This thesis investigates the degradation mechanisms of the Chamfering Unit used in carbon pultrusion process necessary for the production of the crimps for V236 wind turbine. The study analyzes how vibrations are generated during the operation and how their evolution reflects the health status of the unit. Combining physical insights with systematic data recording the work has the aim of estimating the RUL ( Remaining Useful Life). Frequency domain techniques like FFT and H-FFT are implemented in Matlab to extract relevant diagnostic features. In parallel a Python based acquisition support deployed on LattePanda platform and interfaced with an ESP32 microcontroller enabled machine state recognition.