Early fault detection of three-phase induction motor
Induction motors are most widely used electrical machines for industrial, domestic and commercial applications, due to their robustness. Induction motors are undoubtedly reliable but we cannot avoid the possibility of failure. Several faults affect the efficiency and life of an Induction M
2025-06-28 16:32:17 - Adil Khan
Early fault detection of three-phase induction motor
Project Area of Specialization Artificial IntelligenceProject SummaryInduction motors are most widely used electrical machines for industrial, domestic and commercial applications, due to their robustness. Induction motors are undoubtedly reliable but we cannot avoid the possibility of failure.
Several faults affect the efficiency and life of an Induction Motor. One of the most widely occurring faults is bearing fault. Estimating the remaining useful life (RUL) of a bearing gives operators an efficient tool in decision making by quantifying how much time is left until functionality is lost.
Project ObjectivesThe objectives of this project is to successfully predict RUL (Remaining Useful Life) of induction motor with the help of bearings’ vibrational data. The two core parts of our project are,
- Detection of anomaly from the vibrational data of bearings under test.
- RUL (Remaining Useful Life) prediction of induction motor using LSTM based models.
For this project, we first acquire high speed data from NI-FPGA module, the acquired data is later sent to the neural network running simultaneously with the FPGA module and processing the incoming data. With the help of LSTM models, we predict the RUL by detecting any anomalous behavior in the analyzed data.
Benefits of the Project1) Predictions about when the equipment is likely to fail.
2) Significant decrease in unplanned downtime.
3) Reduced maintenance cost.
Technical Details of Final Deliverable- Development of a high-speed data acquisition system.
- A pre-trained machine learning (recurrent deep neural net) model to predict machine failure using time-series data from edge node.
| Item Name | Type | No. of Units | Per Unit Cost (in Rs) | Total (in Rs) |
|---|---|---|---|---|
| Total in (Rs) | 50000 | |||
| Single Phase Motor | Equipment | 1 | 7500 | 7500 |
| Bearings with housing | Equipment | 5 | 400 | 2000 |
| Bearings simple | Equipment | 20 | 60 | 1200 |
| Shafts | Equipment | 5 | 1000 | 5000 |
| Base | Miscellaneous | 1 | 8000 | 8000 |
| Jetson Nano | Equipment | 1 | 18000 | 18000 |
| SD Card 64GB | Miscellaneous | 1 | 1200 | 1200 |
| Thermocouple | Miscellaneous | 1 | 200 | 200 |
| Encoder | Equipment | 1 | 3000 | 3000 |
| Accelerometer | Equipment | 1 | 2500 | 2500 |
| Cooling Fan | Equipment | 1 | 1400 | 1400 |