Establishment of spectral subtraction-based algorithm for experimental modal analysis under operating condition

Che Ku Eddy Nizwan, Che Ku Husin (2022) Establishment of spectral subtraction-based algorithm for experimental modal analysis under operating condition. PhD thesis, Universiti Malaysia Pahang (Contributors, Thesis advisor: Mohd Fairusham, Ghazali).

ir.Establishment of spectral subtraction-based algorithm for experimental modal analysis.pdf - Accepted Version

Download (463kB) | Preview


Experimental modal analysis (EMA) is one of the techniques used to identify the dynamic properties of a structure. In the EMA procedure, the input force and vibration response measurement for transfer function calculation is carried out under shut-down conditions to prevent unmeasurable forces induced by the operating machinery. Under the operating condition, the presence of unmeasurable forces causes an error in the transfer function calculation due to incomplete information of input forces. Therefore, output response from the measurement requires a filtering process to suppress the response from operating forces and retain only the response from the artificially induced force. A new filtering algorithm adopted from a spectral subtraction filter was developed to preserve the impact response in impact-based EMA. The basis of this filter was to identify the amount of ambient magnitude contained in the measured vibration spectrum and utilised the spectral subtraction filter to suppress the ambient features. An effective ambient magnitude parameter was introduced to improvise the gain factor in the spectral subtraction filter. This parameter was calculated based on resolved ambient vector in the direction of impact response at each frequency band. The direction of ambient response at each frequency band was estimated based on the reconstructed artificial ambient. The artificial ambient was reproduced based on pre-triggered data in measured vibration response. In this procedure, the number of pre-triggered data was set as equivalent to the post-triggered data, and the artificial ambient was reconstructed from the summation of Fourier series of the pre-triggered data with twice the original length. The artificial ambient was selected from the second half of the reconstructed signal. From the research observation, a higher consistency level with approximately less than 5% deviation for the input ambient response produce higher similarity in artificial ambient with respect to actual ambient. This finding correlated with the efficiency of the filtering process, whereby the input with a higher similarity artificial response suppresses a more accurate amount of ambient and retains more than 90% of the baseline features. Under the operating condition, the measured frequency response function (FRF) showed non-identical features at operating frequencies as compared to the baseline data. The utilisation of the filtering process based on the modified spectral subtraction filter had successfully restored the essential features in FRF and suppressed unwanted ambient consequences. The findings indicated that the modified spectrum subtraction filter could be utilised to extract the clean impact response under consistent operating conditions. It was suggested this method be applied prior to FRF calculation to enhance the EMA procedure under ambient operating forces.

Item Type: Thesis (PhD)
Additional Information: Thesis (Doctor of Philosophy(Mechanical Engineering)) -- Universiti Malaysia Pahang – 2022, SV: IR. DR. MOHD FAIRUSHAM BIN GHAZALI
Uncontrolled Keywords: pectral subtraction-based algorithm, experimental modal analysis
Subjects: T Technology > T Technology (General)
T Technology > TJ Mechanical engineering and machinery
Faculty/Division: Institute of Postgraduate Studies
Faculty of Mechanical and Automotive Engineering Technology
Depositing User: Mr. Nik Ahmad Nasyrun Nik Abd Malik
Date Deposited: 13 Mar 2023 02:13
Last Modified: 14 Sep 2023 06:57
Download Statistic: View Download Statistics

Actions (login required)

View Item View Item