Development of Backscatter X-Ray Imaging …...Background and Outline • Overview of X-Ray Techniques • A backscatter x-ray (BSX) imaging system was received fr«:>m University

Development of Backscatter X-Ray Imaging Techniques for Space Vehicle Applications

Bence B Bartha

Dale Hope

PaulVona

Martin Born

Tony Corak

United Space Alliance, LLC

CopYright © 2009 by United Space Alliance, LLC These materials are sponsored by the National Aeronautics and Space Administration The U S Government retains a paid-up, nonexclusive, Irrevocable worldwide license In such materials to reproduce, prepare, derivative works, distribute copies to the public, and perform publll .. ly and display publicly, by or on

behalf of the U S Government All other rights are reserved by the copYright owner

UnIted SD8ce Alliance

https://ntrs.nasa.gov/search.jsp?R=20110008202 2020-07-24T06:44:53+00:00Z

Background and Outline

• Overview of X-Ray Techniques

• A backscatter x-ray (BSX) imaging system was received fr«:>m University of Florida for development and testing purposes, current slystems from NucSafe

• Current BSX shuttle applications

• Development with NASA for Constellation applications

• Development of system toward foam applications

• Conclusions and future work

Page 1

UnIted Soace AllIance

X-Ray Background

• X-Rays produced from tungsten anode target

• Energy of electrons produced from cathode determines energy of xrays

• Number of electrons determines number of xray photons

FtLAMENT CONN(CTlONS

Page 2

Fllamef1t FocusIttg CUp

United SDace Alliance

Types of Photon Scatter

• Rayleigh

• Pair production

• Photoelectric Effect

• Compton Scatter

'Y

Nucleus /'" E Ie·.:: lrorl (e -)1

p J0tj"\j

r------... "".,~,; ~J {

r'ni-n qd

(II

Positron (e+)

Page 3

II o 0

United Soace Alliance

X-Ray Imaging Techniques

• Through Transmission

• Scatter Imaging (BSX)

• Backscatter uses detectors on same side as source

• Beams usually highly collimated

• Detectors more sensitive than through transmission technique

• Beam rastered across part with backscatter technique

Through Transmission X .. Ray Imaging

Detector (Film, digital panel)

X .. Ray source

99% Transmission

Object

Backscatter X .. Ray Im4aglng

Detector

X-Ray source 1-------,4 J Detector

1 % Scatter

Page 4


•

•

•

• •

Current Backscatter X-Ray Applications X-Ray techniques used to look for flaws in materials such as castings, forgings, welds, insulation, tiles

BSX X-Ray techniques are currently used for inspecting cargo, at borders and airports

Can also look for voids, lack of fusion, inclusion, corrosion, disbonds, composition, thickness, water intrusion

Used to inspect External Tank

Initial development for Constellation program applications

Orion CEV

American SCience and Engineering

Shuttle External Tank

Page 5


Backscatter X-Ray Experimental Setup

X ... Y Stepper motor stage (0 52mm step size)

• Yxlon x-ray source

• Maximum settings 110kV 20mA

• Four Nal detectors

• 5 5mm spot size

• LabVlew stage control and DAQ

Page 6

United 5Dace Alliance

Backscatter X-Ray Experimental Setup

• Spot size focused to 1 mm with lead apertures

• Collimator sleeves have 60mm travel length

• Lead fins are Inside collimator sleeves and can be rotated to block out primary backscatter x-ray signals

• The Source and the detectors raster across the Imaging area of Interest

Page 7

X-Ray Source Head

[\ \)\ ,/, \ ,i {~1;v ;(, "","'0 ,.,

~ {(' {- /

UnIted SD8ce Alliance

Collimator Setting BSX Imaging Basics

Page 8

• High collimator selttlngs allow photons from all depths to enter the detector


Collimator Setting BSX Imaging Basics

\

\

Blocked

•••••••••• Unblocked

Page 9

• Low collimator settings block out photons scattered from the surface and a given depth into the subsurface

United SDace AllIance

Fin Rotation BSX Imaging Basics

, , , , , I I

I

I I

I I

I

-------~

Page 10

• Fin setting of 0 degrees with respect to the source allows both primary and secondary backscatter photons to enter the detector

• Fin setting of 90 degrees with respect to the source block ()ut primary backscatter signal


Edge Effect BSX Imaging Basics

X-ray beam Primary backscatter Secondary Backscatter

, \ ,

\ , \ \

$-------

Page 11

, I

~

• Near specimen edges a lower signal IS collected due to the Inability of the photons to backscatter to the detector


Shadowing Effect BSX Imaging Basics

X-ray beam Primary backscatter Secondary Backscatter

\

I

\ I \ I \ I ,

1\ \

~-------I I

--t-----, I

;)

Page 12

• A shadovvlng effect near objects IS

produced due to tertiary backscatter

• Edge effects disappear near the center of the sample

• There would be no shadowing effect If the detector IS exactly at the location ()f the source

~I

United 50ace Alliance

BSX Imaging Basics

\ \

\ \

I

\ \

\ \

\

\!t-------I I I I

--i-----~ , I

~

Page 13

• The best signal IS produced when the source IS over the object

I

'D8°£8~£ur o..!}

UnIted Soace Alflance

Edge and Shadowing Effect BSX Imaging Basics

I I I I

--i-----~ , I

~

Page 14

• Shadowing effect IS produced again In the detector directly above the object

• Edge effect IS produced In the detector off of the edge of the object

D


Aluminum Wedge Block Setup

~ Aluminum standard wedge block with thickness from 1.S9mm (1/16in) to 1S.9mm (S/Sin) thickness in 1.S9mm (1/16in) increments

~ Flat bottom holes (FBH) of 3.1Smm (1/Sin) diameter with depths from 1.27mm (O.OSOin) 1S.24mm (0.60in)

~ Each area was imaged using the BSX detector with 100kV, 20 rnA, Smm/s scan speed 1 mm pixel size, 0 degree fins, and a source to object distance of 127mm (Sin)

Page 15


Wedge Thickness Results

I ~ Detectors showed significant variations in photon counts for the for thinner wedges

~ As expected it was difficult to resolve the thickness differences for the thicker wedges

~ The results from the four detectors were summed up to attempt to enhance the contrast between the different thickness wedges in one image

- 50 E 100 E 150 -> 200

250

- 50 E 100 E 150 -> 200

Detector 1

r r '

-100 0 100 X (mm)

Detector 3

250"""----100 0 100

X (mm)

x 104 Detector 2

~ 2 - 50 E 100 E 150

1 -> 200 6" 250 i£J

-100 0 100 X (mm)

X 104 Detector 4

25 - 50 2 E 100 1 5 .s 150 1 > 200 250"'"--__

-100 0 100 X (mm)

Sum Detectors

. ,

X 104

9

8

7

x 104

2

1

-100 0 100 X (mlm)

Page 16


Wedge Thickness vs Photon Counts

The data was fit assuming that the photon counts decay exponentially to 0 with decreasing thickness

counts =exp{-A/Thickness} where {A is a constant}

The change in each wedge thickness was detectable up to 15.9mm aluminum thickness

It is difficult to use the entire data set for the fit since the photon counts change near all the edges of the wedges

11

Log Counts vs Thickness

~ o~ 0/ ·

g ,

o

o

tn 10 5 +' C ::s o u -.E

2 4 6

o Detector Sum -Fit

8 10 12 thickness (mm)

o

14 16 18 20

UnIted 5Dace AllIance

Flat Bottom hole Results

The FBH results for each detector shows that tit is difficult to detect each FBti without significant image processing due to the 1 mm spot size of the source being cllose to the diameter of the flat bottom holes

Detector 1 x 104 Detector 2 x 10

4

':, 3.5 ~ M i

1

----. 50 ----. 50 ; 4 E 100 3 E 100 ~

!

'3 S150

2.5 S150

Jt

2 f ""<~

> 200 1.5 > 200 2 250 250 ' ,

-100 0 100 -100 0 100 X (mm) X (mm)

Detector 3 x 104 Detector 4 x '10

4

----. 50 " ----. 50 4 E 100 '"

';v;, 4 E 100 )

S150 S150 3

> 200 ': " , 2 > 200 2 250

fA"td i 250 #~ "<-...,. ~

-100 0 100 -100 0 100 X (mm) X (mm)

Page 18

UnIted 50ace Alliance

FBH Image Enhancement

The background value was first subtracted from each Image and the contrast was stretched to bring out each FBH In more detail

The 1 27mm and the 2 54mm deep FBH were the most difficult to detect, and can only be slightly seen with detector 1

Various filters were used to enhance the Images to detect all of the flat bottom holes

A Kalman filter was used on the stack of Images obtained from the four detectors which brought out the the 1 27mm deep FBH

Multiple scans, smaller than 1 mm apertures and a finer than 1 mm scanning resolution are options to enhance the capability of the BSX system to detect smaller flaws

Page 19


LDDU Experimental Setup

• LDDU scans were conducted with a step size of 0.52mm, scan speed of 2.6mm/s, 90degree fin rotation, 112mm source to object distance and LDDU tilt angles of 0, 11, and 23 degrees

• Three areas of interest were investigated in the RTV seams that joined the PICA pieces together

Page 20

United 50ace Alllal1ce

LDDU Regions of Interest

Page 21

United SD8ce Alliance

RTV hole position and orientation

• RTV voids exist outside of the pre-cured RTV region

• The O.5in holes are shown to be partially filled

• It must be noted that the voids labeled O.125in are realistically at the most O.0625in which is near the limit of the current system configuration of O.039in source beam diameter (1mm)

o 5" VOids

Page 22


LDDU BSX 23° Tilt Scan Area 3

• Void detected near subsurface of non pre-cured RTV section of the LDDU I I

Page 23


LDDU Through Transmission X-Ray Setup

• Through Transmission X-Ray scans were conducted at 130kV and 60in source to object distance on the two pre-cured RTV areas and one non pre-cured RTV area for three separate angles

• The LDDU was set flat on the table for the 90 degree scans

Page 24


LDDU Through Transmission X-Ray Area 1 Results

• Results compare well to BSX Results

• A dlsbond In the RTV seam IS detected and may explain why the 0 51n RTV vOids were filled In dUring processing

• 0 and 23 degree results show the honeycomb substructure In the Image that Interferes with the detection of some of the RTV vOids

Page 25



• RTV voids are difficult to detect with the through transmission x-ray due to the density and the thickness of the compression pad

• Backscatter x-ray results clearly show the 0.5 and 0.25in diameter voids that are not detectable with the through transmission x-ra

""'" ~

t PI 0 ~ p£<@t J

1'0"

, !

, .

Page 26



• Void in the non pre-cured RTV was detected with both x-ray systems near the surface of the PICA

Page 27


Avcoat Components

• Apollo AS-202 (8-1966) Avcoat plugs

• New Avcoat calibration block

• Apollo AS-202 Avcoat heat shield components

Page 28

UnIted 50ace Alliance

A197314230005 Cells at angle to x-ray beam

,------,

BSX

Avcoat Plugs

117kV

17mA

1mm Aperture

Page 29

A1973142.30006

UnIted S08ce Alflance

Avcoat Reference Block

114kV

15mA

1mm Aperture

Page 30

Sum of 4 dete'ctors and filter

Able to detect all flat FBH and gap fills


Apollo sample A19731423006

• Apollo sample A19731423006 was scanned to provide any availclble information on repairs to the TPS material as well as any appare~nt defects from the original manufacturing. Evaluations to be determined.

Page 31


APOLLO SAMPLEI SCANNED ARIEA

Page 32


Large Scan Area 1 mm Spot Resolution

·Posslble vOid In Avcoat Honeycomb Cell

90kV

30mA

o 75mm Aperture

! -1

Page 33

UnIted SDace AllIance

Small Scan Area 0.S2mm Step Size 20 Degree Tilt 100kV

30mA

o 75mm Aperture

-Ability to detect honeycomb liner

- ossible steel honeycomb core detected as darker regions with detectors 1 and 4 near edge of specimen

Page 34

UnIted Soace Alliance

Small Scan Area O.52mm Step Size 45 Degree Tilt

•

•

100kV

30mA

o 75mm Aperture

Tilt allows to see honeycomb to penetration depth limit of system at 100kV

More energy' IS needed to detect Inalcatlons at avcoat honeycomb and steel facesheet Interface

Page 35


Small Scan Area 1 mm Spot Resolution Large Aperture

100kV

30mA

1mm Aperture

• Detector 4 shows slight indication of underlying steel honeycomb near edge of Apollo piece

Page 36


Small Area 45 Degree Tilt Near Edge

•

100kV

30mA

1mm Aperture

Slow Scan 2 6mm/sec

Depth of penetration is approximately 0.5 inch maximum with optimum settings

Page 37

I j

UnIted Soace AllIance

Apollo sample A19731423005 • Apollo sample A19731423005 was scanned to provide any available Information on repairs to the' TPS material as well

as any apparent defects from the original manufactUring Evaluations to be determined

Page 38



'. Area 1

Page 39


SCAN PARAMETERS

• Apollo sample A19731423005 was scanned 0.52mm step size with 0.75mm aperture at a scan rate of 10mm/s

• X-Ray parameters were 115kV and 15mA

• Fin settings were at 90 degrees with sleeves at 1.25" height

Detector 1

Page 40

-3814

""U :r o o ::J

-1990 ("'0

o c ~ VI

-167


IMAGE RESULTS Area 1

• Images were processed using FFT band pass filter and contrast enhancement

• Able to see through Avcoat sample using 115kV unlike in previous Avcoat Piece

Page 41

DR

j l

t

Area

United SORce Allionce


} 1&<4,,,, "1; {f

v & "'w'4$

Page 42

UnIted 5Doce AllIance

BSX Image results Area 2 and 3 Combination of scans show large triangular area due to wooden shim underneath sample

Page 43


Conclusions and Future Work

• Conclusions

- Single sided nature of BSX allows for x-ray of large complex components

- Not necessary to transmit x-rays through object

- BSX system able to detect voids and other indications in valrious heat shield materials

- Increased resolution is necessary to obtain resolution on the scale of DR

• Future Work

- Development needed to scan along curved surfaces

• The x-ray beam needs to be parallel to honeycomb cell!; to image core and minimize interference from cell wall

- System calibration is necessary to transition technology to floor operations

• Correct kV energy must be chosen to image each matelrial and not the honeycomb substructure, or wooden shims

- Modifications and development necessary to quantitatively ldetect composition and depth information with current setup

Page 44

Umted Soace Alliance

Documents

Development of Backscatter X-Ray Imaging …...Background and Outline • Overview of X-Ray Techniques • A backscatter x-ray (BSX) imaging system was received fr«:>m University