Split-spectrum processing: analysis of polarity threshold algorithm for improvement of signal-to-noise ratio and detectability in ultrasonic signals

Shankar, P.M.; Karpur, Prasanna; Newhouse, V.L.; Rose, Joseph L.

doi:10.1109/58.16976

Cited by 59 publications

(31 citation statements)

References 21 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Other methods, such as the split spectrum processing technique (Bilgutay et al 1989;Shankar et al 1989) as well as the more general class of synthetic frequency diversity algorithms (Ericsson 1994) also show some potential. Time-averaged mean and mean-squared values were effective when the back surface response was relatively strong but were relatively ineffective for low amplitude signals (Hargreaves 1988).…”

Section: Ultrasonic Classification Of Centrifugally Cast Stainless mentioning

confidence: 99%

“…Time-averaged mean and mean-squared values were effective when the back surface response was relatively strong but were relatively ineffective for low amplitude signals (Hargreaves 1988). However, split-spectrum based polarity thresholding (Shankar et al 1989), either alone or in combination with other processing techniques such as minimization (Newhouse et al 1985) appears to significantly improve detectability in CASS specimens. A maximum entropy model of spectral analysis also had limited success (Hargreaves 1988).…”

Section: Ultrasonic Classification Of Centrifugally Cast Stainless mentioning

confidence: 99%

See 1 more Smart Citation

Ultrasonic Characterization of Cast Austenitic Stainless Steel Microstructure: Discrimination between Equiaxed- and Columnar-Grain Material ? An Interim Study

Ramuhalli¹,

Good²,

Diaz³

et al. 2009

View full text Add to dashboard Cite

Executive SummaryUltrasonic nondestructive evaluation (NDE) and inspection of cast austenitic stainless steel (CASS) components used in the nuclear power industry is neither as effective nor reliable as is needed. With current ultrasonic methods these limitations are in large part due to the detrimental effects of wavemicrostructure interactions on the interrogating ultrasonic beam and interference that results from ultrasonic backscatter. The root cause of these phenomena is the coarse-grain microstructure inherent to this class of materials. Some ultrasonic techniques are found to perform better for particular microstructural classifications and this has led to the hypothesis that an ultrasonic inspection can potentially be optimized for a particular microstructural class. For optimized ultrasonic technique selection, methods will be needed to reliably classify the microstructure in-situ, which is then used to guide the selection and optimization of the inspection. This document summarizes scoping experiments that investigate potential in-situ ultrasonic methods for classification and/or characterization of the material microstructures in CASS components, when making measurements from the outside surface of a pipe.The focus of this preliminary study was to evaluate ultrasonic measurement methods to determine if responses from different known microstructures can be differentiated and hence if in-service characterization of cast austenitic stainless steel (CASS) is potentially feasible. On the basis of an initial literature evaluation, two ultrasonic parameters, (i) the time-of-flight ratio between shear and longitudinal waves and (ii) the attenuation for normal incidence longitudinal waves, were selected for investigation. Scoping experiments were performed to determine the ability of these measured parameters to discriminate between different microstructures in CASS components. The objective was to determine if a more thorough and staged exploration would be justified in progressing toward the real-time characterization of CASS for use as feedback to optimize current or new ultrasonic in-service inspection methodologies. With this objective in mind, measurements were restricted to techniques that potentially should be robust if carried forward to an eventual field implementation.The first parameter investigated was a time-of-flight ratio of a normal incidence shear wave to that of a normal incidence longitudinal wave (TOFRSL). The ratio removes dependency on component thickness which may not be accurately known or reported in the field. The second parameter was the attenuation of a normal incidence longitudinal wave. The selected CASS specimens used for the experimental study were five equiaxed-grain material samples and five columnar-grain material samples, and these were used for a two-class discrimination problem.TOFRSL estimates and a threshold algorithm classified all 10 material samples correctly and indicated a potentially reliable and robust technique. Qualitative longitudinal wave attenuation estimate...

show abstract

Section: Ultrasonic Classification Of Centrifugally Cast Stainless mentioning

confidence: 99%

Section: Ultrasonic Classification Of Centrifugally Cast Stainless mentioning

confidence: 99%

Ultrasonic Characterization of Cast Austenitic Stainless Steel Microstructure: Discrimination between Equiaxed- and Columnar-Grain Material ? An Interim Study

Ramuhalli¹,

Good²,

Diaz³

et al. 2009

View full text Add to dashboard Cite

show abstract

“…In order to enhance SNR, researchers have developed many signal processing methods to suppress grain noise. The split spectrum processing [4][5][6] and wavelet transform de-noising methods [7][8][9] are widely researched techniques but show limited improvement [10]. If we assume that the ultrasonic properties of the grains are unaffected by ageing [11], grain noise is coherent and an opportunity then exists to use a structural health monitoring (SHM) approach in which a baseline reading is subtracted from the current reading, thus removing the grain noise and producing a residual signal which reveals the presence of defects.…”

Section: Introductionmentioning

confidence: 99%

Feasibility and Reliability of Grain Noise Suppression in Monitoring of Highly Scattering Materials

Liu

Pamel

et al. 2017

J Nondestruct Eval

View full text Add to dashboard Cite

A feasibility study on grain noise suppression using baseline subtraction is presented in this paper. Monitoring is usually done with permanently installed transducers but this is not always possible; here instead monitoring is conducted by carrying out repeat C-scans and the feasibility of grain noise suppression by subtracting A-scans extracted from the C-scans is investigated. The success of this technique depends on the ability to reproduce the same conditions for each scan, including a consistent stand-off, angle, and lateral position of the transducer relative to the testpiece. The significance of errors are illustrated and a 3D cross correlation is used which enables the same lateral position to be located within successive C-scans. The experimental results show that a noise reduction of around 15 dB is obtained after baseline subtraction, which will significantly improve the defect detection sensitivity. In practice however, successive C-scans may be conducted at different temperatures and with different transducers of similar specifications but a varying frequency response. Compensation techniques to reduce the impact of such variations are then presented and their effectiveness is verified experimentally. It is shown that it is feasible to obtain an overall improvement of around 10 dB in the signal to noise ratio via baseline subtraction, where a temperature difference of up to 10 • C and a peak frequency shift of as much as ±250 kHz from a baseline value of around 7 MHz can be tolerated. However, this improvement was obtained in laboratory conditions with no changes to the surface of the specimen due to oxidation or corrosion. It is shown that differences in tem-B Yuan Liu y.liu14@imperial.ac.uk 1 NDE Group, Department of Mechanical Engineering, Imperial College London, Exhibition Road, London SW7 2AZ, UK perature and transducer frequency response are more difficult to compensate for than changes in test geometry and position.

show abstract

“…Understanding these noise signals is essential to improving the detection process. For example, there exist algorithms [1][2][3] that have been developed specifically to reduce the ultrasonic grain noise. These models use signal processing techniques such as split-spectrum processing, spatial averaging, and bandpass filtering to enhance the signal-to-noise ratio.…”

Section: Introductionmentioning

confidence: 99%

Predicting ultrasonic grain noise in polycrystals: A Monte Carlo model

Yalda

Margetan

Thompson

1996

The Journal of the Acoustical Society of America

View full text Add to dashboard Cite

A Monte Carlo technique is described for predicting the ultrasonic noise backscattered from the microstructure of polycrystalline materials in a pulse/echo immersion inspection. Explicit results are presented for equiaxed, randomly oriented aggregates of either cubic or hexagonal crystallites. The model is then tested using measured noise signals. Average and peak noise levels and the distribution of the noise voltages are studied as the density of grains changes. A Monte Carlo technique is described for predicting the ultrasonic noise backscattered from the microstructure of polycrystalline materials in a pulse/echo immersion inspection. Explicit results are presented for equiaxed, randomly oriented aggregates of either cubic or hexagonal crystallites. The model is then tested using measured noise signals. Average and peak noise levels and the distribution of the noise voltages are studied as the density of grains changes.

show abstract

Split-spectrum processing: analysis of polarity threshold algorithm for improvement of signal-to-noise ratio and detectability in ultrasonic signals

Cited by 59 publications

References 21 publications

Ultrasonic Characterization of Cast Austenitic Stainless Steel Microstructure: Discrimination between Equiaxed- and Columnar-Grain Material ? An Interim Study

Ultrasonic Characterization of Cast Austenitic Stainless Steel Microstructure: Discrimination between Equiaxed- and Columnar-Grain Material ? An Interim Study

Feasibility and Reliability of Grain Noise Suppression in Monitoring of Highly Scattering Materials

Predicting ultrasonic grain noise in polycrystals: A Monte Carlo model

Contact Info

Product

Resources

About