Dataset Repository |
Acoustic dataset was collected on a single stage reciprocating type air compressor placed at the Department of Electrical Engineering Workshop. Specifications of the air compressor are as follows:
Dataset comprises of 8 states which includes healthy state and 7 faulty states namely Leakage Inlet Valve (LIV) fault, Leakage Outlet Valve (LOV) fault, Non-Return Valve (NRV) fault, Piston ring fault, Flywheel fault, Rider belt fault, and Bearing fault. To collect recordings from all these states, faults were seeded into the air compressor. This Dataset can be downloaded from following link. If using Dataset, please cite paper given below.
Please cite the following paper:
The text format is:
Nishchal K. Verma, R. K. Sevakula, S. Dixit and A. Salour, Intelligent Condition Based Monitoring using Acoustic Signals for Air Compressors,
IEEE Transactions on Reliability, vol. 65, no. 1, pp. 291-309, 2016.
This dataset has also been used by MathWorks for extracting VGGish features.
This dataset is a collection of feature matrices formed by analysisng aforementioned dataset in two domains i.e. frequency and time-frequency. Features represent the signal at a much lower dimension. In total there are 7 such matrices. More details please refer paper mentioned below.
Please cite the following paper:
The text format is:
R. Thriukovalluru, S. Dixit, R. K. Sevakula, Nishchal K. Verma and A. Salour, Generating Feature Sets for Fault Diagnosis using Denoising Stacked Auto-encoder,
IEEE International Conference on Prognostics and Health Management, Canada USA, pp. 1-7, June 2016. (Accepted)
We took 8 different training datasets, i.e total 256 acoustic recordings for each state namely Healthy, LIV were taken over a long period of time. Collected dataset was analysed and features were extracted in two domains i.e. frequency and time-frequency. Features represent the signal at a much lower dimension. In total there are 7 such matrices. More details please refer paper mentioned below.
Please cite the following paper:
The text format is:
R. K. Sevakula, A. Shah and Nishchal K. Verma, Data Preprocessing methods for Sparse Auto-encoder based Fuzzy Rule Classifier
, IEEE Workshop on Computational Intelligence: Theories, Applications and Future Directions (IEEE WCI 2015), India, pp. 1-6, Dec. 14-17 2015.
The entire experimentation was performed with 3-AxisCNC EMCO Concept Mill 105. HSS twist drill bit of diameter 9 mm was used for drilling holes in the work piece made of Mild steel. For extensive experimentation, given a drill bit state, for each pair of varying feed rates and cutting speed combinations, a single vibration recording of 8 seconds was taken. Feed rate was varied as 4 mm/min, 8 mm/min and 12 mm/min, and Cutting speed was varied as 160rpm, 170rpm, 180rpm, 190rpm and 200rpm; giving a total of 15 combination pairs.
Please cite the following paper:
The text format is:
Nishchal K. Verma, R. K. Sevakula, S. Dixit and A. Salour, Data Driven Approach for Drill Bit Monitoring
, IEEE Reliability Magazine, pp. 19-26, Feb. 2015.