BASF-DR Techniques for Corrosion Inspection

DR techniq technique ues s for corrosion inspection

Pro roje ject ctio ion n Ra Radi diog ogra raph phy y

aapplica ppli cation tion stre pplica ppli cation tion „„on on strea streaam“ strea m“

common comm on re sea aarch pro ject ct common comm on rese rese re sea rch proje proje pro ject ct BAS BA SSFF and aand BA M BAS BA nd BAM BAM BA M:: computer computerbased basedradiographic radiographic wall wallthickness thicknessmeasurement measurement

contents: contents:

--basi ba sic ccpri ppri ncip iples les basi ba sic princ rinc ncipl iples es --soft so ftwa ware re fo rreval ev alua uati tion on soft so ftwa ware refor for fo eval ev alua uati tion on and anddocumentation documentation --corros cor rosion ion ddepth epth h evalua eva luatio tion n corros cor rosion iondept de pth evaluat eval uation ion --„corro „co rrosio sion nnmappin map ping“ g“ „corro „co rrosio sion mappin map ping“ g“

Zscherpel Zscherpel Alekseyc Alekseychuk, huk, Ewert, BAM VIII.3 VIII.3

NDT 2003, BINDT, September 2003, Worcester, U.K.

DR techniq technique ues s for corrosion inspection

Pro roje ject ction ion Radio diogr gra aphy

Corrosion and Se Seads

Zscherpel Zscherpel Alekseyc Alekseychuk, huk, Ewert, BAM VIII.3 VIII.3

Example xamples s on Corro Corrosion sion and Erosi Erosion on

Corrosion

Erosion


DR techniq technique ues s for corrosion corro sion inspe insp ection

History: ra radiogra diographic phic wa wall ll thickne thickness ss eva valua luation tion • ap appl plie ied d since since ca ca..

BASF BASF drawing drawing from 1954

1950 • nearly nearly no literat literature ure until 1980 • growing importance in chemical industry caus ca used ed by on on-stream stream app applic licati ation on • since since 199 1996 6 pro projec jects ts BASF und BAM for phys ph ysic ical al st stud udie ies s • standardization open op en (IAE (IAEA A Draf Draft) t)

Zscherpel, Alekseychuk, Ewert, BAM VIII.3


DR techniques for corrosion inspection

pipe line

classical classical set-up: set-up: radiation source, Ir 192

typical typical exposure: exposure:

film, imaging plate, or flat panel

insulation profile line




Projection Radiography

film

on-site exposure, Ir 192 Zscherpel, Alekseychuk, Ewert, BAM VIII.3

Example : distribution station

source

radiograph NDT 2003, BINDT, September 2003, Worcester, U.K.



visual evaluation of wall thickness

w

error sources: - wrong density range - inaccurate edge detection (unsharpness) - wrong magnification factor Zscherpel, Alekseychuk, Ewert, BAM VIII.3



Principle of projection radiography tangential penetration of pipe wall f - film focus distance r - outer pipe radius R - outer radius of insulation w‘ - wall thickness projection on detector plane

pipe insulation

w` measured on fil m Zscherpel, Alekseychuk, Ewert, BAM VIII.3



program :

exact magnification correction :

⎛ r w' ⎞⎟ ⎜ ( f − R ) − ⎜ ( f − R ) 2 − r 2 f ⎟ ⎝ ⎠ w = r − ⎛ r w' ⎞⎟ ⎜ 1+ − 2 2 ⎜ ( f − R ) − r f ⎟ ⎠ ⎝

practical approximation :

⎛ R ⎞ w = ⎜⎜1 − ⎟⎟ w' ⎝ f ⎠

2

4

] 3 % [ r 2 o r r e 1 0 2

4

6

8

10

f/2r Zscherpel, Alekseychuk, Ewert, BAM VIII.3



Projection Radiography film digitisation

PC based wall thickness measurement

imaging plate

flat panel

• single program for evalution of all digital images, no size limitation • support for all detectors by external 16 bit LUT tables for linearisation (gray value proportional to radiation intensity)






overview of measurement equipment at BASF : RT exposure

scanner

PC protocol

screen/film : film digitisation RCF + F8 (16 bit, 300 dpi) (7% of exposure time of Pb+D7) Zscherpel, Alekseychuk, Ewert, BAM VIII.3

software for measurement (algorithms) and documentation

data base



16 bit software platform

wtScope.exe

transfer transferof ofimage imageprocessing processingknow-how know-howto toWindows WindowsPC PC (ScanManager (ScanManager1992 1992-1995, -1995,SUN SUNworkstations) workstations)::

--fully fully16 16bit bitdata datadepth depthfor forevaluation evaluationand anddisplay display --unlimited unlimitedimage imagesizes, sizes,limited limitedonly onlyby bymain mainmemory memory(>512 (>512MByte) MByte) --all alldetectors detectors(digitized (digitizedfilm, film,imaging imagingplates, plates,flat flatpanel paneldetectors) detectors) via viaexternal external16 16bit bitLookup LookupTables Tables(LUT) (LUT)linearizable, linearizable,ASCII ASCIItext text --fast fastrekursive rekursivemoving movingaverage averagefiltering filteringfor forarbitrary arbitrarysizes sizesof ofhigh highand and low lowpass passfilters, filters,convolution convolutionwith withvariable variablekernels kernels(ASCII (ASCIItext) text) --sub-sampling sub-samplingfor forimage imagedisplay displaywith withtrue trueaveraging averaging --error errorestimation estimationfor formeasured measuredvalues valuesand and„simple“ „simple“geometric geometriccorrection correction --quick quickdocumentation documentationof ofmeasurements measurementsby byprotocol protocolgeneration generationand and transfer transferto toWord Word --protection protectionof ofapplication applicationby byindividual individualencrypted encryptedlicense licensekeys keys Zscherpel, Alekseychuk, Ewert, BAM VIII.3





input exposure parameters wall thickness result

table with measurement regions tangential wall thickness algorithm

wall thickness deviation in measurement region Zscherpel, Alekseychuk, Ewert, BAM VIII.3



synchronous display of positions in profile and image !

position correction by hand, if algorithm fails ! Zscherpel, Alekseychuk, Ewert, BAM VIII.3





PC generated measurement protocol



Projection Radiography PC based wall thickness measurement Algorithms :

problem: correct detection of wall positions characteristic points, curve fitting, CT reconstr.

inner wall of pipe max. outer wall of pipe

projected wall thickness w‘ Zscherpel, Alekseychuk, Ewert, BAM VIII.3


DR techniques for corrosion inspection Comparison of accuracy of PC based algorithms

new wall thickness algorithm with adaptive smoothing for improved detection of inner wall position : Accu rac y DN 200 s teel tes t pipe

Accuracy DN 150 steel test pipe

0,6

0,6

0,4

0,4

0,2 m m / g 0 n u h c -0,2 i e w b A -0,4

0,2 m m / 0 g n u h c i e -0,2 w b A -0,4

BASF alt BASF neu CTPWT

-0,6

BASF alt BASF neu CTPWT

-0,6

-0,8

-0,8

0

2

4

6

8

10

wall thickness / mm


0

2

4

6

8

10

12

wall thickn ess / mm



algorithm detects edges, visual not seeable in image :

extension of application range in comparison to visual evaluation Ir 192, DN200:

6

11 mm

Co 60, DN300:

14

24 mm




local corrosion : wall thickness differences in penetration direction

source

pipe

erosion w-Δw intensity change, ΔI detector

absorption law :

Iw = I0 e -µw

relative relativemeasurement measurement:: 1. Calibration : 2. Measurement : Zscherpel, Alekseychuk, Ewert, BAM VIII.3

µ-effective absorption coefficient

2 step algorithm

µ eff = ln ( D REF / DIQI ) / ΔwIQI

Δw = ln ( DREF / DMESS ) / µ eff NDT 2003, BINDT, September 2003, Worcester, U.K.




linearization of film curve and (log.) imaging plate scanner by external 16 bit LUT [Log(gray value proportional to radiation dose)]

digitized digitizedfilm, film,step stepwedge, wedge,160 160kV: kV: imaging imagingplate, plate,pipe pipe(DN (DN150), 150),IrIr::




Measurement of local corrosion 1. Calibration (µ-eff estimation) test pipe DN 100 x 5,6 mm pit 1 Ø R3 x 1mm (Tool-Ø / depth) pit 2 Ø R3 x 2 mm pit 3 Ø R3 x 3 mm

LUT for linearisation

reference point

known wall thickness at reference point known wall thickness at measurement point measure point




Measurement of local corrosion 2. depth measurement

pitting depth: 3mm

known wall thickness at reference point (unchanged)

measure point

reference point

result : local penetrated wall thickness in mm (error ca. 0.2 mm)





evaluation of corrosion depth in inner third of pipe diameter possible

error < 0,5 mm





simulation

G.-R.Tillack, C. Bellon

separable investigations on influence of : • wall thickness • radiation source • scattered radiation • unsharpness Zscherpel, Alekseychuk, Ewert, BAM VIII.3



Simulation:

visual position of edge in image and profile differ !




problem :

local corrosion in wall, tangential projection

wall thickness on both sides of pipe 5 mm, only profile form differ ! simulation Zscherpel, Alekseychuk, Ewert, BAM VIII.3



median filtering in direction of symmetry :

measurement after mapping :

corrosion pit

median filter perpenticular (1x49) >2x flaw size difference image pseudo color representation Zscherpel, Alekseychuk, Ewert, BAM VIII.3

corrosion 5 mm -> complete wall ! NDT 2003, BINDT, September 2003, Worcester, U.K.


real pipe, inner corrosion, insulated

corrosion mapping after median high pass (gray value prop. penetrated wall thickness, 1x300 points) : + 10 mm

- 0 mm - 5 mm




corroded pipe with weld

corrosion mapping after median high pass (gray value prop. to penetrated wall thickness 1x200 points) : + 5 mm - 0 mm - 5 mm




validation

application range, limits for pipe inspection

w

L

Ewert, Onel BAM VIII.3

wall thickness


material: steel NDT 2003, BINDT, September 2003, Worcester, U.K.



Summary

• program for PC based routine measurements on images from all radiographic detectors, tangential wall thickness measurement and local corrosion, simple documentation • wall thickness range enlarged compared with visual evaluation in projection radiography • measurement of corrosion depth (in centre of pipe) • 3D simulation tool for case studies on recognibility • Corrosion mapping with median high pass along symmetry axis of th e pipe allows presentation of wall thickness differences relative to the nominal wall thickness of the pipe

inspection technique not standardised, written practice necessary ! standard development will be started in 2004 Zscherpel, Alekseychuk, Ewert, BAM VIII.3


BASF-DR Techniques for Corrosion Inspection

Recommend Documents