Research Activities >>Ultra Sonic NDE >>
 
     
  Phased Array NDT
  Phased Array Imaging on Pipes for Fatigue Crack Imaging
  Single Element SAFT for improved detection of defects   
  Matched Filter based SAFT
     
  Guided Waves NDT
  Pipe Support Inspector using Circumferential Guided Waves
  Circumferential Guided Waves Simulations
  Storage Tank Floor Corrosion Detection and Imaging
  Guided wave Imaging of Complex Thin Ti Welds
  Time Reversal Techniques for Lamb Wave Imaging
     
  Electromagnetic Transduction
  EMATs based Guided Waves Applications
  EMF Transduction for non-conducting non- magnetic materials
     
  Time of Flight Diffraction Technique
  TOFD for Thin Structures
  TOFD for Complex Structures
  Simulation of TOFD using Ray Tracing
  Shear Wave TOFD (S-TOFD) Technique
  Spectral Element Method (SEM) Simulation of TOFD
  Point Source Correlation Technique (PSCT) for TOFD
  POD Models for TOFD
     
 
    Phased Array NDT :
     
    Phased Array Imaging of Fatigue Cracks in Pipes
 
 
  
   This work involves the experimental sizing of fatigue crack profiles that are initiated from artificially made circumferential starter notches in stainless steel (SS) pipes, of 169 mm outer diameter and 14.33 mm thickness, that were subjected to cyclic bending loads in a four point bending load arrangement using two NDE methods: (a) Phased Array Ultrasonic Technique, and (b) Alternating Current Potential Drop (ACPD) techniques.

The Phased array ultrasonic technique was successfully used to monitor and estimate crack growth in two SS pipes subjected to fatigue loading in a four point bending setup. The crack profiles estimated using Phased array, were also compared with the beach marks generated purposely on the fracture surface and the crack profiles estimated using ACPD technique and a good correlation was observed. Phased array data was found to oversize the crack size for the first pipe while ACPD technique data was closer to the beach mark data. But in the case of the second pipe, the phased array estimate of the crack profile matched better with the beach marks as compared to the ACPD data. A two-dimensional ray tracing model was developed to simulate the phased array experiments on pipes and the simulation results were compared with the experimental results. The results obtained from experiments confirm the validity of the simulations. The simulations were then extended to study the signals obtained for three typical crack configurations.


Measured size of the crack at the notch centre
Measured size of the crack at the notch centre
 
 

Cracked Pipe Sample Cross-section
Cracked Pipe Sample Cross-section
 
Fatigue Loading Setup
Fatigue Loading Setup
Crack Profile Imaging

Crack Profile Imaging

   
 

Reference: L. Satyarnarayan, DM. Pukazhendhi, K.Balasubramaniam,  C.V.Krishnamurthy, D.S. Ramachandra Murthy, “Phased Array Ultrasonic Measurement of Fatigue Crack Growth Profiles in Stainless Steel Pipes”, ASME Trans. Journal of Pressure Vessel Technology in press available online (2007)

     

 

 

 
   
Single Element SAFT for improved detection of defects
 

  
A SAFT algorithm was used to synthetically focus the beam off line to obtain the position of the defect. Since the position of the notch tip can be ascertained using SAFT and the thickness of the specimen is known, the difference between the specimen thickness and the position of the notch from the front wall gives the estimated size of the notch.  The experimental B-scan image of the large bottom surface notches (20%, 40%, 60% and 80% of the thickness) and the corresponding SAFT processed B-scan image is given in Figure 1.

The single element scanning feature was then applied on a 24.5 mm (98 % of thickness) bottom notch in the simulation domain to investigate if the technique could be successfully used to image and size near through wall notches also.
The 24.5 mm near through wall notch was simulated using FDTD and was imaged for single element phased array scan. It was seen that the front wall correction algorithm had to be applied to the defect image to separate the notch signal from the front wall echo. The SAFT algorithm was then successfully applied to image and size the defect. The B-scan images of the 24.5 mm (98 % of the thickness) near through wall notch before and after front wall correction and the application of SAFT processing for the FDTD simulated and the experimental cases is shown in Figure 2  respectively.
 
 

Figure 1.B-scan image of the bottom notches of four sizes and the SAFT processed image
Figure 1.B-scan image of the bottom notches of four sizes and the SAFT processed image
 
Figure 2.FDTD simulated B-scan image of 98 % of thickness near through

Figure 2.FDTD simulated B-scan image of 98 % of thickness near through
wall defect (24.5 mm)




 
Srivatsan V., Krishnan Balasubramaniam, and N. V. Nair, “Artificial Neural Network Based Algorithm for Acoustic Impact Based Nondestructive
Process Monitoring of Composite Products” AIP Conf. Proc. 657, 1651 (2003)
   
 
     
    Matched Filter based SAFT
 
 
  
  The image of the defect from a finite sized transducer is different from the image obtained using a point source and this image is the convolution of the spatio-temporal response characteristics of the transducer with the image of the defect.
The finite transducer characteristics can be expressed as a convolution of the transducers electrical characteristics and the Spatial Impulse Response (SIR) of the transducer in the given domain. The SIR is the impulse response of a finite transducer in the medium considered.

A matched filter based approach is presented, which takes into account the SIR of a finite sized phased array transducer, applied on B-scan images of bottom surface notches to perform the Synthetic aperture imaging.

In the matched filter approach we compute a template for each image reconstruction point, and calculate the reconstructed intensity at that point by carrying a weighted addition of the original image with weights specified by the template. This can be described as a correlation of the B-scan image using a pixel dependent mask. The mask chosen in this case is the model calculated received signals for a reflector positioned at the point considered. This leads to smooth reconstructions and increases the SNR in the image.The proposed algorithm was validated by comparing against the SAFT algorithm. A 1.5 mm diameter side drilled hole at a depth of 10 mm in a PH TOOL standard block (IOW Beam profile block: 7075-T6 Aluminum) was imaged using the single element electronic scan feature of the phased array. The B-scan obtained as well as the imaged results using the matched filter and the standard SAFT algorithm is shown in Figure 1. It was seen that the image reconstructed using the Matched filter compared well with the SAFT reconstructed image of the defect.


(d) The image reconstructed using the SAFT algorithm.

(d) The image reconstructed using the SAFT algorithm.
Comparison of Matched filter and SAFT algorithm.

 
 

a) Schematic of the scan setup
a) Schematic of the scan setup
 
(b) B-scan taken using the single element scan feature of the phased array system
(b) B-scan taken using the single element scan feature of the phased array system.
(c) The image reconstructed using the matched filter.

(c) The image reconstructed using the matched filter.        

   
 

Reference:1. L. Satyanarayan, A. Muralidharan, C.V. Krishnamurthy, K. Balasubramaniam, “Synthetic Aperture Imaging of Surface Cracks using finite aperture transducer: A Matched Filter Approach”, Journal of Nondestructive Testing and Evaluation (Under Review).

     

 

 

 
 
    Guided Waves NDT
     
   
Circumferential Guided Waves Pipe Support Inspector
 

  
Long pipelines are normally rested on supports at intermediate distances. The support regions of pipelines are like crevices that are prone to corrosion due to the presence of key ingredients needed to accelerate corrosion like water, minerals, and the stress concentration in presence of a crack. The pipelines generally experience different types of corrosion, (i.e. corrosion spread along the circumferential and longitudinal direction of the pipe). A pinhole type pitting corrosion, which is much localized in nature, is commonly observed at the support regions and is very difficult to detect prior to leakage. Even if the pipe surface at these locations appears to be normal at the visible portion, the condition of the pipe at the hidden region (bottom portion), where it is in contact with the support, needs to be monitored. Among the various techniques used for the detection of corrosion-like defect in pipelines, ultrasonic NDE plays a major role. For performing the ultrasonic inspection in such inaccessible regions (i.e. at support locations), these heavy pipes have to be lifted out of the supports (using cranes) which involves complete shut down of the flow lines and the risk of failure during the lifting operation which would have been already weakened by corrosion. Hence, there is a critical need for an alternate method of nondestructive inspection of the hidden portion of the pipelines at supports.

The generation of higher order ultrasonic guided wave modes propagating in the cylindrical direction and the interaction of these modes with defects located in the pipe support regions was investigated using both finite element modeling and experiments. The experimental studies were carried out using ultrasonic cylindrical guided wave clusters comprising of S0, A0, S1, A1 and A2 modes, on pipe specimens of different diameters, from 6-24 inches in diameters, with simulated pipe support region defects such as EDM notches and pin holes of varying depths, using three different modes of generation (a) conventional 1 MHz, (b) conventional 2.25 MHz, and (c) 2.25 MHz linear phased array transducers. These modes were observed to be generated with significantly reduced dispersion at 2.25 MHz and hence found to be useful for inspection of small foot print defects such as pin-holes. The amplitudes of the experimentally obtained signals were successfully used to size the defects. It was also observed that the signals obtained from the 2.25 MHz phased array transducer was significantly less contaminated by noise in comparison to the signals obtained from the conventional 2.25 MHz transducer. The EDM notches and the 1.5 mm diameter pin holes could be detected and sized even when the wedge was circumferentially located at 140° from the defect.
 
The GUI of the pipe support inspector with image representation of corrosion on a pipe.
 
The GUI of the pipe support inspector with image representation of corrosion on a pipe.
 

Typical pipe support corrosion causing leakage.
Typical pipe support corrosion causing leakage.
 
Pipe Crawler Design
Pipe Crawler Design

 




   
   
 

References: K. Shivaraj, K. Balasubramaniam, C.V. Krishnamurthy. and R. Wadhawan, Ultrasonic circumferential guided wave for pitting type corrosion imaging at inaccessible pipe support locations, ASME Trans. J. Pres. Vessel Tech, in press, (2007).

L. Satyarnarayan, J. Chandrasekaran, B.W. Maxfield, K. Balasubramaniam, Circumferential Higher Order Guided Wave Modes for the Detection and Sizing of Cracks and Pin Holes in Pipe Support NDT&E in press available online (2007)

 
     
    Circumferential Guided Waves Simulation
 
 
  
  The image of the defect from a finite sized transducer is different from the image obtained using a point source and this image is the convolution of the spatio-temporal response characteristics of the transducer with the image of the defect.
The finite transducer characteristics can be expressed as a convolution of the transducers electrical characteristics and the Spatial Impulse Response (SIR) of the transducer in the given domain. The SIR is the impulse response of a finite transducer in the medium considered.

A matched filter based approach is presented, which takes into account the SIR of a finite sized phased array transducer, applied on B-scan images of bottom surface notches to perform the Synthetic aperture imaging.

In the matched filter approach we compute a template for each image reconstruction point, and calculate the reconstructed intensity at that point by carrying a weighted addition of the original image with weights specified by the template. This can be described as a correlation of the B-scan image using a pixel dependent mask. The mask chosen in this case is the model calculated received signals for a reflector positioned at the point considered. This leads to smooth reconstructions and increases the SNR in the image.The proposed algorithm was validated by comparing against the SAFT algorithm. A 1.5 mm diameter side drilled hole at a depth of 10 mm in a PH TOOL standard block (IOW Beam profile block: 7075-T6 Aluminum) was imaged using the single element electronic scan feature of the phased array. The B-scan obtained as well as the imaged results using the matched filter and the standard SAFT algorithm is shown in Figure 1. It was seen that the image reconstructed using the Matched filter compared well with the SAFT reconstructed image of the defect.


 
 

FEM Model for CGW simulation
FEM Model for CGW simulation
 
 
   
 

Reference: 1. L. Satyarnarayan, J. Chandrasekaran, B.W. Maxfield, K. Balasubramaniam, Circumferential Higher Order Guided Wave Modes for the Detection and Sizing of Cracks and Pin Holes in Pipe Support NDT&E in press available online (2007)

     

 

 

 
   
     
   
Storage Tank Floor Corrosion Detection and Imaging
 

   The leakage of storage tank containing any chemical product poses high risk to environment and human safety apart from the actual economic loss of the product stored. These components have to be inspected periodically to assess the corrosion damage in order to prevent catastrophic failures.

In the case of storage tanks, since the region of the annular plate near the shell-to-bottom fillet weld ( both on the top and bottom side )  is subjected to metallurgical change due to the irregular heating and cooling  arising from the welding process, it makes the region more vulnerable to corrosion when compared to other regions. It is also observed that the maximum stress in a tank bottom exists at the toe of the inside shell-to-bottom fillet weld at the annular plate. These may result in stress corrosion cracks and/or pitting corrosion on the liquid side of the annular plate and subsequently cause leakage.

Artificial EDM notches of varying depths i.e., 25%, 50% and 75% of the thickness were drilled to simulate stress corrosion crack like defects in an 11 mm thick MS tank floor sample. Fig 1 shows the constructional arrangement. Fig 2 shows photo of the tank bottom plate sample with defects. The defects were machined the bottom side of the specimen and placed exactly below the shell-to-tank fillet weld.

The A-scan signals and hence the B-scan images were obtained using conventional 2.25 MHz, 1 inch diameter L- wave probe mounted on an appropriate wedge.  The B-scan image of the three defects is shown in Fig 3 along with the energy plot of the defect amplitude for the corresponding A-scan signal. The energy plot is the integrated squared amplitude of the multimode defect signal within the time-gated region. It was observed that the energy of the defect amplitude was proportional to the flaw size. The largest flaw (75% of thickness) had the highest peak followed by the intermediate (50% of thickness) and then by the smallest flaw (25% of thickness).
 
 
 

Figure 1. The constructional arrangement of a storage tank
Fig 1. The constructional arrangement of a storage tank
 
Figure 2. Photo of the tank bottom plate sample with defects
Fig 2. Photo of the tank bottom plate sample with defects

 


Figure 3. B-scan image of the defects

Fig 3. B-scan image of the defects
with energy plot
   
   
 

Reference: 1. K. Balasubramaniam, L. Satyanarayan, J. Chandrashekaran, B. W. Maxfield, “Imaging Hidden Corrosion Using Ultrasonic Non-dispersive Higher Order Guided Wave Modes”, Review of Progress in Quantitative NDE, July 22 – July 27, 2007 Colorado School of Mines, Golden, Colorado, USA. 

 
     
    Guided wave imaging of Complex Thin Ti Welds
 
 
  
  The titanium shells of the storage bottles used in the satellites are welded using Electron-Beam welding technique that leaves a bead on the outer surface. The bottles are in the solution treated condition and carry optimum burst strength factor befitting aerospace requirement. Since the post weld inspection procedure calls for only techniques that can inspect from the outside only, the presence of the bead caused difficulties when traditional NDE methods are used. Currently, the inspection procedure calls for the machining of the bead and subsequent ultrasonic C-scan inspection. The Machining of the beads is a time consuming and cost-inefficient procedure that is not desirable. Hence, the development of an NDE technique for the inspection of weld defects in the presence of the bead is a critical need.

The advantages of the proposed guided wave technology can be summarized as:

    • Use of multimode, guided, plate waves provides a global inspection technique, and hence, the time of inspection is significantly less.

    • The multi-mode Lamb waves can be utilized to selectively evaluate different regions/cross-sections (as a function of depth).

    • The thin nature of the walls allow for the guided waves to be generated and used for flaw detection.
 

Guided wave based method of inspection was carried out on a thin  Titanium sample of thickness 1.6 mm and diameter of 872 mm with a central EB weld . Non contact immersion pulse echo technique was used at a frequency of 1 MHz. The setup consists of modular components controlled by the PC. Flaws like lack of fusion, porosity and cluster of pores were successfully imaged using guided waves that were generated using oblique incident wave. The use of an azimuthally angle allowed for the imaging to be more sensitive. All the defects were detected using Radiographic Testing also. The scanning resolution was 0.5mm.  The B-scan images show the defects and the weld line reflections along the length of the weld.

 
 
Experimental Setup
Experimental Setup
 
Defect Free Region
Defect Free Region
 
1 mm porosity Region
1 mm porosity Region
   
 

Reference: 1. L. Satyarnarayan, J. Chandrasekaran, B.W. Maxfield, K. Balasubramaniam, Circumferential Higher Order Guided Wave Modes for the Detection and Sizing of Cracks and Pin Holes in Pipe Support NDT&E in press available online (2007)

     

 

 

 
   
     
   
Time Reversal Techniques for Lamb Wave Imaging
 

   The basic theory of the time-reversal process is based on property of the acoustic wave equation in a lossless medium; in the most general case, the acoustic wave equation is a differential equation where the time-derivative operator appears only at the second order.

The time reversal method consists of two stages. In the first stage as illustrated in Fig. 1(a), waves are generated by a single source and the associated responses are recorded by an array of sensors, called a time reversal array surrounding the boundary. In the second stage as illustrated in Fig. 1(b), the signal is time reversed and retransmitted into the medium by the same time reversal array, which acts as a source array that is, the part that is recorded first is sent back last and vice versa. The retransmitted signal propagates back through the same medium and refocuses approximately on the source.

A 2D FDTD simulation study was carried out for a rectangular wave guide (plate) of size 740 mmx40mm with one source on the left-hand side and an array of 13 equally spaced receivers on the right-hand arranged perpendicularly with respect to the source. The velocity of sound and density of water was assumed to be 1500m/s and 1000kg/m3 respectively.

In FDTD the path of the sound wave propagating can be continuously seen in time domain. Thus FDTD is able to record the wave propagating and reflecting simultaneously without the change in the algorithm. Fig.2(a) displays visually the sound wave propagating in time domain excited by a source and signal being recorded by receiver array at the location shown by the dotted line.

In the next step the recorded signals are reversed in time and re-transmitted in the same simulation model at the same location were recording is done. The result is that the wave focuses back on the acoustic source as shown in Fig. 2(b)

The input signal in Fig.3(a) when compared to signal received at the source after recording and time reversal of the signals at the receiver array as shown in Fig.3(b)  is similar in shape but slightly lesser in amplitude. This shows the focusing ability of the time-reversed wave.


 

(b) Time reversed Propagation
Figure 2: Snapshot of sound pressure distribution after 2-milli sec

 

Figure1(a): Step I: waves are generated by a single source

Figure1(a): Step I: waves are generated by a single source

 
Figure1(b): Step II: the recorded signals are time reversed and retransmitted
Figure1(b): Step II: the recorded signals are time reversed and retransmitted

 


(a) Forward Propagation

Figure1(a): Forward Propagation

 
(b) Received signal at the sourceFigure 3: A-scan signal obtained at the source before and after time reversal

(b) Received signal at the source
Figure 3: A-scan signal obtained at the source before and after time reversal

   
   
 

Reference: 1. K. Balasubramaniam, L. Satyanarayan, J. Chandrashekaran, B. W. Maxfield, “Imaging Hidden Corrosion Using Ultrasonic Non-dispersive Higher Order Guided Wave Modes”, Review of Progress in Quantitative NDE, July 22 – July 27, 2007 Colorado School of Mines, Golden, Colorado, USA. 

 
   
 
   Electromagnetic Transduction
     
    EMATs for Lamb Wave Defect Detection
 
 
  
  EMAT technology is an advanced Ultrasonic Non Destructive Testing method that differs from the piezoelectric transducers in the way the sound is generated. In conventional method a piezoelectric crystal is used to convert electrical energy into mechanical vibration. But in EMAT technology, the sound generation has been done by means of electromagnetism.

An EMAT consists of a magnet and a coil of wire and relies on electro-magnetic acoustic interaction for elastic wave generation. Using Lorentz forces and magnetostriction, the EMAT and the metal test surface interact and generate an acoustic wave within the material.

The main two important components for designing EMAT are high frequency coil and magnet. The coils are fabricated by Printed circuit technique (PCB). The magnets are permanent Neodymium-Iron-Boron (Ne-Fe-B) sintered magnets with flux density around 0.35 Tesla. The Meander coil EMAT induces the Lorentz force parallel to the surface, whose directions change alternately with the Meandering period. This force distribution generates Rayleigh waves traveling along the surface and simultaneously the longitudinal and shear vertical (SV) waves traveling obliquely into the specimen.

The Meander coil EMAT can generate Lamb waves in thin materials. The Lamb wave has been used for thickness measurement in thin plates and defect detection in stainless steel weld joints. For these measurements, the fundamental lamb wave modes (A0 and S0) have been used. The thickness measurement has been done on an aluminium plate with four different thicknesses. The phase velocity dispersion curves shift right as the thickness decreases. The slope of the excitation line is just the wavelength of the EMAT. The frequency corresponding to the excitation line crosses the dispersion curve for a particular mode is called peak frequency. The peak frequency can be measured for each thickness by using frequency sweep method. The relationship between peak frequency shifts and thickness changes can be used for guided wave thickness measurement. The amplitude reduction indicates that the presence of defect in the sample, because the loss of material induces the loss of energy of sound waves.  The percentages of thickness reduction correspond to the two defects, evaluated by the frequency shift method.
 
EMAT Probe Developed at CNDE
EMAT Probe Developed at CNDE
 
Typical Guided Wave Signals
Typical Guided Wave Signals
 
Guided wave based thickness measurement using frequency shifting.
Guided wave based thickness measurement using frequency shifting.
Guided wave based thickness measurement using frequency shifting.
 
   
   

 

 

 
     
   
EMF Transduction in non-conducting non-magnetic materials
 

   The excitation of elastic waves in conducting materials (metals) is well known and widely studied. In non-magnetic materials and under most conditions, in magnetic materials, linear models give an exact description of the excitation and, through reciprocity, accurately describe the inverse process or reception of elastic waves. Many different configurations for both generation and reception of elastic waves have been studied in work related to EMATs (Electromagnetic Acoustic Transducers), EMUSs (Electromagnetic Ultrasonic Sensors), MsS (Magnetostrictive Sensors) and other manifestations of the coupling of energy from an electromagnetic field into the conduction electrons or magnetic domain structure. Through the conduction electrons and magnetic domains that are very tightly coupled to the atoms or molecules of the material in question, the above mechanisms have been shown to result in the generation of elastic forces which, in turn, can often generate elastic waves.

The precise elastic mode(s) that is (are) generated depends upon the spatial distribution of the exciting forces and the material boundary conditions. Thus, the different configurations that have been studied can be used to generate bulk longitudinal and shear waves, surface waves and a variety of guided waves in metals. This note reports preliminary experimental results for a mechanically bonded electrodynamics transducer used to excite elastic waves in non-conducting materials such as Perspex and alumina where none of mechanisms mentioned above can be responsible for the forces generating the elastic waves.

Many different coil and magnet configurations have been used to generate elastic waves in non-conducting rods and tubes as part of study to develop practical, low-cost devices for measuring the viscosity of liquids. For these studies, coils were wound around the rod or tube (the rod would be immersed in a liquid or the tube filled with a liquid) and bonded with an epoxy or similar adhesive. A permanent magnet was placed over the coil or a set of permanent magnets was placed near the coil; the magnetic field configuration depended upon the force direction needed to generate the required/desired elastic mode.


 

Experimental Setup for EMF
Experimental Setup for EMF
 
Alumina rod used for EMF generation with coil and Permanent Magnet exciter
Alumina rod used for EMF generation with coil and Permanent Magnet exciter.

 

Typical GW signals obtained using EMF
Typical GW signals obtained using EMF
 
 

 
   
 
   Time of Flight Diffraction Technique
     
    TOFD for Flaw Sizing in Thin Structures
 
 
  
  Some of the difficulties with TOFD inspection of thin sections can be listed as below:

The number of mode converted signals reaching the receiver transducer increases with decreasing thickness of the specimen.

As the specimen thickness decreases the spacing between lateral wave and backwall echo decreases. The crack tip echo (due to L-wave) always lies in the region between lateral and backwall echoes and hence temporal resolution plays a determining role when dealing with thin section. The signals overlap making TOF calculations difficult.

As the thickness of the sample reduces, the critical flaw size also consequently decreases. This leads to the reduction in the separation between the two tip diffracted echoes wish normally overlap of the tip diffracted signals, leading to difficulties in TOF calculations.

The problem of inspection of thin sections can be tackled by paying attention to the following details:

Optimizing experimental parameters (distance between probes, angle of the probe and frequency);
Reducing the distance between probes (transmitter and receiver),
Using an alternate signal processing techniques, such as Embedded Signal Identification Technique (ESIT) discussed here.

By using higher frequency (>5 MHz) the resolution of the B-scan image can be increased.

An Embedded Signal Identification Technique (ESIT) was developed for signal identification and automatic defect sizing for TOFD of thin plate-like structures. Experiments were conducted on 10 mm-thick aluminum and mild steel calibration samples with EDM notches (24 different types) and on thin managing steel weld samples (around 7 mm thickness) with fatigue cracks of different sizes. Manual and automatic defect sizing algorithms were compared on these samples at different frequencies and probe angles. The results using automatic defect sizing show better accuracy relative to manual sizing. The experimental results were compared using a commercial TOFD system and found to be comparable for 10 mm-thick calibration samples and works better for thin sections having realistic fatigue cracks.
 
 

Flaw sizing used ESIT algorithm
Flaw sizing used ESIT algorithm
 

Flaw sizing used ESIT algorithm

Flaw sizing used ESIT algorithm

Flaw sizing used ESIT algorithm
Sizing of 3.2 mm fatigue crack in 7.2 mm thickness weld sample; (a) ESIT B-scan image; (b) echo separated signals at the point of cursor and (c) zoomed portion of the defect using 5 MHz, 60° probe angle and 26 mm distance between probes
   
  1. Baskaran, G, Balasubramaniam, K and Lakshmana Rao, C, Ultrasonic TOFD flaw sizing and imaging in thin plates using Embedded Signal Identification Technique (ESIT) Insight 46 537-542 (2004).
    2. Gokul Swamy, K. Balasubramaniam, and G. Baskaran, “A Point Source CorrelationTechnique for Automated Flaw Identification and Sizing Using
    TOFD Inspection,” Materials Evaluation, 63(4) 425-29 (2005)
   
   

 

 

 
     
   
TOFD for Complex Structures
 

   
   Sizing of a surface breaking or embedded crack in a simple geometry by TOFD technique has been well established (1,2). However, sizing a crack in a complex geometry component, where there is seldom a reference lateral wave or back wall echo signal, is often difficult. Inspection of complex geometry component. One such complex geometry component is a steam turbine rotor shaft.

The application of TOFD technique to thin sections as well as complex geometry components requires the wave propagation based simulation of the TOFD technique. Thin geometries are encountered in plates and shells. The complex geometry inspection applications are many   Simulation gives ideas about the expected results from experiments where real experiments are not possible or are difficult to conduct. Further, simulation helps us to derive the optimum choice of experimental parameters.  Using these simulations the optimal probe positions to inspect entire region of T-butt welds was obtained.

Inspection for estimating the depth of a surface-breaking crack in a solid rotor shaft of a steam turbine or in a similar component has to be carried out from the outer surface of the rotor. Considering the sources of the inadvertent and to some extent unavoidable variations associated with this kind of manual inspection method, it is better to conduct as many trials as is practically feasible to have a better accuracy. This study shows that ultrasonic TOFD  inspection for the said problem estimates the depth of the surface-breaking cracks with reasonable accuracy.


 
Experiments on Rotor Shaft Samples.
Experiments on Rotor Shaft Samples.
Experiments on Rotor Shaft Samples.
 



Possible locations for the TOFD Probes
Possible locations for the TOFD Probes
 
FEM Simulation of waves during TOFD
FEM Simulation of waves during TOFD

 

   
  1. Baskaran, G, Lakshmana Rao, C, and Balasubramaniam, K, "Simulation of TOFD technique using finite element method, Insight in press (2007)
    2. S K Nath, K Balasubramaniam, C V Krishnamoorty and B H Narayana, “Sizing of surface-breaking cracks in complex geometry components
    by ultrasonic Time-of-Flight Diffraction (TOFD) technique” Insight 49(4) 11-17 (2007)
 

 
     
    Simulation of TOFD using Ray Tracing
 
 
  
  Some of the applications of TOFD simulation are :

Application of TOFD techniques to complex structures (e.g. T-butt weld)

Predicting the B-scan diffracted signal curve shape for different orientation of defects.
Signal superposition studies  (effect of probe frequency, number of cycles, probe angle and probe separation)
Selection of probe angle for maximum diffracted amplitude at the surface from the defect.

The TOFD technique is simulated using ray theories.  The different cases are compared with experimental results and other possible cases are extrapolated.  The GUI model works almost in real time and gives important ideas about choice of experimental parameters for inspection of thin sections, near surface inspections and inspection of inclined defects more accurately.The A-Scan and B-scans are simulated using a Graphical User Interface (GUI) model developed in LabVIEW. The simulated results are applied to

Defect signal identification for vertical defects
Inspection of inclined defects
effect of pulse width or probe frequency on experimental results
Near surface inspection.
 
 
Schematic of Simulation domain
Schematic of Simulation domain
Lab view based GUI used for simulation
Comparison of experimental and simulation results for a crack.
   
  1. Baskaran, G, Balasubramaniam, K and Lakshmana Rao, C, Ultrasonic TOFD flaw sizing and imaging in thin plates using Embedded Signal Identification Technique (ESIT) Insight 46 537-542 (2004).
    2. Gokul Swamy, K. Balasubramaniam, and G. Baskaran, “A Point Source CorrelationTechnique for Automated Flaw Identification and Sizing Using
    TOFD Inspection,” Materials Evaluation, 63(4) 425-29 (2005)
   
   

 

 

 
     
   
Shear Wave TOFD (S-TOFD) Technique
 

   
  Ultrasonic Time of Flight Diffraction (TOFD) for sizing defects is based on the time of flight of the diffracted echo that is generated when a longitudinal wave is incident on a crack tip. This technique has the limitation during near-surface inspection due to signal superposition.  Here, this limitation is overcome by using the shear wave-diffracted signal (instead of longitudinal wave) and hence called S-TOFD.  Experiments were conducted on samples with defect tip closer to the surface of a flat plate sample to illustrate the utility of the S-TOFD technique. An increase in the flaw sizing accuracy, by using the shear wave diffracted echoes from the tip and through the application of a signal processing technique (ESIT), was demonstrated.

Here, an S-TOFD approach extends the technique to include the diffracted shear wave signals.  The S-TOFD technique proposed here will find application for improved sizing of defect in a) shallow cracks that are surface breaking and b) embedded cracks in thin (< 10 mm) structures and c) for cracks that are almost through the cross section (the crack tip is near the inspection surface). The effective use of the proposed S-TOFD technique will require careful selection of experimental parameters when compared to traditional TOFD, due to the increase in the number of signals that must be considered.

 

Actual Depth (mm)

Measured Depth / (%error)

Manual-L

Manual-S

L + ESIT

S+ESIT

0.5

0.36
(28)

0.41
(18)

0.54
(8)

0.52
(4)

1.75

1.49
(14.28)

1.50
(14.29)

1.69
(3.43)

1.72
(1.71)

1.75

1.53
(12.57)

1.59
(9.14)

1.67
(4.57)

1.73
(1.14)

4.51

-

4.56
(1)

-

4.49
(0.4)

2.63

-

2.57
(2.28)

-

2.64
(0.38)

 

Schematic of TOFD experiment
Schematic of TOFD experiment
for defect sizing.
 
Schematic of the L and S Diffracted
Schematic of the L and S Diffracted
Schematic of the L and S Diffracted
Wave and signals.

 


Typical result for a embedded crack in a thin sample of aluminum.
Typical result for a embedded crack in a thin sample of aluminum.
 
   
  1. G. Baskaran, Krishnan Balasubramaniam, and C. V. Krishnamurthy, Inspection Using Shear Wave Time of Flight Diffraction (S-TOFD) Technique,
    AIP Conf. Proc. 820, 97 (2006)
    2. G. Baskaran, K. Balasubramaniam and C. Lakshmana Rao, Shear-wave time of flight diffraction (S-TOFD) technique, NDT & E International 39
    458–467, (2006)
 

 
     
    Point Source Correlation Technique (PSCT) for TOFD
 
 


A technique for automatic flaw location and sizing using ultrasonic Time of Flight Diffraction (TOFD) technique was explored. In any TOFD inspection, the crack tip echo (es) always occur between the lateral and back wall echoes from the specimen. With decreasing thickness of the specimen, the difference in time of flights between the lateral wave and the reflected bottom surface (back wall) signal from the specimen gets reduced.  Hence, there is a possibility of the signals from the various sources (crack tip diffracted echoes, the lateral wave echo and back wall echoes) interfering with each other.

This presents a difficulty in measuring the time of flight of the tip-diffracted waves for defects in thin specimens (thickness < 12mm). Here the crack tips are modeled as point sources of diffracted waves in a homogenous, isotropic medium. The diffraction arcs are modeled using a ray-based approach and the modeled arcs are correlated with the experimental B-Scan data. The points of high correlation provide information about the location of the crack tips. A statistical echo separation procedure to isolate the diffraction arcs in the B-Scan image is discussed. This work also addresses the issue of application of this TOFD technique to thin section (<12 mm), wherein the echoes from the various sources (lateral wave, back surface reflection, diffraction from crack tips, etc) interfere with each other making it difficult to identify diffracted signals from the flaw tips.  

   

Actual Defect Size (mm)

Calculated Defect Size (mm)

3.5

3.3

6

5.8

6 (inclined)

5.6

 

 

3.5

3.6

6

6.1

6 (inclined)

5.7

 
Comparison of calculated defect size using PSCT with actual sizes in 10 mm AL sample with EDM defects.
 
 
Simulated arcs centered at three different points of different depths on the B-Scan.


Simulated arcs centered at three different points of different depths on the B-Scan.
Here the depth of A<B<C.
 
TOFD Setup for collecting PSCT data
TOFD Setup for collecting PSCT data
   
  1. G. Baskaran, Krishnan Balasubramaniam, and C. V. Krishnamurthy, Simulation of Ultrasonic Technique Using Spectral Element Method, AIP Conf.
    Proc. 820, 111 (2006).
    2. Baskaran, G., Lakshmana Rao, C., Balasubramaniam, K. Simulation of wave interaction with arbitrarily oriented defects in thin sections using spectral element method American Society of Mechanical Engineers, Pressure Vessels and Piping Division (Publication) PVP, 5, Pages 237-
 

 
     
    SEM Simulation of TOFD
 
 


   Numerical simulation of ultrasonic wave propagation using methods such as Finite Element or Finite Difference is computationally expensive particularly when (a) structural dimensions are high, (b) inspection at higher frequencies (due to short wavelengths), and (c) in complex materials that are not isotropic. This paper discusses a numerical technique, which is similar to FEM, but works in frequency domain and has advantage of more accurate results in quick computational time called the Spectral Element Method (SEM). When the second order partial differential wave equation transformed to frequency domain by Continuous Fourier Transform, the wave equation transforms to ordinary differential equation (ODE) that has exact solution.

SEM is similar to FEM, but works in frequency domain and has the advantage of giving more accurate results in quick computational time.  Conventional finite elements treat the dynamic load induced by the mass and rotational inertia of the beam as concentrated loads and moments applied at the ends of the element.  Even though the structural joints may be far apart, many elements must be used if the inertia distributions are to be modeled accurately.  Therefore, the number of elements required to do a dynamic problem is substantially larger than that required for the equivalent static one [3]. The spectral analysis formulates an element, which treats the distribution of mass and rotational inertia exactly.  Only one spectral element need to be placed between any two joints, substantially reducing the total number of degrees of freedom in the system.  SEM formulation uses the fast Fourier transform (FFT) algorithm to transform the distributed parameters from time domain to frequency domain and vice versa [1]. When the second order partial differential wave equation transform to frequency domain by Continuous Fourier Transform, the wave equation transforms to ordinary differential equation (ODE) that has exact solution.  In SEM, the structure is divided into number of waveguides with connectivity at the joints. The energy in a waveguide is directed along its length.  The waveguide that has one degree of freedom (rod element) simulates the longitudinal motion. The waveguide with two degrees of freedom (modified rod element) simulates the shear wave and a waveguide with three degrees of freedom (beam element) simulates the flexural motion.  The beam wavegudes were developed using Euler- Bernoulli theory and Timoshenko beam theory for different applications.  
 

SEM solution of structure with horizontal defect. a) structure with horizontal defect,

SEM solution of structure with horizontal defect. a) structure with horizontal defect,
b) waveguide solution and
c) idealization.
 
Simulated signals using SEM for different length of horizontal defect
Simulated signals using SEM for different length of horizontal defect
Comparison of 2-D FEM with 1-D SEM
Comparison of 2-D FEM with 1-D SEM
   
  1. G. Baskaran, Krishnan Balasubramaniam, and C. V. Krishnamurthy, Simulation of Ultrasonic Technique Using Spectral Element Method, AIP Conf.
    Proc. 820, 111 (2006).
    2. Baskaran, G., Lakshmana Rao, C., Balasubramaniam, K. Simulation of wave interaction with arbitrarily oriented defects in thin sections using spectral element method American Society of Mechanical Engineers, Pressure Vessels and Piping Division (Publication) PVP, 5, Pages 237-
 

 
     
    POD of TOFD Inspection
 
 


   The experimental probability of detection (POD) and probability of sizing (POS) curves, for the detection and sizing of surface- breaking cracks in complex geometry components, by manual ultrasonic Time of Flight Diffraction (TOFD) system has been developed. In the development of POD and POS curves, it has been assumed that the signal response i.e. ‘â’ values for a crack size ‘a’ have a normal distribution. The experimental parameters namely the probe angle and probe center spacing have been altered and consequent effect on both POD and POS has been noticed. The curves obtained in this work can be useful in risk-based inspection planning for complex geometry components e.g. steam turbine rotor shaft using manual TOFD system.

The POD and POS curves have been generated from experimental data. These curves can be used in actual field application under similar test conditions. In the present study, two experimental parameters namely probe angle and probe center spacing have been varied and resulting effect on the POD and POS curves is clearly noticed. This will help in choosing superior set of experimental parameters during actual field application. In the future work, other variables namely operator, probe frequency, geometry of the block, notch sizes etc. can be changed to generate the resulting POD and POS curves. Gain of the TOFD inspection system should be increased for detecting deeper defects. Deeper defects mean defects having depth more than  in POD.  Means the estimate of defect size at 50 % POD.

 

(A) 60° probe with beam axis incident on the tip of 10 mm deep    notch (= 69)
(B) 60° probe with beam axis incident on the tip of 15 mm deep notch (= 109)
(C) 70° probe with beam axis incident on the tip of 10 mm  deep notch (= 72)
(D) 70° probe with beam axis incident on the tip of 15 mm deep notch (= 85)

 

SEM solution of structure with horizontal defect. a) structure with horizontal defect,

POD Curves

Simulated signals using SEM for different length of horizontal defect

POS Curves

 
   
  1. S.K.Nath, Krishnan Balasubramaniam, C.V.Krishnamurthy, B.H.Narayana,Reliability assessment of manual Ultrasonic Time of Flight Diffraction (TOFD) inspection for complex geometry components, Communicated.






















































































































































































































































































































































































































































































































































































































































































































































































































































































  .© Copyright 2007. IIT - CNDE All Rights Reserved.  
Designed & Hosting by IndiaDirect