Shock and Vibration / 2019 / Article / Tab 6

Research Article

Ontology-Based Noise Source Identification and Key Feature Selection: A Case Study on Tractor Cab

Table 6

Implementation of SWRL rules for feature extraction.

Rule 1Feature 1 (ignition frequency check)
Checking whether the parotic noise signal is related to the ignition frequency
Input the ignition frequency (specific numerical value); if the rule matches, it is inferred that the parotic noise signals are related to a specific vibrating part. The feature of this correlation is represented by the “ignition frequency”

Rule 2Feature 2 (peak frequency check in the noise source)
Checking whether the peak frequencies in the noise source correspond to the peak frequencies in the vibration source
Rule 2.1 (specially for/applies to the engine and cab)
Rule 2.2 (specially for/applies to the air filter, muffler, and back panel)
Input the peak frequency (specific numerical value); if the rule matches, it is inferred that the parotic noise signals are related to a specific vibrating part. The feature of this correlation is represented by the “peak frequency”

Rule 3Feature 3 (correlation check and identification)
Calculating the correlation between the parotic noise signal and the location signal of the tractor and identifying the locations exhibiting a specific correlation (four different degrees of correlation)
Rule 3.1 (determining the degree of correlation)
Input the correlation coefficient (specific numerical value); if the rule matches, it is inferred that the parotic noise signals are related to a certain position. The feature of this correlation is represented by the “correlation”
Rule 3.2 (determining the degree of correlation and the locations)
Rule 3.2 is divided into four rules corresponding to Levels 1, 2, 3, and 4
Rule 3.2.1 (applies to Level 1)
Rule 3.2.2 (applies to Level 2)
Rule 3.2.3 (applies to Level 3)
Rule 3.2.4 (applies to Level 4)
Input the correlation coefficient (specific numerical value); if the rule matches, it is inferred that the parotic noise signals are related to a specific vibrating part. The feature of this correlation is represented by the correlation. The degree of correlation is Level 1, 2, 3, or 4