Bearing Diagnostics

  • 50% of rotating machinery failures related to bearings failure

  • Why are they hard to diagnose? Nonlinearity

  • The objective is to develop an algorithm that can monitor bearings health and generalize it for various operating conditions ( load and speed)

  • Bearings with various conditions and various severities were investigated

  • A comparison with Envelope Analysis, wavelet and FFT was performed

  • 99% accuracy achieved in classifying the bearing fault then determining the fault severity

  • Publications: CSNDD | JVA | JVC | CSNDD | ND (book) | TN (book)

Gear Diagnostics of Sikorsky Helicopter Engine

  • A project in collaboration with UTRC

  • The objective is to build an algorithm that can identify faults in one of the gear-boxes using noisey vibration signals

  • Configuration: healthy, root crack on 1 tooth, root crack on 5 teeth, missing tooth

  • 6 DOF model was built and various root tooth crack cases were modeled

  • 99% accuracy achieved in identifying the defect in the gear-train

  • Publications: PHM | JSEA

Control of Nonlinear Fault Simulator

  • NFS was built at VCADS as a test bed for our developed diagnostic algorithms

  • NFS has a configurable and adjustable design

  • NFS can achieve any desired input trajectory

  • Various nonlinear phenomena can be simulated

Modeling and Control of KUKA iiwa 7 R800

  • 7 DOF manipulator

  • This project is divided into three parts: 1) Kinematic and dynamic modeling 2) Trajectory planning 3) Control

  • A sliding controller is used to achieve the desired motion

Sensor Fusion Using CNN-LSTM Network

  • What is the best way to integrate data from various sources?

  • A integrative deep learning approach to fuse data from different sensors

  • The algorithms is based on two parts:

        • CNN are used to extract features from different time serieses

        • LSTM are used to rank and select the extracted features from different sources

Physics-Informed AI

  • A physics-informed approach was developed (defect-free model was built)

  • Physics-Informed features were extracted using residual and cross-sample-entropy analysis

  • Random forest model integrated with SHAP algorithm was implemented to rank the extracted features

Crack Detection in Rotating Shaft

  • A project in collaboration with University of Uberlandia, Brazil

  • Two cracks with different severities were introduced using crack propagator

  • Mutual information was used to rank features

  • Only three features were required to distinguish between the faults

  • 100% accuracy was achieved

  • Publications: IFTOMM (book)

Electro-hydrulic Servo Actuator Diagnostics

  • 11th order dynamical model of a two-stage servo actuator system

  • Two parametric faults were studied

      • Increased friction between spool and sleeve à wear, silting, oil

      • Degradation of the armature permanent magnet à temperature

  • Single and simultaneous faults

  • High noise

  • Publications: PHM

Fault Detection in Electric Motors

  • A project in collaboration with Western Michigan University

  • Detect inter-turn short circuit (ITSC)

  • Early diagnosis is critical (1 second data)

  • ITSC fault causes overheating result in catastrophic failure

  • Noise in sensor reading

  • 100% accuracy was achieved

Embedding Dimension Diagnostic Method

  • How can we measure a system dimensionality?

  • A novel method was developed during my doctoral work

  • Based on the information in the embedding dimension of a given signal

  • Capable of distinguishing various nonlinear response within the system (chaos, multi-periodic, etc.)

  • Applied to gear-train setup to identify cracks

NovaVent Emergency Ventilator

  • Worked with Villanova team lead by Dr. Nataraj collaborating with local companies and hospitals to develop Novavent, a low-cost ventilator for the treatment of COVID-19 patients

  • News: Forbes| kyw | MediaRoom | KeystoneEdge

  • Learn More

Phase Space Topology Family of Methods

  • Industrial machinery applications reported nonlinear phenomena such multi-periodic, quasi periodic and chaotic that were originating from defects or due to their nonlinear nature.

  • How can we characterize system response and get insight using nonlinear dynamics?

  • A novel family of methods was developed during the my doctoral work

  • Based on extracting information from the phase space domain

  • Generalized on various range on real mechanical and electrical systems

  • A patent was filled

  • Publications: VETOMAC| JVET | Grant

Ranked Recurrence Diagnostic Method

  • Based on integrating mutual information and recurrence quantification analysis

  • Recurrence plots reveals all the times when a dynamic system visits roughly the same area in the phase space (repeating different kinds of behavior)

  • Publications: JSEA

Transfer Learning for Nonlinear Pendulum

  • Develop a transfer learning approach to keep the fault identifier updated when system changes

  • Structural change in the system

          • Large amount of data before structural change

          • Limited data after structural change

  • Change in operation scenario

          • Large amount of data for previous operation scenario

          • Limited data for new operation scenario