Initial funds released: INR 101.3 million
In recent decades, there has been a significant shift in the approach towards developing novel materials technologies that have the potential to revolutionize the way components are created. This fundamental point of departure is based on the replacement of costly “trial and error” methods of finding new materials - mostly by modifying the chemical composition (by annealing, for instance), by a modern mechanics-centric, geometry-oriented, thermodynamically consistent and computational/data-driven approach to systematically seek new material configurations with specific desirable characteristics or properties. In order to assume a leading role in this paradigmatic shift, ISRO has recently funded the Centre of Excellence in Advanced Mechanics of Materials at the Indian Institute of Science to explore such new possibilities of design and manufacturing for space applications. The purpose of the CoE is to confront this grand design challenge in space applications (and beyond) and take sustained leadership in this area.
The mission of the CoE is to put together a research program:
Major Research Aims – The Big Picture
Our aim is to identify and apply these principles and package them into software toolboxes for scientists and practitioners.
Title: “An Integrated Computational and Experimental Approach to Structural Design for Ballistic Impacts and Blasts; grant sponsor: DRDO, New Delhi;
Released Funds: Rs. 8.14 Crores. The collaborating Institutes are: IISc (the nodal institute), IIT Kharagpur and NIT Goa.
In recent decades there has been a significant shift in the approach towards developing novel materials technologies that have the potential to revolutionize the way components are created. This shift is based on the replacement of “trial and error” methods of finding new materials towards a computational data drive approach to seek new materials with specific characteristics or properties. The US government has given a huge boost to these efforts through the “materials Genome” initiative. In tune with this trend, the aim of this inter-institutional, collaborative, research initiative is to initiate a research exercise on novel and rational design methodologies of lightweight metallic/ceramic/cellular composite armours for use in military vehicles other structures of interest to the Indian defense establishment under ballistic/hyper-ballistic impacts. By extensively incorporating experimental observations within modern mathematical models, which incorporate among others nonlinear material constitutions of the impacting and the impacted bodies as well as appropriate stabilization or regularization strategies to deal with extreme strain localization and/or material fracture, the major objective is to update or modify and hence improve these models so as to enable better predictability of such computer models, thereby eschewing, to quite a non-trivial extent, the complex and costly empirical route to design. A successful execution of the proposed research plan has led to substantial innovations both in aspects of the experimental measurements and in several aspects of the computational (discretization/model updating/ filtering/ optimization/ parallelization) schemes to be employed.
Grant sponsor: NERC-GCRF, UK,
Funds: INR 18.5 lakhs (IISc’s share in this joint project between the University College London and IISc);
Duration: 1 year; Started: December 15, 2016; Completed: December 14, 2017.
Achievements: Large uncertainties in the possible sources of tsunamis around India are key issues for tsunami risk assessment for the coast of India. In 2004, the lack of awareness and preparedness to a possible tsunami arising from the Sumatra-Andaman (S-A) fault unfortunately contributed to the death of around 15,000 people on the Eastern coast of India, with a catastrophic economic impact on poor and fragile local communities. This proposal’s first aimed at improving the scientific understanding that underpins prevention strategies such as hazard mapping for urban planning. Its second aim was to explore the processes that would enable the incorporation of scientific assessments into policies and practice, in order to ethically and efficiently protect vulnerable and poor communities from future tsunamis. Its third aim was to identify the hurdles that would hamper the integration of science, policies and practice, and propose specific strategies towards pragmatic decision-making under severe uncertainties. Using a stochastic inversion scheme, we have identified the possible sources of future tsunamis in the western coast (e.g. tsunami wave amplifications by submarine landslides or diversion of seismic waves through splay faults) and the precise effects of data uncertainty (or noise) on the predicted solutions.
Grant Sponsor: Mitsubishi-Hitachi Power Systems (Japan),
Funds: Rs. 1.01 Crores,
Duration: 5 years; Start Date: August 1, 2015.
Aims and Scope: Steam turbines operate at high pressure and temperature. Hot turbine components are therefore exposed to creep loading. Additionally, the transient events, e.g. start-up, shut down causes fatigue loading because of the induced thermal stresses especially in the thick-walled components like rotors, stators, casing etc. Extensive creep and fatigue loading may eventually lead to crack initiation followed by its growth, which would certainly limit the life of the turbine components. Thus the optimisation problem here involves finding out the relevant control strategies such that the start-up is as fast as possible whilst maintaining the thermal stresses developed in the components lower than a prescribed strength (decided from the concepts of low cycle fatigue of materials). This time dependent control strategy should be implemented incrementally utilising the feedback from the system (in form of the developed thermal stress data), thus building a closed loop type feedback control architecture. The goal in this project is to find out automatically using feedback control the strategies that would optimize the start-up time whilst maintaining the necessary constraints on load generation and also several structural strength related constraints.
Grant Sponsor: NAL, Bangalore,
Funds: Rs. 8.34 lakhs;
Duration: 1 year; completed in March, 2013.
Achievements: The inability of the wings to produce the desired lift at high angles of attack owing to flow separation is a major cause that limits aircraft performance. Indeed, boundary layer separation causes the aircraft to stall thereby severely penalizing the aerodynamic performance. This calls for devising appropriate flow control techniques that detect and control flow separation even before it occurs. Such control strategies aimed at overcoming the aerodynamic stall of air-vehicles would enable them to operate at high angles of attack, take-off and land at lower speeds and shorter runways, be more fuel efficient, and increase the maximize lift achievable. Towards this, a recently developed flow control scheme involves the application of Electro-Active Polymer (EAP) films which can deform out-of-plane in the shape of dimples under the application of suitable electric fields. The fluid flow control scheme may thus comprise of several EAP films (along with electrodes) distributed across the surface of the object subject to fluid flow control. By way of an important precursor to a successful implementation of this control strategy, we have employed the smooth DMS-FEM for an accurate numerical determination of the deformed EAP film shapes under applied electric fields. The inverse problem of determining the input voltage to the electrodes towards attaining a preferred deformed film profile of the EAP has also been solved.
Grant Sponsor: LRDE, Bangalore,
Funds: Rs. 10.0 lakhs;
Duration: 1 year; started: November, 2012; completed: November, 2013.
Aims: This project marks the first one in a proposed series of 3 projects, wherein have undertaken the solution of the problem of single or multiple target tracking in the presence of sea clutter. The completed work has essentially been aimed at developing a class of methods, based on a few particle filters (PF-s) devised at the Computational Mechanics Lab, IISc, and testing against numerical simulations for the detection of the target (slow moving or otherwise; assumed to be a single target) in the presence of sea clutter. We have also made substantive headway into the problem of data association (posed as an optimization problem) in the presence of multiple targets.
Grant Sponsor: Indian National Centre for Ocean Information System (INCOIS), Hyderabad;
Funds: Rs. 35 lakhs;
Duration: 3 years; Started on: February, 2011; Completed on: February, 2014
About this project: Ongoing studies on earthquake/tsunami recurrence indicate that the previous tsunamigenic earthquake in the Andaman-Sumatra subduction zone that had affected the east coast of India may have occurred about 1000 years ago. The predecessor of the 1945 tsunami is not yet dated; but the threat from a future earthquake cannot be discounted. The studies subsequent to the 2004 tsunami have highlighted the need for quantifying the sources of tsunamis in the Indian Ocean and develop models of tsunami generation and inundation.
The primary accomplishments of the project are:
Grant Sponsor: ISRO-IISc Space Technology Cell;
Funds: Rs. 10.2 lacs;
Duration: 1.5 years; started: April 1, 2011; completed: October, 2012.
Accomplishments: We have formulated stochastic computational strategies to relieve the Kalman filter estimates of their sensitivity of process and measurement noises. Such sensitivity has been widely reported to be a major bottleneck is applying the Kalman or Extended Kalman filters (KF or EKF respectively) to large dimensional system identification problems, typically encountered in Civil and Aerospace engineering applications. Our main tool has been a non-convex optimization scheme (a stochastic search scheme) to adaptively arrive at optimal initial state covariance, process noise covariance and measurement noise covariance matrices so the parameter estimates could be obtained in a robust and automated way. A second accomplishment of the work has been to develop a gain-optimized version of the KF as well as the ensemble version of the KF. The advantage with this approach is that, one directly deals with the gain matrix without having to separately treat the three covariance matrices.
Grant Sponsor: DST (India) and NOW (Netherlands);
Funds: INR 51.72 lakhs (Indian side) + 1,82,000 Euros (Dutch side);
Duration: 3 years; Completed: March 31, 2016.
Accomplishments: Rheumatoid arthritis (RA) is a chronic inflammatory disease which affects synovial joints affecting over 3 million people in Europe and up to 10 million in India. The disease is progressive and severely debilitating and has a large economic impact as it affects people in their relatively younger years. There has been great progress in RA management where therapeutic agents are being developed to arrest disease progression. There is however a compelling need to diagnose the disease in early stages when irreversible joint damage has not yet set in, and to monitor progression so as to tune the expensive therapies. The current practice of imaging indicators of rheumatoid arthritis (RA) suffers from severe shortcomings. Conventional X-ray imaging is only sensitive to changes in established RA long after irreversible joint damage has occurred. Magnetic Resonance Imaging (MRI) is responsive to early-stage inflammations of the synovium (synovitis) but is expensive and poorly accessible. Ultrasound Doppler imaging shows promise in following synovitis, but is not sensitive. Photoacoustic imaging has great potential in addressing several issues in diagnosing RA and monitoring therapeutic interventions. The technique is able to detect and image blood vessels with high resolutions using relatively inexpensive technologies. With this we can visualize vascularization accompanying inflammation of the synovium which is a sensitive and early predictor of joint disease. In the accomplished research, we have built on the expertise and infrastructure available in the two main laboratories to:
Grant Sponsor: CSIR;
Funds: Rs. 12 lakhs;
Duration: 2 years; completed in March, 2010.
Accomplishments: The essence of this project was to introduce, for the first time, a family of particle or ensemble-based stochastic filters for elastography. The focus of the project was therefore to highlight the superior performance of these stochastic filters over deterministic, regularized quasi-Newton reconstruction schemes. It has indeed been found that several variants of the stochastic filter, e.g. the ensemble Kalman filter, perform substantially better than the deterministic scheme against small variations in the algorithmic parameters as well as variations in the noise intensities in the measured data. In another novel approach, developed as part of this project, a handshake strategy, we have used the regularized normal equation for parameter updates (used with the deterministic Newton-like schemes) within the observation model of a particle or ensemble Kalman filter. The numerical evidence clearly suggests that the handshake scheme outperforms even the standard stochastic filtering schemes for elasticity profiling of soft tissue organs.
Grant Sponsor: Indira Gandhi Centre for Atomic Research (IGCAR), Kalpakkam;
Funds: Rs. 18 lakhs;
Duration: 3 years; Completed on: October, 2013
Accomplishments: We have evolved novel and effective methodologies for obtaining statistical information about failure of large dimensional dynamical systems, especially shell structures (modeling, for instance, components of the piping systems of interest to the Nuclear Power Industry), under earthquake-like excitations. The work included studies on failure rate estimations of components of such structures under normal operating conditions as well as under design basis events during the life of these components. The three main phases of this project are as follows. The first is the computational modeling of shell structures as Cosserat surfaces with elastic-plastic material behaviour using a novel, smooth finite element method. The second phase involves a novel form of stochastic control, along with the Girsanov transformation of probability measures, to set up a Monte Carlo simulation strategy with just a few sample runs for the prediction of failure probability and thus addresses the important issue of failure rate estimation. The integration of these two developments on a common software platform, with adequate user-interfaces in the form of professionally developed pre- and post-processing capabilities, constituted the third and final phase.
Grant Sponsor: RESPOND, ISRO, Govt. of India;
Funds: Rs. 6.5 lakhs; Duration: 3 years; completed in March, 2009.
Accomplishments: We have developed a framework to treat the highly nonlinear problem of membrane wrinkling using a family of mesh-free methods (all developed at the Computational Mechanics Lab, IISc) and the Cosserat point method. While the mesh-free methods are well known to have significant numerical advantages over the FEM owing to the dispensation of a mesh-based domain discretization, the Cosserat point method is particularly helpful in reducing (or even bypassing) the numerical ill-conditioning that typically arises in the computational approaches dealing with ultra-thin membranes of interest to the Indian space scientists. During the course of this work, such advantages have been adequately demonstrated via numerical simulations of ultra-thin membranes of different shapes. In particular, the Cosserat point approach has, for the first time, enabled the simulation of wrinkled membranes using 3D Cosserat brick elements (without precipitating any thickness locking). Some of the programs, developed as part of this work have also been transferred to the VSSC, Trivandrum, where they are presently being used.
Grant Sponsor: Ministry of Information Technology, Govt. of India;
Funds: Rs. 17 lakhs; Duration: 3 years;
Other Investigator: Prof. R M Vasu (Department of Instrumentation); completed in March, 2009.
Accomplishments: This work constituted an attempt at reconstructing the elasticity profile of soft tissue organs based on the gathered displacement data owing to non-invasive palpation of these organs. Given that a cancerous region in such organs is often characterized via a relatively higher elasticity modulus (vis-à-vis the surrounding normal tissue matter), a suitable reconstruction of the distribution of the elasticity (shear) modulus has enabled a non-invasive scheme towards early detection/onset of cancer. Mainly deterministic reconstruction schemes, based on regularized quasi-Newton schemes, have been used for purposes of reconstruction. For validating the reconstruction scheme, soft-tissue mimicking phantoms (made of poly-vinyl alcohol) have been produced and the displacement data has been gathered through an optical imaging procedure. Reconstruction of the elasticity profiles of such phantoms, wherein one or two lump-like inhomogeneities were introduced based on controlled freezing-and-thawing exercises, have provided credence to the feasibility of this approach towards cheap and early detection of cancer.
Grant Sponsor: Naval Science and Technological Laboratory (NSTL), Visakhapatnam;
Funds: Rs. 10 lakhs; Duration 2 years; started in December, 2006; completed in December, 2008.
Accomplishments: We have developed a family of particle filters for target tracking in 2D only based on noisy measurements of the bearing angles. In particular, a bootstrap filter, a Gaussian sum filter and a semi-analytical filter have been developed and implemented for this purpose. Bearings-data supplied by the NSTL have been used within these filters to track (estimate) the 2D locations of the object being targeted. A specific difficulty with this problem is the presence of high non-Gaussianity in the observation data. Accordingly the Gaussian sum filter, which can most efficiently model (and account for) the non-Gaussian features in the process or measurement noises, has been observed to perform the best among the three above filters. While the semi-analytical filter also has performed very well, the bootstrap filter has shown unacceptably high sensitivity to the process and measurement noises. The relevant programs, developed in MATLAB, have been transferred to the NSTL.
Grant sponsor: Indian National Centre for Ocean Information Services;
Funds: Rs. 23 lakhs,
Duration : 3 years; started in September, 2008; completed in September, 2011.
Achievements of the project: This was the precursor to the other related projects (see project nos. 3 and 7). As part of the work, we had, for the first time, initiated mesh-free discretization approaches for the computation of tsunami propagation based on shallow water wave models in 1D and 2D. Herein, we had also been working on novel variation reducing numerical integration strategies for temporal propagation of the semi-discretized shallow water equations. In the process, the variation diminishing term had been identified as one that, as anticipated, parabolized the shallow water equations, which were first order hyperbolic. This also yielded a physically consistent approach for introducing the so-called artificial viscosity term, for which a physical justification was hardly available. We had applied the methods to a few tsunami propagation scenarios with the ocean bathymetry consistent with the available data from the Indian Ocean. The results had indeed been encouraging and they essentially formed a foundation for the more advanced work on these lines and research collaborations with the Department of Statistical Science, University College London, UK.
Grant Sponsor: Cranes Software International Ltd.,
Funds: Rs. 16 lakhs; Duration: 2 years; started in May, 2006; completed in May, 2008.
Accomplishments: The primary aim of the work has been by way of value addition to the NISA family of commercially available finite element programs owned by Cranes Software. In the process, during the first year of the project, an adaptive version of the implicit Newmark method was developed and implemented within NISA towards automatic time stepping for numerical integration of structural dynamic systems that exhibit regimes of sensitivity to time step sizes. A typical case in point is a structure subjected to localized impact loads. While the structure may be insensitive to time steps before and well after the duration of the impact load and far away from its point of application, the step sizes have to be considerably smaller during and immediately after the application of the impact force. The adaptive scheme, incorporated within NISA, has been demonstrated to treat such cases with efficiency and accuracy. In addition, during the second year of the project, a mesh-free approach, combined within a finite element based framework, has been used to obtain the response of solids with one or more propagating crack(s). Specifically, the mesh-free scheme has been adopted around the crack tip in order to accurately capture the stress filed without numerical errors owing to mesh distortion (typically encountered in most FEM codes). The discretization in the rest of the domain is via the usual FEM, available with NISA.
Grant Sponsor: BRNS, Govt. of India;
Funds: Rs. 12 lakhs;
Duration: 3 years; started in January, 2005; completed in January, 2008.
Accomplishments: A novel numerical scheme for simulating extreme events (such as the first passage of a system response beyond a rare or critical level has been proposed and successfully implemented based on a Girsanov transformation of probability measures. The emphasis has been on applications involving reliability studies of shell structures (typically present as nuclear containment vessels) under stochastically dynamic loading conditions (e.g. earthquakes). Towards this, the system equations (which are typically in the form of partial differential equations) are first spatially discretized through a finite element method to yield a system of stochastic differential equations (SDE-s). Then a stochastic Newmark method (also developed as part of this project) is employed to time integration of the semi-discretized SDE-s. A control term is used in the Newmark scheme to induce the system response to breach the requisite critical levels, which would have otherwise not been crossed by most (nearly all) of the simulated system response trajectories. In order to correct for this the introduction of this control term in a weak sense, a Girsanov correction is applied so as to obtain practically relevant estimates of statistical moments (or probability mass information) for those rare orbits that cross the critical levels. The work has therefore been a non-trivial extension of the more classical concept of importance sampling in the context of reliability analyses of structures.
Grant Sponsor: ISRO-IISc Space Technology Cell;
Funds: Rs. 5 lakhs;
Duration: 2 years; started in February, 2004; completed in February, 2006
Accomplishments: A new finite element strategy has been proposed to study the nonlinear dynamics of wrinkled membranes (or, more specifically, the dynamic variations of the wrinkled or slack zones) within a tension field framework. The finite element analyses have been performed through plane stress elements that do not resist compression, i.e., the zones where compressive stresses tend to appear wrinkle immediately. Accordingly, the problem is generally ill-conditioned, especially as slack zones are formed, and hence the semi-discretized equations have been regularized through a small, Tikhonov regularization parameter. Following this, the numerical integration of the regularized semi-discretized (geometrically) nonlinear system of equations has been accomplished through a transversal linearization scheme that does not necessitate forming the tangent stiffness matrix at each linearization step. In this sense, the proposed approach has been found to be efficient and accurate. These facts have also been verified through extensive numerical simulations on thin membranes, typically used as antennas in a variety of space structures.