HUMS2023 Data Challenge Result Submission
Team Name: Deakin-MLDS
Team Members: Dhiraj Neupane, Lakpa Dorje Tamang, Ngoc Dung Huynh, Mohamed Reda
Bouadjenek and Sunil Aryal
Institutions: School of IT, Deakin University, Waurn Ponds, VIC, Australia
Publishable: Yes
1. Summary of Findings
We implemented a simple method for early detection in this research. The implemented methods are
plotting the given mat files and analyzing scalogram images generated by performing Continuous Wavelet
Transform (CWT) on the samples. Also, finding the mean, standard deviation (STD), and peak-to-peak (P2P)
values from each signal also helped detect faulty signs. We have implemented the autoregressive integrated
moving average (ARIMA) method to track the progression.
In summary, the earliest signal distortion was seen on the file Day022_Hunting_SSA_20211209_124241.mat
for sensors Ip-1 and RR-4. For sensor RF-2, the earliest signal fault was detected in
Day022_Hunting_SSA_20211209_141330.mat. The files Day027_Hunting_SSA_20220118_111018.mat and
Day027_Hunting_SSA_20220118_111317.mat show clear faulty patterns at all four channels. All the
observations state that the fault signals were first seen on Day 22, file
Day022_Hunting_SSA_20211209_124241.mat, in channels 1 and 4. For three out of four sensors' data,
except for RL-3, the fault was detected on Day 22, which was not seen further until Day 27. On Day 27, files
Day027_Hunting_SSA_20220118_111018.mat and Day027_Hunting_SSA_20220118_111317.mat, the
crack was observed clearly. The progression curve, drawn after finding five subsequent values of P2P after
Day 27 using ARIMA, shows the abrupt increase in the values after Day 27.
Table 1 Summary of Analysis Results
