Indian Institute of Technology Guwahati

About: Indian Institute of Technology Guwahati is a education organization based out in Guwahati, Assam, India. It is known for research contribution in the topics: Catalysis & Computer science. The organization has 6933 authors who have published 17102 publications receiving 257351 citations.

Analytical solutions for the stiffness and damping coefficients of squeeze films in MEMS devices with perforated back plates

Abstract: Closed-form expressions for the stiffness and the damping coefficients of a squeeze film are derived for MEMS devices with perforated back plates. Two kinds of perforation configurations are considered—staggered and matrix or non-staggered configuration. The analytical solutions are motivated from the observation of repetitive pressure patterns obtained fromnumerical (FEM) solutions of the compressible Reynolds equation for the two configurations using ANSYS. A single pressure pattern is isolated and further subdivided into circular pressure cells. Circular geometry is used based on observed symmetry. Using suitable boundary conditions, the Reynolds equation is analytically solved over the pressure cells. The complex pressure obtained is used to identify the stiffness and damping offered by the pressure cells. The stiffness and damping forces due to pressure cells within a pattern are added up separately. In turn, the stiffness and damping due to all the patterns are summed up resulting in the stiffness and damping forces due to the entire squeeze film. The damping and spring forces thus obtained analytically are compared with those obtained from the FEM simulations in ANSYS. The match is found to be very good. The regime of validity and limitations of the analytical solutions are assessed in terms of design parameters such as pitch to air gap, hole length to diameter and pitch to hole radius ratios. The analysis neglects inertial effects. Hence, the results are presented for low values of Reynolds number.

Improving Robustness of Machine Translation with Synthetic Noise

Abstract: Modern Machine Translation (MT) systems perform remarkably well on clean, in-domain text. However most of the human generated text, particularly in the realm of social media, is full of typos, slang, dialect, idiolect and other noise which can have a disastrous impact on the accuracy of MT. In this paper we propose methods to enhance the robustness of MT systems by emulating naturally occurring noise in otherwise clean data. Synthesizing noise in this manner we are ultimately able to make a vanilla MT system more resilient to naturally occurring noise, partially mitigating loss in accuracy resulting therefrom.