Condition monitoring of a turbine flowmeter using a deconvolution method

In this paper, time-frequency distribution (TFD) is calculated using a deconvolution method to represent the collected signals for an operating normal and faulty flowmeters for the objective of detecting any existing flaws. The final result is obtained as a function of time and frequency. The deconvolution method can be defined as the process of recovering the input to some system from its input given information about that particular system. This type of approach is used in the field of time-frequency analysis to enhance the resolution of the time-varying spectrum. In this experiment, a fault is introduced by slightly cutting a part of three of the rotor blades to simulate erosion then data is collected for the two cases representing the normal and faulty flowmeters.