Resumo:
Currently, several welding techniques join metallic materials, with solid-state welding, such as Rotary Friction Welding (RFW), standing out. This method generates heat and plastic deformation through the rotation of parts under compression, creating a bond at temperatures below the melting point, resulting in reduced distortions and residual stresses. Among the RFW variants, Inertia Friction Welding (IFW) and Continuous Drive Friction Welding (CDFW) stand out, differing in the method of generating frictional energy. Although CDFW is a proven process, it presents challenges such as uneven heating and low material diffusion in the heat-affected zone (HAZ), affecting roughness and the coefficient of friction. Traditional RFW quality control methods use destructive testing, which does not allow for real-time testing. As an alternative, monitoring vibration signals during the process provides information about the thermal, metallurgical, and deformation phenomena occurring. Although studies have already explored the use of vibration signals for quality diagnostics, there is still no specific approach applied to quality control in CDFW. This work aims to fill this gap by using vibration signals to monitor the quality of CDFW in cylindrical A36 carbon steel parts. For this, the Design of Experiments (DOE) was adopted, with experiments based on a factorial matrix. The vibration signals, acquired by an accelerometer, were normalized and filtered, and later analyzed using the Fast Fourier Transform (FFT) and Short-Time Fourier Transform (STFT). These techniques allowed the identification of the time-frequency-energy characteristics of the signals. In addition, a method of segmentation and decomposition of raw signals was proposed using Empirical Mode Decomposition (EMD), with statistical analysis of the resulting intrinsic mode functions (IMFs). A two-factor Analysis of Variance (ANOVA) correlated the signal data with process quality parameters, such as flash symmetry, axial shortening, thermomechanically affected zone (TMAZ) width, and the average grain diameter of the TMAZ. The results confirmed that changes in process parameters significantly affect vibration signals, allowing their use to characterize weld quality. This innovative approach not only enables rapid quality verification but also contributes to automation and advanced monitoring in Industry 4.0.
