Datasets
Standard Dataset
Vibration measurements generated with contactor switching using FBG
- Citation Author(s):
-
DanielBenetti
UTFPREduardoDureckUTFPR - Submitted by:
- Daniel Benetti
- Last updated:
- Mon, 06/05/2023 - 07:34
- DOI:
- 10.21227/22eb-8n60
- Data Format:
- License:
152 Views- Categories:
- Keywords:
Abstract
The dataset consists of measurements of different stages of degradation in low-voltage contactors used for industrial purposes. The measurements were obtained with fiber Bragg grating (FBG) sensors that detect the dynamic deformation generated in switching under different load conditions and internal components. The posted dataset was preprocessed and separated into two different events. The signal is segmented and reduced from the original measurement (separated into opening and closing). Furthermore, two sets of measurements were obtained. The first is used for training and validating the machine learning model, while the second consists of a test set obtained from the contactor reassembly for the final generalization test.
The use of dataset is composed of five states of degradation of the internal components, being separated into 1- Normal state (NS); 2- Two contacts in intermediate degradation stage (W2C); 3 - All contacts in critical degradation (CC); 4- Pressure springs in intermediate degradation stage (WS); 5- Pressure springs in critical degradation stage (CS). Each folder referring to the training and validation data of the model contains twenty-five measurements with three types of drive: 1- Induction motor 15 Hp; 2- Resistance bench; 3- No load. As for the test data, it is composed of one hundred measurements in the normal state (NS), twenty measurements with two contacts in an intermediate condition of degradation (W2C), and twenty measurements with the springs in the critical state (CS). Data can be grouped differently, extracting features for events separately on opening and closing or by load type.
Comments
need for training vibration for motor