-
Elevator Health Monitoring System using an Inertial Measurement Unit (May 2019 - February 2021)
Elevators are an essential amenity of any multistoried building nowadays. They are frequently used especially in commercial buildings, hotels, hospitals, and apartments. Of them, geared traction-type elevators are the most common. In this configuration, the passenger car is pulled using steel rope and pulley or sheave.
​
However, the elevator riding experience is largely dependent on the proper erection, installation, operation, and maintenance of the elevator. Moreover, Elevator performance deteriorates over time. Faulty motor, brake control equipment and car leveling mechanism, over-speed or under-speed, over travel, poorly functioning buffer, corroded guide rails or derailed car from the guide rails, uneven acceleration or retardation, etc. of an elevator can create an unpleasant and undesirable user experience. This thesis proposes the use of a low-cost Inertial Measurement Unit (IMU) for inspection and monitoring of elevator health with the help of the acceleration profile of the elevator journey in both vertical and lateral directions. It also compares the performance of a Savitzky-Golay filter and a Kalman Filter on the raw acceleration data for eliminating high-frequency noise while maintaining the actual shape of the profile. The profile in the vertical direction can infer about the health of the motor and brake, while the lateral acceleration profile can identify the source of undesirable vibrations. Finally, different elevator parameters are also extracted in this study to check the compliance with the local building regulations (Bangladesh National Building Code 2015).
​
This is a final year thesis supervised by Professor Dr. Md. Zahurul Haq and submitted to the Department of Mechanical Engineering, Bangladesh University of Engineering and Technology in partial fulfillment of the requirements for the degree of Bachelor of Science in Mechanical Engineering.
Video: Elevator Health Monitoring Using An Inertial Measurement Unit - Undergraduate Thesis Presentation
Video: The Double Servo Setup for the self-calibration of the orientation of the IMU sensor using Feedback Control
![Figure_comparison_sgfilter_AccX_5tog_M=1](https://static.wixstatic.com/media/5c899a_c74df371647749e78ff25bfbc615e418~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_c74df371647749e78ff25bfbc615e418~mv2.png)
![Figure_comparison_sgfilter_AccY_5tog_M=1](https://static.wixstatic.com/media/5c899a_69a3447681e746bf8b1039f9abb34308~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_69a3447681e746bf8b1039f9abb34308~mv2.png)
![Figure_comparison_sgfilter_AccZ_5tog_M=1](https://static.wixstatic.com/media/5c899a_69c37180c2eb4a8b860c69e63882465a~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_69c37180c2eb4a8b860c69e63882465a~mv2.png)
![Figure_comparison_sgfilter_AccX_5tog_M=1](https://static.wixstatic.com/media/5c899a_c74df371647749e78ff25bfbc615e418~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_c74df371647749e78ff25bfbc615e418~mv2.png)
Figure: S-G Filter comparison for different window sizes
![Figure_comparison_kalmanfiltered_AccX_qz](https://static.wixstatic.com/media/5c899a_5683c322382f40ad8e1e20fe201b1e86~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_5683c322382f40ad8e1e20fe201b1e86~mv2.png)
![Figure_comparison_kalmanfiltered_AccY_qz](https://static.wixstatic.com/media/5c899a_78e0715f860c4c789e70a3fb4b205e2f~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_78e0715f860c4c789e70a3fb4b205e2f~mv2.png)
![Figure_comparison_kalmanfiltered_AccZ_qz](https://static.wixstatic.com/media/5c899a_361b7d6aafec48d3ad1868a1d2f608f5~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_361b7d6aafec48d3ad1868a1d2f608f5~mv2.png)
![Figure_comparison_kalmanfiltered_AccX_qz](https://static.wixstatic.com/media/5c899a_5683c322382f40ad8e1e20fe201b1e86~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_5683c322382f40ad8e1e20fe201b1e86~mv2.png)
Figure: Kalman Filter comparison for different Process Noise Covariance
![SGvsKalman_AccX](https://static.wixstatic.com/media/5c899a_aee4eca66f434ebf916807aeecb64df8~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_aee4eca66f434ebf916807aeecb64df8~mv2.png)
![SGvsKalman_AccY](https://static.wixstatic.com/media/5c899a_b8701cd973734c86847ac7384eb7ff40~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_b8701cd973734c86847ac7384eb7ff40~mv2.png)
![SGvsKalman_AccZ](https://static.wixstatic.com/media/5c899a_89ff9172a21844b8960b43ea24b5927e~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_89ff9172a21844b8960b43ea24b5927e~mv2.png)
![SGvsKalman_AccX](https://static.wixstatic.com/media/5c899a_aee4eca66f434ebf916807aeecb64df8~mv2.png/v1/fill/w_980,h_463,al_c,q_90,usm_0.66_1.00_0.01,enc_avif,quality_auto/5c899a_aee4eca66f434ebf916807aeecb64df8~mv2.png)
Figure: Comparison between S-G filtered and Kalman filtered acceleration profile