Frequency response functions calculations in vibration tests are strongly influenced by the type of excitation signal used to drive the test specimen. Random signals are frequently used although they present filter leakage problems and require longer test times than pseudorandom signals. Pseudorandom signals are periodic in the time window and do not present filter leakage problems. When measurement noise is not a major concern, pseudorandom excitation signals require only one input output data sample, and hence, the test times decrease significantly when compared to random signals which often require an average of 100 or more sets of data. This paper shows the efficiency of pseudorandom over random excitation in experimentally determining a system's FRF once a given baseband is selected. The FRF's frequency resolution is controlled by the Δf selected and is the same for both excitation methods. The paper shows that once the startup transients have vanished, the pseudorandom test requires only one sample period. Therefore, valuable test time can be saved in order to meet specific test requirements.
展开▼
