In the field of machine prognostics, vibration analysis is a proven method for detecting and diagnosing bearing faults in rotating machines. One popular method for interpreting vibration signals is envelope demodulation, which allows a technician to clearly identify an impulsive fault source and its severity. However incipient faults -faults in early stages - are masked by in-band noise, which can make the associated impulses difficult to detect and interpret. In this thesis, Wavelet De-Noising (WDN) is implemented after envelope-demodulation to improve accuracy of bearing fault diagnostics. This contrasts the typical approach of de-noising as a preprocessing step. When manually measuring time-domain impulse amplitudes, the algorithm shows varying improvements in Signal-to-Noise Ratio (SNR) relative to background vibrational noise. A frequency-domain measure of SNR agrees with this result. / Includes bibliography. / Thesis (M.S.)--Florida Atlantic University, 2014. / FAU Electronic Theses and Dissertations Collection

Orientador: Alberto Luiz Serpa / Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Engenharia Mecânica / Made available in DSpace on 2018-08-23T13:16:44Z (GMT). No. of bitstreams: 1 Ferlin_AmadeuSoares_M.pdf: 11538250 bytes, checksum: df676e0c073182186a89891ff432c041 (MD5) Previous issue date: 2013 / Resumo: Esta dissertação discute aspectos do monitoramento do processo de soldagem por Friction Stir Welding (FSW) de chapas de alumínio utilizando acelerômetros fixados à superfície das chapas. Os sinais adquiridos durante o processo de soldagem foram processados utilizando dois métodos diferentes - a Transformada de Fourier de Tempo Curto (STFT) e a Transformada Wavelet - que foram capazes de detectar e localizar falhas criadas através da remoção de material nas chapas. A aplicação destas técnicas nos sinais adquiridos durante o processo de soldagem permitem a identificação de mudanças no comportamento destes sinais através de resultados gráficos contendo informações de frequência e de tempo. Através destas informações é possível determinar o momento em que as mudanças de comportamento ocorreram e, consequentemente, prever as posições correspondentes à ocorrência da falha no respectivo cordão da solda / Abstract: This dissertation discusses aspects of the Friction Stir Welding (FSW) process monitoring of aluminum plates using accelerometers attached to the plates surface. The signals acquired during the welding process were processed using two different methods - Short Time Fourier Transform (STFT) and Wavelet Transform - that were able to detect and locate failures created by removing material of the plates. The application of these techniques on the signals acquired during the welding process allows the identification of changes in the signals pattern by graphical results containing frequency and time information. Thus it is possible to determine the moment at which the changes in the behavior occurred and therefore predict the positions corresponding to the occurrence of the failure in the weld line / Mestrado / Mecanica dos Sólidos e Projeto Mecanico / Mestre em Engenharia Mecânica

A dissertation submitted to the Faculty of Science, University of the Witwatersrand, in fulfilment of the requirements for the degree of Master of Science. Johannesburg, 2015 / The Bushveld Complex (BC) is the largest known layered intrusion. This suite of rock crop out in northern South Africa to form the Western, Eastern and Northern Limbs. Most research carried out focuses on the mineralized horizons in the Rustenburg Layered Suite (RLS) of the BC. This study presents a large database of wireline geophysical logs across a substantive part of the stratigraphy of the RLS. These consist of density and magnetic susceptibility datasets sampled at 1 cm. The major lithologies of the RLS intersected in the boreholes presented are gabbro, gabbronorite, norite and anorthosite whose density histograms reveal that they are predominantly normally distributed, with density averages of 2.86-2.91 g/cm3. The lithologies consist of mainly two minerals, pyroxene and plagioclase. In general, the average density increases with an increase in pyroxene. The distribution of the magnetic susceptibility for these lithologies has a large variation from SI to 13.2 SI, which is typical of layered intrusions. Susceptibility distributions are also multi-modal, asymmetric and not normally distributed, which makes the average magnetic susceptibilities less representative of the lithologies. Cross-correlation plots between density and magnetic susceptibility for several boreholes show that the above-mentioned lithologies form clusters (circular to elliptical), which typically overlap. This has been further investigated using k-means classification, to automatically detect these clusters in the cross-correlation plots and to compare these with those created by lithologies. The comparison shows some degree of correlation, implying that physical properties can be used to identify lithologies. This is particularly true for the Eastern Limb. However the classification has not been effective in all of the boreholes and often becomes complicated and an inaccurate representation of lithology log. This occurs in boreholes in which there is an overlap in the physical properties of the abovementioned lithologies. Analysis on the density and magnetic susceptibility data has also been carried out using wavelet analysis at individual locations across the BC. This has revealed multi-scale cyclicity in all of the boreholes studied, which is attributed to subtle layering created by variations in modal proportions between plagioclase and pyroxene. In addition to this, since layering is generally ubiquitous across layered intrusions, this cyclicity can be assumed to be present across the entire BC. This technique may become increasingly important should the cyclicity in physical property data correlate with reversals in fractionation trends since this may suggest zones of magma addition, whose thickness or III volumes can be quantified using wavelet analysis. This could be an important contribution since the current perspective on magma addition in the RLS is that four major additions have formed this 8 km thick suite of rocks, as opposed to smaller periodic influxes of magma. Wavelet-based semblance analysis has been used to compare the wavelengths at which the cyclicity occurs across boreholes. A comparison of wavelengths of this cyclicity shows that boreholes in the northern Western Limb show positive correlation in the density data at wavelengths >160 m and 20-60 m, while those further south show correlations at wavelengths of 120-200 m and 60-80 m. Boreholes of the Eastern Limb show positive correlation in the density and magnetic susceptibility data at wavelengths of 10-20 m, 20-30 m and 5m. These positive correlations across boreholes in density and magnetic susceptibility respectively, may imply that cyclicity may be produced by a chamber-wide process for several kilometres of the BC.