LOW-COST DEVICE FOR FAULT DIAGNOSIS IN BEARINGS BASED ON THE HILBERT-HUANG TRANSFORM

In order to monitor the condition of machinery complex industrial environments, high-cost equipment is required for signal acquisition and processing. However, low-cost sensor nodes with high processing capability are a potential solution to improve diagnostic systems. This paper presents a low-cost device for fault diagnosis based on the vibration response in rotating machines with the implementation of the Hilbert-Huang transform (HHT) analysis to extract the main characteristics of the signal. HHT, used to analyze non-linear and non-stationary signals, incorporates an Empirical Mode Decomposition (EMD) process. Processing is carried out in an embedded system to acquire vibration response data and extract signal characteristics that allow condition monitoring. As a result of local processing in the vibratory measurement device in an embedded system, it is achieved to decompose the signal in order to the characteristic failure in the bearing and transmit the alarm to a hub. This eliminates the need for a central diagnostic system and reduces the total cost of the system. Keywords: vibration detection; fault diagnosis; empirical mode decomposition (EMD); local processing, embedded systems.