Predictive maintenance is the current maintenance philosophy in the manufacturing industry. Vibration Monitoring of machines and the subsequent analysis of the recorded spectrum (Vibration Spectrum Analysis) is the latest reliable technique for achieving greater reliability on equipment.
Vibration monitoring is the most reliable method of evaluating the general condition of a rotating machine. The spectrum analysis provides clear and measurable information, and allows the engineer to identify the condition of the machine. The method is applied by using special sensors placed on machines that measure acceleration, velocity and displacement. Spectrum analysis receives the signal from the sensors and processes it by identifying the frequencies displayed using digital filters and signal transformations with Fourier analysis (FFT analysis), Wavelet analysis and Hilbert-Huang analysis (HHT analysis) in computer software. The use of each transformation depends on the problem we seek to investigate in each machine.
When a machine element is damaged or fails, a specific frequency will be stimulated in the spectrum by displaying greater value. In fact, each spectrum of the rotating machine produces vibrations with distinct characteristics in specific frequencies. These vibrations can be measured and compared to benchmarks in order to detect and make diagnosis of the failure.