top of page
This site was designed with the
.com
website builder. Create your website today.Start Now
Some Research Projects
Dynamic data driven modeling of tracked order vibration in Rolls Royce engine (PhD Thesis)

  • Tracked order vibration has been the industry standard to monitor vibrations in Rolls Royce engine

  • Challenging to develop system-level physics based models for tracked order vibrations that can accurately account for the effect of current as well as the historical operational conditions

  • Lack of baseline model hinders accurate diagnosis as well as prognosis.

  • Current thesis work designs and implements a data-driven framework using wavelet-based dynamic mode decomposition in conjunction with multiscale detrended cross-correlation to model the tracked order vibration.

  • Evaluated the performance of our in-house methodology (WDMD) with off the shelf machine learning framework and WDMD was found to be 10% more efficient in capturing the overall trend.

  • Improved modeling accuracy with interesting findings

  • Because of the confidential nature of the work, results cannot be uploaded currently.

  • Research Collaborators: Rolls Royce Derby, Rolls Royce Germany
     

Publications

  1. Krishnan, M., Jin, R., Sever, I. A., and Tarazaga, P. A., 2019, "Data Based Modeling of Aero Engine Vibration Responses", IMAC XXXVII A conference and Exposition on Structural Dynamics, FL, Jan 28-31. (https://link.springer.com/chapter/10.1007/978-3-030-12676-6_34)

  2. Krishnan, M., Gugercin, S., Sever, I., and Tarazaga, P. A., 2020, "Dynamic data driven modeling of aero engine response," IMAC XXXVIII A Conference and Exposition on Structural Dynamics, TX, Feb 10-133. (https://link.springer.com/chapter/10.1007/978-3-030-47638-0_30)

  3. Krishnan, M., Sever, I., and Tarazaga, P. A., 2020, "Determining interdependencies and causation of vibration in aero engines using multiscale cross-correlation analysis," IMAC XXXVIII A Conference and Exposition on Structural Dynamics, TX, Feb 10-13. (https://link.springer.com/chapter/10.1007/978-3-030-47638-0_29)

  4. Krishnan, M., Gugercin,S,  & Tarazaga, P. A. “Wavelet based Dynamic mode decomposition. (Manuscript ready for submission to MSSP)

  5. Krishnan, M., Sever, I., and Tarazaga, P. A., 2021, “Wavelet based dynamic mode decomposition for aero engine vibration”, AIAA, Scitech – 2021 (Manuscript submitted for clearance).

  6. Krishnan, M., Jin, R., Sever, I. A., Gugercin,S. & Tarazaga, P. A. "Data driven modeling of Aero Engine Vibration Responses". (Manuscript preparation)

  7. Krishnan, M., Gugercin,S,  Sever, I. A.,. & Tarazaga, P. A. “Wavelet based DMD for modeling the tracked order vibration response using a cascading dynamical systems approach". (Manuscript preparation)

​

Patent:

Cascading dynamical system with WDMD and Delay DMD towards modeling tracked order vibration response (Planning)

​

​

​

2019-01-23-s20_digital_ghost_twin_indust
Capture5.PNG
Capture6.PNG
Capture7.PNG
Data-driven approach to investigate the interaction of Anechoic Structural Traveling Waves with fluid Environment

  • A mechanical wave is generated as a result of an oscillating body interacting with the well-defined medium and it propagates through that medium transferring energy from one location to another. 

  • The ability to generate and control the motion of the mechanical waves through the finite medium opens up the opportunities for creating novel actuation mechanisms. 

  • Waves with desired characteristics are generated by the excitation of piezoelectric elements through coupled system dynamics.

  • This project is aimed at developing a data-driven model for the interaction of anechoic structural traveling wave with a fluid environment.

  • The data-driven methodology is developed using experimentally obtained Frequency response data using the responses collected by a hydrophone of a beam excited under water using PZT patches as shown in the figure.

  • Vector-Fitting (VF) algorithm is used to build a state-space model of the dynamical system from steady state FRF data

  • The results of the work are summarized in the video.

​

Publications

  1. Krishnan, M., Malladi, V. S., & Tarazaga, P. A. (Under Review in Mechanical systems and Signal Processing). Leveraging a data-driven approach to simulate and experimentally validate aMIMO multiphysics vibroacoustic system

​

​

​

Capture3.PNG
Experimental Data-Driven Determination of Dispersion Characteristics of Structures 

  • Collaborated in developing a data-driven technique to estimate dispersion relations that are experimentally tedious

  • Utilized easy-to-measure Frequency Response Functions (FRFs) to develop a numerical model of the structure under test.

  • Constructed group velocity curves through a series of numerical simulations.

  • Validated the experimental data-driven approach by analytical calculation for a one-dimensional homogeneous beam

  • Designed and performed an experiment using 3D laser vibrometer to measure steady-state flexural as well as longitudinal FRF's

Capture4.PNG
Project |03

 

3dlaser.jpg

​​Publications

  1. Malladi V.V.N.S., Albakri., Krishnan. M., Tarazaga P.A., and Gugercin S., (Under Review in Mechanical Systems and Signal Processing) "Estimating Experimental Dispersion Curves from Steady-StateFrequency Response Measurements"

​

​

​

Investigation of propagation of Traveling waves in a buckled structure

  • Utilize the buckling property of structures to generate high amplitude traveing waves

  • Dynamically exciting buckled structure to continuously switch between its two potential well to create high amplitude traveling wave which can be used for propagation or drag reduction

  • Experimentally validation using computer vision and laser vibrometer.

  • Resulted in a spinoff project: Manipulating the structure using smart materials to induce flutter

Picture1.png
Picture2.jpg
Real-time damage detection of structures

  • Structural damage detection involving condition assessment, fault diagnosis and prognosis of civil, mechanical and aerospace infrastructure has garnered significant attention and a wealth of literature exists in this area

  • Traditional offline damage detection schemes have disadvantages

  • This work proposes a real-time damage detection framework using time-varying auto regressive(TVAR)  modeling in conjunction with recursive principal component analysis (RPCA)  to detect spatio-temporal structural damage in real time.

Capture11.PNG

​​Publications

  1. Krishnan, M., Bhowmik, B., Hazra, B., & Pakrashi, V. (2018). Real time damage detection using recursive principal components and time varying auto-regressive modeling. Mechanical Systems and Signal Processing, 101, 549-574.
  2. Bhowmik, B., Krishnan, M., Hazra, B., & Pakrashi, V. (2019). Real-time unified single-and multi-channel structural damage detection using recursive singular spectrum analysis. Structural Health Monitoring, 18(2), 563-589.
  3. Krishnan, M., Bhowmik, B., Tiwari, A. K., & Hazra, B. (2017). Online damage detection using recursive principal component analysis and recursive condition indicators. Smart Materials and Structures, 26(8), 085017.
  4. Bhowmik, B., Krishnan, M., Hazra, B., & Pakrashi, V. (2017). Online damage detection using recursive principal component analysis. Procedia engineering, 199, 2108-2113.

​

​

​

IMG_1355.JPG
Miniature underwater robot: Traveling wave
​

  • The purpose of this work is to develop a state of the art miniature underwater robot using the traveling wave generated in the fins as the propulsion mechanism. 

​​

  • Research colleague: Dr.Sriram Malladi, Sheyda Davaria, Dr. Pablo Tarazaga

Picture2.jpg
bottom of page