Every machine and structural component functions under dynamic or variable loading in real life. Cracks or other damage may result from the various loading situations are encountered to be most dangerous which indirectly leads to failure of the component. The physical characteristics of the structure, particularly its rigidity and flexibility, may alter when the fracture first appears. Mode shapes and natural frequencies will be directly altered by the dynamic changes. Stainless steel shaft is considered as the specimen to evaluate the natural frequencies using experimental (Natural Frequency Test) and finite element analysis. An evolutionary system called the BAT algorithm is employed in this work to determine the site of damage based on the bats’ echo locations. The BAT algorithm considers the first three natural frequencies obtained through testing and FEA methodologies as input. Dynamic analysis of the fractured shaft is used to produce a data pool, which is then trained using the suggested methods to determine the depth and length of the crack. FEA and experimental analysis are compared with the results of the suggested technique. The percentage error fall within realistic range.