Research

Vibration Control, Monitoring, Reliability Analysis of Wind Turbines

Last few decades have witnessed tremendous increase in energy demand to support the global socio-economic development. Depleting fossil fuel and ever increasing carbon footprint have forced mankind to search for alternate green source of energy. In this context, renewable and clean sources like sun, wind, biological process, waves etc have become viable alternatives for power generation. Of all these possibilities, wind energy has remained a popular source and the net power production in this sector have undergone a steady growth. Wind turbines are of two types depending upon their location i.e. onshore and offshore. We are particularly interested on offshore wind turbines where marine environment has combined effects of wind and wave on the turbines which are often supported by monopiles, jacket/tripod, semi submersible platform, barge, tension legged platform and spar buoy types and experience significant structural vibrations. Besides tower motion, modern multi-megawatt offshore wind turbines have large flexible blades, which also undergo significant vibration caused by extreme loading. Due to this reason, blades suffer fatigue and need regular maintenance. 

In this context, we are interested on the following topics: Digital Twin, Computer Vision-based Monitoring, Vibration Control, Fatigue Life Estimation, Reliability Analysis, Effect of Climate Change, Fault Detection/Diagnosis/Prognosis of Rotating Machinery, Anomaly Detection, Machine Learning based Condition Monitoring.

High-Dimensional Reliability Analysis

Due to the increase in complexity of engineered structures, reliability analysis has become an essential tool to maintain integrity in the system. The reliability analysis is used to quantify the failure probability of the structure under the structural system and external excitation uncertainties.  In general, traditional reliability based design optimization methods often suffer from the curse of dimensionality for high-dimensional problems. Here we are interested on the dimension reduction problems using the deep learning-based framework such as generative models for efficient  reliability analysis  of the complex structure  such as wind turbines,  reactor building  of  a nuclear power plant  etc.

Physics-Informed Deep Learning

Most of the surrogate models are data-driven and the data representing the physics of a stochastic system is often not captured by data-driven surrogate models. In this scenario, Physics-informed machine learning is a viable tool in which the governing physical laws are included in the loss function of the neural network such as initial, boundary conditions for solving FEM models. We are particularly interested on the application of physics-informed machine learning for solving reliability analysis problem. In general, reliability analysis is carried out to estimate the failure probability in element level or system level considering a threshold value. Different algorithms are developed for reliability analysis of structures with an aim of reducing the computational costs. 

Seismic Isolation, Energy Harvesting

Vibration control of structures gained considerable attention from researchers and engineers in the last two decades as the traditional designs failed to meet the ever-increasing and stringent performance demand. The control strategies for this purpose are broadly classified into passive and active/semi-active systems. Different types of dampers are used to mitigate seismic/ wind induced vibration such as TMD, TLD, TLCD, Inerter, Base Isolation. We aim to develop an efficient strategy for controlling the seismic/wind induced stochastic response of a high-rise building and at the same time, we are focused on how to convert the vibrational energy into electrical energy (energy harvesting). 

Nonlinear Dynamics, Volterra Series

Limit cycle oscillations (LCO) are the sustained oscillations, occurred due to the interaction between aerodynamic and structural forces in presence of the nonlinearities. For airfoil LCO, the main aspects are the sub-critical and super-critical oscillations because super-critical oscillation affects the fatigue damage whereas due to discontinuity behaviour of sub-critical LCO, the airfoil may experience the unexpected dynamic behaviour. Therefore, control and health monitoring are essential components to ensure safety during its operation to protect them against vibration induced structural damages. Here we are interested on Chaos, Bifurcation, Stability related problems of wings. 

Volterra and Wiener series are two classes of polynomial representations of nonlinear systems. They are perhaps the best understood and most widely used nonlinear system representations in signal processing and system identification. A Volterra or Wiener representation can be thought of as a natural extension of the classical linear system representation. In addition to the convolution of the input signal with the system's impulse response, the system representation includes a series of nonlinear terms that contain products of increasing order of the input signal with itself. It can be shown that these polynomial extension terms allow for representing a large class of nonlinear systems which basically encompasses all systems with scalar outputs that are time-invariant and have noninfinite memory.  We are particularly interested on the solving random vibration problems using Volterra series and machine  learning.

Resiliency Assessment of Infrastructure Network

Will be updated soon..