A computerized Fast Fourier Transform system has been used to analyse vibration measurements sampled from a 100-ton electric arc furnace. It has been the hypothesis that patterns in the vibration data would correlate to specific events in the electric arc furnace during the melting process. The theory was that the vibration patterns would be most powerful in multiples of 50 Hz, as the electric arcs operate with a frequency of 50 Hz. It was concluded that the multiples of 50 Hz were dominant. Investigation was made regarding how the amount of scrap affects the amplitude in the vibrations after a defined amount of energy input. It was confirmed that the intensity of the vibrations decreased with increasing mass and basket volume. Another discovery was the M-shaped pattern at the beginning of the melting process. This pattern was statistically analysed. It was found that 71 % of the 41 charges showed an M-shaped correlation. The appearance of this M-shape was analysed regarding power usage, and steel-type. It was concluded that the steel type affected the appearance of the M-shape. The occurrence of flat-bath was also investigated. The theory was that the vibration data would be fairly constant with the occurrence of flat-bath. It was discovered that the vibration data experienced a somewhat constant behaviour towards the end of the melting process in approximately 57 % of the 41 charges. Difficulties were encountered regarding detection of patterns, and correlating them to specific conditions, due to the many parameters that affect the vibration measurements from charge to charge. It was concluded that vibration analysis are unlikely to be used as an absolute way to foresee every event in the electric arc furnace during each charge. But can rather serve as a statistical tool, upon which decisions of how the melting process should be conducted could be based on.