This Master Thesis explores the application of Fault Detection and Diangosis (FDD) algorithm on a hydraulic crane setup. Extended Kalman Filter (EKF) and Spectral analysis were developed and tested, to see the role and the relationship between the leak magnitude and the chance of it being identified and possibly isolated. Laboratory experiments were conducted, followed by verification of the mathematical model and further offline model and signal based simulations. The findings in the research illustrate how the magnitude of the leak influences whether or not the fault can be detected and isolated. Choosing a threshold constant for the leak coefficients played a crucial role in determining the faults. The selected threshold proved to hold for leaks of level 4 and higher, while being inaccurate for lower levels. The results proved the superiority of the EKF over the spectral analysis. After further insight, it was concluded that spectral analysis is a very quantitative method that requires a lot of pre-analyzing for determining the faults, while EKF provides trustworthy results over the artificial leak level of 4 (0.1293 – 0.9853 l/min) at 25ºC. Further research is prompt of re-thinking the EKF structure, with a possibility of implementing the leaks as a state with leak coefficients as an updating parameter estimates.