the Institute are engaged in teaching, research and industrial consultancy.

IIT Madras

Option discovery and skill acquisition frameworks are integral to the functioning of a hierarchically organized Reinforcement learning agent. However, such techniques often yield a large number of options or skills, which can be represented succinctly by filtering out any redundant information. Such a reduction can decrease the required computation while also improving the performance on a target task. To compress an array of option policies, we attempt to find a policy basis that accurately captures the set of all options. In this work, we propose Option Encoder, an auto-encoder based framework with intelligently constrained weights, that helps discover a collection of basis policies. The policy basis can be used as a proxy for the original set of skills in a suitable hierarchically organized framework. We demonstrate the efficacy of our method on a collection of grid-worlds evaluating the obtained policy basis on downstream tasks and demonstrate qualitative results on the Deepmind-lab task. © 2021, Springer Nature Switzerland AG.

Department of Computer Science and Engineering

World University root node

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Option Encoder: A Framework for Discovering a Policy Basis in Reinforcement Learning

In this study, we computationally corroborate the flow of rock glaciers against borehole measurements, within the context of a model previously developed (2020). The model is, here, tested against the simulation of the sliding motion of the Murtel-Corvatsch alpine glacier, which is characterized in detail in the literature with internal structure description and borehole deformations measurement. The capability of the model to take into account the composition of the rock glacier, as a mixture of ice and rock and sand grains with the local impact of pressure and heat transfer, results in the accurate detection of the internal sliding. With careful calibration of the model parameters, the computed numerical solution of the model reports a relative error of 1.8% and of 0.3% in the reproduction of the measured shear zone velocity and of the ratio of measured shear zone deformation over top surface deformation, respectively. Furthermore a deeper understanding of the role of the model parameters involved in the simulation of such a process is also gained and we discuss the same in detail. © 2021 The Author(s)

Fulltext

International Journal of Non-Linear Mechanics

Computational corroboration of the flow of rock glaciers against borehole measurements

The most critical component in an aerospace tank pressurisation system is a pneumatic pressure regulator with helium gas as the medium. The current paper focused on dynamic analysis of high pressure, spring-loaded, cryogenic medium pressure regulator. In order to evaluate its dynamic behaviour, a nonlinear model has been developed based on the fundamental equations by considering various forces–pressure forces, spring forces, bellow forces and Coulomb friction forces. The simulation is carried out for nominal operating conditions of the pressure regulator using commercial available AMESIM simulation tool. The effect of damping orifice size and downstream ullage volume are studied and obtained results are presented in the paper. Also experimental tests have been carried out for the same parameters to validate the simulation model using a proto model developed in-house with emphasis on machining of various precise parts of the high pressure regulator. These parts are manufactured in-house using high precision surface grinding, jig boring, electro discharge machining process, etc. The machined parts are realised to achieve precise sliding diametrical clearances between 32 microns and 50 microns, eccentricity within 20 microns and average surface roughness of 0.4 microns. The dimensional parameters of the machined parts are evaluated using the co-ordinate measuring machine (CMM), Talyrond and Talysurf equipments. The simulation results and experimental test results are found to be in good agreement. The developed model serves as a tool for optimisation of the regulator performance parameters. © 2020 Informa UK Limited, trading as Taylor & Francis Group.

Advances in Materials and Processing Technologies

Simulation of cryogenic high pressure regulator dynamics

Microstructural evolution during the strain-induced phase transformation of austenite in an Austempered ductile iron (ADI) under various thermomechanical processing conditions is studied in the present study. An alloyed ductile iron is taken as the base material, and thermomechanical treatment is carried out on a Gleeble 3800 thermomechanical simulator coupled with dilatometry. The effect of deformation on the austempering process has been studied by microstructure characterization using optical microscopy (OM), scanning electron microscopy (SEM), and X-ray diffraction (XRD) techniques. The variations in retained austenite volume fraction and its carbon content with respect to different austempering times are analyzed to study the effect of strain-induced transformation of austenite. It has been observed that the thermomechanical treatment significantly influences the phase transformation kinetics during the austempering process. The thermomechanical treatment produced a martensite free ausferritic microstructure for all austempering times with a high volume fraction of carbon enriched retained austenite as compared to the conventional heat treatment. © 2021 Trans Tech Publications Ltd, Switzerland.

Department of Metallurgical and Materials Engineering

Materials Science Forum

Microstructure evolution during strain-induced transformation of austenite in an austempered ductile iron (ADI)

The primary objective of this project is to increase the fatigue life of aluminium wheels and to prevent the fatigue failure of the wheel above the bolt hole region. In the testing and validation of wheels, the cornering fatigue test (CFT) plays an important role in deciding the stiffness of the wheel. In CFT testing, the wheel will be fixed in the bottom flange, and the bending moment will be applied, according to the load rating of the wheel multiplied by its corresponding test factor. CFT testing results of aluminium wheels showed crack above the bolt hole region in most of the cases. To prevent this mode of failure, and also to improve the fatigue life of the wheel further, it has been found that increasing the localized hardness above the bolt hole region will improve the fatigue life in CFT. Laser peening is a surface hardening technique by which the surface hardness of the material will be enhanced, along with which hardness and compressive residual stresses will be imparted on the surface in which the treatment is done. The depth of the high magnitude compressive stresses will greatly improve the material’s resistance to fatigue failure. In aluminium wheels, laser peening was done circumferentially around the bolt hole region to induce localized compressive stress and increase the hardness in that area. The wheel in which laser peening was done was tested on CFT and showed a significant improvement in life. © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd 2021.

Department Of Chemical Engineering

Lecture Notes in Mechanical Engineering

An innovative method of preventing bolt hole crack in aluminium wheels by using laser peening technique

In the present investigation, water based multi-walled carbon nanotube (MWCNT) dispersed nanofluid was produced and used as a cutting fluid in small quantity cooling lubrication (SQCL) assisted grinding of hardened AISI52100 steel. The work material was ground by an alumina wheel. The SQCL technology utilizes atomized aerosol of liquid grinding fluid. Use of MWCNT nanofluid aerosol in place of soluble oil was found to be substantially beneficial in reduction of specific energy and development of compressive residual stress on ground surface. The steady and lower value of the force ratio (tangential to normal grinding force, Ft/Fn) was suggestive of the superior lubrication and grit sharpness retention ability of the aerosol of MWCNT nanofluid. Higher compressive residual stress on the ground surface under MWCNT nanofluid environment suggests lesser grinding zone temperature, which can be attributed to its superior wettability, lubricity and heat dissipation ability of the MWCNT nanofluid. The consumption rate of grinding fluid in SQCL was varied. It could be reduced to 200 ml/h for MWCNT nanofluid to obtain an almost similar or better effect to that of 500 ml/h for soluble oil. Lesser surface tension of MWCNT nanofluid, compared to its counterpart, seemed to facilitate wetting, thus improving lubrication and heat transfer capability further. © 2021 Taylor & Francis Group, LLC.

Machining Science and Technology

On grinding force ratio, specific energy, G-ratio and residual stress in SQCL assisted grinding using aerosol of MWCNT nanofluid

In the aero engine rotor-bearing systems, a design modification to move the critical speed away from resonance zone and to improve the unbalance response is a practical problem for designers. The objective functions are selected as amplitude of vibration due to unbalance and second critical speed. The two selected objectives are found to be conflicting, and hence, multi-objective optimisation is used to solve this problem through numerical formulation. Diameters of the stepped shaft and stiffness of the two bearings are selected as design variables. Weight is set as a constraint, and additional constraints based on the initial design are introduced. Objectives resulting from multi-objective optimisation and corresponding design variables are clustered separately using k-means clustering algorithm. The novelty of the work is in empowering the designer to recognise the hugely diverse designs resulting in identical solutions and achieving manufacturing sovereignty in choice of design through clustering in objective and design space and their mapping. © 2021, Springer Nature Singapore Pte Ltd.

Clustering in Pareto Front: Application on an Aero Engine Rotor-Bearing System for Improved Design

In any proton-exchange membrane fuel cell, the bipolar plates grab attention because of the high production cost and heavyweight. Hence, the open-cell nickel foams are considered as an alternative to these grooved plates. The reliability of the metallic foams used as flow distributors and gas diffusion layers plays a vital role in the fuel cell's overall performance. Fretting wear damage of the metal foam at strut/strut interface and strut/supporting frame interface due to the vehicular vibrations and pressurized fuel flow is expected to affect the performance and is investigated. This paper discusses the fretting wear behaviour of nickel foam strut that constitutes open-cell nickel foam based on laboratory tests. The experiments are performed by using two different contact configurations: steel ball on wrought nickel flat and nickel strut on wrought nickel flat. The test results reveal the fretting friction coefficient's dependence on the normal load, surface roughness and contact geometry. Although the nickel strut versus nickel flat fretting wear tests showed a low coefficient of friction, severe damages were observed on the nickel struts’ worn surfaces compared to the wrought nickel flat. Scanning electron microscope images of worn scars on nickel foam struts indicate the presence of shallow grooves, craters, micro-cracks and delamination craters at higher loads. © IMechE 2021.

Department of Electrical Engineering

Proceedings of the Institution of Mechanical Engineers, Part J: Journal of Engineering Tribology

Fretting wear behaviour of nickel foam struts used in fuel cell applications

Propane, n-butane, and their mixtures are commonly used as fuels for various applications. Reliable numerical simulations of flames involving these fuels are most essential to study their efficient use in domestic and industrial burners. Such simulations require a suitable reaction mechanism containing the necessary kinetics to predict key combustion characteristics in flames accurately. However, the use of detailed kinetic mechanisms for these computations is formidable in multidimensional simulations, owing to their large sizes. To address this, in this work, a compact mechanism (45 species, 392 reactions) for propane, n-butane, and their mixtures, suitable for multidimensional simulations is derived and comprehensively validated against available experimental data in flames. The main oxidation pathways retained in this compact model are showcased. One-dimensional (1D) computations of premixed and nonpremixed flames are carried out using FlameMaster and Chemkin-Pro. Two-dimensional (2D) calculations are performed within ANSYS-FLUENT. Here, multicomponent diffusion, thermal diffusion, and radiation submodels are used. Laminar flame speeds, extinction strain rates, and amounts of major species in premixed flat flames are validated using 1D computations. Temperature and major species in propane and n-butane jet diffusion flames are verified in 2D simulations. Further, 2D computations using the compact mechanism also predict the laminar burning velocity of premixed flames of n-butane and propane mixtures established over a Bunsen burner. The ability of the compact mechanism to predict desired targets in premixed and nonpremixed flames is analyzed from a kinetic basis. Validation using 2D simulations demonstrates the usability of the compact mechanism in multidimensional computations. © 2021 Wiley Periodicals LLC

International Journal of Chemical Kinetics

Reduced mechanism for flames of propane, n-butane, and their mixtures for application to burners: Development and validation

In the background of designing a polymer nanocomposite based flexible strain sensors, homogenized dispersion of nanofillers remain challenging due to the trade-off between upholding of inherent characteristics of nanofillers and stable dispersion of nanofillers in the polymer matrix. This study provides an approach of balancing the stable dispersion of nanofillers and maintain the inherent characteristics of nanofibers by the method of polymer functionalization. Herein, two oppositely charged polymers polystyrene sulfonate sodium salt (PSS) and poly-diallyl-dimethyl-ammonium chloride (PDDA) are used separately to functionalize the multiwalled carbon nanotubes (CNT) and it dispersed in poly (vinylidene difluoride) (PVDF) polymer matrix to get flexible nanocomposite strain sensor to investigate both electrical conductivity and the gauge factor. PSS functionalized CNT (PSS-CNT) shows better dispersion than the PDDA-CNT in PVDF matrix. Besides, the electrical conductivity of PVDF/PSS-CNT (0.44 ± 0.02 Scm−1) is about 28 times higher than the PVDF/PDDA-CNT, and the gauge factor of PVDF/PSS-CNT (9.8 ± 0.3) is 2-fold higher than the PVDF/PDDA-CNT. The stable dispersion of PSS-CNT is due to the repulsive nature existing between the individual CNTs caused by the attachment of charged polymer functional groups as same as intrinsic functional groups of the pristine CNT. This work signifies the role of synergy between the charged particles in the functional groups and charged nature of the filler influencing the performance of flexible strain sensors. The detailed investigation using various techniques reveals significance of selection of suitable polymers for functionalization by showing its influence in the context of designing the high-performance flexible strain sensor. © 2021