Anirban Guha

STRUCTURAL HEALTH MONITORING OF A CANTILEVER BEAM USING SUPPORT VECTOR MACHINE

In this article, the effectiveness of support vector machine (SVM) is examined for health monitoring of beam-like structures using vibration-induced modal displacement data. The SVM is used to predict the intensity or location of damage in a simulated cantilever beam from displacements of the first mode shape. Twelve levels of damage intensities have been simulated at 12 locations, and six levels of white Gaussian noise have been added, thereby obtaining 1,008 simulations. About 90% of these are used for training the SVM, and the remaining are used for testing. The trained SVM is able to predict damage intensity and location of all the training set data with nearly 100% accuracy. The test set data reveal that SVM is able to predict damage intensity and damage location with errors varying from 0.28% to 4.57% and 0% to 20.3%, respectively, when there is no noise in the data. Addition of noise degrades the performance of SVM, the degradation being significant for intensity prediction and less for damage location prediction. The results demonstrate the use of SVM as a powerful tool for structural health monitoring without using the data of healthy state.

A simple single variable shear deformation theory for a rectangular beam

This paper proposes a simple single variable shear deformation theory for an isotropic beam of rectangular cross-section. The theory involves only one fourth-order governing differential equation. For beam bending problems, the governing equation and the expressions for the bending moment and shear force of the theory are strikingly similar to those of Euler–Bernoulli beam theory. For vibration and buckling problems, the Euler–Bernoulli beam theory governing equation comes out as a special case when terms pertaining to the effects of shear deformation are ignored from the governing equation of present theory. The chosen displacement functions of the theory give rise to a realistic parabolic distribution of transverse shear stress across the beam cross-section. The theory does not require a shear correction factor. Efficacy of the proposed theory is demonstrated through illustrative examples for bending, free vibrations and buckling of isotropic beams of rectangular cross-section. The numerical results obtained are compared with those of exact theory (two-dimensional theory of elasticity) and other first-order and higher-order shear deformation beam theory results. The results obtained are found to be accurate.

Microstructural damage based micromechanics model to predict stiffness reduction in damaged unidirectional composites

Prediction of the residual stiffness of the carbon fiber reinforced polymer composite, subjected to fatigue loading, can be performed using some of the phenomenological models. However, it is still a challenge to find the stiffness based on the known microstructural damage state (that was developed irrespective of the load history). In this work, two micromechanics-based models were developed to predict reduction in the stiffness of the damaged composite. Fiber crack density and interface debonding was used to define the microstructural damage state of the composite. These models account for the fiber crack density in the form of change in either geometry (equivalent ellipsoid model) or material property of the fiber (reduced stiffness model). The microstructural damage state in the unidirectional carbon fiber reinforced polymer composite, obtained from the on-axis tension–tension fatigue loading, was used to validate the models. The results from reduced fiber stiffness model were compared against experiment and finite element analysis for the given microstructural damage. The stiffness obtained using reduced fiber stiffness model was in good agreement with that obtained from the experiment. However, reduced fiber stiffness model underestimated reduction in stiffness compared to finite element analysis.

Numerical Study and Experimental Validation of Effect of Varying Fiber Crack Density on Stiffness Reduction in CFRP Composites

Representative volume element (RVE) has commonly been used to predict the stiffness of undamaged composite materials using finite element analysis (FEA). However, never has been an independently measured true microstructural damage quantity used in FEA to predict composite stiffness. Hence, in this work, measured fiber crack density in unidirectional fiber composite (generated using controlled fatigue loading) was used to predict reduction in stiffness using a RVE. It was found that the stiffness changes with change in depth of the volume element along the fiber direction and asymptotically reaches a constant value beyond a critical length called representative depth. It was argued that this representative depth should be more than the minimum of two characteristic length scales, twice of ineffective length and average length of broken fibers. Effective stiffness obtained from FEA of the optimum-sized RVE was in excellent agreement with the experimental results for given microstructural damage state.

Briquette Compacting Machine: A Design for Rural Applications

This paper enumerates the design and synthesis of a mechanism for a fuel briquette compacting machine. The briquettes are made of a mixture of husk (rice or wheat or any other) and animal waste in appropriate proportions and compacted in the machine. Several experiments were initially conducted to arrive at the right proportions of water, husk, and dung, and to determine the force to compact the biomass. The pellets were dried in the sun and burnt in a stove. The final specifications of the machine were arrived based on these simple trials. There were several challenges in the machine development. The biomass mixture tends to cake and harden if the machine is left idle for long. It is not desirable for a briquette to crumble both in a wet as well as a dry state. The mechanism has to handle these requirements and as the machine, is to be manually operated frictional effects and any tendency to jam ought to be minimized. The engineering drawings of the machine are being freely distributed to rural mechanics who are desirous of replicating the machine, after observing the machine in action at CTARA at IIT Bombay. Several machines have thus been built and are operational in many villages. The paper covers such aspects and how a compact machine was arrived at through synthesis of an appropriate mechanism that is inherently not easily prone to “jamming”. The synthesis is based on techniques derived from symmetric coupler curve generation.

Equation-based domain knowledge utilization into neural network structure and learning

The use of neural networks as black boxes, though useful for modeling complicated industrial systems, has some limitations. No physical interpretation can be given to sections of the trained network. Incorporation of domain knowledge into neural network attempts to address this lacuna. Most of the attempts in this direction have been in the area of data classification in which sub-classes created with the help of domain experts have led to better neural networks. This work attempts to incorporate domain knowledge into the structure of a neural network for solving a regression problem—that of a piston pump leakage prediction. It shows a way in which prior knowledge about subsystems, in the form of equations, can be used to create a neural network for modeling the entire system. This approach significantly outperforms a traditional feed forward neural network. As a key contribution, this approach allows physical interpretation of the neurons which can aid in troubleshooting and anomaly detection.

Mechanistic model for fiber crack density prediction in cyclically loaded carbon fiber-reinforced polymer during the damage initiation phase:

Prediction of the fiber crack density (as one of the microstructural damages) for unidirectional fiber-reinforced polymer composite under monotonic tensile load, using strength models, has been reported in the literature. However, the microstructural damage prediction for a fiber-reinforced polymer subjected to fatigue loading is still a challenge. In this work, a progressive damage initiation model was developed to predict the fiber crack density in carbon fiber-reinforced polymer composite subjected to fatigue loading. A stochastic model was used for modeling the fiber fatigue strength. Reduction in effective life of the fiber was modeled using linear Miner’s rule. Effect of fatigue strength parameters on fiber crack density was found to be considerable compared to the effect of interface shear strength. At a low number of cycles, fiber crack density obtained from the model was in good agreement with the experimentally measured fiber crack density.

Kinematic analysis and design optimisation of a surgical rod cutter for shearing of Ti6Al4V rods

Titanium rods used as spinal implants need to be cut to a suitable length in the operating room. The mechanism used to cut them needs to be entirely manually operated since the interference of electrical and electronic components with the sensitive electronic components cannot be predicted. The mechanism also needs to be as small as possible since it needs to be decontaminated in an autoclave prior to its use. This work explores different options for such a mechanism. The best design from a scarce patent literature was optimised to obtain a 56% reduction in dimension. The manual force was verified with a mechanism simulator (ADAMS). A material model in DEFORM allowed the experimentally determined shearing force to be simulated. This led to the prediction of change in the mechanism’s dimensions with change in the rod’s dimensions.