journal of non destructive testing and evaluation

Volume 11

issue 2

September 2012

from the Chief Editor

DDDDDrrrrr. Krishnan B. Krishnan B. Krishnan B. Krishnan B. Krishnan BalasubramaniamalasubramaniamalasubramaniamalasubramaniamalasubramaniamProfessor

Centre for Non Destructive EvaluationIITMadras, Chennai

[email protected]@gmail.com

URL: http://www.cnde-iitm.net/balas

This years annual get-together of ISNT, the NDE2012, to be held in Ghaziabad(Delhi), will bring forth the complete flavor of the ISNT community. The Journalof Nondestructive Testing and Evaluation team welcomes all delegates to thisgala technical festival. The NEW Editorial Board joins me in congratulating allthe ISNT Award Winners for the 2012. I would like to recognize thecontributions of the out-going members of the Editorial Board;Drs. D.K. Bhattacharya, P.Kalyanasundaram, T. Jayakumar, K. Viswanathan,K. Rajagopal, MVMS Rao, J. Lahiri, KRY Simha, Shri. Vaidyanathan, Shri RameshParikh, Shri. Srinivasa Rao, Shri. G. Ramachandran. The Journal also welcomesthe new members of the Editorial Board; Dr. CV Krishnamurthy,Dr. O. Prabhakar, Dr. MT Shyamsunder, Dr. B. Venkataraman, Dr. H. Wolf,Dr. K. Srinivas and Shri. P. Nanekar. In addition, several new internationalexperts in the fields of NDT have agreed to serve in the board. It is envisagedthat the newly constituted board will further enhance the quality andreadership of the Journal. In addition, efforts are underway to include theJournal in the different citations indices.

In this edition, the BASICS section is focused on Theoretical Modeling Methodsand how it applies to NDT. Today, the theoretical models have taken the formof VIRTUAL NDT that the new generations of NDT Engineers feel comfortableand may immensely benefit. The HORIZONS describes the use of a hybridmethod of using high power ultrasound to vibrate and consequently heat thecrack surfaces inside materials and observe the indications using infra-redcamera in a transient mode. The technical articles in this issue cover a widerange of topics in NDT including adhesive bond inspection using ultrasonics,solutions to the eddy current benchmark problems as posed by the WorldFederation of NDE Centers, composite inspection using guided ultrasonic wavemodes, and weld inspection for rails using phased array technology.

I would like to touch base on one of the topics of interest to all of us i.e. fromwho will we buy our NDT instruments and products? Unlike the past, few Indiaentrepreneurs are coming forefront with products that can compete with theinternational brands viz. Technofour, Modsonic, PulseEcho, Electro-MagfieldControls, Dhvani Research, EEC, etc. These companies face severe adversitiessuch as limited market for their products in India, competition frominternational brand that have high marketing budgets, lack of echo systems tofoster innovation and invention, a Government policy that is suspicious of smallbusinesses, etc. However, it may be the right time to ask ourselves a fewquestions.

Are Indian products getting due recognition from the NDT users in India?

Is there a case to be made for creating an eco-system for encouragingentities that develop indigenous products?

Should ISNT an ISNT members take a lead in incubating new companiesthat can provide quality Indian made products?

Should ISNT lobby the Government and the stakeholders to providepreferential treatment of Indian made products?

What is the right mix between Indian made products vs. imported ones?

May be the time to ponder is NOW!

2

vol 11 issue 2 September 2012 Journal of Non Destructive Testing & Evaluation

I S N T - National Governing Council

Chapter - Chairman & Secretary

PresidentShri P. Kalyanasundaram

President ElectShri V. Pari

Vice-PresidentsShri D.J. Varde

Shri Swapan ChakrabortyShri N.V. Wagle

Hon. General SecretaryShri R.J.Pardikar

Hon.Jt.SecretariesShri Rajul.R.ParikhDr.B.Venkatraman

Hon. TreasurerShri S.SubramanianHon. Co.Treasurer

Shri Sai Suryanarayana

Immediate Past PresidentShri K.Thambithurai

Past PresidentShri Dilip.P.Takbhate

MembersShri Anil V.Jain

Shri Dara E. RupaShri D.K.Gautam

Shri Diwakar D.JoshiDr. Krishnan Balasubramaniam

Shri Mandar A. VinzeShri B.B.Mate

Prof. G.V.PrabhugaunkarShri B.K.PangareShri M.V.Rajamani

Shri Samir K. ChoksiShri B.K.Shah

Shri S.V.Subba RaoShri Sudipta Dasgupta

Shri R.K.SinghShri A.K.Singh(Kota)

Shri C.AwasthiShri Brig. P.Ganesham

Shri Prabhat KumarShri V.SathyanShri P.Mohan

Shri R.SampathMs. Hemal Thacker

Shri A.K.SinghiShri T.V.K.KidaoShri B. Prahlad Dr. BPC Rao

Dr.Sarmishtha Palit Sagar

Permanent InviteesShri V.A. Chandramouli

Prof. S. RajagopalShri G. Ramachandran

and All Past Presidents,All Chapter Chairmen / Secretaries

Ex-officio MembersChairman NCB,Secretary NCB,Treasurer NCB,

Controller of Examination NCB,President QUNEST,Secretary QUNEST,Treasurer QUNEST

AhmedabadShri D.S. Kushwah, Chairman,NDT Services, 1st Floor, Motilal Estate,Bhairavnath Road, Maninagar,Ahmedabad 380 028. [email protected] Rajeev Vaghmare, Hon. SecretaryC/o Modsonic Instruments Mfg. Co. Pvt. Ltd.Plot No.33, Phase-III, GIDC Industrial EstateNaroda, Ahmedabad-382 330 [email protected]

BangaloreProf.C.R.L.Murthy, ChairmanDept. of Aerospace Engg,Indian Institute of Science, Bangalore 560012Email : [email protected]

ChennaiShri R.Sundar, Chairman,First Floor, North Wing, PWD Office Complex,Chepauk, Chennai - 600 [email protected], [email protected] RG. Ganesan, Hon. Secretary,Chief Executive, BETZ Engineering & Technology49, Vallalar Street, Adambakkam, Chennai - 600 [email protected] ; [email protected]

DelhiShri A.K Singhi, Chairman,MD, IRC Engg Services India Pvt. Ltd612, Chiranjiv Tower 43, New Delhi [email protected] M.C. Giri, Hon.Secretary,Managing Partner, Duplex Nucleo EnterpriseNew Delhi [email protected]

HyderabadShri N. Saibaba, Chairman,Chief Executive, Nuclear Fuel Complex,ECIL PO, [email protected] ; [email protected] M.N.V. Viswanath, Secretary,Dy. Manager, Quality Assurance-Fuels,CFFP Building,Nuclear Fuel Complex,ECIL PO, [email protected] ; [email protected]

JamshedpurDr N Parida, Chairman,Senior Deputy DirectorHead, MSTD, NML, Jamshedpur - 831 [email protected]. GVS Murthy, Hon. Secretary,MSTD, NML, [email protected] / [email protected]

KalpakkamDr. B. Venkatraman, ChairmanAssociate Director, RSEG, & Head, QAD,IGCAR, Kalpakkam 603 [email protected] B. Dhananjaykumar, Hon.SecretaryReprocessing Group, IGCAR,Kalpakkam – 603 102 [email protected]

KochiShri CK Soman, Chairman,Dy. General Manager (P & U),Bharat Petroleum Corporation Ltd. (Kochi Refinery),PO Ambalamugal 682 302. [email protected] V. Sathyan, Hon. Secretary,SM (Project),Bharat Petroleum Corporation Ltd.(Kochi Refinery),PO Ambalamugal-682 302 [email protected]

KolkataShri Swapan Chakraborty, ChairmanPerfect Metal Testing & Inspection Agency,46, Incinerator Road, Dum Dum Cantonment,Kolkata 700 028. [email protected] Dipankar Gautam, Hon. Secretary,4D, Eddis Place, Kolkata-700 [email protected]

KotaShri R.C. Sharma, ChairmanQAS, RAPS - 5 & 6, PO AnushaktiRawatbhata 323 303 [email protected] S.K. Verma, Hon. Secretary,TQAS, RAPS - 5 & 6, PO AnushaktiRawatbhata 323303. [email protected]

MumbaiShri.S.P.Srivastva, Chairman303, Lok Centre, Marol Maroshi Road,Andheri (East), Mumbai 400 059Email: [email protected] ,[email protected] ; [email protected] Samir K. Choksi, Hon. Secretary,Director, Choksi Brothers Pvt. Ltd.,4 & 5, Western India House, Sir P.M.Road,Fort, Mumbai 400 001. [email protected]

NagpurShri Pradeep Choudhari, ChairmanParikshak & Nirikshak, Plot M-9, LaxminagarNagpur - 440 022Mr. Jeevan Ghime, Hon. Secretary,Applies NDT & Tech Services,33, Ingole Nagar, B/s Hotel Pride, Wardha Road,Nagpur 440 005. [email protected]

PuneShri BK Pangare, ChairmanQuality NDT Services, Plot BGA, 1/3 Bhosari, GeneralBlock, MIDC, Bhosari, Pune- 411 [email protected] BB Mate, Hon Secretary,Thermax Ltd., D-13, MIDC Ind. Area, RD AgaRoad, Chinchwad, Pune- 411 [email protected]

SriharikotaShri V Ranganathan, Chairman,Chief General Manager , Solid Propellant Plant,SDSC – SHAR, Sriharikota – [email protected] B KarthikeyanHon. Secretary, ISNT Sriharikota ChapterSci/Eng. NDT/SPROB,SDSC – SHAR, Sriharikota – [email protected]

TarapurShri PG Bhere, Chairman,AFFF, BARC, Tarapur-401 [email protected] Jamal Akhtar, Hon.Secretary,TAPS 1 & 2, NPCIL, Tarapur. [email protected]

TiruchirapalliR.J. PardikarAGM, (NDTL)BHEL Tiruchirapalli 620 014. [email protected] L. Marimuthu, Hon. Secretary,HA-95, Anna NagarTiruchirapalli 620 026. [email protected]

VadodaraShri P M Shah, Chairman,Head-(QA) Nuclear Power Corporation [email protected] S Hemal Thacker, Hon.Secretary,NBCC Plaza, Opp.Utkarsh petrol pump, Kareli Baug,Vadodara-390018. [email protected]

ThiruvananthapuramDr. S. Annamala Pillai, ChairmanGroup Director, Structural Design & Engg Group,VSSC, ISRO, Thiruvananthapuram [email protected]. Binu P. ThomasHon. Secretary, Holography section, EXMD/SDEG,STR Entity, VSSC, Thiruvananthapuram 695 [email protected]

VisakhapatnamShri Om Prakash, Chairman,MD, Bharat Heavy Plate & Vessels Ltd.Visakhapatnam 530 012.Shri Appa Rao, Hon. Secretary,DGM (Quality), BHPV Ltd., Visakhapatnam 530 012

Contents

Chief EditorProf. Krishnan Balasubramaniam

e-mail: [email protected]

Co-EditorDr. BPC [email protected]

Managing EditorSri V Parie-mail: [email protected]

Editorial BoardDr N N Kishore, Dr. CV Krishnamurthy,Dr. O. Prabhakar, Dr. MT Shyamsunder,Dr. B. Venkataraman, Dr. H. Wolf,Dr. K. Srinivas, Shri. P. NanekarShri B Ram Prakash

10

16

24

26

9

About the cover page:

Advisory PanelProf P Rama Rao, Dr Baldev Raj,Dr K N Raju, Sri K Balaramamoorthy,Sri V R Deenadayalu, Prof S Ramaseshan,Sri A Sreenivasulu, Lt Gen Dr V J Sundaram,Prof N Venkatraman

ObjectivesThe Journal of Non-Destructive Testing &Evaluation is published quarterly by the IndianSociety for Non-Destructive Testing forpromoting NDT Science and Technology. Theobjective of the Journal is to provide a forumfor dissemination of knowledge in NDE andrelated fields. Papers will be accepted on thebasis of their contribution to the growth ofNDE Science and Technology.

Journal of Non DestructiveTesting & Evaluation

Published byShri RJ Pardikar,General Secretary on behalf ofIndian Society for Non Destructive Testing (ISNT)

The Journal is for private circulation to membersonly. All rights reserved throughout the world.Reproduction in any manner is prohibited. Viewsexpressed in the Journal are those of the authors'alone.

Modules 60 & 61, Readymade GarmentComplex, Guindy, Chennai 600032Phone: (044) 2250 0412Email: [email protected] at VRK Printing [email protected]

Volume 11 issue 2September 2012

Computed Tomography (CT)using X-rays has become verypopular in many field includingmedical, science, and NDT. Inconventional CT technique, the X-ray or Gama-ray sources are usedto penetrate the object underexamination and the data collectedby moving the sample is used toreconstruct the 3D image of theobject. Recently, the use ofemission tomography has beenshown to have several advantagesfor some key applications. Inemission tomography the radiationfrom within the body is used toreconstruct the 3D image of theobject. In the cover page, theemission computed tomography(CT) reconstruction of bowl isshown indicating blockages inpassage..

(Courtesy: QAD section of Indira GandhiCentre for Atomic Research, Kalpakkam)

37

42

47

54

60

Chapter News

Basics - Modelling and simulation forNondestructive Evaluation

Horizon - Sonic Infrared Imaging – Emerging NDETool

NDE Events

NDE Patents

NDE Puzzle

Technical Papers

Quality Assessment of Composite AdhesivelyBonded joints by Non-linear Ultrasonic MethodR.L. Vijayakumar, M.R. Bhat and CRL Murthy

Solution to the third eddy current benchmarkproblem of WFNDE centersS. Thirunavukkarasu, B. Purna Chandra Rao, S. ShuaibAhmed and T. Jayakumar

Characteristics of turning Lamb modes in compositesub-laminatesC. Ramadas, Krishnan Balasubramaniam, Avinash Hoodand C.V. Krishnamurthy

Rail Weld Inspection using Phased Array UltrasonicsGirish.N.Namboodiri, Krishnan Balasubramaniam,T.Balasubramanian, Jerry James and Sriharsha

Probe

34

Madras Metallurgical Services (P) LtdMetallurgists & Engineers

Serving Industries &Educational

Institutes for thepast 35 years

24, Lalithapuram street, Royapettah, Chennai 600014Ph: 044-28133093 / 28133903 Email: [email protected]

A-3, Mogappair Indl. Area (East) JJ Nagar,Chennai 600037 Phone 044-26564255, 26563370

Email: [email protected];[email protected] www.kidaolabs.com

KIDAO Laboratories

Scaanray Metallurgical Services(An ISO 9001-2000 Certified Company)

NDE Service ProviderProcess and Power Industry, Engineering andFabrication Industries, Concrete Structures,

Nuclear Industries, Stress Relieving

Electro-Magfield Controls & Services &LG Inspection Services

Plot 165, SIDCO Industrial Estate, (Kattur)Thirumullaivoil, Vellanur Village, Ambattur Taluk

Chennai 600062 Phone 044-6515 4664 Email: [email protected]

We manafucture : Magnetic Crack Detectors, Demagnetizers,Magnetic Particles & Accessories, Dye Penetrant Systems etc

Super Stockist & Distributors: M/s Spectonics Corporation, USAfor their complete NDT range of productrs, Black Lights, Intensity

Meters, etc.

Betz Engineering &Technology Zone

An ISO 9001 : 2008 Company

Call M. Nakkeeran, Chief Operations,Lab: C-12, Industrial Estate, Mogappair (West), Chennai 600037

Phone 044-2625 0651 Email: [email protected] ;www.scaanray.com

49, Vellalar Street, near Mount Rail Station, Chennai 600088Mobile 98401 75179, Phone 044 65364123Email: [email protected] / [email protected]

International Training Division21, Dharakeswari Nagar, Tambaram to Velachery Main Road,Sembakkam, Chennai 600073www.betzinternational.com / www.welding-certification.com

NABL Accredited Laboratory carrying out Ultrasonictest, MPL and DP tests, Coating Thickness and

Roughness test. We also do Chemical and Mechnical testsMetallographyStrength of MaterialsNon Destructive TestingFoundry Lab

Classifieds

Shri. K. Ravindran, Level IIIRT, UT, MT, PT, VT, LT, ET, IR, AE, NR and VA

Southern Inspection ServicesNDT Training in all the

following eleven Methods

No.2, 2nd Floor, Govindaraji Naicker Complex,Janaki Nagar, Arcot Road,

Valasaravakkam, Chennai 600 087Tamil Nadu, India

Phone : +91 44-2486 4332, 2486 8785, 4264 7537E-mail: [email protected] and

[email protected] Website: www.sisndt.com/www.ndtsis.com/

wwwpdmsis.com

01J, First Floor, IITM Research Park, Kanagam Road, Taramani,Chennai 600113 India Phone : +91 44 6646 9880

Dhvani R&D Solutions Pvt. Ltd

E-mail: [email protected] www.dhvani-research.com

Transatlantic Systems

Support for NDT ServicesNDT Equipments, Chemicals and Accessories

Call DN Shankar, Manager14, Kanniah Street, Anna Colony,

Saligramam, Chennai 600093Phone 044-26250651 Email: [email protected]

• Inspection Solutions - CUPS, TAPS, CRISP, TASS• Software Products - SIMUT, SIMDR• Training - Guided Waves, PAUT, TOFD• Services & Consultancy - Advanced NDE, Signal Processing

- C-scans, On-line Monitoring

NATIONAL CERTIFICATION BOARDANNOUNCEMENT

ISNT – Level IIICertification Programme

January – February, 2013Pune, India

Dr. M.T. ShyamsunderController of Examinations

National Certification Board, ISNT

All payments shall be made through the means of acrossed Demand Draft drawn favouring

“NCB - ISNT” and payable at “Chennai”.Cheques will not be accepted.

The last date for receipt of application alongwith payment is 21st December, 2012.

Course Director and contact person at Pune:

Shri Bhausaheb K. PangareCourse Director

C/o. M/s. Quality NDT ServicesPlot No BGA 1/1,2,3, Bhosari General Block,

MIDC, Bhosari, Pune-411026Ph.: 020-27121843/27119490/8600100700

[email protected] / [email protected] /[email protected]

Indian Society forNon - Destructive Testing

(Regd. Society: S. No. 49 of 1981) Module No. 60 & 61, Garment Complex,

SIDCO Industrial Estate, Guindy, Chennai 600 032 Tele : 044-22500412, 044-42038175

E Mail: [email protected] , [email protected] ISNT Invites nominations/applications for the

National NDT Awards fromIndian Nationals for theirsignificant contributions

and excellence in the fieldof NDT and the Best

Chapter award from all thechapters of ISNT. Anannouncement to this

effect is already circulatedto eligible members /

chapters.

The various categories andawards are listed in every

issue of the JNDE.

The nominations /applications are to be sent

in the prescribed formwhich can be downloadedfrom www.isnt.org.in and15 copies of the same areto be sent to the following

address by10th September, 2012.

Shri Dilip P. TakbhateChairman,

Awards Committee,Indian Society for

Non Destructive TestingModule No.60 & 61,

Third Floor,Readymade Garment Complex,

SIDCO Industrial Estate,Guindy, Chennai – 600 [email protected]

I S N TANNUALAWARDS

CHAPTER NEWSOther Activities:EC Meeting held on 9th July 2012 at ISNT, Mumbai Office.EC Meeting held on 10th August 2012 at ISNT, Mumbai Office.AGM- 2012 on 8th September 2012 at Hotel Chakra, Sakinaka.Around 180 members had attended the Meeting and we also hada Lecture on Stress Management by Dr Shriniwas Kashalikarduring AGM.EC Meeting held on 28th September 2012 at ISNT, MumbaiOffice.18 Life Members and 1 Associate Member were enrolled to ISNT,Mumbai Chapter during this period.

7. Pune

Technical Talk:Three Technical Lectures cum events were held during the year.1. CE Marking by Shri. Hiremath- Hon.Jt. Secy of Pune Chapter.2. ISNT NDE Stalwart Shri.K.Viswanathan on NDE in Indias

Research Programme on 8.8.2012.3. NDE Awareness Programme for the third year Engineering

Students of Kolhapur Institute Of Technology (KIT) for 4commonly used NDE methods viz. RT, UT, MT, PT in fourParallel sessions. Lecture on Career Opportunities in NDEwas also delivered.

Other Activities:Membership has marginally increased around 165 nos.

8. Sriharikota

Other Activities:New Executive Committee has been elected for the term 2012-2014.

9. Trivandrum

Other Activities:Dr.V.R.RAVINDRAN NDT Vigyan Purashkar: ISNT Trivandrumchapter called for NDT V.R.Ravindran Purashkar in memory ofDR. R.Ravindran for students of Academic Institutions in Keralaand Members of Trivandrum Chapter pursuing their students inany Institute of India. The Award is for Best Doctoral Thesis orProject Report or Technical Paper or Innovation in the areawhich directly or indirectly benefiting Non-Destructive Testingand Evaluation. One issue of Image, the technical bulletin of thechapter was released in June, 2012.AGM was conducted on 23rd June, 2012 at Classic Avenue,Trivandrum:Dr. B. Venkatraman, Associate Director IGCAR delivered the MRKurup memorial lecture.Annual Technical Meet: Sri. Anil Kesavan, NLPTA certified trainerdelivered the Annual Technical Meet lecture on 23rd June 2012.The topic Why we behave the way we do.Seminar on He leak detection: A seminar on He leak detectionand its techniques was organized by the chapter jointly with M/s Agilent Technologies India Ltd. on 10th Sep, 2012.

10. Trichy

Technical Talk: Invited and joint lectures:· One day workshop in NDT for the Students of J J collage

of Engg. Technology Trichy on 27th September 2012.Courses & Exams:· Package program jointly with WRI and PSG College of

tech, Level-II in RT, UT, MT & PT.· Certified Radiographer course RT-I in association with BARC,

Mumbai.· One Year NDT Training Program for BHEL Employee

wards in association with BHEL Educational Society, Level-II in RT,UT,MT & PT

· Surface NDT methods Level II in PT & MT(09/07/12 to16/07/12)

1. ChennaiCourses & Exams Conducted :1. In-house on UT Level-II (ASNT) Kochi from 02.07.2012

to 10.07.2012.2. RT Level-II (ISNT) course from 09.07.2012 to 22.07.2012.3. UT Level-II (ASNT) course from 20.07.2012 to 29.07.2012.4. RT Level-II (ASNT) course from 31.07.2012 to 08.08.2012.5. MT & PT Level-II (ASNT) course from 03.08.2012 to

12.08.2012.6. RT Level-II (ASNT) course from 17.08.2012 to 26.08.2012.7. UT Level-II (ISNT) course from 03.09.2012 to 15.09.2012.EC Meeting held on 01.07.2012EC Meeting held on 15.08.2012AGM was celebrated on 14.07.2012 at ISNT, Conference Hall,Chennai.

2. Delhi

Core committee meeting was conducted on 25.08.2012 at IndianCoffee house connaught Place all about NDE 2012 and itspromotion activities.

3. Hyderabad

Technical Talk:Dr. N. Kondal Rao Memorial Lecture “The pervasive importanceof materials: Biological to Nuclear” was delivered by PadmaVibhushan Dr. R. Chidambaram, Principal Scientific Advisor toPM New Delhi on 14/07/2012.One day Work shop was held on HLT by Mr.Peter Nico Palenstyn,USA, which was jointly organised by ISNT HC & M/s AgilentTechnologies at Hyderabad.Courses & Exams Conducted :Hyderabad Chapter conducted courses & Examinations on variousmethods like VT,PT,ET,UT for Level-I during August to October,2012.VT Level-I conducted from 21-24 Aug-2012LPT Level-I conducted from 27-30 Aug-2012ET Level-I conducted from 11-17 Sept -2012UT Level-I conducted from 25th Sep 01st Oct 2012.Other Activities:2nd EC meeting was held on 25/08/2012

4. Kolkata

Courses and Exams Conducted:1. MPT Level II conducted from 15th 18th July, 20122. PT Level II conducted from 22nd 24th July, 20123. RT Level II conducted from 1st 7th August, 2012-11-124. UT Level II conducted from 18th 24th September, 2012.Other ActivitiesThree Life Membership application were received during thisperiod and the chapter has around 209 members.

5. Kalpakkam

Courses and Exams Conducted:Level I for LT & Level II Courses and Exams were conductedfor UT, LT, VT, PT

6. Mumbai

Courses and Exams Conducted:1. ECT Level- II examination was conducted at NDTS on 28th

July 2012.2. Welding Inspector examination was conducted on 2ndSep,

2012 at ITT, Mahim.3. General NDT Course for ONGC Engineers from 10th

September 2012 to 14th September 2012 at Hotel Atithi.

9

vol 11 issue 2 September 2012Journal of Non Destructive Testing & Evaluation

10


Dr Prabhu RajagopalAssistant Professor and Associate, Centre for Nondestructive EvaluationIIT Madras, Chennai 600036, T.N., India

Basics

Modelling and simulationfor NondestructiveEvaluation

Abstract

This paper provides a rapid summary of the background, motivations and keyfeatures of modelling methods that underlie wave-based Nondestructive Evaluation(NDE) techniques. Different domains where modelling methods are used in theNDE process are first described in brief. This is followed by a quick descriptionof different paradigms in the modelling of wave generation and scattering,together with their applications and limitations. The basis of various well-known analytical, approximate and purely numerical procedures used for modellingof NDE phenomena are discussed. The article takes a descriptive rather thanrigorous and mathematical approach, and aims to provide the reader with aready summary with necessary reference material provided for further andadvanced studies of the subject.

illuminate the volume of the material. Uponinteraction with an artefact, waves arescattered back and are then picked up eitherby the same transducer as in the pulse-echomode in Figure 1(a) or by a receivingtransducer in Figure 1(b). The total fieldf total in the medium is usually then consideredas the sum of incident and scattered fields,

f total = f incident + f scattered (1)

Underlying this depiction are some basicabstractions we have made: that waves cantravel in rays towards an artefact, that thematerial is homogenous enough not to disturbthe rays in their path, the artefact is largeenough to obstruct some of the rays causingthem to reflect or scatter, and that a receivingmechanism lies in the path of such scatteredrays.

Understanding this process is essential tothe success of the NDE method in practice,as we are looking for optimal positioning ofthe probing and receiving mechanism, atechnique that is not disturbed by the materialitself and is rather only sensitive to anydefective regions, and some relationship ofthe defect features upon the measurements.In order to obtain this information we mustalso have insight into how the materialresponds to the transduction mechanism andthe wave considered and how types ofdefects interact with the waves.

Finally, a summation such as described inEq.1 is possible only if the wave field canbe assumed to behave linearly in the materialconsidered and therefore we must knowwhat regimes and excitation frequencies ortypes lead to linear or non-linear behaviour.Thus modelling, analysis and simulation arenecessary for even a basic deployment andsuccess of an NDE method. Accuratemodelling often helps explain counter-intuitive results, design experiments, andpredict defect severity based on practicalmeasurements.

regarding the modelling of waves in materialsthat underlie such wave-based NDE methods.The approach taken is to introduce variousabstractions and methods, rather than arigorous treatment (see for example,Rajagopal [1]).

II. INTRODUCTION

Modelling essentially consists of anabstraction which helps us to understand,describe and make predictions of a physicalphenomenon considered and as such, is almostinstinctive to the way we think. Let usconsider two simple inspection scenarios asshown in Figure 1, representing the pule-echo and pitch-catch modes of inspection ofa wave-based NDE method. Waves arelaunched from a transducer in both cases and

I. BACKGROUND

Nondestructive Evaluation (NDE) methodsmake use of techniques and procedures thatcan provide insight into the integrity ofstructures and materials. Well-known NDEmethods include simple ones such as dyepenetrant testing, remote visual inspectionas well as more advanced wave-basedtechniques such as ultrasonic testing,radiographic testing, eddy current testing andmagnetic particle inspection. Often in orderto obtain information regarding the volumeof a structure, NDE procedures require theuse of wave-based methods such as ultrasonictesting instead of surface preparationtechniques such as dye-penetrants. In thisarticle we will look at some basic concepts

Fig. 1 : Schematic showing (a) the pulse-echo and (b) pitch-catch modes of inspection of a wave-based NDE method.

11


Basics

and Feshbach (McGraw-Hill, 1953) [17] setsan excellent rigorous basis of the variousmethods of analysis. A thorough review ofexact and approximate methods inelectromagnetics and acoustics can also befound in the book by Bowman et al. (North-Holland, 1969) [18]. Here we only outlinethe methods in a summary manner.IV.1. Variable separation: The directapproaches to modelling of wave phenomenaessentially aim to solve partial differentialequations (see Eq.2) describing them, subjectto some initial or boundary conditions:conditions occurring at material boundaries,or at defects. The problem of solving partialdifferential equations can be simplified toone of solving several ordinary differentialequations, if solutions which are products ofseveral functions each depending on onlyone variable can be found [19]. This formsthe basis for the simplest modelling techniqueknown as the wave function expansionmethod (also called the eigenfunctionexpansion or the variable separation method),which seeks variable-separable solutions tothe wave equations as described in Eq. 2. Ifa defect is bounded by constant coordinatesurfaces of a standard coordinate system inwhich the wave equations permit suchseparation, this method can be applied. Thesolutions are obtained as expansions or linearcombinations of eigenfunctions of theresulting ordinary differential equations,which are often special functions with knownproperties. The unknown coefficients in theexpansions are then evaluated from thesystem of linear equations resulting fromapplying appropriate continuity conditionsat the boundaries of defects.However this kind of separation in variableis possible only for a limited number ofcoordinate systems: for the scalar waveequation, this happens only for 11 coordinatesystems, enumerated in the book by Harker(p 67, Ch 3) [20]. Of these 11 systems,only 6 permit a similar decomposition forthe vector wave equation, namely, therectangular, circular, elliptic and paraboliccylindrical, spherical and the conical systems[21]. Under some special conditions whencylindrical symmetry is satisfied, morecoordinate systems such as parabolic, oblatespheroidal and prolate spheroidal systemsalso join the list. This means that only defectswhose contours can be captured by thissmall set of coordinate systems can bemodelled by using the wavefunctionexpansion method. Even then, for geometriesother than the cartesian, cylindrical orspherical, the analysis gets too rigorous toextract results for practical applications.IV.2 Integral representations andreciprocity: Physically, the scattering ofelastic waves comes about from an obstacleto waves acting as a secondary source, asdescribed by the Huygens Principle. Thus

wavelength for resolution of image artefacts[9,10].III.4 Reliability: Quantification of errors inpredictions is crucial to the usefulness of theresults obtained by NDE methods andmodelling plays a very important part ofthis. This is usually done through probabilityof detection (PoD) curves which plot theprobability that a defect of a certaincharacteristic dimension is detected againstincreasing values of the dimension chosen[11]. Although much of PoD data is obtainedby painstaking experimentation, modellingand simulation leading to Model-assisted orMA-PoD has become an important part ofthis process [12,13]. Modelling can provideguidance to explaining PoD curves, as wellas in going about the PoD curve generationprocess through results for cases whereexperiments are difficult or time-consuming.III.5. Remaining life prediction: Althoughusually not considered part of the domain ofNDE, it is important part to obtain anestimate of active remaining life of acomponent diagnosed with defects, in orderto decide if it must be replaced or cancontinue operation. Fracture mechanics isone of the main techniques used for modellingthe growth of cracks in defectivecomponents. The predictions are based on anumber of approaches that consider linear ornon-linear behaviour for defects under appliedloading conditions [14,15].

IV. MODELLING TECHNIQUESThe study of wave propagation in materialsand scattering by artefacts is at the heart ofmany modelling procedures, and providesdata for other modelling routines such asimaging and reliability analysis. Wave-basedNDE methods make use of acoustic,ultrasonic or electromagnetic waves, as suitedto different applications. Sound waves in airor non-viscous liquids and two-dimensionalelectromagnetic wave problems can bestudied using a single scalar potential withthe wave equation,

(2a)

General electromagnetic wave problemsrequire a single wave equation involving avector potential Ψ

→ :

(2b)

which leads to three scalar equations, oneeach in the three directions where Ψ

→ is

defined.

The text by Jones (Clarendon Press, 1986)[16] provides an overview of methods inthese areas, while the classical text by Morse

III. MODELLING DOMAINS

Modelling is applied to several domains ofthe NDE process, where it serves differentpurposes as described below in this section.III.1 Wave generation and reception: Thisrequires an understanding of the transductionmechanism used in the probes used for theNDE technique under consideration. Forexample, piezoelectricity is often used inultrasonic transducers, and properties ofpiezoelectric crystals are modelled to gaininsight into the excitation of particular (shearor longitudinal) modes and frequencycharacteristics (see for example, Auld [2]).The generating transducer is often consideredthe primary source for waves inside thematerial, whereas the receiving transducer issensitive to waves scattered by artefacts inthe material which act as secondary sources.The source for this terminology is the well-known Huygens’ principle due to the DutchPhysicist Christiaan Huygens who firstdemonstrated this effect for light waves in amaterial that acts as a discontinuity for thewave can act as a secondary source: thusgrain boundaries in a material often scatterwaves causing the generally observed randomnoise in ultrasonic measurements. Cracks orvoids act as collections of secondary sourcesdistributed over a surface, whereas featuressuch as material changes contribute secondarysources distributed across a volume. Recentadvances in transduction such as the use ofregular and conformable phased arrays requiremore sophisticated models to predict wavegeneration and reception [3,4].III.2 Wave propagation and scattering :This involves a representation of how amaterial supports a wave and how defects inthe material can interact with the waves. Forexample, we can often assume that metallicmaterials are homogenous and isotropic on alarge scale, and on the other hand, materialssuch as composites have non-isotropiccharacteristics.III.3 Signal and image analysisA practical goal is to correlate experimentalmeasurements with defect signatures. Thisusually follows from investigation of wavescattering by defects and may requireextensive parametric studies involvingdifferent defect features (see for example,[5,6]). Imaging methods are also often beused to obtain practically useful maps ofdefect locations that can be read directly byoperators [7,8]. The image space is a map ofthe actual space and obtaining the relationbetween these is the domain of imaginganalysis. Imaging analysis can vary fromsimple methods to advanced procedures thatmay be used to resolve and distinguishartefacts mutually or from material features.Recent advances (termed super-resolutionmethods) are helping to break the traditionalRayleigh-limit barrier of half the operating

12


Basics

In the Kirchhoff approximation the surfaceof the obstacle is divided into illuminatedand shadow regions (illustrated in Figure 2),denoted for example by S+ and S_ , and thetotal field is approximated as:

(f total)+ = f incident + f incident

+ f reflected ; (f total)– = 0 (3)

where the term (ftotal)+ can be called the‘geometrical ray’ field

This way, the Kirchhoff approximationmodels the scattering behaviour as if at eachelement of the scattering surface, incidentplane waves interact with unboundedinterfaces having the same surface normal.Like the Born approximation with which itshares a number of features, the Kirchhoffapproximation has been used widely inscattering and inverse problems Achenbach[38] Chapman [39, 40] Schmerr [41]. Thelast quoted work [41] contrasts the Bornand the Kirchoff approximations andcomments on their limitations. Schmerr etal., [42] presented a framework unifyingthese two approximate methods.

The GR field has its limitations- it isdiscontinuous at the boundaries of shadows(defined by Snells law) and vanishes totallyin the shadow region. In reality, energy iscontinuously radiated into the geometricalshadow of the obstacle by waves whichtravel around its surface (see Figure 2 above),causing diffraction. The Kirchhoffapproximation recovers the first singlydiffracted field, but even this becomesincorrect in the presence of sharp edges. TheGeometrical Theory of Diffraction (GTD)[43], has been formulated to include the fulleffects of diffraction within the generalframework of ray theory. Keller [44] firstdeveloped the method rigorously for scalarwaves and the method was later extended toelastodynamics by Resende (2D case) and ingeneral by Achenbach and co-workers [37].

The GTD presents a correction to the totalscattered field of the form:

f total = f GR + f diffracted (4)

material region in the absence of defects.Versions of this (Born) approximation areused in many imaging algorithms. In thesimilar quasi-static approximation, fields andtheir gradients on or inside scatterers areapproximated by those due to a static loadequivalent to the low-frequency load applied.The Born approximation works best in back-scattering while the quasi-staticapproximation is applicable in general in thelong-wavelength limit [34-36].

IV.3.2 High frequencies: In the high-frequency limit, ray methods provide anexcellent route to approximate solutions forscattering problems (The book by Achenbachet al.(Pitman Books, 1982) [37] for example,covers ray methods for elastic waves indetail). Ray methods are based onconstructing high frequency series solutionsto the governing wave equations (Eq. 2)

involving terms of the type , where

ω is the circular frequency, h is acharacteristic dimension of the defect and cis the wave speed, which would be validasymptotically as ωh/c→∝. In practice, theyare known to give useful results even atwavelengths comparable to h and the resultscan be extended to the time-domain.

Physically, such expansions have a simplegeometric interpretation in terms of rays andthe leading term is just what would bepredicted by geometrical wave theory, whilesubsequent terms offer corrections to it.Simple geometrical ray theory involvingreflection and refraction often yields accurateresults close to the specular or near-speculardirections this is the usual representation ofwave phenomena in the NDE context, asillustrated in Figure 1. The standardgeometrical ray (GR) theory can also beused directly to solve scattering problemswhen the obstacles considered do not havesharp edges.

The GR field is quite accurate in describingcertain problems such as backscattering fromsmoothly curved objects with a curvaturelarger than then incident wavelength. Butwhen the edges of the object begin to havea strong influence leading to a sharply definedshadow, edge diffraction becomes importantand standard ray theory becomes inadequate.Approximate solutions to the scattered fieldcan then be obtained through the integralrepresentation formulas, taking the predictionby ray methods as an estimate of the totalscattered displacement on the obstacle. Asimple way to improve the accuracy is touse the GR field as an approximation for thetotal scattered field on the obstacle in therepresentation integrals yielding the famous‘Kirchhoff approximation’.

the response of a system of defects toincident waves can be thought of as arisingdue to a set of secondary sources distributedon the surface of or in the volume of thedefective region. The mathematical basis forthis principle as applicable to acoustic (scalar)waves is given by Helmholtzs integralformulas in the steady state and Kirchhoffsgeneralization for arbitrary time-dependence[21]. Based on analogous results for vectorand elastic waves, it is possible to derive analternative integral representation for thescattering problem, which is more direct andintuitively closer to the physics (see Paoand Varatharajulu [22] and Gubernatis et al.[23] for example, for a discussion of elasticwave scattering).

Reciprocity concepts relate loading-responsepair configurations in a medium and as suchform the basis of a number of related conceptsin mechanics and electrostatics, such as theprinciple of virtual work. Reciprocitytheorems in conjunction with the Greensfunction present another powerful and elegantroute in arriving at integral relations for wavescattering problems. Tan [24, 25],Varatharajulu [26], Kino [27], Auld [28], Kinoand Khuri-Yakub [29] and more recently,Achenbach (Cambridge University Press,2003) [30] and Achenbach [31] provide acomprehensive overview and highlight thesignificance of reciprocity relations. Theexcellent book by de Hoop (Academic Press,1995) [32] also provides extensive furtherreference in this regard. Both approacheslead to integral relations connecting theincident and the scattered waves, and can besolved either analytically or through integralequations that can be solved numerically.

IV.3 Approximate analytical andnumerical methods: The integral equationrepresentation of the scattering problemprovides a convenient starting point for anumber of approximate analytical and purelynumerical solution methods.(see Hackman(Academic Press, 1993) [33] for a review ofsome exact and approximate analytical workon elastic wave and acoustic scattering).Approximate methods are normallyapplicable to low- or high-frequency regimesof the waves considered. Here low and highare relative terms, describing the extent ofdefects of interest against the wavelengthconsidered low frequencies being the domainwhere defects are small compared to thewavelength and high frequencies are wherethe wavelength is several times smaller thanthe defect dimension.

IV.3.1. Low frequencies: In the limit of lowfrequency or long wavelength, a commonapproximation is to assume that the fieldand its gradient inside (volume defects) oron the surface (voids, cavities and cracks) ofscatterers is the same as the response of that

Fig. 2 : Illustration of illuminated andshadow regions as consideredin the Kirchhoff approximation

13


9. M. Fleming, Far-field Super ResolutionImaging, PhD Thesis, MechanicalEngineering Department, Imperial CollegeLondon, 2008, London, UK.

10. T Hutt and F Simonetti, Experimentalobservation of super-resolution imagingin highly attenuative media, Review ofProgress in Quantitative NondestructiveEvaluation, eds D.O. Thomson and D.E.Chimenti, American Institute of PhysicsConference Proceedings, 1430, pp. 739-746, 2012.

11. W.D. Rummel, Probability of DetectionAs a Quantitative Measure ofNondestructive Testing End-To-EndProcess Capabilities, MaterialsEvaluation, ‘Back to Basics’ series,(1998). Available online at: http://www.asnt.org/publications/materialseval/basics/jan98basics/jan98basics.htm

12. S.N. Rajesh, L. Udpa and S.S. Udpa,Numerical Model Based Approach forEstimating Probability of Detection inNDE Applications, IEEE Transactionson Magnetics (1993), VOL. 29(2): p.1857-60.

13. R.B. Thompson, A unified approach tomodel-assisted determination ofprobability of detection, Review ofProgress in Quantitative NondestructiveEvaluation, (eds) D.O. Thomson and D.E.Chimenti, American Institute of PhysicsConference Proceedings, 975: p. 1685-1692, 2008.

14. D. Broek, Elementary EngineeringFracture Mechanics, Kluwer AcademicPublishers, Dordrecht, 1986.

15. T.L. Anderson, Fracture Mechanics -Fundamentals and Applications, 3rdEdition, Taylor and Francis Group, 2005.

16. Jones, D.S., Acoustic and ElectromagneticWaves. 1986, Oxford: Oxford UniversityPress.

17. Morse, P.M. and H. Feshbach, Methodsof Theoretical Physics. 1953: McGraw-Hill Book Company.

18. Bowman, J.J., T.B.A. Senior, and P.L.E.Uslenghi, eds., Electromagnetic andacoustic scattering by simple shapes.1969, Amsterdam: North-HollandPublishing Company.

19. Hopf, L., Introduction to the DifferentialEquations of Physics. 1949, New York:Dover Publications.

20. Harker, A.H., in Elastic waves in solids.With applications to NondestructiveTesting of Pipelines. 1988, IOPPublishing Ltd and British Gas plc:Bristol, England and Philadelphia, USA.

21. Martin, P.A. and G.R. Wickham,Diffraction of Elastic Waves by a Penny-Shaped Crack: Analytical and NumericalResults. Proceedings of the Royal

wave generation and scattering werepresented. The basis of various well-knownanalytical, approximate and purely numericalprocedures used for modelling of NDEphenomena were discussed, together withtheir applications and limitations. Referencematerial further and advanced studies of thesubject are provided at the end of the article.

ACKNOWLEDGEMENT

Part of the literature review featured in thispaper was carried out when the author wasa PhD student at Imperial College London,between 2003-2007.

REFERENCES1. P. Rajagopal, Towards Higher Resolution

Guided Wave Imaging: Scattering Studies,PhD Thesis, Mechanical EngineeringDepartment, Imperial College London,2008, London, UK.

2. Auld, B.A., Acoustic fields and waves insolids. Vols. 1 and 2. 1973, 1990, Florida:Robert E. Krieger Publishing Company.

3. Russell, J., Long, R. and Cawley, P.Development of a membrane coupledconformable phased array inspectioncapability, Review of Progress inQuantitative NDE, Vol 29,DO Thompson and DE Chimenti (eds),American Institute of Physics, pp. 831-838, 2010.

4. Russell, J., Long, R. and Cawley, P.Development of a twin crystal membranecoupled conformable phased array forthe inspection of austenitic welds, Reviewof Progress in Quantitative NDE, Vol30, DO Thompson and DE Chimenti(eds), American Institute of Physics, pp.811-818, 2011.

5. Alleyne, D.N. and P. Cawley, Thequantitative measurement of Lamb waveinteraction with defects, IEEETransactions on Ultrasonics,Ferroelectrics and Frequency Control,39(3): p. 381-397.

6. Ditri, J.J., Utilization of guided elasticwaves for the characterization ofcircumferential cracks in hollowcylinders. Journal of the AcousticalSociety of America, 1994. 96(6): p. 3769-3775.

7. Holmes, C, Drinkwater, BW & Wilcox,PD., Post-processing of the full matrixof ultrasonic transmit-receive array datafor non-destructive evaluation, NDT andE International, 38(8): p. 701-711, 2005

8. Davies J, Cawley P, The Application ofSynthetic Focusing for Imaging Crack-Like Defects in Pipelines Using GuidedWaves, IEEE Transactions on ultrasonics,ferroelectric and frequency control, 2009,56, : p. 759-771.

The term representing , the diffracted field,offers only an insignificant correction to theGR field in the backscatter and speculardirections, but contributes strongly in theforward direction. itself is further constructedfrom components consisting of primary orthe first edge diffraction and secondary ormultiple diffraction:

(5)

These diffracted fields are constructedanalogously to the reflected fields of thegeometrical ray method and the scatteredamplitudes related to the incident onesthrough Diffraction coefficients. Appropriatecanonical problems (for example, for crackproblems, the canonical problem is that ofelastic wave scattering from a semi-infinitestraight-edge crack; for scattering from convexsurfaces, the canonical problem is that frominfinite cylinders see [45]) whose solutionsare known are selected and the Diffractioncoefficients are obtained by comparing theray solution with them. In the usual raymanner, the geometry and the curvature ofthe wave front as well as the defect are thenincorporated through the edge conditions,which lead to the generalized Snells Law fordiffraction.

IV.3.Intermediate frequencies: In theintermediate frequency regime, severalnumerical methods, including FiniteDifference (FD) [46,47] Finite element(FEM) [48-50] and t-matrix (see [51] for anexcellent description of the t-matrix method)[52-55] methods have been used. The Finiteelement method scores over the t-matrixmethod in that it can treat pulses in generaland is not a single frequency method like thelatter. Harumi and Uchida [56] provide agood review of various numerical studies,mainly the FEM. Numerical methods arevery versatile in that they solve the scatteringproblem for any frequency regime andgeometry and in this sense, can be seen moreas experimental simulations than analyticalsolutions. A key issue though, is that theyoften dont provide generic results as theytend to be constructed for specific cases.However with ever advancing computerpower and the wide availability of robustcommercial packages in recent years, coupledwith increasing complexity of practicalinspection, numerical modelling has becomewidespread in recent years.

V. CONCLUSIONS

This paper provided a summary of themotivations and key features of modellingmethods that underlie wave-basedNondestructive Evaluation (NDE)techniques. Following a brief description ofthe domains where modelling methods areused, different paradigms in the modelling of

Basics

14


February, 1991).

46. Harker, A., Numerical modelling of thescattering of elastic waves in plates.Journal of Nondestructive Evaluation,1984. 4(2): p. 89-106.

47. Harumi, K., Okada, Hiaso., Saito,Tetsuo., and Fujimori, Toshiaki.,Numerical experiments of reflection ofelastic waves by a crack or an ellipticcylinder. IEEE Ultrasonics Symposium,1982: p. 1064-1069.

48. Datta, S.K., Fortunko, C.M., and King,R.B., Sizing of surface cracks in a plateusing SH waves. IEEE UltrasonicsSymposium, 1981: p. 863-867.

49. Datta, S.K., A.H. Shah, and C.M.Fortunko, Diffraction of medium and longwavelength horizontally polarized shearwaves by edge cracks. Journal of AppliedPhysics, 1982. 53(4): p. 2895 - 2903.

50. Abduljabbar, Z., S.K. Datta, and A.H.Shah, Diffraction of horizontallypolarized shear waves by normal edgecracks in a plate. Journal of AppliedPhysics, 1983. 54(2): p. 461 - 472.

51. Pao, Y.-H., Mathematical theories of thediffraction of elastic waves, inProceedings of the first internationalsymposium on ultrasonic materialscharacterization held at NBS,Gaithersburg, Md. , June 7-9, 1978(National Bureau of Standards specialpublication 596, ‘Ultrasonic materialscharacterization’), H. Berger and M.Linzer, Editors. 1980, National Bureauof Standards (USA).

52. Hackman, R.H., and Todoroff, DouglasG., An application of the spheroidal-coordinate-based transition matrix: Theacoustic scattering from high aspect ratiosolids. Journal of the Acoustical Societyof America, 1985. 78(3): p. 1058-1071.

53. Varatharajulu, V., and Pao, Yih-Hsing.,Scattering matrix for elastic waves. I.Theory. Journal of the Acoustical Societyof America, 1976. 60(3): p. 556-566.

54. Visscher, W.M., A new way to calculatescattering of acoustic and elastic wavesI. Theory illustrated for scalar waves.Journal of Applied Physics, 1980. 51(2):p. 825-834.

55. Visscher, W.M., A new way to calculatescattering of acoustic and elastic wavesII. Applications to elastic wavesscattered from voids and fixed rigidobstacles. Journal of Applied Physics,1980. 51(2): p. 835-845.

56. Harumi, K., and Uchida, M., Computersimulation of ultrasonics and itsapplications. Journal of NondestructiveEvaluation, 1990. 9(2/3): p. 81-99.

crack. Journal of Applied Physics, 1978.49(5): p. 2599-2604.

35. Gubernatis, J.E., et al., The Bornapproximation in the theory of thescattering of elastic waves by flaws.Journal of Applied Physics, 1977. 48(7):p. 2812-2819.

36. Jain, D.L., and Kanwal, R.P., The Bornapproximation for the scattering theoryof elastic waves by two-dimensionalflaws. Journal of Applied Physics, 1982.53(6): p. 4208-4217.

37. Achenbach, J.D., A.K. Gautesen, and H.Mcmaken, Ray methods for waves inelastic solids. 1982, Pitman Books Ltd:London.

38. Achenbach, J.D., et al., Diffraction ofultrasonic waves by penny-shaped cracksin metals: theory and experiment. Journalof the Acoustical Society of America,1979. 66(6): p. 1848-1856.

39. Chapman, R.K. and J.M. Coffey, Atheoretical model of ultrasonicexamination of smooth flat cracks, inReview of Progress in Quantitative NDE,D.O. Thompson and D.E. Chimenti,Editors. Vol. 3. 1984, Plenum: New York.p. 151-162.

40. Chapman, R.K., Ultrasonic reflectionfrom smooth flat cracks: Exact solutionfor the semi-infinite crack, in CEGBReport NW/SSD/RR/14/81. 1981, N.D.TApplications Centre, Scientific ServicesDepartment, Central Electricity Board,U.K., North Western Region.

41. Schmerr Jr, L.W., Song, Sung-Jin., andSedov, Alexander, Ultrasonic flaw inversesizing problems. Inverse Problems, 2002.18: p. 1775-1793.

42. Schmerr Jr, L.W., A. Sedov, and C.-P.Chiou, A unified constrained inversionmodel for ultrasonic flaw sizing. Researchin Nondestructive Evaluation, 1989. 1:p. 77-97.

43. Karal Jr, F.C., and Keller, Joseph B.,Elastic wave propagation in homogeneousand inhomogeneous media. Journal of theAcoustical Society of America, 1959.31(6): p. 694-705.

44. 108. Keller, J.B., and Karal Jr, Frank C.,Geometrical theory of elastic surface-wave excitation and propagation. Journalof the Acoustical Society of America,1964. 36(1): p. 32-40.

45. Podil’chuk, Y.N., Y.K. Rubtsov, and P.N.Soroka, Geometrical theory of diffractionin the scattering of harmonic elastic wavesby smooth convex cavities. InternationalApplied Mechanics, 1991. 27(2): p. 131-140 (Translated from PrikladnayaMekhanika, Vol. 27, No. 2, pp. 2635,

Society of London. Series A,Mathematical and Physical Sciences,1983. 390: p. 91-129.

22. Pao, Y.-H. and V. Varatharajulu, Huygens’principle, radiation conditions, andintegral formulas for the scattering ofelastic waves. Journal of the AcousticalSociety of America, 1976. 59(6): p. 1361-1371.

23. Gubernatis, J.E., E. Domany, and J.A.Krumhansl, Formal aspects of the theoryof the scattering of ultrasound by flawsin elastic materials. Journal of AppliedPhysics, 1977. 48(7): p. 2804-2811.

24. Tan, T.H., Reciprocity relations forscattering of plane, elastic waves. Journalof the Acoustical Society of America,1977. 61(4): p. 928-931.

25. Tan, T.H., Far-field radiationcharacteristics of elastic waves and theelastodynamic radiation condition.Applied Scientific Research, 1975. 31(5):p. 363 - 375.

26. Varatharajulu, V., Reciprocity relationsand forward amplitude theorems forelastic waves. Journal of MathematicalPhysics, 1977. 18(4): p. 537-543.

27. Kino, G.S., The application of reciprocitytheory to scattering of acoustic wavesby flaws. Journal of Applied Physics,1978. 49(6): p. 3190-3199.

28. Auld, B.A., General electromechanicalreciprocity relations applied to thecalculation of elastic wave scatteringcoefficients. Wave Motion, 1979. 1: p.3-10.

29. Kino, G.S., and Khuri-Yakub, B.T.,Application of the Reciprocity theoremto Nondestructive evaluation. Researchin Nondestructive Evaluation, 1992. 4:p. 193-204.

30. Achenbach, J.D., Reciprocity inElastodynamics. 2003, Cambridge:Cambridge University Press.

31. Achenbach, J.D., Reciprocity and relatedtopics in elastodynamics. AppliedMechanics Reviews, 2006. 59: p. 13-32.

32. de Hoop, A.T., Handbook of radiationand scattering of waves : acoustic wavesin fluids, elastic waves in solids,electromagnetic waves. 1995, London:Academic Press.

33. Hackman, R.H., Acoustic scattering fromelastic solids, in Physical Acoustics, A.D.Pierce and R.N. Thurston, Editors. Vol.XXII. 1993, Academic Press Inc: NewYork.

34. Domany, E., Krumhansl, J.A., and Teitel,S., Quasistatic approximation to thescattering of elastic waves by a circular

Basics

15


WWWWWant to Aant to Aant to Aant to Aant to Advt in our Jdvt in our Jdvt in our Jdvt in our Jdvt in our Journal ?ournal ?ournal ?ournal ?ournal ?

WWWWWant to rant to rant to rant to rant to reach oeach oeach oeach oeach ovvvvver 6000 memberser 6000 memberser 6000 memberser 6000 memberser 6000 membersacross India?across India?across India?across India?across India?

Advertisement tariffAdvertisement tariffAdvertisement tariffAdvertisement tariffAdvertisement tariff

Classified advt (Business card) : Rs. 600/-

BackCover (Multi Colour) : Rs. 22,000/-

Inside Cover - Front (Multi colour) : Rs. 11,000/-

Inside Cover - Back (Multi Colour) : Rs. 11,000/-

Inner Pages Full (Multi Colour) : Rs. 8,000/- US $ 300

Inner Pages Full (Black & White) : Rs. 6,000/- US $ 200

Inner Pages Half (Black & White) : Rs. 3,000/- US $ 125

Panel Display : Rs. 1,200/- US $ 50

** S** S** S** S** Sererererervice vice vice vice vice TTTTTax 12.36% extraax 12.36% extraax 12.36% extraax 12.36% extraax 12.36% extra

Mr. V. PariManaging Editor

Journal of Non Destructive Testing & EvaluationModules 60 & 61, Readymade Garment Complex,SIDCO Industrial Estate, Guindy, Chennai 600032

Phone: (044) 2250 0412Email: [email protected] ; [email protected]

SSSSSubscribe to JNDTEubscribe to JNDTEubscribe to JNDTEubscribe to JNDTEubscribe to JNDTEJournal subscription

(India & SAARC countries)includes ISNT membership & subscription to

Journal of Non Destructive Testing & Evaluation

Payments should be made byDemand Draft, drawn favoring

“JNDE – ISNT”, payable at Chennai.

Journal Subscription forR & D establishments and educational institutions Rs. 700/- per annum and US $ 60

Service Tax 12.36% extra

16


Dr. C V KrishnamurthyCentre for NDE and Department of Physics, IIT MadrasChennai 600036 Tamilnadu Email ID: [email protected]

Sonic Infrared Imaging– Emerging NDE Tool

being phase-sensitive, allowing flawdepth to be retrieved from the phaselag.

We shall look at three representativecases before we come to what causesthis heating to take place. Referencesgiven at the end provide additionalmaterial.

Case 1: Metals

Sonic IR imaging technique used was ashort sound pulse (50 – 200 ms) of highfrequency (typically 20–40 kHz ultrasonicwelding generator (power ~ 1 kW)). Itwas applied on the surface of the objectunder inspection through a guncontaining a piezoelectric transducer thatcoupled to the specimen through the1.3-cm dia tip of a steel horn. This single-pulsed technique required only a fewtens of milliseconds to acquire asequence of images of the entire timeevolution of the heating process. The20 kHz thermoelastic heating and coolingvariations associated with the sonicpulse were averaged out over theintegration time (~ 1 ms) of the IRcamera. Figure 1 is an example of howa small (0.7 mm long) tight crack in anAluminum plate sample shows up in thethermal image. The crack was initiatedfrom a saw cut by fatiguing the samplein a cyclic-loaded mechanical tensiletesting machine.

The single pulse technique is distinctfrom thermoelastic temperature variations

Thermosonics or sonic IR is an emergingthermal nondestructive testing NDTmethod that is based on detecting theheat generated by defects in a materialwhen a powerful ultrasonic excitation isapplied it. The method has been effectivein the detection of defects such ascracks and delaminations in metallic,ceramic, and polymer matrix composites.The major advantage of this techniqueis that large structures and structureswith complex curvature can be evaluatedrapidly and accurately – this advantagecomes about from three features, namely(1) every material, irrespective of theshape, can be mechanically excitedthrough acoustic/elastic vibrations and/or waves, (2) vibrations and/or wavescan interrogate large volumes, and (3)the heat generated can be imaged usinginfrared cameras that offer full fieldvisualization.

Although vibrothermal NDE techniqueswere pioneered in the early 1980s, dueto lower sensitivity of the IR cameras, ofthe order of 0.2°C, available at that time,successful practical applications werereported only from the late 1990s.Considerable progress has since beenmade on various aspects, particularly inmodelling and simulations. We alsocome across different names such asvibro-thermography, thermosonics, sonicIR, and thermoelastic stress analysisunder which investigations are reported.

In vibrothermography, the structure ismechanically excited with frequenciesranging from a few Hz to tens of kHz. Inthe low frequency range the structure isphysically attached to a mechanicalexcitation source, such as a dynamicshaker. In the higher frequencies (kHz)range the structure is in contact with apiezo-shaker or an ultrasonic horn. Inthermosonics/sonic IR, an ultrasonichorn, of the type used for plastic

welding, is usually used in contact modewith the excitation typically at tens ofkHz. In this pulse sonic thermography,the pulse is applied for a short time(normally less than 1 s) to producelocalized heating at the defect, whichcan be revealed by a high speed infraredcamera through monitoring its effect onthe surface temperature distribution. Inthis way very quick defect detection canbe achieved but a relatively highultrasonic excitation power, that couldpotentially damage the inspected part,is usually needed to heat the defect. Avariant is the ultrasound lock-inthermography, where the high frequencyultrasonic wave used to excite the partis amplitude modulated at a very lowfrequency (a few tens of millihertz) andthe recorded sequence of images isprocessed in the frequency domain,producing two images: an amplitudeimage and a phase image. This thermalimaging technique has the advantage of

Fig. 1 : (left panel) Optical micrograph of a fatigue crack in a 3-mm-thick aluminiumalloy bar. The bottom edge of the saw cut used to initiate the crack is seen(black) just above the top of the crack, which is approximately 0.7 mm long andclosed. (right panel) Selection of four frames from a sequence of sonic IR images(3 – 5 µm) of the crack. The top left image was taken prior to turning on thesonic excitation, the top right immediately following the excitation pulse, andthe bottom left and right images taken at two later times during the 50 ms sonicexcitation.[from Favro et al., Rev. Sci. Instrum., 71 (2000) 2418-2421]

Horizon

17


HORIZON

that are synchronous with loading cyclesas it averages over such time scalesleaving only irreversible processes tobe imaged. It is faster than lock-inthermography since the latter utilizesvery low frequency (a few tens ofmilliHertz) sinusoidal amplitudemodulation of the acoustic source,coupled with video lock-in IR imaging atthat very low frequency requiring severalminutes to produce processed lock-inimages.

Case 2: Graphite/epoxycomposite structures

The material used in this study consistedof 5208 epoxy reinforced with woven([0/90]4s) IM7 carbon fiber. Thethickness and fiber volume fraction ofeach panel are nominally 3.0 and 0.55mm, respectively. The specimens are 25mm wide and 216 mm long. The notchedsamples have a hole of 6.2 mm diameterat the geometric center.

Figure 2 shows the schematic of theexperimental set-up. A 20 kHz ultrasonicwelding gun (max. Power 1 kW) is usedto insonify the sample with a 200 mspulse of sound. The tip of the weldinghorn is pressed against the samplethrough an intermediate couplingmaterial (leather).

Figure 3 shows the sonic-IR images fora notched sample and an un-notchedsample at various load cycles. It is seenfrom the figures that damage was presentbefore the samples were subjected tofatigue loadings. However, additionaldamage occurs near the circular notchas the samples were subjected to cyclicloading. The damage mechanisms arepredominantly delamination anddistributed matrix cracks. It is observedthat the damage areas increased withthe fatigue load cycles as expected. Inthis case, the initial damage due tomachining the samples from a largecomposite plate appears to have initiatedfurther damage.

A finite element analysis of a notchedsample was carried out and was foundto validate the assumption that themating surfaces at the fatigue damagedo slide during the application of thesound pulse.

Case 3: Cast Iron Structures

Figure 4 shows two batches of severaldefective (size and location unknown)

reconstructed using four equidistantdata points for each modulation cycleand the resulting infrared imagessequence processed in the frequencydomain by means of a Fouriertransformation tuned to the frequencyof amplitude modulation. This processled to the computation of a magnitudeand a phase image that are used topresent the relevant information aboutsubsurface flaws.

The results shown in Figure 5, 6 and 7suggest that a high power excitation isnot necessary to heat the flaw. Because

and sound turbocharger housing parts,made from spheroidal graphite cast iron(EN-GJS-400). For validation purpose,each sample was later examined by 3Dcomputed tomography (CT).

By amplitude modulating the highfrequency ultrasonic excitation with alow frequency lock-in signal (typicallybetween 0.01 and 1 Hz), a periodicaltemperature pattern was generated at thecomponent’s surface, and recorded by ahigh-resolution infrared camera overseveral modulation periods. The surfacetemperature pattern was then

Fig. 2 : Schematic diagram of sonic-infrared imaging technique. [from Mian et al.,Composites Science and Technology 64 (2004) 657–666].

Fig. 3 : (Left panel) Sonic-IR images for an un-notched sample at (a) 0 cycle, (b) 3000cycles, and (c) 13, 000 cycles.(Right panel). Thermosonics images for a notchedsample (a) at 0 cycle, (b) 3000 cycles, (c) at 13, 000 cycles. [from Mian et al.,Composites Science and Technology 64 (2004) 657–666].

18


HORIZON

MECHANISMS, ISSUES ANDCHALLENGES

A widely held view of the process ofheat generation is the following:

Sound waves at low frequenciespropagate with appreciable amplitudeover distances much longer than awavelength. For typical complex-shaped

coupling between the test specimen andthe acoustic horn normally causes theexcitation of harmonics andsubharmonics of the driving frequency,resulting in a broadband excitation thatenhances the thermosonic signal. It isbelieved that rather than the ultrasoundintensity, it is this “hammering” effectthat actually produces resonance of thepart and heats the flawed region.

of unavoidable losses due to theimperfect coupling between sonotrodeand sample, not all the nominal power isactually transferred to the part. Typically,the effective acoustic power injected intothe part, especially when componentswith complex geometry are considered,is only a small fraction (between 0.5%and 10%) of the nominal one. Moreimportantly, the nonlinearity of the

Fig. 4 : (left panel) Cast iron turbocharger housing parts: (a) type I and (b) type II. (middle panel) Schematic illustration of ultrasoundactivated lock-in vibrothermography. (right panel) Experimental setup for concurrent measurement of the heating response andthe vibration response of cast iron components by means of infrared thermography and laser Doppler vibrometry. [fromMontanini, Freni, and Rossi, Rev. Sci. Instrum. 83, 094902 (2012)]

Fig. 5 : Vibration profiles measured by laser Doppler vibrometry on sound and defective samples.Thermo-acoustic spectrum and itsderivative obtained by exciting a defective part with an ultrasonic sweep in the 15–25 kHz range.

Fig 6 : Flow chart of the phase image processing algorithm used for defects sizing. [from Montanini, Freni, and Rossi Rev. Sci. Instrum.83, 094902 (2012)]

19


HORIZON

One of the challenges in all thesemethods is in producing a consistentcontact between the source (say, horn)and the sample to obtain a repeatableexcitation so as to detect the damage ofinterest consistently. Several materialssuch as Teflon, Aluminum, Copper, andeven materials like card stock and leatherhave been used to produce a repeatablecontact with some degree of success. Arelated issue is to determine the locationof clamping points on the sample, thepressure applied at the clamps, and thepressure applied at the horn tip as thesehave been reported to affect results.

Further, when the transducer is coupledto a sample through a nonlinear couplingmaterial, the vibration in the sampleshows more a much more complicatedform, as shown in Figure 8b measuredusing a laser Doppler vibrometer. Thevibrometer beam is reflected from thesurface of the sample in the vicinity ofthe defect, and the surface velocity ofthe sample in the direction parallel tothe laser beam is recorded in a computer.A recorded waveform of an uncoupledultrasonic gun, in this case one designedto produce nominal 40kHz vibration, isshown in Figure 8a.

The difference between these twosystems (one being the isolatedtransducer, the other being the combinedsystem of transducer/coupling material/specimen) can be better seen throughthe corresponding spectra of thewaveform. Figure 8c, the FourierTransform of the waveform in Figure 8a,shows a single 40 kHz response,indicating that horn is producing thispure frequency only. However, Figure8d, which is a Fourier Transform of thewaveform in Figure 8b, shows manyharmonics of the 40 kHz fundamental.Thus, the vibration at the site of thecrack may be quite different from whatone might have expected from theknown output of the horn. Thus thenonlinear, high amplitude excitationcould lead to crack opening or closuremaking the inspection not entirely non-destructive. To obtain a detailedunderstanding of the sonic heatingeffect, one must take these complicationsinto account.

We look at a few controlled studiesaimed at resolving some of these issues.

discharge-machining (EDM) notches andflat-bottom holes do not generatevibrothermographic indications becausethey have no rubbing surfaces.

Theories of frictional heating haveinspired simulations, but littlequantitative experimental data has beenavailable to validate those theories, inpart because of the large amount ofexperimental variability. For metals, thereis still substantial debate over therelative significance of friction versusother mechanisms such as plasticdeformation. Some numerical simulationssuggest that plastic deformation issignificant and others suggest it is not.It has been argued that frictional rubbingbetween crack faces is not responsiblefor heat generation or energydissipation, but that crack heating isentirely due to interactions in theelastoplastic region of a crack. Despitenumerous finite element simulations andtheoretical explanations, no definitiveexperimental validation of either theoryhas been presented and no single theoryhas been universally accepted to dateto explain the source—or sources—ofheat generation in vibrothermography.Given the nature of excitation, it ispossible that a combination of two ormore mechanisms may occur.

industrial parts reflections from variousboundaries of the specimen introducerepeated conversions among thevibrational modes, leading to a verycomplicated pattern of sound within thespecimen. This sound field completelyinsonifies the regions under inspectionduring the time that the excitation pulseis applied. If a subsurface interface ispresent, say a fatigue crack in a metal,or a delamination in a compositestructure, the opposing surfaces at theinterface will be caused to move by thevarious sound modes present there. Thecomplexity of the sound is such thatrelative motion of these surfaces willordinarily have components both in theplane of the crack and normal to it. Thus,the surfaces will ‘‘rub’’ and ‘‘slap’’against one another, with a concomitantlocal dissipation of mechanical energy.This energy dissipation causes atemperature rise, which propagates inthe material through thermal diffusion.This dissipation is monitored throughits effect on the surface temperaturedistribution. The resolution of theresulting images depends on the depthof the dissipative source as well as onthe time at which the imaging is carriedout.

It is worth noting that traditionalsynthetic defects such as electro-

Fig. 7 : Ultrasound activated vibrothermographic NDT phase angle measurements oncast iron turbo housing parts (batch II): (a) gray level phasegram of sample IIaat 0.2 Hz modulation (defective, front view); (b) gray level phasegram of sampleIIa at 0.1 Hz modulation (defective, front view); (c) gray level phasegram ofsample IIa at 0.05 Hz modulation (defective, front view); (d) gray level phasegramof sample IIb at 0.05 Hz modulation (sound, front view); (e) gray levelphasegram of sample IIa at 0.5 Hz modulation (sound, back view); and (f) x-rayradiography showing the buried multi-segmented porosity flaw. All tests wereperformed with fc = 22 019 Hz, P = 264 W (12%), number of preheating/heating cycles: 0/1, and PTFE coupling.[from Montanini, Freni, and Rossi Rev.Sci. Instrum. 83, 094902 (2012)]

20


edge tension (ESET) samples. Anannealing heat treatment was used torelieve closure stresses along the cracks.These cracked samples were then loadedinto the mounting apparatus and vibratedusing an ultrasonic welder that generatedhigh vibrational amplitudes at afrequency of 20 kHz. The images on thetop panel of the middle panel in Figure9 shows the heat generation of apropagating crack. Image processing wasused to isolate regions of heatgeneration in the image shown at thebottom. The lower image shows someheat generation along the crack and twodistinct regions of significant heatgeneration (indicated by white arrows)near the crack tip. There are two primaryregions of heat generation due tobranching in the crack as it grew, the leftheating region correlates to a branch inthe crack that did not continue to growwhile the right branch continued to grow,indicated by the heating region on thecenter-right side of the image, andindicated by the arrow on the right sideof the figure. Heat generation was foundto occur more than 0.5 mm past the cracktip where plastic deformations weretaking place, giving strong evidence ofplasticity-induced heat generation.Despite the likelihood of some heat nearthe crack tip due to friction during theportion of the vibration cycle where thecrack faces were in contact, the majority

EXAMPLE #1: HEAT GENERATIONDUE TO FRICTION

The left panel in Figure 9 shows IR crackheating and microscopy images for acrack in a titanium (Ti 6-4) sample. Thecrack was grown to a length of 8.0 mmin three-point bending with an R-ratio(min/ max stress) of 0.5 and maximumstress of 772MPa. Regions of heatgeneration (in vibro-thermography)correlate exactly to regions wherefracture surface damage was observedwhen the crack was broken open andobserved. If frictional rubbing occurs, itcan cause modifications (damage) torubbing asperities on surfaces such asfretting, plastic deformation, melting, etc.The left panel in Figure 9 shows acorrelation between the heat generatedfrom a crack and the damaged regionson the crack faces caused by intensefrictional rubbing.

EXAMPLE #2: HEAT GENERATIONDUE TO PLASTIC DEFORMATION

Planar cracks were grown in 2024aluminum eccentrically loaded single

Fig. 8 : (Top left) Waveform due to a 40 kHz horn not coupled to any sample. Notethat the amplitude is constant after it reaches a steady state. The source was setfor 800 milliseconds. (Top right) Waveform when the transducer is coupled toa target through a compliant coupling material. The vibration in the sampleshows a more complicated form. The source was turned on for 200 milliseconds.(Bottom left) Fourier Transform of the waveform shown in the top left panelindicates the single source frequency. (Bottom right) Fourier Transform of thewaveform shown in the top right panel shows the harmonic frequencies of thefundamental one. [from Han X., Favro L., and Thompson R.L.,, CP657, Reviewof Quantitative Nondestructive Evaluation Vol. 22, ed. by D. O. Thompson andD. E. Chimenti (2003) 500-504]

Fig. 9 : Testing heat generation mechanisms in metals and composites [from J. Renshawet al., NDT&E International 44 (2011) 736–739](Left panel) Correlation of frictional heat generation to crack face fretting inTitanium (Ti 6-4), (top,left) raw heating data after 1.0s of excitation;(top,center)processed heating at a isolating regions of heat generation; (top,right) frettingobserved on crack faces; and (bottom) a schematic of the crack and samplegeometry.(Middle panel-top) Raw infrared (IR) heating data compared to (bottom)processed IR data isolating regions of heat generation along the crack and at thecrack tips.(Right panel) The CFRP sample (bottom) containing an array of drilled holes,(left) schematic of a CFRP sample containing drilled holes with arrows showingthe direction of applied stress, and (right) observed infrared heating in the bardue to viscoelasticity and the stress concentration at the holes after 1.0s ofexcitation.

HORIZON

21


stresses to observed heating.

Figure 10 shows the schematic of theexperimental set up. A vibration sourcelocated off-center excites motion in thespecimen from the front (crack side). Thespecimen is mounted using rubber pins.A laser vibrometer records motiondirectly behind the crack, while aninfrared camera images crack heating.

As heating depends on both crack lengthand stress level, a useful way to evaluatedetectability was based on crackdetectability surface - crack heating as afunction of length(h) and stress (s). Sucha surface fit based on a simple powerlaw ΔT = c1η

c2σc3 (neglecting thecamera noise) is shown in Figure 11 (seea Video in the cited reference). Thesurface is colored red beyond 50mK, wellabove the detectability threshold of amodern IR camera and gray elsewhere.If the crack length and stress are withinthe red region then the crack shouldbe detectable. Surface fits givec2=2.8 ± 0.2, 870:2, c3=1:73 ± 0.06 fortitanium and c2 = 1.9 ± 0.2, c3 = 1.49 ± 0.05for Inconel. These experiments indicatethat crack heating increases both withcrack length and dynamic vibrationalstress level.

Case 4: Non-contact excitation

The basic principle of this newtechnique shown on the left in Figure12 is the excitation of a material withhigh amplitude acoustic waves withoutcontact between the specimen andexcitation source, and measuring thechange in the temperature caused bythe interaction of acoustic waves with

substantial promise for finding cracks inmetals, to date no quantitative relationhas been shown betweenvibrothermographic heating and keycharacteristics of the crack such aslength or morphology. Because thevibration is usually applied atfrequencies on the same order as naturalresonances, vibrothermographic crackdetection performance is a strongfunction of both the overall shape ofthe specimen and the location of thecrack within the pattern of resonances.

Quantitative studies to evaluate therelationship between vibrothermographiccrack heating, crack size, and vibrationalstress have been carried in a series oftests on 63 specimens each of Ti-6-4titanium and Inconel 718 at three differentsites with different equipments. In theseexperiments the specimens were tunedto resonate in their third order flexural(bending) mode at approximately 20 kHz.The tuned specimens naturally vibratenear the excitation frequency, minimizingthe nonlinear hammering effect and thepresence of other frequencies. Since theresonant mode shape is known it ispossible to calculate the motioneverywhere in the specimen from themotion at a single point, such as can bemeasured with a laser vibrometer. Thetuned specimens and resonant vibrationsmean that the experiments described hereare far more controlled than in the typicalvibro-thermography experiment. Motionis, for the most part, at a singlefrequency and a single resonant mode.From the mode shape, frequency, andmotion measured at a single point thevibrational stresses at the crack can becalculated so as to directly relate those

of the heat generated in this experimentappears to have been the result of plasticdeformations past the crack tip.

In both these examples, the vibrationalstresses used were small in comparisonto the fatigue limit of the material (i.e.less than 40% of the materials fatiguelimit) to avoid causing additional damageto the specimen, such as crack growthwhich has been observed at highvibrational stress levels . After testingwas completed on each sample, thecracks were observed and their lengthswere carefully measured to checkwhether crack growth had occurred ornot. After a surface examination, thecracks were broken open to observe therubbing crack faces using optical andscanning electron microscopy to checkfor friction-induced damage to therubbing surfaces.

EXAMPLE #3: HEAT GENERATIONDUE TO VISCOELASTICITYHeat generation has been observed atartificial delaminations in compositeswhen vibrated at certain resonantfrequencies. Simulations seem to indicatethat viscoelastic heating could be adominant mechanism in polymercomposites. However it is possible thatadditional viscoelastic heating may bepresent in the vicinity of a defect due tothe high stress concentrations aroundsuch defects when they are vibratedintensely, as in vibrothermographicstudies.

The images on the right panel of Figure9 show viscoelastic heating in a carbonfiber reinforced composite (CFRP)measured using an infrared camera. Holeswere drilled in the bar shown at thebottom of figure to generate stressconcentrations primarily above andbelow the holes when vibrated in thedirection indicated in the image. Appliedvibrational stresses were kept low toavoid plastic deformation. When thesample containing the drilled holes wasvibrated, the stress concentrationsaround the holes generated heat abovethe baseline heating of the excitedresonance in a pattern consistent withthe expected stress concentrationaround a stressed hole, as shown in theIR heating image in the figure.

EXAMPLE #4: ROLE OFVIBRATION STRESS ON CRACKSIZE

While vibrothermography has shown

Fig. 10 : (left) Experimental configuration (right) Specimen geometry[from S.D. Hollandet al., NDT&E International 44 (2011) 775–782]

HORIZON

22


maximum temperature on the oppositeside of the sample. Although cracks canbe detected and identified using NCATS,it would be necessary to scan the entirestructure point by point.

In summary, the primary sources of heatgeneration in a vibrating crack includefrictional rubbing, plastic deformations,and viscoelasticity in and around defectsdepending on the material, type ofdefect, and the applied vibrational stresslevel. Frictional rubbing occurs,especially in cracks, and is evidencedby alterations or damage to rubbing crackfaces. Plastic deformations in the plasticzone of a propagating crack may alsooccur at high vibrational stress levelsand will generate heat in addition tofriction, especially in regions beyond thecrack tip. Viscoelastic heating can alsooccur depending on the material and isrelated to the vibrated material’sproperties and level of appliedvibrational stress. Viscoelastic heatingis increased in regions of stressconcentration and does not require arubbing interface to generate heat.Improved understanding of the sourcesof heat generation in vibrothermographywill help to design robust testingprocedures to implementvibrothermography in industry as wellas improve issues with experimentalrepeatability. It appears that multiple heat-generating sources must be taken intoaccount for successful implementationof vibrothermography as anondestructive evaluation technique.

REFERENCES

1. L.D. Favro, Xiaoyan Han, ZhongOuyang, Gang Sun, Hua Sui, and R.

resistance appear to be reasons thatthere was no observable change intemperature at the horn tip during theexcitation at all input power levels anddurations tested. The right panel ofFigure 12 shows the time-temperatureprofiles indicating that heat generationappears to persist beyond the durationof the acoustic excitation.

An example of the application of NCATSon a defective aluminium wheelcomponent is shown in Figure 13. Theheat generation in the region shown inFigure 13 is considered to be acombination of internal friction inaluminum alloy (providing thebackground) and frictional heating dueto the rubbing of the crack faces withexcessive heat due to crack face rubbinghighlighting the presence of the crack.The combination of internal frictionmechanisms, its relaxation behavior, andthe thermal gradient in the samplethickness impact the maximumtemperature and time required to attain

the material. The middle panel of Figure12 shows the calibration of the horndisplacement with input power indicatingthe linear response unlike what is foundin the usual vibrothermography (seeFigure 8). When acoustic wave pulseencounters a boundary with a material aportion of the energy is reflected fromthe boundary and another portion istransmitted into the material. As thetransmitted portion propagates throughthe material a portion is absorbed by thematerial and a portion continues topropagate until it reaches anotherboundary at which point the processbegins again. This process continuesuntil the total pulse energy is reflected,absorbed, or transmitted by the material.

The length of the acoustic horn has beendesigned to be equal to a quarterwavelength, with the tip being an anti-node for maximum longitudinaldisplacement. The combination of thetip being an anti-node and the tipmoving freely in air with minimum

Fig. 11 : Graphical animation of the crack detectability surface for titanium: Crackheating as a function of crack length and dynamic vibrational stress. Red datapoints come from ISU, yellow from PW, and green from GE. Thesemitransparent surface fit is colored gray below the threshold of easydetectability (50 mK) and red above the threshold. Supplementary materialrelated to this article can be found online at doi:10.1016/j.ndteint.2011.07.006.[from S.D. Holland et al., NDT&E International 44 (2011) 775–782]

Fig. 12 : (left) Schematic of the non-contact acousto-thermal signature (NCATS) experimental setup. (middle) Displacement of thehorn with increasing input power. (right) NCATS time–temperature signatures at different input powers to the horn. Distancebetween horn and sample: 300 ìm. Duration of excitation: 250 ms. Sample: Ti-6Al-4V [from Sathish et al. Rev. Sci. Instrum.83, 095103 (2012)]

HORIZON

23


National NDT Awards No. Award Name Sponsored by 1. ISNT - EEC M/s. Electronic & Engineering Co., Mumbai

National NDT Award (R&D)

2. ISNT - Modsonic M/s. Modsonic Instruments Mfg. Co. (P) Ltd.,National NDT Award (Industry) Ahmedabad

3. ISNT - Sievert M/s. Sievert India Pvt. Ltd., Navi MumbaiNational NDT Award (NDT Systems)

4. ISNT - IXAR M/s. Industrial X-Ray & Allied RadiographersBest Paper Award in JNDE (R & D) Mumbai

5. ISNT - Eastwest M/s. Eastwest Engineering & Electronics Co.,Best Paper Award in JNDE (Industry) Mumbai

6. ISNT - Pulsecho M/s. Pulsecho Systems (Bombay) Pvt. Ltd.Best Chapter Award for Mumbaithe Best Chapter of ISNT

7. ISNT - Ferroflux M/s. Ferroflux ProductsNational NDT Award (International recognition) Pune

8. ISNT ISNTNational NDT Award forYoung NDT Scientist / Engineer

9. ISNT - Lifetime Achievement Award

Note-1: The above National awards by ISNT are as a part of its efforts to recognise and motivate excellence in NDT professionalenterpreneurs. Nomination form for the above awards can be obtained from ISNT head office at Chennai, or from the chapters.The filled application are to be sent to Chairman, Awards Committee, Indian Society for Non-destructive Testing, Module No. 60 &61, Readymade Garment Complex, SIDCO Ind. Estate, Guindy, Chennai-600 032. Telefax : 044-2250 0412 Email:[email protected]

L. Thomas, Infrared imaging ofdefects heated by a sonic pulse, Rev.Sci. Instrum. 71, (2000) 2418

2. Ahsan Mian, Xiaoyan Han, SarwarIslam, Golam Newaz, Fatigue damagedetection in graphite/epoxycomposites using sonic infraredimaging technique, CompositesScience and Technology 64 (2004)657–666

3. B. Hosten, C. Bacon, and C. Biateau,Finite element modeling of thetemperature rise due to thepropagation of ultrasonic waves inviscoelastic materials and experimentalvalidation, J. Acoust. Soc. Am. 124,(2008) 3491–3496

Fig. 13 : (left) Position of the IR camera, acoustic horn, and the wheel component toimage the crack. (right) NCATS imaging of a crack in aluminum wheelcomponent [from Sathish et al. Rev. Sci. Instrum. 83, 095103 (2012)]

4. Marco Morbidini and Peter Cawley,The detectability of cracks usingsonic IR, J. Appl. Phys. 105, (2009)093530

5. Xiaoyan Han, L.D. Favro andR.L.Thomas, Sonic IR Imaging ofdelaminations and disbonds incomposites, J. Phys. D: Appl. Phys.44 (2011) 034013

6. Stephen D. Holland, Thermographicsignal reconstruction forvibrothermography, Infrared Physics& Technology 54 (2011) 503–511

7. Jeremy Renshaw, John C.Chen,Stephen D. Holland, R. BruceThompson, The sources of heat

generation in vibrothermography,NDT&E International 44 (2011) 736–739

8. Stephen D.Holland, Christopher Uhl,Zhong Ouyang, TomBantel, Ming Li,William Q. Meeker, John Lively, LisaBrasche, David Eisenmann,Quantifying the vibrothermographiceffect, NDT&E International 44 (2011)775–782

9. R. Montanini, F. Freni, and G. L. Rossi,Quantitative evaluation of hiddendefects in cast iron components usingultrasound activated lock-invibrothermography, Rev. Sci. Instrum.83, (2012) 094902

10.Shamachary Sathish, John T. Welter,Kumar V. Jata, Norman Schehl, andThomas Boehnlein, Development ofnondestructive non-contact acousto-thermal evaluation technique fordamage detection in materials, Rev.Sci. Instrum. 83, (2012) 095103

HORIZON

NDE eventsWe hope that this new feature added to the journal during the last yearhas been useful for the readers in planning their activities in terms ofpaper submissions, registering for seminars, etc. Please send yourfeedback, comments and suggestions on this section [email protected]

December 2012

National Seminar on NDE 2012December 10- 12, 2012 ; New Delhi, Indiahttp://www.nde2012.org

The 2nd International Workshop and Congress oneMaintenanceDecember 12- 14, 2012 ; Lulea, Swedenhttp://www.emaintenance2012.org/index.html

January 2013

2013 API Inspection Summit & ExpoJanuary 7-10, 2013 ; Galveston Island Convention CenterGalveston, Texas, USAhttp://www.api.org/inspectionsummit

The Annual Reliability and Maintainability Symposium2013January 28-31, 2013 ; Rosen Shingle Creek Resort andGolf Club Orlando, FL USAhttp://www.ndthub.com/ndt-events/the-annual-reliability-and-maintainability-symposium-2013

February 2013

PNAA’s 2013 Aerospace ConferenceFebruary 12-14, 2013 ; Lynnwood, WA, USAhttp://www.pnaa.net/events/annual-conference/82-events/192-2013-aerospace-conference

First International Symposium on Optical CoherenceTomography for Non-Destructive TestingFebruary 13-14, 2013 ; Linz, Austriahttp://www.ndthub.com/ndt-events/first-international-symposium-on-optical-coherence-tomography-for-non-destructive-testing

8th International Conference on Advances in Metrology –AdMet-2013.February 21-23, 2013 ; National Physical Laboratory,New Delhi, India.http://www.admetindia.org

INFRARED THERMOGRAPHY SURVEYLowest Rates for Thermography Job with highest quality

Thermography charges per day is Rs.12,000 for Maximum 100 points and 8 Hours JobThermography will be done & analyzed by the qualified experts

Contact:Mr. K. RAVINDRAN

SOUTHERN INSPECTION SERVICES,#.2, IInd FLOOR, GOVINDARAJI NAICKER COMPLEX,

JANAKI NAGAR, ARCOT ROAD, VALASARAVAKKAM,

CHENNAI 600 087. TAMILNADU, INDIA

TEL: +91 44 - 24864332 / 24868785 / 42647537

Email ID: [email protected]

Website: sisndt.com, www.ndtsis.com, www.pdmsis.com

24


NDE patentsWe hope that the section on NDE Patents, which featured in the last few issues of thisjournal has continued to trigger your curiosity on this very important topic ofIntellectual property. We continue this section with a few more facts on patents and alisting of a few selected NDE patents. Please send your feedback, comments andsuggestions on this section to [email protected]

Compiled by Dr. M.T.Shyamsunder, GE Global Research, Bangalore, India

associated with vessels and/or

buildings, using nondestructive

evaluation (NDE). Characterizing the

dielectric properties of materials

associated with a target can be

executed by directing primary

microwave energy from a source

towards a target, receiving secondary

microwave energy signals returned

from the target, analyzing the

secondary microwave energy signals,

and characterizing dielectric properties

of materials associated with the target

based on analysis of the secondary

signals. A system for non-

destructively characterizing a target

material’s dielectric properties can

includes a microwave energy source,

a waveguide for directing microwave

energy towards a target, a receiver

for receiving microwave signals

reflected off targets, an analyzer for

assessing the difference between

incident and reflected microwave

signals to determine the presence of

corrosion within a target or a targeted

area, and an indicator for providing

results of analysis.

Inventors: Bray; Alan V.

(Spicewood, TX), Schmidt; Gary R.

(Austin, TX), Cuevas; Alfonso (Austin,

TX), Dube; Victor (McDade, TX)

UNITED STATES PATENT6,373,245April 16, 2002Method for inspecting electricresistance welds usingmagnetostrictive sensors

Abstract

A method and apparatus is shown for

implementing magnetostrictive sensor

techniques for the nondestructive

UNITED STATES PATENT7,521,926April 21, 2009

Method for testing a component in a

non-destructive manner and for

producing a gas turbine blade

Abstract

The invention relates to a method for

the nondestructive testing of a

component, in which corrosion regions

close to the surface, composed of

oxidized carbides or sulfided base

material close to the surface, are

determined by means of an eddy

current measurement. This allows the

blades or vanes to be ground down

and/or separated out in particular prior

to a complex process of cleaning and

coating the gas turbine blade or vane.

Inventors: Beck; Thomas (Panketal,

DE), Reiche; Ralph (Berlin, DE),

Wilkenhoner; Rolf (Kleinmachnow,

DE)

Assignee: Siemens

Aktiengesellschaft (Munich, DE)

UNITED STATES PATENT6,674,292January 6, 2004Microwave corrosion detectionsystems and methods

Abstract

Corrosion, mold and moisture can be

detected under outer layer of

structures, such as surfaces

UNITED STATES PATENT8,270,556September 18, 2012Apparatus for forming stresscorrosion cracks

Abstract

An apparatus for forming stress

corrosion cracks comprises a heating

unit which includes a conductive

member and a heating coil disposed

adjacent to the conductive member to

generate steam pressure in the tube

specimen, an end holding unit, and a

control unit for controlling the heating

unit and the end holding unit. The

stress corrosion cracks occurring in the

equipment of nuclear power plants or

apparatus industries during operation

can be directly formed in a tube

specimen using steam pressure under

conditions similar to those of the actual

environment of nuclear power plants,

thus increasing accuracy for analysis

of properties of stress corrosion cracks

which are in actuality generated,

thereby improving reliabil ity of

nuclear power plants or apparatus

industries and effectively assuring

nondestructive testing capability,

resulting in very useful industrial

applicability.

Inventors: Lee; Bo Young (Goyang,

KR), Kim; Jae Seong (Goyang, KR),

Hwang; Woong Ki (Seoul, KR)

Assignee: Industry-University

Cooperation Foundation Hankuk

Aviation University (Gyeonggi-do, KR)

Continuing our endeavor to provide you updates on NDE and Inspectionrelated patents, listed below are a few recent patents from a variety ofdifferent areas related to Nondestructive Evaluation and Inspection whichwere issued by USPTO in the last few years. If any of the patents are ofinterest to you, a complete copy of the patent including claims anddrawings may be accessed at http://ep.espacenet.com/

26


28


Assignee: Dacco Sci, Inc. (Columbia,

MD)

UNITED STATES PATENT6,294,912September 25, 2001Method and apparatus fornondestructive inspection ofplate type ferromagneticstructures usingmagnetostrictive techniques

Abstract

A method and apparatus is shown for

implementing magnetostrictive sensor

techniques for the nondestructive

evaluation of plate type structures

such as walls, vessels, enclosures, and

the l ike. The system includes

magnetostrictive sensors specifically

designed for application in conjunction

with plate type structures or pipes that

generate guided waves in the plates

or pipes which travel threrethrough in

a direction parallel to the surface of

the plate or pipe. Similarly structured

sensors are positioned to detect the

guided waves (both incident and

reflected) and generate signals

representative of the characteristics

of the guided waves detected that are

reflected from anomalies in the

structure such as corrosion pits and

cracks. The sensor structure is

longitudinal in nature and generates

a guided wave having a wavefront

parallel to the longitudinal axis of the

sensor, and which propagates in a

direction perpendicular to the

longitudinal axis of the sensor. The

generated guided waves propagate in

the plate within the path of the

propagating wave. The reflected

waves from these abnormalities are

detected using a magnetostrictive

sensor. Shear horizontal waves may

also be created by rotating the

magnetic bias 90.degree. and used for

similar inspection techniques. Pipes,

which act as curved plates, may also

be inspected as well as electric

resistance welds therein.

Inventors: Kwun; Hegeon (San

Antonio, TX)

adhesive tape corrosion sensor which

is utilized under actual field or

laboratory conditions in detecting

coating and substrate degradation

using Electrochemical Impedance

Spectroscopy (EIS) of coated or

uncoated metal structures has been

developed. The invention allows for

broad applicability, flexibility in

util izing the sensor in various

environments without structural

compromise and the ability to inspect

and evaluate corrosion of the actual

structure, regardless of the size,

shape, composition, or orientation of

the structure. The electrodes may be

removed once a measurement is

made or remain in the original fixed

position so that subsequent

measurements may be made with the

same electrode. The nondestructive

sensor apparatus is comprised of a

pressure-sensitive adhesive tape that

consists of a conductive film or foil and

conductive adhesive overlapping

another pressure-sensitive adhesive

tape that consists of a conductive film

or foil and non-conductive adhesive.

The conductive tape serves as the

sensing element or device. The non-

conductive tape serves as the lead

between the sensing element and the

point of measurement. In an

alternative configuration, the tape with

the conductive adhesive may be used

alone, acting as both sensor electrodes

and the lead to the point of

measurement. The metal structure or

other substrate being sensed or

evaluated for degradation serves as

the working electrode. This two

electrode sensing device is responsive

to water uptake, incubation, and

corrosion by measuring differences in

impedance spectra. The invention can

readily detect, quantify and monitor

coating and metal degradation from

its earliest stages, well before any

visual indication of corrosion appears,

under both laboratory and field

conditions.

Inventors: Davis; Guy D.

(Baltimore, MD), Dacres; Chester M.

(Columbia, MD), Krebs; Lorrie A.

(Baltimore, MD)

evaluation of plate type structures

such as walls, vessels, enclosures, and

the l ike. The system includes

magnetostrictive sensors specifically

designed for application in conjunction

with plate type structures or pipes that

generate guided waves in the plates

or pipes which travel therethrough in

a direction parallel to the surface of

the plate or pipe. Similarly structured

sensors are positioned to detect the

guided waves (both incident and

reflected) and generate signals

representative of the characteristics

of the guided waves detected that are

reflected from anomalies in the

structure such as corrosion pits and

cracks. The sensor structure is

longitudinal in nature and generates

a guided wave having a wavefront

parallel to the longitudinal axis of the

sensor, and which propagates in a

direction perpendicular to the

longitudinal axis of the sensor. The

generated guided waves propagate in

the plate within the path of the

propagating wave. The reflected

waves from these abnormalities are

detected using a magnetostrictive

sensor. Shear horizontal waves may

also be created by rotating the

magnetic bias 90.degree. and used for

similar inspection techniques. Pipes,

which act as curved plates, may also

be inspected as well as electric

resistance welds therein. In addition,

steel sheet butt welds may be

inspected with this technique.

Inventors: Kwun; Hegeon (San

Antonio, TX), Kim; Sang Young (San

Antonio, TX)

Assignee: Southwest Research

Institute (San Antonio, TX)

UNITED STATES PATENT6,328,878December 11, 2001Adhesive tape sensor fordetecting and evaluatingcoating and substratedegradation utilizingelectrochemical processes

Abstract

A portable and nondestructive

NDE PATENTS

29


Inventors: Polly; Daniel R. (Oxnard,

CA)

Assignee: The United States of

America as represented by the

Secretary of the Navy (Washington,

DC)

UNITED STATES PATENT4,826,650

May 2, 1989Ultrasonic examination ofreactor pressure vessel topguide

Abstract

In a boil ing water reactor, an

apparatus and process for ultrasound

inspection of the top guide is

disclosed. The top guide constitutes a

lattice of stainless steel bars overlying

the core plate and being assembled

at confronting grooves with the lattice

mounted at the side edges to the

reactor pressure vessel. This lattice

braces the upper ends of the vertically

supported fuel assemblies in their

requisite orientation and spaced apart

relation to enable among other things

the required spatial interval to be

maintained for control rod moderation

of the reaction. Because of the

proximity of the top guide to the fuel

assemblies, the individual bars making

up the lattice need to be checked for

cracking, especially that cracking

produced by irradiation assisted stress

crack corrosion. With a defined cell in

the lattice emptied of its contained and

adjoining fuel assemblies, there is

disclosed an ultrasound test for

cracking. A sound transducer on a first

special frame sweeps horizontally

across the top of a bar interrogating

the bar with vertical ultrasound waves

for detecting horizontal cracks.

Similarly, a sound transducer on a

second special frame sweeps

vertically across the side of a bar

interrogating the bar with angularly

incident horizontal ultrasound waves

for detecting vertical cracks.

Nondestructive testing of the lattice

assembly occurs without required

disassembly.

Abstract

A nondestructive method and

apparatus for optical detection and

monitoring corrosion in structures

normally inaccessible to light and

observation. An optical fiber coated

with a corrosion sensitive compound

is embedded in the structure. Tapped

Bragg gratings of different Bragg

periods are spaced along the fiber and

refract a narrow bandwidth

component of a broad beam light

pulse transmitted through the fiber.

Due to corrosion, the refracted

components are reflected by the

compound and their amplitudes are

detected and displayed for each

narrow bandwidth.

Inventors: Perez; Ignacio M.

(Yardley, PA), Agarwala; Vinod S.

(Warminster, PA)



Secretary of the Navy (Washington,

DC)

UNITED STATES PATENT4,927,503

May 22, 1990Method for assessment ofcorrosion activity in reinforcedconcrete

Abstract

The probe is a nondestructive testing

device for locating and measuring

corrosion activity in reinforced

concrete structures by direct detection

of electrochemical current flow. The

device consists of a surface probe

valved to present alternative

measurement paths when measuring

the probe potential with respect to a

remote reference electrode, allowing

the measurement of IR drops

associated with corrosion of

reinforcement “rebar”. By grid surveys

of concrete structures, areas suffering

internal corrosion (the primary cause

of marine concrete deterioration) can

be located and the level of corrosion

activity determined.

Assignee: Southwest Research

Institute (San Antonio, TX)

UNITED STATES PATENT5,854,492December 29, 1998Superconducting quantuminterference device fluxmeterand nondestructive inspectionapparatus

Abstract

A nondestructive inspection apparatushaving a SQUID is made with compactconfiguration and is capable ofdetecting a metallic or non-metallicmetal for defects, corrosion, and thelike, by forming the SQUID and amagnetic field applying coil on thesame substrate. The SQUID comprisestwo Josephson junctions, a washer coilconnected to the Josephson junctionsto form a superconducting loop, shuntresistors, a damping resistor, and afeedback modulation coil, all of whichare formed from a superconductingthin film on a supporting substrate. Amagnetic field applying coil is formedon the same supporting substrate witha superconducting thin film or anormal conducting metal thin film. Themagnetic field applying coil, whichgenerally has plural turns around theSQUID, applies a dc or ac magneticfield to a sample. The change inmagnetic field caused by a defect inthe sample is detected by the washercoil, and the position and size of thedefect may thus be determined. Sincethe magnetic field applying coil isintegrated on the same substrate asthat on which the SQUID is formed,the apparatus may be made compact.

Inventors: Chinone; Kazuo (Chiba,JP), Morooka; Toshimitsu (Chiba, JP),Nakayama; Satoshi (Chiba, JP),Odawara; Akikazu (Chiba, JP)

Assignee: Seiko Instruments Inc.(JP)

UNITED STATES PATENT5,646,400July 8, 1997Corrosion detecting andmonitoring method andapparatus

NDE PATENTS

31


Inventors: Richardson; David L.

(Los Gatos, CA), Clark; Jack P. (San

Jose, CA), Patterson; Peter M.

(Livermore, CA), Perry; Richard W.

(San Jose, CA)

Assignee: General Electric Company

(San Jose, CA)

UNITED STATES PATENT4,145,251March 20, 1979Corrosion measuring apparatusfor radioactive components

Abstract

Remotely manipulatable probe and

apparatus for positioning a corrosion

thickness sensing transducer over

selected areas of the surface of a

radioactive component submerged in

a pool of water for radiation shielding.

BACKGROUND In known types of

nuclear power reactors, for example

as used in the Dresden Nuclear Power

Station near Chicago, Ill., the reactor

core comprises a plurality of spaced

fuel assemblies arranged in an array

capable of self-sustained nuclear

fission reaction. The core is contained

in a pressure vessel wherein it is

submerged in a working fluid, such as

light water, which serves both as

coolant and as a neutron moderator.

Each fuel assembly comprises a

removable tubular flow channel,

typically of approximately square

cross section, surrounding an array of

elongated, cladded fuel elements or

rods containing suitable fuel material,

such as uranium or plutonium oxide,

supported between upper and lower

tie plates. The fuel assemblies are

supported in spaced array in the

pressure vessel between an upper

core grid and a lower core support

plate. The lower tie plate of each fuel

assembly is formed with a nose piece

which fits in a socket in the core

support plate for communication with

a pressurized coolant supply chamber.

The nose piece is formed with

openings through which the

pressurized coolant flows upward

through the fuel assembly flow

channels to remove heat from the fuel

elements. A typical fuel assembly of

this type is shown, for example, by B.

A. Smith et al. in U.S. Pat. No.

3,689,358. An example of a fuel

element or rod is shown in U.S. Pat.

No. 3,378,458. Additional information

on nuclear power reactors may be

found, for example, in “Nuclear Power

Engineering”, M. M. El-Wakil, McGraw-

Hill Book Company, Inc., 1962. While

the various reactor components are

thoroughly factory tested before being

placed in the reactor, there is a

continuing need for in-service

inspection equipment which can

rapidly and conveniently verify the

integrity of or detect any anomalies

in such components at the reactor site,

particularly after such components

have been subjected to reactor

service and have, therefore, become

radioactive. Such radioactive

condition of used components

requires remotely operable equipment

which can examine such components

under water to protect the test

equipment operators from radiation.

A particular need is inspection

equipment which can provide a

nondestructive examination and

quantitative indication of corrosion

formation, such as oxide formation,

on such reactor components. It is

particularly desirable to provide

corrosion measurement of removable

reactor components which potentially

have a relatively long service life such

as fuel assembly flow channels. For

example, as mentioned above, each

fuel assembly is surrounded by a

removable tubular flow channel. While

the normal service life of a fuel

assembly in the reactor core is in the

order of four years, the flow channel

can be removed and reused on a

replacement fuel assembly in the

absence of excessive corrosion or

other defects. Previous methods of

determining the extent of channel

corrosion involved the cutting up of a

channel and the shipping of samples

of corroded portions to a laboratory

for examination. This approach

resulted in destruction of potentially

reusable channels and an undesirable

expenditure of time and money. Thus

there is a need for remotely operable,

nondestructive corrosion measuring

equipment for determining whether or

not a radiated component is fit for

further service. Fuel assembly

channels are normally formed of a

zirconium alloy made up of two U-

shaped members welded together.

They are usually factory processed by

autoclaving (exposure to high

temperature steam) to form a thin,

tight protective oxide surface film of

deep gray or black color. In service

oxide corrosion usually occurs at local

areas, expecially at portions which

have been exposed to highest

temperatures and neutron flux

density, and develops as clusters of

pin point spots or nodules of corrosion

which are light grey or white in color

and which thus give the local area a

“salt and pepper” appearance. As

such corrosion progresses, the

nodules expand in area and

eventually coalesce to form a

continuous oxide corrosion film or

sheet over the local area. Continued

corrosion results in a thickening of the

oxide film and eventual spalling, that

is, a flaking off of the oxide particles.

Under present procedures, the

channel is removed from service

before spalling is expected to occur

to avoid contamination of the coolant

with the oxide particles. Measurement

of thickness of the corrosion film can

be used to preduct the onset of

spalling. Measurement of corrosion

thickness can also be used to indicate

the effectiveness of heat treatment

and other processes used to provide

improved corrosion resistance. It is

also desirable to examine other local

areas of the channel such as weld

seams, for indications of corrosion.

Therefore it is an object of the

invention to remotely and

nondestructively measure formation

of corrosion on a radioactive

component. It is another object of the

invention to provide a corrosion

thickness sensing means which readily

and remotely can be positioned over

NDE PATENTS

32


33


UNITED STATES PATENTH2,127October 4, 2005Corrosion detection bydifferential thermography

Abstract

A pulsed heat energized system for

detecting corrosion or similar

oxidation products located

intermediate a layer of metal and an

overlying layer of paint or other metal

protective material. The system

employs a continuing stream of

radiant thermal energy pulses

impinging on the external surface of

the coated metal and responds to the

phase angle difference between a

waveform representing the energy

pulses and a waveform representing

the undulating temperature response

of the paint surface to the energy

pulses. Use of the system for detecting

corrosion of a military or other aircraft

in order to avoid stripping the aircraft

for corrosion inspection and correction

is contemplated. Enhanced

independence of the corrosion

detection from measurement

variations and earlier detection of

corrosion presence are achieved with

respect to other corrosion

arrangements.

Inventors: Byrd; Larry W. (Huber

Heights, OH)



Secretary of the Air Force

(Washington, DC)

body portion transmits light and

refracts light at its conical surface, the

operator can, in effect, see through

and beneath the probe to position the

transducer over the desired local area

of the component being examined. In

the illustrated embodiment, the body

is surrounded by a ring of metal of

sufficient weight to provide a desired

force of the resiliently mounted

transducer against the surface under

examination.

Inventors: Qurnell; Frank D. (San

Jose, CA)

Assignee: General Electric Company

(San Jose, CA)

UNITED STATES PATENT7,822,273October 26, 2010Method and apparatus forautomatic corrosion detectionvia video capture

Abstract

In order to prevent human injury

during the inspection of tanks carrying

caustic material this device has been

created which will allow a tank to be

inspected remotely. A camera is used

to take an image of the tank surface

and software is incorporated into the

method that will analyze the surface

of the tank for certain corrosion

characteristics. This data is complied

in an easily readable for the operator.

Inventors: Arcaini; Gianni

(Jacksonvil le, FL), Arpa; Aydin

(Jacksonville, FL), Ruan; Yanhua

(Jacksonville, FL), Kuchi; Prem

(Jacksonville, FL)

a selected area of a radioactive

component. Equipment is

commercially available which uses an

eddy-current technique for indicating

the distance between a transducer and

a conductive surface. The transducer

includes a coil which is energized by

a high frequency current. Magnetic

flux from the coil produces eddy

currents in the conductive surface.

Thus the power or energy supplied by

the coil to produce the eddy currents

is also proportional to the distance

between the transducer and the

conductive surface. This displacement

dependent variation in power is

detected by suitable electronic

circuitry and converted to a calibrated

display or recording of the distance

between the transducer and the

conductive surface. Thus such a

device can be used to measure the

thickness of a nonconductive coating

on a metal. It is another object of the

invention to utilize an eddy-current

technique to remotely measure

thickness of corrosion on the surface

of radioactive components. SUMMARY

These and other objects of the

invention are achieved by a

transducer containing probe,

suspended at the end of a manually

manipulatable pole, which can be

visually positioned over selected areas

of a radioactive component

submerged to a suitable depth in

shielding water. The probe comprises

a body portion formed of a transparent

material and having the general shape

of a frustum of a cone, the transducer

being resiliently supported in a central

bore of this body portion. Since the

Dr. Rameswar Das doyen of NDT Society of India / ISNT passed away on July 14th, 2012.Dr. Das was born in Dhaka and did his Post Graduation in Chemistry before completing his Doctorate in Bio-Chemistry from Indian Institute of Science, Bangalore.He worked for Oriental Chemical Works at Kolkata and he was instrumental in development of various sectors likeRailways, Defense, Aviation and Nuclear etc.Dr. Das is a founder Member of many Professional Societies including ISNT and contributed immensely for thegrowth of ISNT. He has also received many Awards in his career.His passing away is a great loss to ISNT. Our heartfelt condolences to the bereaved members of his family for theirreparable loss.May his soul rest in peace.

ISNT Colleagues

OBI

TUA

RY

NDE PATENTS

We hope you enjoyed solving the“Wordsearch Puzzle” which waspublished in the last issue. We arestill receiving entries from thereaders and will be announcing theWINNERS and also publish thesolution in the next issue.

In this issue, we have anotherWordsearch puzzle to continuestimulating your brain cells! Wehope you will find this sectioninteresting, educative and funfilled. Please send your feedback,comments and suggestions on thissection [email protected]

Introduction

The “Word Search Puzzle”, contains morethan thirty (30) words related to MagneticParticle Testing. These includetechniques, terminologies, phenomenon,famous people, etc. These words arehidden in the puzzle and may be presenthorizontally, vertically, diagonally in aforward or reverse manner but alwaysin a straight line.

Instructions

- All you have to do is identify thesewords and mark them on the puzzlewith a black pen

- Preferably you may take aphotocopy of the Puzzle sheet andmark your answers on that (see themarked example)

- Once completed please scan youranswered puzzle sheet as a PDFfile and email the scanned sheet [email protected] with yourname, organization, contactnumber and email address

Rules & Regulations

- Only one submission per person isallowed

- The marked answers should belegible and clear without anyscratching or overwriting

- The decision of the Editor-in-Chief,Journal of NDT &E is final andbinding in all matters

The correct answers and the names ofthe prize winners will be published inthe next issue

NAME : ___________________________

ORGANIZATION : __________________________

PHONE : ___________________________

EMAIL ID : ___________________________

ndt puzzlendt puzzleConceptualized & Created by

Dr. M.T. Shyamsunder,GE Global Research, Bangalore

WORDSEARCH PUZZLEMAGNETIC PARTICLE TESTING

34


NCB – ISNT

ANNOUNCES

ISNT LEVEL – III CERTIFICATIONPROGRAMME

AT PUNE(7TH JANUARY to 17TH FEBRUARY,

2013)

The Announcement for the ISNT Level-III Certification programme has been

displayed in the web “www.isnt.org.in”with complete details along with

application form. Interestedparticipants may avail this opportunity

for better prospects.

Last date for receipt of application formalong with payment isDecember 21, 2012

37Technical Paper

Vol. 11, Issue 2 September 2012 Journal of Non destructive Testing & Evaluation

Quality Assessment of Composite Adhesively Bondedjoints by Non-linear Ultrasonic Method

R L Vijayakumar1, M R Bhat and CRL MurthyDepartment of Aerospace Engineering, Indian Institute of Science, Bangalore-560012, INDIA

Email: [email protected]

ABSTRACTQuality assessment of adhesive joints is an involved and challenging non destructive evaluation problem, both during their production andlife cycle. The load carrying capacity of an adhesive joint is governed by the properties of the adhesive layer itself as well as the interfacecharacteristics. Therefore an important task of NDE is the development of techniques to characterize the bond quality or even to measurethe bond strength of the joints. Since an adhesive layer can be treated as a soft interface between two components, it is expected thatnonlinear effects arise from the propagation of an ultrasonic pulse. Binding forces are nonlinear and cause a nonlinear modulation ofultrasonic waves transmitted through the bonded region. As a consequence, the generated higher harmonics of an insonified monochromaticwave will yield information about the adhesive bonds. In case of resonance the amplitude of strain in a soft interface layer is stronglyincreased and therefore, the layer considerably contributes to the amplitude of the second harmonic. The non linear behavior of such alayer and its influence on nonlinear parameter (β) was studied experimentally. Nonlinear ultrasound inspection was carried out on adhesivejoints with Carbon Fiber Reinforced Plastic (CFRP) adherends, the quality of the epoxy adhesive layer was varied by adding differentamounts of poly vinyl alcohol. It was found that the non linear parameter (β) increased with increasing amount of degradation, giving anindication of degradation.

Keywords: composites, adhesive joints, nonlinear ultrasound, bond quality.

1. INTRODUCTION

Adhesive bonding plays an important role in the assemblyof aerospace and automobile structures and has augmentedor replaced the conventional joining methods like weldsand rivets. Though there are many advantages of adhesivebonding like reduced stress concentration, uniform stressdistribution, improved fatigue strength etc, they do havesome disadvantages as well such as susceptibility todegradation when exposed to moisture and temperature.The quality of the joint may degrade over a period of timeleading to catastrophic failure of the structure. Thisemphasizes the role of non destructive inspection methodsin keeping the structures in good health.

Ultrasonic wave based methods have been widely used inthe field of non-destructive testing of adhesive joints [1,2]. However, most of these conventional ultrasonic methodsare aimed at detection , location and evaluation of grossdefects like delaminations and voids by way of studyingthe reflection and transmission characteristics of the wavesat the boundaries of the discontinuities. They are but lesssensitive to evenly distributed micro-cracks, pores ordegradation. Moreover, general degradation in strength maybe found in apparently flawless joints. Therefore a NDEmethod that can reflect material degradation in anadhesively bonded joint at any stage of service becomesmore important. Ultrasonic wave based techniques with anon-conventional approach can be a powerful tool for nondestructive inspection where the characteristics of itspropagation are directly related to the properties of material.It is known that material failure is usually preceded bysome kind of nonlinear mechanical behavior before

significant material damage occurs [3, 4]. If ultrasonicwaves are made to propagate through the material at thisstage a strong non linear effect may be generated due tothe non linear properties of that material. Thus, it could beexpected that the degree of material degradation can beevaluated by measuring the ultrasonic wave parametersthat are affected by this non-linearity.

The classical theory of nonlinear wave propagation in solidshas been discussed and presented by Truell and Hikata[5,6]. Most of the research on nonlinear phenomenon hasfollowed along these lines. The nonlinearity of ultrasonicwave means that the second or higher order frequencycomponent exists besides the fundamental component whenthe wave propagates through a degraded medium. Themagnitude of these higher order components is related toproperties of material as well as the wave amplitude andpropagation distance. Therefore, the magnitude of thehigher order component will appear differently in normaland degraded material, when amplitude of the wave andpropagation distance is maintained the same. This tendencyis known from previous research carried out on metals forfatigue damage evaluation [7, 8].

In the last few years the NDE community has turned itsattention to investigating the possibility of using non linearacoustic techniques to measure properties of adhesive joints[9, 10]. A detailed theoretical description on non linearpropagation of waves in layered media has been given byHirsekorn and Brekhovskikh [11, 12]. The basic parameterthat has to be evaluated in this context is the nonlinearparameter (β). There are two approaches to obtain thisparameter. First is from studying the second harmonic

38 Technical Paper

Journal of Non destructive Testing & Evaluation Vol. 11, Issue 2 September 2012

generation and the second from acousto-elastic effect [9].The latter requires application of stress and measurementof very small change in sound velocity in a thin adhesivelayer and therefore does not seem to be very practical.Thus, the method of second harmonic generation ispreferred by most of the researchers.

The objective of this study is to utilize the second harmonicmethod in characterizing CFRP-Epoxy adhesive joints withvaried bond quality and to show the correlation betweennonlinearity and adhesive degradation with the help ofexperimental results. The mechanism of second harmonicgeneration during propagation of ultrasonic wave throughthe degraded joint is firstly shown on the basis of nonlinearelasticity. RITEC high power nonlinear ultrasonic systemwas used to generate and measure the second harmonicfrequency component. Series of experiments were carriedout in order to obtain a correlation between the amplitudeof the second harmonic frequency and the adhesivedegradation. Sets of specimens with different degrees ofdegradation were prepared and tested. Experimental resultsshowed that nonlinear acoustic effect can be used as aneffective tool for the evaluation of degradation of adhesivejoints.

2. MATERIALS AND SAMPLES

Single lap shear joints were prepared as per ASTM D5868 standard using carbon fiber reinforced plastic (CFRP)material as substrates and a two part epoxy adhesive;Araldite AV138M / Hardener HV 998. The CFRP adherendwas fabricated using 14 layers of unidirectional carbonprepreg CP150ns with each layer having a thickness of0.18mm. All the layers were stacked in the 0° directionand cured in an autoclave according to the curing cycle(60°C for the first 30 minutes, 125°C for 90 minutes and7bar external pressure with vacuum) suggested by theprepreg manufacturer. The composite laminate thusobtained had a thickness of 2.5mm. It was subjected tonon-contact, water immersion ultrasound scanning using a5 MHz focused transducer to ensure that the substratematerial was free from gross defects and anomalies. Thelaminate was then cut to the required size of 101.6mm x25mm, surface preparation in the region to be bonded wascarried out according to ASTM D 2093 standard. An areaof 25mm X 25mm was bonded using the two part epoxysystem in which the resin and hardener were mixed in theratio of 100:40 as per the manufacturer’s recommendationand cured for 24 hours at room temperature. A uniform

bond-line thickness of 0.76mm was maintained using amold specially designed and fabricated for the purpose.The dimensions of the resulting composite lap shear jointsare shown in Fig. 1.

Five different sets of samples having six samples in each(total of 30 samples) were prepared; the quality of theepoxy resin was degraded adding different amount of polyvinyl alcohol (PVA). While the healthy samples (H) werefree of PVA, others were denoted as P10 (with 10% PVAby total weight of the epoxy-hardener mix), P20 (20% PVA),P30 (30% PVA) and P40 (40% PVA).

3. NONLINEAR ACOUSTIC EFFECT INMATERIALS

The nonlinear behavior of materials can be explained usingthe nonlinear version of Hooke’s law as shown in Eq. (1)[7, 8]

σ = Eε(1+βε+ . . . ) (1)

Where ‘E’ is Young’s Modulus and ‘β’is a higher orderelastic coefficient commonly known as nonlinear parameter.This relationship has been approved experimentally formetallic materials by some researchers [6]. The contributionof material degradation towards higher order harmonics isvery small and hence all higher order terms except thesecond, can be neglected. In order to explain the generationof higher order harmonic waves, consider the case wherea single frequency ultrasonic longitudinal wave is incidenton one side of a bar with degradation and received on theother side. If ‘A1’ is the amplitude of the initial soundpressure and ‘ω’ is the angular frequency and ‘k’ is thewave number; the incident longitudinal wave can beexpressed mathematically as [13]

uo = A1cos(kx–ωt) (2)

where ‘uo’ is the initial displacement of the excited wave.If the effect of attenuation is neglected, then the equationof motion for longitudinal planar wave in a material canbe represented as

(3)

Where ‘ρ’ is the density of the medium, ‘u’ is thedisplacement, ‘x’ is the propagation distance of the soundwaves in the medium, ‘σ’ is the stress and ‘t’ is the time.Using Eq. (1) and (3) and the relationship between strainsand displacements, one can obtain the nonlinear waveequation for displacement ‘u’

(4)

In order to obtain a solution, let us assume the displacement‘u’ to beFig. 1 : Single Lap shear joint

39Technical Paper


u = uo+u1 (5)

Where ‘uo’ is the initial excited wave and ‘u1’ is the firstorder solution. If ‘uo’ is the single frequency sinusoidalwaveform then the solution to the second order will be[14]

u = uo+u′

= A1cos(kx–ωt) – β—8

A12k2xs in 2(kx–ωt) (6)

The second term in Eq. (6) represents the second harmonicfrequency component, as a result we can explain howsecond order harmonic component occur through thepropagation process [15, 16]. The magnitude of the secondorder harmonic component (A2) depends on ‘β’ whichrepresents the nonlinear elastic characteristics of thematerial and is closely related to the degradation of thematerial. Therefore, a measure of the magnitude of ‘β’ canbe utilized to evaluate the change of the material’sproperties or degradation.

4. EXPERIMENTAL DETAILS

A schematic diagram of an experimental setup to measurethe magnitude of second harmonic component in thereceived ultrasonic waves is shown in Fig. 2. Experimentswere carried out using an ultrasonic pulser-reciever RITECRPR-4000 instrument which adopts the heterodyne signaldetection technique in order to reduce additive electricalnoise effectively. In through transmission technique whilea 1MHz contact PZT transducer was used as a pulser, a2MHz contact transducer was used as a receiver to pickup the second harmonic. Both the transducers were coupledto the CFRP adherend surface using glycerin as couplantand a spring clip was used to hold both the transducers inposition at a constant contact pressure.

An input signal with varying amplitudes was given to thepulser. In order to avoid overlap of the echoes within theadherends, the wave was restricted to 6 cycles. On theother hand, the burst was long enough to ensure interferenceeffects in the adhesive layer. The resulting signal from the2MHz receiver was amplified using a preamplifier (60dB),

while an internal band pass filter in the range 800 kHz to2.5MHz was used to ensure that the noise is minimized.Figure 3 shows the received signal which was thentransmitted to the digital oscilloscope where the signalwas digitized for further signal processing. The amplitudesof the first (A1) and second harmonic components (A2)were recorded at different excitation voltages ranging from100 to 400 volts, by means of Fast Fourier transformation(FFT).

Though in such a set up the nonlinear effects may havecontributions from many components such as delay line ofthe transmitting and receiving transducers, couplant film,adherend, and adhesive layer itself, in the present work,considering all the other parameters unaltered, and onlythe material properties of the adhesive being degraded usingPVA, any change in the nonlinear parameter can beattributed to adhesive layer degradation. An analyticaldescription in first order can be achieved by using a volumemodel approach that takes the interference in the adhesivelayer partially into account [9], since A1 >> A2, onlyinterference of the fundamental wave is considered thisimplies

(7)

Where, ‘λ’ denotes the wavelength and ‘x’ a distance equalto thickness of the adhesive layer. It is clear from theabove equation that ‘βá[A2/(A1)2]’, hence a highermagnitude of A1 contributes significantly towards themagnitude of second harmonic. However, a high magnitudefundamental harmonic can be obtained in the condition ofresonance which occurs if Eq. (8) is satisfied for n = 0,1,2…

(8)

It needs to be noted that in the case of resonance theincrease of A1 amounts to a factor of about 3.5. As A2 issquarely proportional to A1, its value can be high. Therefore,it appears to be promising to observe the nonlinearproperties of an adhesive layer by studying the secondFig. 2 : Experimental arrangement for nonlinear ultrasonic

inspection of adhesive joints

Fig. 3 : Received time domain signals for a healthy sample anddegraded P40 sample

40 Technical Paper


harmonic. The appearance of ‘λ’ in Eq. (7) implies thatthe second harmonic must also depend on the soundvelocity. However, since the influence of wavelength isseen even in numerator i.e., on A1, a simple prediction ofits influence on second harmonic is not possible. Figure 4shows the variation of the magnitude of the secondharmonic (A2) with fundamental harmonic (A1)2, atdifferent input voltage values. As shown, the amplitude ofA2 increases with an increase in the magnitude of input,since at higher voltages A1 becomes high and contributessignificantly towards A2. It can also be seen that the slopeof the variation also increases with increased PVApercentage, indicating the influence of degradation on thepropagated ultrasonic wave and higher harmonics.

The slopes of these curves give a measure of nonlinearityparameter (β) which was computed and plotted againstPVA percentage as shown in Fig. 5, it can be observed thatthe nonlinearity parameter (β) increases with increase inPVA percentage indicating degradation. It should be notedthat the nonlinear parameter for severely degraded (P40)sample increases by a factor of 7 as compared to a healthysample.

The samples were loaded in a testing machine till failureto determine its strength. Figure 6 shows the variation of‘β’ with the average shear strength of the CFRP adhesivejoint samples, as shown the nonlinear parameter decreaseswith increasing strength. It is reported that [9] the influenceof the nonlinearity at the interface of the joint due to strongintermolecular forces of attraction is negligibly smalltowards changes in β, hence any significant changes in βcan be attributed to loss of stiffness in the bond line dueto adhesive degradation.

The results obtained clearly shows the influence ofdegradation on nonlinear modulation of the input ultrasonicwaves. The nonlinear parameter (β) increases with increasein degradation. Higher nonlinear parameter implies lowerstrength. It should however be remembered that there area number of factors which influences the nonlinearparameter (β), including joint properties like bond-line

thickness, density of adhesive, type of adhesive used andits properties, attenuation in the medium adherends etc,‘β’ is also influenced by type of transducer used, electronicnoise, couplant, pressure applied and so forth.

The results obtained in these studies though are for CFRPadhesive joints prepared, it can be expected to be applicableto other types of polymer composites as well. Despite aninternal band pass filter used in the RITEC RPR-4000pulser- receiver, the received signal appeared to be quitenoisy. However, since all the other parameters remainedfairly constant except for the adhesive property, changesin ‘β’ could be measured and attributed to the degradationof the adhesive joint.

5. CONCLUSION

Experimental investigations were carried out usingnonlinear ultrasonics approach to study degradation in theadhesive layer of a CFRP-epoxy-CFRP adhesive lap joint.The results obtained show very interesting and encouragingcorrelation between nonlinear parameter and thedegradation in the adhesive joint. The acoustic pulse getsnonlinearly modulated due to degradation in the adhesive,leading to an increase in the nonlinear parameter (β) withincreased degradation. These results can give a quantitativeassessment of the health of a bonded joint. Though theresults obtained are for the CFRP-epoxy-CFRP adhesivelap joint specimen, the approach should be applicable toother types of bonded joints. However, since non linearparameter ‘β’ in a realistic situation may also depend upon

Fig. 4 : Variation of the second harmonic amplitude (A2) withfundamental harmonic (A1)

2 at different levels of inputvoltage.

Fig. 5 : Variation of the nonlinearity parameter (β) with PVApercentage

Fig. 6 : Variation of the nonlinear parameter (β) with Average bondstrength of the CFRP joints

41Technical Paper


a lot of other parameters like adhesive properties, geometry,transducers used, noise etc., these factors can influence ‘β’in a real structure and the contribution of each of theseparameters on nonlinear modulation of a propagatingultrasonic wave needs to be understood.

REFERENCES

1. R.D. Adams, B.W. Drinkwater, Nondestructive testing ofadhesively-bonded joints, NDT&E international 30, (1997), 93-98.

2. Shuo yang, Lan Gu, R. F Gibson. Nondestructive detection ofweak joints in adhesively bonded composite structures,Composite structures, 51, (2001) 63-71.

3. A. Sutin, Nonlinear acoustic nondestructive testing of cracks,14th International Symposium on Novel Aromatic Compounds(ISNA), (1996), 328-334.

4. J.K. Na, J.H. Cantrell, W.T. Yost, Linear and nonlinear ultrasonic properties of fatigued 410Cb stainless steel, in: D.O.Thompson, D.E. Chimenti (Eds), Review of progress in QNDE,15, 1996, pp.1347-1356.

5. R. Truell, C. Elbaum, B.B. Chick, Ultrasonic Methods in SolidState Physics, Academic Press, New York, (1969), pp. 38-63

6. A. Hikata, B.B. Chick, C. Elbaum, Dislocation contribution tothe second harmonic generation of ultrasonic waves, Journal ofApplied physics, 36 (1) (1965) 229-338.

7. K.Y.Jhang, K.C.Kim, Evaluation of material degradation usingnonlinear acoustic effect, Ultrasonics, 37 (1999) 39–44.

8. R.K.Oruganti, R.Sivaramanivas, T.N.Karthik. Quantification offatigue damage accumulation using non-linear ultrasoundmeasurements. International Journal of Fatigue, 29 (2007) 2032-2039.

9. M.Rothenfusser, M.Mayr, J.Baumann. Acoustic nonlinearities inadhesive joints Ultrasonics 38 (2000) 322-326.

10. P.P.Delsanto, S.Hirsekorn, V.Agostini, R.Loparco, A.Koka.Modeling the propagation of ultrasonic waves in the interfaceregion between two bonded elements, Ultrasonics, 40 (2002)605–610.

11. S. Hirsekorn. Nonlinear transfer of ultrasound by adhesive joints- a theoretical description, Ultrasonics 39 (2001) 57-68.

12] L.M. Brekhovskikh, in: Harcourt, Brace, Jovanovich (Eds.),Wavesin Layered Media, 2nd edition, Academic Press, 1980.

13. J.A. TenCate, K.E. Van Den Abble, Laboratory study of linearand nonlinear elastic pulse propagation in sandstone, Journal ofAcoustical Society of America, 100 (1996) 1383-1391.

14. I.E. Shkolnik, T.M. Cameron, Nonlinear acoustic methods forstrength testing of materials, International Symposium on NovelAromatic Compounds (ISNA), (1996) 316-327.

15. A. Hikata, B.B. Chick, C. Elbaum, Effect of dislocations onfinite amplitude ultrasonic waves in aluminum, Applied Physicsl3 (11) (1963) 195-203.

16. W.T. Yost, J.H. Cantrell Jr., M.A. Breazeale, Ultrasonicnonlinearity parameters and third-order elastic constants of copperbetween 300 and 3 K, Journal of Applied Physics 52 (1) (1981)126-132.

42 Technical Paper


1. INTRODUCTION

The World Federation of Nondestructive Evaluation Centers(WFNDEC) was founded in July 1998 with an objectiveto improve NDE technology and its uniform applicationon a worldwide scale [1]. The founding members ofWFNDEC include leading NDE research centers inArgentina, Belarus, Brazil, China, India, Republic of Korea,Russia, South Africa, Ukraine, and the United States witha permanent secretariat at Centre for NDE, Iowa StateUniversity, Ames, USA. Indira Gandhi Centre for AtomicResearch (IGCAR) is a founding member with activeparticipation in all the WFNDEC activities.

Towards benchmarking of various NDE techniques, theWFNDEC regularly defines benchmark problems inultrasonics, magnetic flux leakage and eddy currenttechniques, to the NDE community for solving by numericalor analytical means. The WFNDEC has announced inFebruary 2012, a new benchmark problem in eddy currenttechnique and this is the third in the series [1]. The firsteddy current benchmark problem was purely based on axi-symmetric model (coil and defect configurations) andinvolves the prediction of impedance changes of a bobbindifferential coil due to groove type defects and supportplates [2]. The second benchmark problem had an axi-symmetric tube coil condition with 3D localized flaws.Both finite element and semi numerical approaches yieldedaccurate prediction of the second benchmark problem [3,4].

The third problem is completely free from axi-symmetry.The objective of the problem is to predict the impedancechanges in a pancake type eddy current coil due to arectangular slot in an Inconel-600 tube at five differentfrequencies and validation of the predictions usingcontrolled experimental measurement data provided by theWFNDEC. The third benchmark problem is challenging.The authors used CIVA software for solving this benchmark

Solution to the third eddy current benchmarkproblem of WFNDE centers

S. Thirunavukkarasu, B. Purna Chandra Rao, S. Shuaib Ahmed and T. JayakumarNondestructive Evaluation Division, Metallurgy and Materials Group

Indira Gandhi Centre for Atomic Research, Kalpakkam 603 102, TN, IndiaE-mail: [email protected]

ABSTRACTThis paper presents solution to third eddy current benchmark problem by World Federation of NDE (WFNDE) Centers. CIVA modelingsoftware has been used for predicting the impedance changes for a pancake type eddy current coil in presence of a rectangular slot in anInconel-600 tube, as defined in the problem. Quantitative comparison of the model predictions has been made with the experimental dataprovided by WFNDE centers. A very good agreement has been observed at lower frequencies. For higher frequencies, transfer functionbased approach has been followed to account for non-linear effects associated with cable and contacts that have not been modeled. Thisapproach has been successfully validated.

Keywords: Eddy current, CIVA software, Inconel-600, tube testing, benchmark problem

problem. This paper presents the definition of the problem,solution method employed, model predictions andcomparison of the model predictions with experimentaldata.

2. DEFINITION OF THE EDDY CURRENTBENCHMARK PROBLEM

The third eddy current benchmark problem has been definedto predict the impedance changes of a pancake type coilat different axial positions due to the presence of arectangular slot (simulated crack) in an Inconel-600 (non-ferromagnetic) tube [1]. The impedance changes are to bepredicted at excitation frequencies of 25 kHz, 50 kHz, 100kHz, 150, kHz and 200 kHz. The cross sectional viewillustrating the benchmark problem is shown in Figure 1.

Fig. 1 : Cross sectional view defining the third eddy currentbenchmark problem.

43Technical Paper


The slot is through wall thickness with small opening of0.085 mm, simulating a crack. The tube, coil, slotdimensions and the test conditions to be used by CIVAmodel are given in Table I.

WFNDEC has organized for performing controlledexperiments and provided impedance change data for modelvalidation purpose. The slot was fabricated by electricaldischarge machining and the impedance changemeasurements were made using an impedance analyzer.Experimental data was acquired with the coil placedsymmetrically over the slot and moved incrementally inthe axial direction in steps of 1.0 mm. The impedancechanges were measured at 25 kHz, 50 kHz, 100 kHz, 150,kHz and 200 kHz and this data was shared with the authorson request.

3. MODELING OF THE BENCHMARKPROBLEM USING CIVA SOFTWARE

As there is no axi-symmetry in the problem geometry withrespect to the coil axis due to curved surface of the tube,2D axi-symmetric modeling is not possible. 3D Finiteelement and boundary element techniques are, in general,computationally intensive [5, 6]. In this context, use ofsemi-analytical approach such as the one used in CIVA isadvantageous.

CIVA is benchmarked software for numerical simulationof ultrasonic, eddy current and radiography techniques.The eddy current module of CIVA software is based onsemi-analytical methods using dyadic Greens functionapproach [4, 7, 8]. In this approach, the interaction betweendefect and electric field generated by the probe is describedwith an integral equation given below.

(1)

where JΩ is the unknown fictitious current density definedin the volume Ω containing the defect and depends on thetotal electric field. This equation is derived from theMaxwell’s equations and is solved numerically using themethod of moments. The solved current density is used forcalculating the probe response or signal from a defect. Theterm J0 in (1) is an excitation term that depends on thetotal primary electric field E0(r) by the probe in the region

Ω containing the defect. The dyad G–—Ωee links the fictitious

current density to the electric field it creates inside Ω. Thecontrast function f(r) in equation (1) is defined by

(2)

where r is the position coordinates, σ0 is the tubeconductivity and σ(r) is the flaw conductivity. CIVA eddycurrent module has been extensively validated through aseries of experiments [8].

The CIVA software version 9.2 was used for solving thebenchmark problem. This version of the software is notcapable of solving 3D coils inside a tube, rather it solvesfor equivalent plate geometry ignoring the curvature.However, the influence of curvature effects is minimal forthe pancake type localized probe. Figure 2 shows the 3Dgeometry of the benchmark eddy problem with absolutepancake type coil and rectangular slot. In order to predictthe impedance changes at different axial positions, the coilwas moved from -24 mm to +24 mm along the axialdirection at a scan pitch of 1 mm with the defect center atthe origin as carried out in the experiments. An optimizedmesh based on trial and error and memory limitation waschosen for solving the problem. The optimized mesh had15 elements along the length, 10 elements along thethickness and 20 elements along the width of the slot.

Table I : Dimensions and material properties defined by WFNDEC

Geometry Material Electrical conductivity, MS/m Tube dimensions, mm Other details

Tube Inconel-600 0.84 OD - 18.990Thickness -1.175 Nil

Coil Copper 59.80 OD - 7.836 Lift-off (λ) - 1.1mmID - 3.058 Number of turns 305Height -1.044(x2-x1 as shown in Fig. 1)

Defect Air filled rectangular 0.00 Length-12.20 85 ìm slot throughWidth-0.085 thickness slotDepth-1.175

Fig. 2 : CIVA model geometry of the eddy current benchmarkproblem.

44 Technical Paper


3. RESULTS

Figure 3 shows the model predicted impedance changesfor the rectangular slot when the coil is scanned in theaxial direction. Figure 3 also shows the experimental datagiven by the WFNDE centers. A difference in the amplitudeand phase angle of the model predictions was observedwith the experimental measurements and this difference ismore prominent at 150 and 200 kHz. There exists a hugedifference when the coil is exactly placed over the middleof the slot. This is attributed to skin effect, lift-off andother nonlinear effects associated with cable as well ascontacts [10].

In order to quantitatively assess the closeness of the modelpredictions with the experimental measurements, the sumof Euclidian distances (D) of individual impedance pointswas used. The computed D-values for different frequenciesare given in Table II. The difference between modelpredictions and experimental measurements increasesmonotonically with increase in the excitation frequency.This increase suggests that the difference is predominantlydue to the nonlinear effects associated with cable impedanceat higher frequencies in obtaining the experimental data.

These unknown effects influence the experimentalmeasurements, which results in the difference between themodel predictions and experimental measurements. Hence,it is appropriate to take into account the effect due to theabove said variables using transfer function approach [9,10].The transfer function T takes the form given below:

MEZTZ Δ×=Δ

(3)

where ΔZE is the experimentally measured impedancechange and ΔZM is the CIVA model predicted impedancechange of the pancake coil. In the present case, ΔZE andΔZM are 49 x 5 matrices with complex impedance data(rows representing scan positions and columns representingfrequency). Hence, the transfer function T should essentiallybe a 49 x 49 matrix. In order to compute the transferfunction, matrix inversion method has been used as shownbelow:

(4)

Thus, to compute T, it is necessary to find the inverse ofthe rectangular matrix ΔZM. In order to find the inverse ofΔZM, singular value decomposition (SVD) has been used[11]. In linear algebra, SVD is used for factorization ofreal or complex matrices. Formally, the SVD of an m×nreal or complex matrix M is a factorization of the formM=UΣV*, where U is an m×m real or complex unitarymatrix, Σ is an m×n rectangular diagonal matrix withnonnegative real numbers on the diagonal, and V* (theconjugate transpose of V) is an n×n real or complex unitarymatrix. The diagonal entries Σ are known as the singularvalues of M. Computing the inverse of M is easy, as theinverse of unitary matrices U and V is the conjugatetranspose of the matrices themselves and the inverse of thediagonal matrix Σ is, nothing but, the reciprocal of thediagonal elements. Hence,

M-1=VΣ-1U* (5)

Figure 4 shows the surface plot of the absolute value ofcomplex transfer function, T computed by matrix inversionmethod using SVD. The shape of the transfer function issimilar to the sinc function. The model predicted impedancechanges were corrected by multiplying this transfer functionmatrix to get impedance changes data equivalent to theexperimental data. Figure 5 shows the corrected modelpredictions and the experimental impedance changes. Ascan be seen, model predictions are in good agreement withthe experimental measurements. The Euclidian distancemetric (D) after accounting for the transfer functioncharacteristics for different frequencies are also shown inTable II. As can be noted, the magnitude of D’s is smalland tending to zero. This indicates that the corrected model

Fig. 3 : Model predicted and experimental impedance changes of the pancake coil due to the rectangular slot at different frequencies.

45Technical Paper


Fig. 4 : Surface plot of the absolute value of the complex transferfunction computed by matrix inverse method.

predictions through the established transfer function areexactly similar to the experimental measurements. It mustbe noted here that T is specific to a probe and it must bedetermined for each and every probe.

In order to further study the transfer function approach, aleave-one out cross-validation strategy has been adopted.In this, one frequency data was not used to compute thetransfer function matrix while model validation wasperformed for all the selected frequencies using the transferfunction matrix. Figure 6 shows the result of the leave-oneout cross-validation strategy when 150 kHz data 200 kHzdata were not used for computing the transfer function. Ascan be observed, a good agreement between experimentaland model predictions exists even when the data from oneof the frequencies was not used for computing T. Thecomputed Euclidian distance metric in both the cases wasfound to be 7.3904 and 13.3177 respectively for 150 kHzand 200 kHz data, which is reasonably good.

Table II : Computed Euclidian distance for differentfrequencies.

Frequency, kHz Euclidian distance (D) D after transferfor model predicted data function correction

25 8.8717 0.0449x 10-4

50 15.3214 0.0876x 10-4

100 26.5983 0.1524x 10-4

150 36.5687 0.2007x 10-4

200 45.8545 0.2380x 10-4

4. CONCLUSION

CIVA modeling software was used for solving the thirdeddy current benchmark problem, towards predicting theimpedance changes of a pancake type eddy current coil atdifferent axial positions due to the presence of a rectangularslot (simulated crack) in an Inconel 600 tube. The transfer

Fig. 5 : Model predicted signals after transfer function correctionand experimental impedance changes of the pancake coilat different frequencies.

Fig. 6 : Results of the leave-one out cross-validation strategy for150 kHz and 200 kHz impedance data not being used forcomputing the transfer function.

46 Technical Paper


function approach has been adopted to account for thevariables that are not considered in the model. The transferfunction was computed by matrix inverse method usingSingular Value Decomposition. The model predictions aftercorrection by transfer function are in good agreement withthe experimental measurements (obtained from the WFNDEcenters).

ACKNOWLEDGEMENTS

Authors thank Shri S.C. Chetal, Director, Indira GandhiCentre for Atomic Research (IGCAR), Kalpakkam for hisencouragement and support. Authors also thank Shri S.Mahadevan, Scientific Officer, NDE Division, IGCAR, formany useful technical discussions during the course ofthis work. Shri S. Shuaib Ahmed is thankful to IGCAR forproviding the research fellowship.

REFERENCES

1. h t t p : / / w w w. w f n d e c . o rg / b e n c h m a r k p r o b l e m s _ f i l e s /2012%20EC%20Benchmark%20Announcement.pdf.

2. Y. Li, Z. Zhang, Y. Sun, L. Udpa and S. Udpa, “NumericalSimulation Results for the Eddy Current Benchmark Problem”,Proceedings of the conference on Rev. of Prog in QNDE, editedby D. O. Thompson and D. E. Chimenti, Plenum, New York,Vol. 21, pp. 1902-1908, 2002.

3. Tian, Y., Li, Y., Udpa L. and Udpa, S., “Simulation of the WorldFederation’s Second Eddy Current Benchmark Problem”,Proceedings of the conference on Rev of Prog in QNDE, Editedby D. O. Thompson and D. E. Chimenti, Plenum, New York,Vol. 22, pp. 1816-1823, 2003.

4. G. Pichenot, C. Reboud, R. Raillon, and S. Mahaut, “Results of2007 benchmark obtained with CIVA at CEA: Prediction of ECTinspection over tubes with 2D and 3D flaws”, Proceedings ofthe conference on Rev of Prog in QNDE, Edited by D. O.Thompson and D. E. Chimenti, Plenum, New York, Vol. 27, pp.1775-1782, 2008.

5. Nathan Ida, “Numerical modeling for electromagneticnondestructive evaluation”, Chapman & Hall Publisher.

6. S. Thirunavukkarasu, B.P.C. Rao, S. Mahadevan, T. Jayakumar,Baldev Raj, Z. Zeng, L. Udpa and S.S. Udpa, “Finite elementmodeling for detection of localized defects using remote fieldeddy current technique”, Journal of Research in NondestructiveEvaluation (Taylor and Francis), Vol.20, No:3, pp. 145-258, 2009.

7. C. Reboud, D. Prémel, G. Pichenot, D. Lesselier and B. Bisiaux,“Development and validation of a 3D model dedicated to eddycurrent nondestructive testing of tubes by encircling probes”,International Journal of Applied Electromagnetics and Mechanics,Vol. 25, pp. 313-317, 2007.

8. C. Reboud, G. Pichenot, D. Prémel and R. Raillon, “2008Benchmark Results: modeling with CIVA of 3D flaws responsesin planar and cylindrical work pieces””, Proc. of the conferenceon Rev of Prog in QNDE, Edited by D. O. Thompson and D.E. Chimenti, Vol. 28, 1915-1921, 2009.

9. B.P.C. Rao and N. Nakagawa, “Validation of boundary elementmodel for eddy current NDE”, Proc. of the workshop onelectromagnetic NDE (eNDE-2004), 2004.

10. B.P.C. Rao and N. Nakagawa, “A study of boundary elementeddy current model validation”, AIP Conference proceedings,Rev of progress in QNDE-2005, Edited by D.O. Thompson andD.E. Chimenti., Vol. 25, 2005.

11. William H. Press, Saul A. Teukolsky, William T. Vetterling, BrianP. Flannery, “Section 2.6, Numerical Recipes: The Art of ScientificComputing (3rd ed.)”, New York: Cambridge University Press,ISBN 978-0-521-88068-8.

47Technical Paper


1. INTRODUCTION

Delamination is one of the damages that debilitate acomposite structure’s performance in flexural andcompression loading. In recent times, many researchers[1] explored the use of ultrasonic Lamb waves [2] forStructural Health Monitoring (SHM) and Non-destructiveEvaluation/Testing (NDE/T) of laminated compositestructures. The main complexity involved with Lamb waves,when employed for NDE and SHM applications, is analysisand interpretation of received Lamb wave signal. This isbecause, during the interaction of Lamb waves withdelamination type defects, Lamb waves exhibitcharacteristics like mode conversion, amplitude reduction,reflection and transmission etc.

Some studies were carried out on the interactionphenomenon between Lamb modes and various defects.Žukauskas and KaŽys [3] investigated both numericallyand experimentally, the interaction of ultrasonic wave witha delamination type defect in GLARE3-3/2 compositesample in the through transmission mode. It was shownthat in a defective zone additional Ao modes were generated.Guo and Cawley [4] studied the interaction of Lamb waveswith delaminations in a cross-ply composite laminate.Investigations were carried out on delaminations using thefundamental symmetric Lamb mode (So-) in a pulse-echoconfiguration. Delaminations were introduced at variousinterfaces between the plies. When So mode encounters adelamination in its propagation path, a reflected wave wasshown to be generated. It was found that, when shearstress is zero at that interface, no wave reflection could beobserved. Karthikeyan et al. [5] and Ramadas et al. [6]studied, both numerically and experimentally, theinteraction of primary anti-symmetric Lamb mode withsymmetric delaminations, by using air-coupled ultrasonictransducers. Symmetrical delaminations were investigated

Characteristics of turning Lamb modesin composite sub-laminates

C. Ramadas2, Krishnan Balasubramaniam1, Avinash Hood2 and C.V. Krishnamurthy3

1Center for Non-destructive Evaluation, Indian Institute of Technology Madras, Chennai-600 036, INDIA2Composites Research Center, R & D E (E), Dighi, Pune-411 015, INDIA

3Department of Physics, Indian Institute of Technology Madras, Chennai-600 036, INDIA

ABSTRACTWhen the fundamental anti-symmetric Lamb mode (Ao) propagating in a sub-laminate encounters a delamination edge, reflection andtransmission into the main laminate and sub-laminate takes place. The Lamb mode propagating from one sub-laminate to the other istermed as ‘Turning Lamb Mode’. Numerical and experimental studies, employing air-coupled transducers, were carried out to understandthe variations in reflection and transmission factors of Lamb modes in the sub-laminates of glass/epoxy quasi-isotropic laminates. Thepower reflection and transmission coefficients of the reflected and turning Lamb modes were also estimated for various interfaces ofdelaminations in three laminates - unidirectional, cross-ply and quasi-isotropic. The variations in transmission factor and power transmissioncoefficient, with respect to the thickness ratio, of ‘Turning Lamb Mode’ were observed to be dissimilar.

Keywords: Turning Lamb mode, Power transmission coefficient, Power reflection coefficient, Air-coupled transducers, GFRP composites.

in pitch-catch mode. It was found that when Ao modeinteracts with a symmetric delamination, it generates anew mode, S-o-, which is confined only to the sub-laminatesi.e. within the delamination region. This mode cannot bedetected in the main laminate, although a mode convertedAo mode (when the So mode interacts with the exit portionof the delamination) can be used to infer the presence ofthe So mode in the delamination.

Zhou and Yuan [7] analytically studied the reflection andtransmission of flexural modes in an isotropic beamcontaining a semi-infinite axial crack. It was observed thatthe power of the reflection and transmission coefficientsdepend on both the frequency and the position of the crackacross the beam thickness. Finally the results were verifiedusing conventional finite element method (FEM). Yuan etal. [8] carried out analytical studies on flexural wavereflection and transmission from main beam to sub-beamsin uni-directional (UD) composite beams having a semi-infinite delamination of open and closed nature. It wasfound that the portion of reflected and transmitted powerdepends strongly on the frequency of incident wave andposition of delamination across beam thickness. The resultswere verified using FEM. Wang and Rose [9] investigatedwave propagation in isotropic beams containing closedsemi-infinite delamination.

A good understanding of the interaction phenomenonbetween Lamb modes and defects helps in efficientimplementation of damage detection techniques. Numericalsimulations give a better picture, on interactionphenomenon, than experiments. Delamination in acomposite laminate divides the main laminate, locally, intotwo sub-laminates. When a Lamb mode is incident at theedge of a delamination, it reflects back and also transmitsfrom one sub-laminate to the other as ‘Turning Lamb Mode’(TLM) [10-11]. Ramadas et. al. [10] estimated the

48 Technical Paper


transmission factors, based on amplitudes, of turning Lambmodes using Hilbert Transform.

The main focus of this work is to study the variations inpower reflection coefficient of Lamb mode in sub-laminatesand power transmission coefficients of TLM with respectto thickness ratio, which is defined as the ratio of thicknessof sub-laminate in which Lamb mode is excited, to themain laminate thickness. Numerical simulations werecarried out on three different laminates, unidirectional (UD),cross-ply (CP) and quasi-isotropic (QI). Transmission andreflection factors, based on wavelet transform, of Lambmodes were estimated for QI laminate and compared withexperimentally obtained values. It was observed that thenature of variations in transmission and reflection factorswith respect to thickness ratio were same. Subsequently,power transmission and reflection coefficients of Lambmodes, when interface of delamination is at variouslocations across the laminate thickness, in all threelaminates – UD, CP and QI, were estimated. There wasfound to be a new trend in the nature of variations inpower transmission coefficients, with respect to thicknessratio, of TLMs.

The organization of this paper is as follows. Section twodeals with numerical simulations carried out on QI. Sectionthree illustrates the experimental work carried out on QIlaminates. Estimates of power transmission and reflectioncoefficients are brought out in Section four. Results anddiscussion is presented in Section five. The paper concludesin Section six with a recapitulation of the importantfindings.

2. NUMERICALLY SIMULATEDTRANSMISSION AND REFLECTIONFACTORS

Fig 1(a) shows the specifications of the model used forsimulations. Numerical modeling was carried out usingfinite element code, ANSYS, on QI glass/epoxy (GFRP)laminate having stacking sequence of [0/±45/90]s.

Table-1 lists the mechanical properties of GFRP material.Since the thickness of each lamina was 0.33 mm, the totalthickness of each laminate was 2.64 mm. Damping wasnot considered in numerical simulations. Excitation pulsemodulated by Hanning window had five cycles with 200kHz as central frequency of excitation. Lamb modeemployed in the investigation was the fundamental anti-symmetric mode (Ao). This mode was excited by givingthe displacement pattern, obtained from DISPERSE [12],across the sub-laminate thickness. The type of elementused for FE modeling was a solid element. There are threetranslatory degrees of freedom (DoF) at each node.Delamination was modeled by demerging the nodes at thatlocation. The type of delamination considered in this workis ‘semi-infinite’ because the delamination started fromone of the edges of the plate and extended up to half theplate length (150 mm), and had only one edge as shownin Fig 1(a). The length of delamination was selected in

such a manner that the arrival time of reflected wave groupsis much higher than those from the delamination edge.The wave groups captured in this work were propagatingtowards and/or from the delamination edge. Since eachlaminate contains a total number of eight plies, stackedsymmetrically with mid-plane, it is possible to have adelamination at any one of the four interfaces across thethickness.

Table 1 : Material properties

Material E11(GPa) E22(GPa) u13 u23 G13(GPa) rkg/m3

Glass/Epoxy 44.68 6.90 0.280 0.355 2.54 1990

The following delineates the numerical simulation carriedout when the interface of delamination introduced wasbetween the plies [90] and [-45]. The transmitter (T) andreceiver (R1) were on the top sub-laminate while the receiver(R2) was placed over the bottom sub-laminate as shown inFig 1(a). Receivers R1 and R2 capture the Ao modegenerated by the transmitter, reflection from thedelamination edge and the TLM, respectively. A-scancaptured at receiver R1 is shown in Fig 1(b). The firstwave group was the incident Lamb mode and the secondone is the reflection of the first from the edge ofdelamination. Over each wave group, wavelet transformwas carried out. The mother wavelet chosen was Morlet.Figs 1(b) and 1(d) show the envelope of the waveletcoefficients and time-frequency representation (usingwavelet transform), respectively of the wave groups, theincident and reflected. Fig 1(c) shows the TLM capturedby the receiver R2. The envelope of wavelet coefficientswas fitted over the TLM. Time-frequency representation isshown in Fig 1(e). From time-frequency representation itwas found that the central frequency was approximately200 kHz and the arrival times of wave groups closelymatched with those of the Ao modes propagating in thesub-laminates.

The reflection factor of the reflected wave group is definedas the ratio of the peak of wavelet coefficients of thereflected wave group (Fig 1(b)) to the incident (Fig 1(b)).Similarly, the transmission factor of the TLM is defined asthe ratio of the peak of wavelet coefficients of the TLM(Fig 1(d)) to the incident (Fig 1(b)). Delamination wasintroduced between each interface of the plies and in eachcase, the reflection and transmission factors were computed.These factors were plotted with respect to the ‘thicknessratio’, which is defined as the ratio of the sub-laminatethickness to the main laminate thickness. From numericalsimulations it was found that, the reflection factor wasfound to decrease with increase in the thickness ratio ,whereas the transmission factor increased with increase inthe thickness ratio as shown in Figs 4(a) and 4(b).

3. EXPERIMENTAL WORK

Experiments were carried out on GFRP quasi-isotropiclaminate having a ply sequence of [0/+45/-45/90]s.

49Technical Paper


Fig. 1 : (a) Model used for numerical simulations, (b) and (c) A-scans obtained at receivers R1 and R2 respectively, (d) and (e) wavelettransforms of A-scans shown in (b) and (c) respectively.

50 Technical Paper


Since the thickness of each ply was 0.33 mm, the totalthickness of the laminate was 2.64 mm. Four specimens ofeach size 300 ´ 200 mm2 were fabricated using the ResinFilm Infusion (RFI) technique. Delamination wasintroduced in between the plies (a) [0] and [+45] (b) [+45]and [-45] (c) [-45] and [90] and (d) and [90] and [90] inthe first, second, third and fourth specimen, respectively.Size of delamination introduced in each specimen was 150´ 100 mm2 as shown in Figure 2. Air-coupled ultrasonictransducers with a central frequency of 200 kHz, providedby Ultran Group, USA, were used for carrying outexperiments. Pitch-catch arrangement was employed.

The angles of air-coupled probes were adjusted to generateand receive Ao modes in the sub-laminates. The distancesof separation between the probes and their locations overthe sub-laminates are shown in Fig 3.

Ultrasonic probes were oriented in three differentconfigurations to capture the incident, reflected and turningLamb modes. Figs 3(a), 3(b) and 3(c) show the time-frequency representation of the incident, reflected and theTLM respectively. The same procedure depicted innumerical simulations was followed here for computationof the transmission and reflection factors. Figs 4(a) and4(b) show the variations in the reflection and transmissionfactors with respect to the thickness ratio.

4. POWER REFLECTION ANDTRANSMISSION COEFFICIENTS

The power reflection and transmission coefficients wereestimated through numerical simulations, assuming planestrain condition, on an assortment of three symmetriclaminates – UD, CP and QI having stacking sequence of[04]s, [0/90/90/0]s and [0/+45/-45/90]s respectively. Thepower associated with the reflected and transmitted Aowave groups was estimated through numerical simulations.The following expression [1] gives the time averaged powerflow, <P>, across any cross-section.

(1)

where, the stresses sxx, txz and displacements u and w arefunctions of x, y and t (time). From equation (1) it is clearthat the integration has to be performed across the laminatethickness in the time interval, to. Power transmissioncoefficient is defined as ratio of power transmitted to theincident power. Similarly, power reflection coefficient isdefined as ratio of power reflected to the incident power.

To begin with, the semi-infinite delamination wasmodeled between the first lamina and the second lamina inUD laminate. The top and bottom sub-laminates were [0]and [07] respectively. Ao mode was excited over [0] sub-laminate by giving appropriate displacement profile taken

Fig. 2 : Quasi-isotropic laminate with semi-infinite delamination.

Fig 3 : (a) Incident wave group at delamination edge (b) reflectedwave group from delamination edge (c) turning wave group.

51Technical Paper


from DISPERSE. There were three receivers, one eachover the top sub-laminate, bottom sub-laminate and mainlaminate. The receiver at the top sub-laminate captured theincident and reflected Lamb modes from the delaminationedge. The receivers at the bottom sub-laminate and themain laminate captured the TLM and the transmitted Lambmode respectively. The average power associated with eachwave group was calculated using equation (1). The powerreflection and transmission coefficients were estimated asdefined above. Now, the delamination was introducedbetween the second and third laminae. Again, the proceduredescribed above was followed to calculate the powerassociated with reflected and transmitted Lamb modes. Thisprocess was repeated for all seven interfaces ofdelaminations in each laminate – UD, CP and QI.

Figures 5 (a), (b) and (c) show the variation in powerreflection and transmission coefficients of the reflectedand transmitted Ao Lamb modes.

Fig. 4 : Variation of (a) reflection factor and (b) transmission factorwith respect to the thickness ratio.

Fig. 5 : Variation of power (a) reflection coefficients, (b)transmission coefficients of TLM and (c) transmissioncoefficients of Lamb mode transmitted into the mainlaminate.

5. RESULTS AND DISCUSSION

Numerical simulations carried out on GFRP QI laminateshowed that the reflection and transmission factors definedbased on the wavelet transform were found to decreaseand increase respectively with respect to thickness ratio.

52 Technical Paper


Experiments were carried out, employing air-coupledultrasonic probes, on GFRP QI laminates containing semi-infinite delamination as shown in Fig 2. The reflection andtransmission factors were estimated in each laminate. Figs4(a) and 4(b) show variation in reflection and transmissionfactors of the reflected Lamb mode and the TLM withrespect to thickness ratio. The nature of variation inreflection and transmission factors in numerical simulationsand experiments was identical. The mismatch between thenumerically simulated and experimentally obtainedreflection and transmission factors was attributed mainlyto attenuation. In numerical modeling, attenuation was nottaken into account. However, this endeavor revealed thatthe trend in variation of reflection and transmission factorswith the thickness ratio in numerical modeling andexperiments is same.

The power associated with the incident, transmitted andreflected Lamb modes was calculated using equation (1)on a variety of GFRP laminates - UD, CP and QI containingsemi-infinite delaminations. Fig 5(a) shows the variationin power reflection coefficient of Lamb mode reflected atthe edge of delamination. The power reflection coefficientdecreased with increase in the thickness ratio. When thethickness ratio was 0.125, the power transmissioncoefficient was around 0.75 in all three laminates – UD,CP and QI. With subsequent increase in the thickness ratiofrom 0.125 to 0.875, the power transmission coefficientdecreased from 0.75 to 0.015 (approx).

An interesting phenomenon was noticed in case of theTLM. The power associated with the TLM was found toincrease when the interface of delamination was movedfrom the top lamina to the fifth lamina. When the interfaceof delamination was in between the fifth and sixth laminae(thickness corresponds to 0.625), the power transmissioncoefficient of the TLM was high as shown in Fig 5(b). Thepower transmission coefficient was found to decrease whenthe interface of delamination was moved further down.Initially, the variation in power transmission coefficientwith respect to thickness ratio (from 0.125 to 0.375) wasvery high. When the thickness ratio was further increasedfrom 0.375 to 0.625, variation in power transmissioncoefficient was low as shown in Fig 5(b). For the thicknessratio from 0.625 to 0.875, variation in power transmissioncoefficient with respect to thickness ratio was again veryhigh as seen in Fig 5(b). Among all three laminates, themaximum value of power transmission coefficient was 0.35in QI laminate, followed by UD and CP laminates, at thethickness ratio, 0.625. When Ao was incident at the edgeof delamination, it transmitted into the main laminatethrough the edge of delamination. Fig 5(c) shows the powerassociated with the Lamb modes transmitted into the mainlaminate. The power transmission coefficient of the Lambmode transmitted into the main laminate was found toincrease with increase in the thickness ratio. The powertransmission coefficients vary from 0.05 to 0.8 (approx)for the thickness ratio variations from 0.125 to 0.875,respectively as shown in Fig 5(c).

The trend in variation in power transmission and reflectioncoefficients in all three laminates – UD, CP and QI for allseven interfaces of delamination is observed to be the same.The reflection factor (based on wavelet coefficients) andthe power reflection coefficient (based on power) of Lambmodes exhibit a decreasing trend with increase in thicknessratio as shown in Figs 4(a) and 5(a), respectively. Thetransmission factor (based on wavelet coefficients) and thepower transmission factor (based on power) of the TLMexhibit dissimilar behavior with increase in the thicknessratio as shown in Figs 4(b) and 5(b) respectively. Thetransmission factor increases with increase in the thicknessratio, whereas the power transmission coefficient reachesthe maximum at thickness ratio 0.625, then starts decreasingwith increase in the thickness ratio.

At a given thickness ratio, the sum of the power reflectioncoefficient and power transmission coefficients is not equalto unity because of the following reason. When Ao modeis incident at a delamination edge, in addition to reflectionand transmission, it also generates a new mode, So, whichalso propagates along with the reflected and transmittedAo modes. Since there is some power associated with Somode as well, the total power carried by the reflected andtransmitted Ao modes is not equal to unity.

This study has revealed the fact that the trend of variationsin power transmission coefficients of the TLM with respectto the thickness ratio is completely different from that ofthe transmission factor.

6. CONCLUSIONS

Numerical and experimental studies carried out on thetransmission and reflection characteristics of Lamb modesin the sub-laminates revealed that amplitudes of the TurningLamb Mode and reflected Lamb mode increase anddecrease respectively with increase in the thickness ratio.Variation in power reflection coefficient with respect tothe thickness was also found to follow a similar trend asthat of the reflection factor. There was an increase inamplitude of the Turning Lamb Modes with every increasein thickness ratio, whereas the power was found to rise tomaximum when the thickness ratio was 0.625, and thenshowed a decreasing trend with increase in the thicknessratio.

ACKNOWLEDGEMENTS

The authors acknowledge the help rendered by Dr. RahulHarshe and Mr. Vinod Durai Swami for fabrication oflaminate by RFI process, all from R&DE (E), Pune. Helprendered by Mr. Janardhan Padiyar from CNDE, IITM inexperiments is acknowledged. The authors gratefullyacknowledge the critical comments given by Prof PeterCawley, Imperial College, London.

53Technical Paper


REFERENCES1. Raghavan A and Cesnik, C. E. S (2007), Review of guided wave

structural health monitoring. Shock Vib. Dig. 39(2), 91-114.

2. Rose, J.L. (1999), Ultrasonic Waves in Solid Media. CambridgeUniversity Press, Cambridge.

3. Žukauskas E and KaŽys R (2007) Investigation of thedelamination type defects parameters in multilayered GLARE3-3 / 2 composite material using air – coupled ultrasonic technique.Ultragarsas (Ultrasound), 62, 44 – 48

4. Guo N and Cawley P (1993), The interaction of Lamb waveswith delaminations in composite laminates, J. Acoust. Soc. Am.94, 2240-46.

5. P. Karthikeyan, C. Ramadas, M. C. Bhardwaj and KrishnanBalasubramaniam (2009), Non-Contact Ultrasound Based GuidedLamb Waves for Composite Structure Inspection: SomeInteresting Observations AIP Conference Proceedings, Rev. ofProg. QNDE (Ed. D. Thompson and D.E. Chimenti) Vol. 281096, 928.

6. C. Ramadas, Krishnan Balasubramaniam, M. Joshi and C.V.Krishnamurthy (2009), Interaction of primary anti-symmetricLamb mode with symmetric delaminations: Numerical andexperimental studied. Smart Mater. and Struct., 18(8), 085011.

7. Zhou Li and Yuan Wanchun (2008), Power reflection andtransmission in beam structures containing a semi-infinite crack,Acta Mechanica Solida Sinica, 21(2), 177-188.

8. Yuan Wan-Chun, Zhou Li and Yuan Fuh-Gwo (2008), Wavereflection and transmission in composite beams containing semi-infinite delamination, J. of Sound and Vibration, 313, 676-695.

9. Wang C H and Rose J L (2003), Wave reflection and transmissionin beams containing delamination and inhomogeneity, J. of Soundand Vibration, 264, 851-872.

10. C. Ramadas, Krishnan Balasubramaniam, M. Joshi and C. V.Krishnamurthy (2011) Numerical and experimental studies onpropagation of Ao mode in a composite plate containing a semi-infinite delamination: Observation of turning modes. CompositeStructures, Vol. 93(7), 1929-1938

11. C. Ramadas, Krishnan Balasubramaniam, M Joshi and C VKrishnamurthy (2011) Reflection and transmission of Lamb wavesin sub-laminates, Proceedings of 16th International Conferenceon Composite Structures, June 2011, Porto, Portugal

12. DISPERSE Software version 2.0.16b (2003), Imperial College,London, UK

54 Technical Paper


Rail Weld Inspection Using Phased Array Ultrasonics

Girish.N.Namboodiri1, Krishnan Balasubramaniam2, T.Balasubramanian1, Jerry James2, Sriharsha2

1 National Institute of Technology, Tiruchirappalli 2 Indian Institute of Technology, Madras

ABSTRACTInspection of rail welds has always been a challenge to the Railways. The conventional ultrasonic methods which are employed now forthe detection of defects are not found to be good enough for defects that exist in different parts of the weld. Phased Array Ultrasonics whichperforms sectorial scanning could be used effectively for detection of defects in rail welds. The analysis of phased array images basicallyconcentrates on defects that are volumetric in rail. The feasibility studies conducted in parts of rail other than welds were promising. Defectindications were seen very much separately from its surroundings. Accurate positioning of the defect is possible. Close lying defects canbe seen separately which assures better resolution. Linear normal scans were very much suitable to detect cracks of complex geometries,as it gives a specific indication pattern each time when it is present.

Keywords: Phased Array, Sectorial Scan, Rail Weld, CracksPACS: 43.38.Hz, 81.70.-q, 43.35.Zc, 81.70.Cv

INTRODUCTION

Most commonly used methods of welding rails in Indiaare alumino thermic welding and flash buttwelding.Alumino thermic welds are used to weld rails inareas where there is a lack of power source. Basicallyalumino thermic reactions which are exothermic chemicalreaction are made use in this process of welding. Strengthof this weld just comes to about 56% of the parent rail. Itis having a high failure rate compared to flash butt weldand the quality of the weld is also poor. Thermite weldsare more prone to corrosion also. A wide variety of defectsalso occur in these types of welds. Porosity, blowholes,cracks; slag inclusions are a few of them. These defectsmay arise at any region of the weld but mostly the defectswhich arise from bottom of the weld are found to be moredangerous. On the other hand flash butt welds are goodquality welds with strength almost equivalent to that ofparent rails. Their failure rate is very less compared toalumino thermic welds. Very few defects namely lack offusion and oxide inclusions arise in them. Requirement ofa power source and a unique mobile flash butt weldingmachine limits its use in remote areas where aluminothermic welds are mostly preferred.

Greg Garcia [1] carried out his research which consistedof 2 phases as a part of the program to determine howphased array ultrasonic technology can be applied towardsthe inspection of rail in service. The phased array researcheffort was performed in conjunction with RD Tech/OlympusNDT, a manufacturer and supplier of ultrasonic and eddycurrent phased array systems. A phased array approach forrail flaw detection and sizing performed by TransportationTechnology Centre Inc. under Federal RailroadAdministration sponsorship has been focused on sizingtransverse defects located in the railhead. The Phased arrayprocess evaluated during this research effort uses an

electronic scanning method of transmitting and receivingultrasonic energy from various locations of the railhead. Adetailed study on the defects that arise at various positionsin the rail was done [2]. Total cost of all rail failures wasalso estimated and found that controlling rail failure couldreduce it to a great extent. A novel damage detectiontechnique based on wave propagation of rails was proposedin the year 2006 which was concentrated on theidentification of structural surface damage on rail structures[3].

At present, conventional ultrasonic probes are used in railweld inspection. A detailed study has been conducted tocheck the probability of detection of this technique. Aperiodic testing of complete weld by hand probing of weldhead/web and bottom flange is done using 0° 2 MHz, 70°

2 MHz, 45°2 MHz and 70° 2 MHz probes. Out of these,the 0° 2 MHz normal probe scanning is aimed at detectionof defects like porosity, blowholes etc. in thermite weldsand for the detection of oxide inclusions as in the case offlash butt welds. The phased array linear scan at 0° usinga 2 MHz probe can perform the same objective in a muchbetter way. Apart from detection of the defects mentionedabove, it could also detect the cracks if present inside theweld. The image pattern generated by the crack helps ineasy characterisation and analysis. The 70° 2 MHz headscan is intended to pick up blowholes, lack of fusion, slaginclusion etc. in the head region of the weld. Phased arraysectorial scan using 2 MHz probe at an angular range of10° to 60° can very well detect these defects in the headregion. The set of angles generated helps to reduce thechance of missing any defects which could arise in betweenthis range. A probing is also performed from the sides ofthe rail using a 2 MHz probe at an angular range of 10°

to 60° for the detection of transverse defects which couldarise inside the weld head region.

55Technical Paper


Detection of defects that arise from the foot of rail weldis a matter of deep concern. Presently a 45° 2 MHzconventional probe is used for the detection of these defects.A particular type of crack which occurs in the form of ahalf moon is a potentially harmful defect which needs tobe detected as soon as possible after its origin. The sectorialscan using the 2 MHz probe at the angular range of 10° to60° serves this purpose to a great extent. Presence of boltholes near the weld region affects the conventional scanningmethod in this case. However the lower angles used inphased array scan helps to overcome this difficulty.

Inspection of the flange region is performed with a 70° 2MHz conventional probe at present. A phased array sectorialscan using 5 MHz probe at an angular range of 60° to 80°

is suitable for replacing the above scan plan. The higherfrequency ensures better sensitivity also in addition to theadvantage of the variable angular range. They can detectdefects in the flange region at their earlier stages itself.Since each indication from a phased array sectorial scangives the exact location of the reflecting surface, it becomeseasy to distinguish between a signal from a defect andthose coming from the sharp corners at the bottom ofwelds.

THEORY

Phased array probes consist of a set of piezoelectricelements arranged in an array. When excited, the elementsproduces ultrasonic waves, which interacts with each otherconstructively or destructively leading to an increase ordecrease in the resultant wave energy respectively. Byvarying the time at which these elements are excited, it ispossible to use these effects to both steer and focus theresulting combined wavefront. This is the basic principlebehind Phased array testing. Software called Focal LawCalculator is used to establish suitable specific delay timesfor firing each group of elements, so that the requiredbeam shape could be generated through wave interactions.It also takes into account the probe and wedgecharacteristics as well as geometry and acoustical propertiesof the material while establishing the focal laws [4].

Electronic linear scan and sectorial scan are special featuresphased array has over conventional UT. In electronic linearscan, the same focal law and delay are multiplexed acrossa group of active elements and the entire elements are

involved in the scanning, with only a set of elementsactivated at an instant. Scanning is done at a constantangle and along the total aperture length. In sectorial scan,the beam is swept through a particular range of angles fora particular focal depth. The beam steering taking placeduring sectorial scan maps the components at appropriateangles optimizes the probability of detection of defects[5]. Electronic focusing optimizes the beam shape andsize at expected defect location and also can improve theSignal to Noise ratio significantly.

EXPERIMENTAL SETUP

A manual inspection of the rail weld using phased arrayultrasonic probes was done. A portable phased arrayequipment was used for the inspection. The experimentalsetup is shown in the Fig. 1.

Table 1 : Groups for inspection with defined values to the variables.

Group Thickness (mm) Wave Type Angle range Elements Range Gain Velocity(degrees) excited (mm) (dB) (m/s)

1 172 Longitudinal 10-60 1-16 200 35 5890

2 172 Longitudinal 10-60 1-16 110 35 5890

3 72 Longitudinal 10-60 1-16 80 30 5890

4(a) 10 Shear 60-80 48-64 15 52 3240

4(b) 20 Shear 60-80 48-64 25 52 3240

5 172 Longitudinal 0 All 200 37 5890

Fig. 1 : Experimental setup.

Scan plans prepared were aimed at achieving a maximumcoverage area of the weld region. Based upon the probeposition, range, gain and angular range used for inspection,5 groups were set up for the complete inspection of theweld. The details of these groups are mentioned in theTable 1.

The setup for each group was saved accordingly in theequipment. At the time of inspection, a shifting from onegroup to another has to be done and the data correspondingto each scan has to be acquired. The analysis of the resultswas done after a set of welds are inspected and basedupon the rejection criteria and sensitivity settings given so

56 Technical Paper


that, welds could be characterised as a) immediatelyreplaceable b) to be kept under observation or c) goodquality weld. For the inspection of the flange region of theweld, the size of the probe should be as small as possible.

Two calibration samples, one for thermite weld and otherfor flash butt weld, were prepared with artificial defectsmade at different positions inside the weld. Side drilledholes and flat bottom holes were drilled at various locationsin these samples to replicate the common defects whicharise in these welds. Three side drilled holes (SDH) weredrilled at depths of 70 mm, 85 mm and 100 mm respectivelyon the thermite weld sample. A 3 mm diameter hole wasdrilled from the top surface to replicate the transverse defectin the head region of the weld. In the flange region, a 3mm diameter SDH was drilled at a depth of 10 mm fromflange surface. A flat bottom hole of diameter 3 mm wasalso drilled at the bottom of the slanting portion of theflange region. In the case of flash butt weld sample, theholes in the web region were put at the region ofintersection of the weld and parent rail since lack of fusionis the most common type of defect here. Calibrations wereperformed on these samples using the artificial defectsmade at various depths. Separate groups were made forthe two types of weldments because of their varying nature.For thermite welds, an additional 10 mm thickness isconsidered in the scan setup while scanning from the topbecause of the layer of reinforcement metal coming at thebottom. In the foot region, inspection was performed in 2

stages a) in the region of flange where thickness is nearlyconstant at 10 mm b) in the slanting region of the flangeregion where the thickness was varying slightly. In thelatter, while setting the thickness in the equipment, anaverage value was set. The flange region scanning had tobe done with a lot of care as most of the critical defectsoriginate from this region. These 2 stages of scanning inthe foot region can provide a nice coverage of this region.

RESULTS

The first group of inspection basically targets the defectsin the foot of the welds. In thermite welds, cracks grow inthe shape of half-moon from the foot. This defect has tobe identified as soon as it is formed. To replicate this typeof defect, a 3 mm diameter flat bottom hole was made inthe foot to a depth of 10 mm. When scanned from the topsurface at the angular range of 10° to 60°, a very niceindication of the defect was obtained at an angle of 17°.The true depth (DA) value and the distance from probesurface to the top surface of defect (PA) helped inpinpointing the defect. A clear indication of the flat bottomhole (FBH) was obtained and was clearly seen in thesectorial scan shown in Fig. 2(b). Similar results wereobtained from the FBH drilled in the flash butt welds also.Scanning needs to be performed from both sides of theweld. The S scan image of weld region without any defectis shown in Fig. 2(a).

Fig. 2 : a) Defect free image b) Indication from FBH of 10 mm depth from the bottom surface.

Fig. 3 : a) Defect free image b) Indications from holes at 70 mm, 85 mm and 100 mm depths.

57Technical Paper


The second group is almost the same as the first group butwith a reduced range of 110 mm to visualize the regionnear the head in a much better way so as to reduce thechance of missing any defect. Three 3 mm diameter sidedrilled holes were made in the web region of the weld atdifferent depths which were seen clearly and separately inthe sectorial image shown in Figure 3(b).This scanningneeds to be performed from both sides of the weld forensuring better coverage area. Figure 3(a) shows the nodefect image. Based upon the values of true depth, forwardposition of the reflector with respect to tip of the wedgeand sound path length, the position of these indicationswere checked and confirmed that the indications were fromthese defects itself. The peak of the indication from 3mmdiameter hole at 70 mm was set to 60 % by varying thegain. The gain was around 35 dB which will be set asconstant for further scanning of thermite welds. The threeholes can be seen separately which assures good resolution.

The third group scanning was done with probe kept at thesides of the rail head. The intention of this scan was topick up the transverse defects which could arise inside thehead region. A 3 mm diameter hole was drilled from thetop surface of the rail head. When scanned from the sides,an indication was obtained from the exact position of thedefect. A sectorial scan was taken from the side of the railat an angle range of 10° to 60°. The indication in this case

was coming from a depth of 55 mm as shown in theFigure 4(b). A no defect image is also shown in Figure 4(a).Scanning was repeated from the other side of the weldalso so as to completely cover the weld head region.

In the fourth group, a 5 MHz probe was used to inspectthe foot region of the rail weld. The surface in this part ofrail is very rough and rusty. Hence cleaning of this surfaceis very much important so as to provide a smooth surfacefor the probe to be placed for inspection. Ultrasonic gel orgrease needs to be used as couplant which should form agood coupling between the probe and surface. To ensurecomplete coverage of the foot region, an angle range of60° to 80° was selected. Probe was kept and S scan wastaken. The side drilled hole at 10 mm depth was seen verymuch clearly. Shear wave was used for inspection as theypossess higher energy than longitudinal waves and wasgiving better indications from the defect regions. The sameprobe was used to catch the flat bottom hole of 5mm depthdrilled at the bottom of the slanted portion in the foot ofthe rail. A sectorial scan at an angular range of 60° to 80°

was taken as shown in Figure 5(b). A very nice indicationwas got at a depth of 9.68 mm from the top of the flangeregion. The image with no defects is shown in Figure 5(a).

The fifth group of scanning is was a normal linear 0°

phased array scan with a 2 MHz probe. The purpose of

Fig. 4 : a) Defect free image b) Indication from hole at 56 mm depth.

Fig. 5 : a) Defect free image b) Indication from a hole of depth 10 mm.

58 Technical Paper


horizontal crack in the web region of the rail weld, whichhas propagated from a nearby bolt hole is shown in Fig. 7.The regions marked in the figure were showing a decreasein the depths from 80 mm to 60 mm.

These 5 groups of scanning ensure a complete coveragearea of the weld. The setups corresponding to each groupof scanning could be preset and focal laws could becalibrated before moving for an inspection. The gain valuesare set by considering the reflection from known defectsat known depths. The data acquired corresponding to eachgroup could be saved and analysis may be done at a laterstage by taking into consideration of the rejection criteria.

CONCLUSION

This paper presents a method of inspection of rail weldsusing phased array ultrasonics which proves to be muchefficient than the existing methods. Based on theexperimental study conducted, following conclusions weremade.

Location of the defects in rail welds could be exactlydetermined using the scanning performed. This helpsin confirming whether the indication is coming fromthe defect inside the rail weld itself or from somesharp corners of the weld.

this scan was to detect the horizontal cracks which ariseinside the weld region. The cracks under linear phasedarray scan produce a particular scan pattern which helps inits identification. The holes drilled at different depths inthe rail weld were clearly seen when the scanning wasperformed from the top flat surface. The probe was placedat the top flat surface. In Figure 6(a), an image taken in ano defect weld region is shown where the back wall echocould be seen clearly. In Figure 6(b), the indication fromholes drilled at depths of 70 mm and 90 mm can be seenclearly.

Horizontal cracks present inside the weld were found togenerate an image pattern similar to the orientation of thecrack. By taking into consideration the regions of higheramplitudes in the indications, an idea of the extent to whichthe crack has propagated can be obtained. Any horizontalcracks lying in the head or web region can easily be detectedby the linear scanning performed using phased array probe.Experiments conducted on samples with cracks at knownlocations provided satisfactory results. Image patternsalmost replicated the cracks present in the sample. Theadvantage which these patterns offer to the inspectors infield is huge. Chance of missing a potentially harmfuldefect like a crack in weld will be decreased to a greatextent thus ensuring better safety. An indication from a

Fig. 6 : a) Defect free image b) Indication from side drilled holes at depths 70 mm and 90 mm respectively.

Fig. 7 : Horizontal crack in the web region of weld.

59Technical Paper


Almost complete coverage of the rail weld region isensured with the help of the different scan positionsproposed.Analysis of the data acquired is quite easy which helpsin arriving at a conclusion easily. Based upon a suitablerejection criteria prepared, decisions could be takenquickly for the rejection or acceptance of the inspectedrail weld.The wide range of angles used in the sectorial scanshelps in reducing the chance of missing a defect.Cracks can be detected much easily using this techniqueas cracks produce a specific pattern which helps in itseasy identification.Presence of bolt holes near the welds will not affectthe scanning in anyway because of the lower anglesemployed in sectorial scan.

REFERENCES

1. Greg Garcia, TTCI, Jinchi Zhang, Olympus NDT, “Applicationof Ultrasonic Phased Arrays for Rail Flaw Inspection”, FederalRail Road Administration, July 2006.

2. D. F. Cannon, K. O. Edel, S. L. Grassie, K. Sawley, “Rail defects:an overview”, Fatigue Fract Engng Mater Struct, 26 (2006), pp.865-887.

3. G. Zumpano and M. Meo, “A new damage detection techniquebased on wave propagation for rails”, International Journal ofSolids and Structures 43 (2006), pp.1023–1046.

4. “Phased Array Testing: Basic Theory for Industrial applications”,Olympus NDT, First edition, November 2010.

5. R/D Tech, (2004), Introduction to Phased Array UltrasonicTechnology Applications, R/D Tech Guideline, 1st ed., R/D TechInc., Quebec, Canada.

6. Azar L, Y. Shi, and S.C. Wooh, “Beam Focusing Behaviour ofLinear Phased Arrays”, NDT and E International, 33 (2000), pp.189-198.

7. “Manual for ultrasonic testing of rails and welds”, Researchdesigns & standards organization, Ministry of Railways,Government of India, Revised-2006, pp. 95-97.

8. U. Zerbst, R. Lundén, K. O. Edel, and R. A. Smith, “Introductionto the damage tolerance behaviour of railway rails – a review”,Engineering Fracture Mechanics 76 (2009), pp.2563–2601.

9. Y. Fan, S. Dixon, R. S. Edwards, and X. Jian, “Ultrasonic surfacewave propagation and interaction with surface defects on railtrack head”, NDT&E International 40 (2007), pp. 471–477.

PROBE

Karuppan , a poor stone breaker had to get up at 4.00 Am every day, 365 days an year,

walk 5 km to reach his work spot and break stones to earn a living. His immense faith in

God and his prayers worked in his favour and so when one day out of sheer frustration

when he wished to be a rich man, God granted the wish and he became a rich man. He

was enjoying his new status and was regarded highly by those surrounding him. One day

he saw the King passing by and all the citizens were venerating the king. Now Karuppan

desired to be a king so that he will be treated like the King and the wish was granted by

God. Days passed as he was enjoying the newly acquired status till it became unbearably

hot one day. Karuppan thought that if he became the Sun, then he can be the master of

the earth. His wish was granted and he became the Sun. He was shining all over the earth

and was generally benevolent. The season changed and rainy season came, during which

clouds formed and were obstructing the Sun’s rays reaching the earth. Karuppan then

wished that the clouds are more powerful than the Sun and wished to be the clouds. Lo

and behold he became the clouds and was travelling all over the earth until he was stopped

by a tall mountain. Karuppan was now sure that the mountains are the mightiest of all and

desired to be a mountain. As in the past his wish was granted and he became a mighty

granite mountain. The granite mountain attracted the humans and they started cutting

the mountain into pieces. Karuppan could not bear the pain and was pretty sure that the

mightiest of all creations is a stone breaker and wished to become the stone breaker.

I recall this story, which I heard when I was young, because some of you are wondering

“why an article like PROBE in a technical journal?” Karuppan was actually longing for

freedom and liberation. Every one of us towards the end of our innings asks the question

“why was I born? Whatever we do in life we end up with that question. The question is

when that question shall be asked. The answer will be the sooner the better, so that we

have no regrets. Let us get back to the origin of the universe. Black Hole -Big Bang – Galaxies

- Life – Animals – Human Beings. This means that there was nothing and energy in the

beginning. The energy was given mass by God particle and the evolution took place.

Therefore the basis for existence is energy. The basic energy manifests itself in different

forms, meaning that there is no difference between us and the materials we use. We are

a miniature universe.

Materials do not behave differently from us. They also get stressed, fatigued, age, react to

the presence of other materials (corrode) like human beings. They also exhibit properties

like toughness, elasticity, hardness, yield etc. If we burn our bodies and analyse we end up

with 12 chemicals. The study of NDT is the study of finding out the truth. Spirituality also

is about finding out the truth.

Both proclaim “Sathyemeve Jeyathe” (Truth Prevails)”.

Ram.

journal of non destructive testing and evaluation

Documents

isnt members

ndt users

fields of ndt

ndt instruments

new editorial board

indigenous products

indian products

formof virtual ndt