Health Monitoring of Three Phase Induction Motor Using Current Signature Analysis and Deep Learning Algorithms

  Induction motors possess one of the most important roles industrially and commercially. The developing faults in the motors can become catastrophic, if remain unanalyzed. This project presents an effective and novel solution to diagnose the major mechanical faults at the early possibl

Project Title

Health Monitoring of Three Phase Induction Motor Using Current Signature Analysis and Deep Learning Algorithms

Project Area of Specialization

Artificial Intelligence

Project Summary

Induction motors possess one of the most important roles industrially and commercially. The developing faults in the motors can become catastrophic, if remain unanalyzed. This project presents an effective and novel solution to diagnose the major mechanical faults at the early possible stage by utilizing two efficient condition monitoring techniques
to effectively deploy the strategies for the predictive maintenance. Primarily it employs the MCSA (Motor Current Signature Analysis), in which the faults are located by the spectral analysis of the particular harmonic components in the line current at specific characteristic frequencies generated by specific faults as the unique rotating flux. Deep Learning has also been utilized, which assesses severity of the fault and operating condition of machine. The induction machine is prepared with a unique design for the implementation of faults. Specifically in this project we’ve analyzed bearing faults that are detected and localized in the results along with the severity assessment of the operational condition due to induced faults.

Project Objectives

• The goal of this project id to design and develop a condition monitoring system for a 3-
  phase induction motor using current signature analysis and vibration analysis using deep
  learning.
• To design and develop condition monitoring system using current sensors and accelerometers.
• To analyze current signatures using spectral monitoring techniques such as FFT and STFT.
• To create a machine learning model of vibration data to for fault prediction using deep neural network.

Project Implementation Method

• Development of embedded monitoring system using FPGA + Processor, current sensors and accelerometers with high-rate data converters.
• Development of embedded software for high-speed data acquisition.  
• Induction motors test-bed development for testing and validating results.
• Data-logging and compilation for post-processing.

Benefits of the Project

• Enhancement of industrial maintenance strategy to predictive maintenance.
• Reduce downtime and costs of maintenance
• Efficient and effective monitoring of operational motors

Technical Details of Final Deliverable

• A complete embedded monitoring system with high-speed data acquisition of current and vibration data.
• Embedded fault monitoring and detection mechanism using current signature analysis.
• Data-logging of motor parameters (current and vibration data) for building deep learning models.
• A real-time inference mechanism on vibration data to predict remaining useful life (RUL) of the machine.

Final Deliverable of the Project

HW/SW integrated system

Type of Industry

Education

Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Industry, Innovation and Infrastructure

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com