What’s important in liner and

medium physical properties

Roman Popil, Ph.D.
Roman@gatech.edu

Institute of Paper Science

Georgia Tech,
Atlanta

Presentation for a general technical audience,

sponsored by AbitibiBowater Coosa Pines Al, 1
 Who is Roman Popil ?
• Georgia Tech/IPST senior research scientist in
Atlanta since 2002
• Principal Investigator of Engineered Packaging
research consortium program
• Manager of IPST Paper Analysis lab for hmm…

contract testing and research – yes, send those

samples to me !
• Been in paper industry R&D since 1986:
MacMillan Bloedel, Honeywell-measurex
• PhD in plasma physics from the University of
British Columbia – Canadian, eh ?
2
Paper Physical Analysis
Testing Services at IPST
Contact  Roman Popil, Ph.D.
Lab Manager
404 894 9722
 Roman@gatech.edu
Testing, seminars, consulting, research…
3
IPST@ GT Paper Analysis Lab

4
What was requested for this seminar
1. Liner and medium properties how they relate to end-use
2. The IPST testing lab capabilities
3. Market rends, use of recycle
4. Lab equipment for tetsing

5
What you will see in this talk ?
Specifically…
• How BCT is related to ECT related to liner and
medium compression strength
• Why caliper and bending stiffness matter as well
• Why RCT is not a good measurement for
product optimization
• How SCT is better for product optimization
• How to use the SCT for product development
• Some basic paper physics for background
• How the TSI with BW and SCT can be used to
predict RCT anyway (if you must have RCT)
• How to implement in an Autoline to predict RCT

6
 Why measure compression
strength of medium or linerboard ?
A warehouse of
stacked boxes, the
customer wants the
boxes to remain
stacked and protect
the contents

What happens instead….

ECT BCT
creep creep
test test

Strains increase until box panels

Linerboard shows a buckling
buckle, fold and collapse –
pattern from vertical applied load 7
stacks topple
United States Rule 41 – since 1991
certificastion for shipping containers
• Old rule
– Burst(Mullen)
– Why ??

• alternative rule
– ECT – much better

The requirement for

ECT allows reduction in
basis weight my
increasing the strength
of the board
8
 What’s going on out there ???
Tests on liners medium
Scandinavia 1 Scandinavia 2 Spain Germany

Reports Basis weight Basis weight Basis weight Basis weight

typical
values Moisture Moisture Moisture Moisture
Caliper Caliper

Reports SCTCD (STFI) SCTCD (STFI) SCTCD (STFI) SCTCD (STFI)

guaranteed
SCTMD (STFI)
values

Tensile stiff. MD Tensile stiff. MD

Tensile stiff. CD Tensile stiff. CD

Burst Burst Burst Burst

Cobb Cobb Cobb Cobb
Scott bond Scott bond
ZDT
Pick resistance

Air permeance Air permeance Air permeance

Roughness Roughness
Friction Friction
9
 Tests on corrugated board

Country Requirement
Belgium Burst, FCT
Denmark Burst, Puncture
FEFCO Burst, ECT ,Puncture
France Burst, ECT
Italy Burst, FCT
South-East Asia Burst, Basis weight
Sweden Burst, ECT, FCT
Switzerland Burst, ECT
USA,Canada Burst, ECT, basis weight
United Kingdom Burst, FCT
Japan Thickness, ECT,FCT

10
Corrugated board weight trend

143#

122# From the magazine:

Solutions! February
102# 2006

11
 Compression strength is a key
property for box performance
Box compression is predicted by the McKee
(1963) equation:
board bending stiffness Box
Board edge compression perimeter
strength

BCT  C( ECT ) 3 / 4 DMD DCD )1 / 4 Z

BCT in lbs
ECT in lb/in
D’s in lb-in
Z = box (length + width) x 2 in inches 12
ECT is related to liner and medium
compression strengths

ECT  C" (2  SCTliner   SCTmedium )

Compression strength
Compression strength of of medium
linerboard
Board take-up factor
(length of medium to liner)
C flute  1.42
B flute  1.25
A flute  1.50
13
ECT Methods: Clamp vs Neckdown
vs Waxed Edge

Sumitomo and Emerson ECT clamps, 200

grit sandpaper on jaw faces, 10 psi L&W neckdown ECT sample
cutter, localizes compressive
pressure from spring clamps, we measure
failure
5 psi – is that ok ?
References: Koning 23, Frank,
References on T 839: Schampfer 10, 15, 38, Eriksson 40
Eriksson 40
14
Bending stiffness of corrugated board
Liner Liner
modulus Board caliper
caliper

2
Eliner t liner h
D board

linerboard tensile strength  Eliner  t liner

15
Measurement of transverse shear
stiffness of board
• usingthe IPST torsion pendulum tester
also requires 4 point measurement of
board bending stiffness

•Ensis Papro “DST” tester measures

the twisting frequency of 4 x 1” board
specimens

•Both methods require cutting and

mounting of samples

16
 How do we measure loss of
transverse shear from crushing ?
• developed at IPST from 1st principles,
calibrated output using weights
• damped twisting motion of a board is analyzed
by a computer program to find the twisting
stiffness
• bending stiffnesses of the test specimen are
found by 4 point L&W tester
• the transverse shear stiffness is calculated
using an equation
• Ensis Papro (New Zealand) also has developed
a similar dynamic twist tester

IPST torsion pendulum

17
 Box performance depends on the
strength of the components
• Historically, strength of board and
components has been measured by:
• Mullen burst
• Ring Crush
• Short Span Compression Test (SCT)
• Edge Crush test (waxed method) also clamp
method

18
 Other alternatives to measure
transverse shear:

MD

The measured area of

BQM from XQ Innovations Australia,
the board must not
measures the resonant frequency of a board
have anything
vibrated along the MD, - no sample prep
underneath it.
required, one button press provides a number
in seconds
19
“CA” type Mullen for testing handsheets and paper up to 100
psi. Calibration foils are 30, 65 and 110 psi
Hydraulic sample clamp actuated
by moving rings down

Electronic
display of max
burst pressure
in psi

Note:
After burst you must
pull the lever back
to neutral, otherwise
the diaphragm will
burst !!

Rubber diaphragm bulges outward as glycerin gets pumped into the

chamber
20
L&W Burst tester for linerboard, Calibration foils are 110 and 160 psi

Ctrl+9 to start a
new
measurement
“PF2”
Spec no. 1 or
100

21
The line of the burst will often be along the MD

22
Burst measures a combination of
properties

From theory considerations the pressure to burst a sheet is

proportional to:

MD and CD tensile strengths, s – MD stretch at

failure, R diaphragm curvature
23
Burst is dependent on many factors

24
 SCT and RCT testing worldwide
• USA
– Some are using only RCT
– other only SCT
– Some both, SCT on low basis weight
and RCT on high basis weight
• Europe
– SCT dominating
• Asia
– RCT almost only
• Australia
– Moving from RCT to SCT
25
Some typical properties for linerboard/
medium set Bending stiffness
Basis Hard
Tests Weight Caliper Density 2 pt L&W (mNm) In-plane Ultrasonic velocities data

(g/m2) (microns) (g/cm3) md cd Ex (GPa) Ey (Gpa) Vxy Vyx G (Gpa)

Roll 96048
14# med 65.82 144.24 0.46 0.33 0.13 4.16 1.74 0.18 0.42 0.98
Roll 96044
16# med 78.16 162.74 0.48 0.55 0.17 4.7 1.13 -0.03 -0.13 0.96
Roll 96038
18# med 86.36 175.6 0.49 0.79 0.28 4.42 1.71 0.14 0.36 1.03
Roll 96033
20# med 94.58 183.65 0.51 0.94 0.36 4.41 2.05 0.17 0.37 1.14
Doug C 26#
med 125.25 252.9 0.5 2.03 0.8 1.85 4.31 0.32 0.14 1.08
Roll 99004
33# med 162.69 329.6 0.49 4.23 2.11 3.64 2.05 0.18 0.32 1.07

42# Liner A 203.97 293.26 0.7 9.31 4.61 7.13 3.54 0.17 0.33 1.94

42# Liner B 209.54 294.22 0.71 8.46 2.95 7.04 2.34 0.05 0.016 1.63

42# Liner C 211.78 318.84 0.67 11.42 3.66 6.46 2.59 0.14 0.34 1.55

Density of medium is 0.5 g/cm3

Density of linerboards are 0.7 g/cm3
26
 Some properties of
linerboard/medium
STFI lbf/in Bending Stiffness
Tests Instron (tensile) data (cd only) (Taber Unit) (gmf-cm) Taber units in mN-m
MD CD
MD Slope CD Slope Ten.Stiff Ten.Stiff
(N/mm) (N/mm) (N/mm) (N/mm) md cd md cd
Roll 96048
14# med 63.606 26.723 445.336 187.124 6.04 1.66 0.73 0.162846 0.071613
Roll 96044
16# med 86.484 24.94 605.6 174.643 6.8 2.82 0.93 0.276642 0.091233
Roll 96038
18# med 75.5 23.057 529.48 161.583 7.96 3.53 1.43 0.346293 0.140283
Roll 96033
20# med 81.783 27.492 573.532 192.661 8.8 4.52 2.15 0.443412 0.210915
Doug C 26#
med 112.549 41.728 788.64 292.429 13.55 12.38 6.43 1.214478 0.630783
Roll 99004
33# med 130.31 60.093 912.764 421.027 19.82 26.79 14.1 2.628099 1.38321

42# Liner A 194.91 88.752 1365.885 622.111 27.12 53.56 31.65 5.254236 3.104865

42# Liner B 182.951 61.354 1281.896 429.92 22.47 50.87 17.96 4.990347 1.761876

42# Liner C 191.789 70.67 1343.845 495.546 23.36 61.03 23.6 5.987043 2.31516

Virgin unbleached 42# softwood kraft

80/20 virgin/recycle
27
Box performance is predicted
based on strength of the
components of the board
Whitsitt
1983, IPC
Related
ECT to
RCT:
(1) ECT = 0.8 * (RCT1 + RCT2 + αRCTMed) + 12
For linerboards > 42 lb/msf
(2) ECT = 1.27 * (RCT1 + RCT2 + αRCTMed) - 6
For linerboards < 42 lb/msf

Figure 1) Whitsitt Relationship of ECT to Composite Ring Crush

Different equations
for different basis
weights…hmmm… 28
ECT predictive models based on
SCT are simpler
ECT  C 2  SCTliner   SCTmedium 
C  0.7
Take-up factor for
Simplified McKee equation the medium ~ 1.42
for C flute
for box compression:

BCT  C  ECT t  Z
C   5.87
Box footprint
perimeter
Corrugated
board caliper 29
 But…the linerboard industry still
sells by the Ring Crush spec
• manual measurement, cannot be
automated
• requires accurate cutting of sample
– can’t easily do handsheets
• handling of the sample
• measurement of sample caliper and
selection of the correct
fixture insert
• it is actually a combination of
compression and…
• bending failure, this is not good for
Ford Highland Park plant –
optimization programs Model T magneto assembly
line 30
Ring Crush method

Sample and holder

between
compression platens
– what could go
wrong ??

31
The trouble of RCT measurement…
Ohh…let’s just send the report
By golly, you’re right !! out anyway – ha,ha.ha!!
These results are
absolutely ridiculous !!

RCT measurement is prone to

many types of errors…some can
be human …
32
The RCT vs SCT “debate”
For RCT:
1. Frank, Benjamin (Tappi J., Sept. 2007) RCT has lower
coefficient of variation, correlation with ECT is better because:
a) test occurs over a larger test piece more
averaging, smaller variability
b) Bending occurs in ECT and box failure
Against RCT:
2. Seth, Fellers etc., SCT is the intrinsic compression strength of
the linerboard,
3. Variability is inherent in the linerboard due to contaminants,
defects, formation
4. SCT is more easily optimized through wet end deification,
furnish changes or refining
5. RCT is not available on in-line automated testing machines.
6. RCT requires more sample preparation and handling.
33
RCT is supposed to predict BCT

BCT test

34
Then which way is the length of the
test strip for CD and MD tests ?

Can you cut the strips exactly along MD or CD, should they be together
or spread apart, how far apart is good enough, what do you do for small
35
handsheets ?
Cutter for 15 mm wide 4 inch span
samples

However, parallel edges not so critical for SCT can use a ratchet
cutter for consecutively adjacent strips handy for handsheet samples
36
Got to use a different cutter for ring
crush though…

Both look the same – need to read the small label ! 37

Ring Crush
Samples are
6 inches long
and 12.5 mm
(1/2”) wide
Use different
center “island”
inserts
depending on
the caliper for
the specimen
38
 Calm cool technicians required with dry hands
Gloves ??
Nah…What for ?! I don’t know about you,
but I am wearing
gloves to protect my
nice new nail job !

39
IPST data ex Mike Schaepe
I sure hope
you like my
data !

What’s this ??

40
 Short Span Compression Test is
SCT
• STFI = SVENSKA
TRÄFORSKNINGSINSTITUTET (Swedish
Forest Products Research Laboratory , now
renamed “Inventia”)

Test method was developed

by Christer Fellers in the
late 1970’s

The short span compression

test is not “STFI”

41
Christer gets another award from Jeffrey Suhling
 Fun Fact (from a student presentation)
• Compression tests are all called *CT, e.g.
FCT, ECT, BCT, but SCT is rather new…
• Many call the short span compression test
SCT as “STFI”, pronounced “stiffy”
• Why is this a bad idea ??
• Answer: ________
Ahem.. um, ah…
my what ?!!!..
Unfortunately, I can
see your “STFI” is way
down today…

42
Corrugated medium comparison – SCT
and RCT
ECT
SCT

IPC, Whitsitt 1985

RCT

ECT increased in accordance with SCT results.

RCT predicted ECT decrease.

Conclusion:
SCT is better suited for low grammage corrugated medium 43
than RCT.
SCT and RCT comparison
Note the difference
between STFI and
RCT (about 2 X)

At very high
basis weights
and calipers, get
slip: more
surface
compression

44
 Effect of grammage and density

RCT

Increased buckling for lightweights

decreases RCT with density
45
Christer Fellers’ slide
 Pressurized
Sample strip clamps
prevent
slipping

CD

0.7 mm gap between

clamping jaws

46
 The L&W SCT
On-board test sample
moisture correction is
available, based on resistance
measurement requires
Simple calibration that is furnish
strip dependent – unused at IPST
insertion, since all samples are
one button conditioned
operation Instrument is checked
Strip edge periodically using a selected
parallelism standard paper sample kept in
not critical a drawer with its data history
for instant comparison

For mill QC use – samples must be conditioned to 50% RH

equilibrium…fastest way  cut 6” x 15mm strips, place in microwave
oven for 10 seconds, hang separated strips in a draft for ~ 2 hours,
then test 47
 After many years of service…
Thermal dot matrix Green LED Keyboard button
printer dots fades, read-out fades, contacts become
rollers wear out missing dots erratic or fail

Unit must be in a conditioned lab, if taken into humid environments

48
(e.g. mill floor) electronics will be affected
SCT works because over the 0.7
mm test span there is no bending
Shows that specific
STFI is
independent of
basis weight and is
a property of the
fiber consolidation

Euler buckling
curves

49
Real compressive failure is
marked by material
breakdown and occurs when
the column height prohibits
buckling

When testing a strip, the crease is

often hard to see – looks like
nothing has happened 50
 STFI or SCT (Short Span Compression Test) is
considered to be the intrinsic compression test strength
for linerboard – this is how you can understand
what affects it through a “Page equation” model for SCT:

Note the
Can affect dependence on fiber
this by low modulus (fibril
consistency angle,species)
refining
and sheet density
(wet pressing)

51
Effects of Processes on Compressive strength using SCT

Excerpted from
“Paper Physics” by
Niskanen

52
Using SCT
Shows the only way to change
SCT is to change the furnish
and beating level of fibers
Recall the Shallhorn Gurnagul
model for SCT…

Graphs excerpted from

the Handbook of
Physical Testing

53
Ring Crush is affected by the bending stiffness and the
caliper of the test specimen so is not a true
compression test

Tappi Journal 1998

Note the dependence on Taber

and caliper in this empirical
relationship

Note that Taber is a bending stiffness

measurement
54
 Why the bending stiffness of linerboard matters
Video observation of ECT tests show:
a) No out-of-plane buckling at higher basis weights
b) Slight buckling at low basis weights for A and C flute boards

Schaepe and Popil

2006

Buckling of
linerboard
facings occurs in
A and C flute,
more so for
lightweights
55
 We can calculate the buckling load for
linerboard using MD and CD Bending Stiffness

We did a lot of work with an L&W 2 point

stiffness tester, a more straightforward
design, cross calibrated instruments with
stainless steel shims.
Taber has historically
been used for 2 point From linear elasticity theory, the conversion
bending stiffness of factor from Taber moment M to stiffness D
liners and medium, is:
Taber measures EI  L 
bending moment M D  M    M 1.67
not stiffness EI/b !
b  3 b 
This means that Taber x 1.67  bending 56
stiffness !!
 Wet pressing increases SCT but
lowers bending stiffness
STFI vs density
Lab data SCT
As density increases using kraft
50
300 gsm y = 40.655Ln(x) + 61.97
handsheets
increases but pressed
bending stiffness
45 R2 = 0.9999
goes down to various densities
40
200 gsm: y = 24.501Ln(x) + 42.411
R2 = 0.9949
35

30

D  Et / 12 3
STFI lb/in

160 gsm: y = 12.587Ln(x) + 25.411

25
R2 = 0.9895

20
what is going on is
15 that the caliper
10 decreases lowering D
100 gsm: y = 10.107Ln(x) + 19.315
R2 = 0.954

5
0.35 0.45 0.55 0.65 0.75 0.85 0.95
Density gsm

57
Wet pressing increases SCT but
lowers bending stiffness
40
As density increases SCT
35 increases but bending stiffness
Geometric mean bending stiffness (mN-m)

goes down
30

10 gsm
25
150 gsm
200 gsm
20
300 gsm D  Et / 12
3

15
what is going on is
10 that the caliper
decreases lowering D
5

0
0.35 0.45 0.55 0.65 0.75 0.85
3
Density (g/cm )
58
 Wet pressing increases density
increases SCT but decreases
bending stiffness
For linerboard: Modulus increases with wet
pressing from increased bonding

Linerboard
3
Et caliper
D
12
SCT  E
59
More buckling occurs with increased wet
pressing so this limits ECT
The ECT buckling model is used to predict the
optimum liner density for ECT
9

Note the Peak

8
ECT occurs
here
7
The ECT buckling
ECT (kN/m)

6
model predicts the
optimal density for
100 gsm
5 160 gsm 160 gsm liner
200 gsm
single wall C flute
4
to be 0.75 g/cm3
3
0.25 0.35 0.45 0.55 0.65 0.75 0.85 0.95 1.05
3
Density (g/cm )

Predicting ECT for handsheets using SCT and bending stiffness

relationships to determine the ECT as a function of density for handsheets
for various basis weights as specified.
60
How can RCT and SCT be related ?
(current IPST ongoing research)
• RCT is a combination of bending and
compression failure
• SCT is compression failure only
• McKee reasoning for plate failure as a
combination of compression and bending takes
the empirical form:
Compressio n failure load 
(Compressio n strength ) b  ( Buckling load ) (`1b )

Related to the bending stiffness which depends on

the structure e.g, round tube, plate, flute shape, etc. 61
Buckling load for a thin walled tube:
From Timoshenko and
Gere: E t
 cr  
3 (1  2 ) r

“E x t ” is tensile
r = 24.2. mm stiffness
t = 0.3 mm
For this formula to
hold need to fit one
l = 6 mm half wave into the tube
height:
l
 1.72 r  t  4.7mm
m
Yep, specimen is high enough to
buckle - so we can use this formula
for buckling of the RCT sample 62
We can get a measure of the ring
buckling load using the L&W TSI

By measuring the speed of sound in paper,

basis weight and the caliper we can
estimate the tensile stiffness
The tensile stiffness is directly proportional
to the buckling load of a thin walled tube.
This can all be done instantly and
automatically…

63
 Combining measurements to
predict RCT

The Autoline for linerboard will have SCT, caliper, basis weight, and a
TSI – these measurements can be combined to fit an RCT model, the
model can be programmed into the Autoline software to provide a
calculated estimated equivalent value for RCT …
But first let’s review a few concepts before we see how to do this…
64
PAPER STUCTURE - Fibers: former wood cells,
OVERSIMPLIFIED typically 1 – 3 mm long, 50
microns wide

ZD

CD SEM x-section

MD

Paper consists of 50 to 30% air

and a network of bonded fibers
aligned predominantly along the
MD – machine direction

For box making applications, only CD

65
SEM surface compression is considered
 Paper as a 3D spring:
In 1D we have Hooke’s law for
the Force F to move a spring
distance x

F=kx

For a 3D solid slab such as paper, we have by analogy, (actually

a good physical analogy is the common cellulosic cleaning
sponge) the pressure or stress σ required to displace the solid by
a strain ε for each principal direction:

 MD  EMD MD  CD  ECD CD  ZD  EZD ZD

The E’s here are the elastic moduli for each of the 3 directions

The ε’s are the strains i.e. relative changes in displacement : ∆ℓ / ℓ 0

66
 E’s and sound speed
To a good approximation:

EMD   VMD
2
(1   12 21 )   VMD
2

ECD   VCD
2
(1   12 21 )   VCD
2

Moduli (MD or CD) = apparent density x velocity squared

We can measure the speed of sound propagating along the plane of the
sheet using a pair of transducers:
?
Volts
time

67
Transmitted pulse Received pulse
 Measuring sound speed in paper
IPC 1970’s Robot Arm difference method technique

Bimorph “paddle” transducers Close up of transducers for “near”

apart for “far” reading reading in shear mode, transducers
are rotated for longitudinal mode

From successive pairs of measurements a 2 distances get

Poisson ration, and shear modulus – 15 minutes per test 68
 IPST Robot Arm – from the 1980’s
Lars Ingman marketed a version through “Robotest”
4 user selectable
measuring
stations

Robot arm end

effector rotates
for polar plots
Takes an hour
for a 2 degree
resolution plot

You know,
I’d rather not
wait an hour
for just one
polar plot,
thanks !!

Sonisys now has an updated version called “UTI”

A typical lab tech’s response 69

L&W TSI comes to the rescue

8 pairs of transducers fixed distance, head comes down at a higher

pressure for basis weights > 100 gsm
Polar plot and V2 produced in 6 seconds but:
Might be influenced by low density and high roughness to produce
underestimates of V2
So…correlation of V2 from TSI vs tensile stiffness has to be verified
for particular sample sets
70
L&W TSI measures VCD2 = TSI_CD
For the buckling load part in the RCT model we need Ext
This is the same as:

BW 2
ECD t  V t  2
V t  BW  TSI _ CD
t
So the buckling load in an RCT model becomes:

  C BW  TSI _ CD

E t
 cr 
3 (1  2 ) r
With C’’ being constant since r and the Poisson ratio’s
do not change appreciably
71
 Modeling RCT in terms of
combined compression and
buckling strengths
The model for ring crush is proposed as:

1b
RCT  C ( SCT ) ( Et ) b

•Based on McKee reasoning for the compression strength of a plate but

instead replacing the buckling load of a plate by that for a thin cylindrical
ring (e.g. Roark’s formulas) where:
•Pcr is the buckling load for a ring is proportional to Et
•Et = tensile stiffness ( modulus E times sheet thickness t) = constant x (v2)
x basis weight
•Can get v^2 ( specific stiffness) from L&W TSI output
• so… can now get an RCT value by measuring SCT, BW and TSI_CD
72
 This shows that TSI * BW = Et when E x t is measured by
the Instron as tensile stiffness
Ultrasonic tensile vs Instron N/mm

1400
y = 1.7524x - 73.305 Ultrasonically
R2 = 0.9829 measured values
1200
are always higher
TSI x BW (N/mm)

than mechanical
1000 equivalents

800
Series1
Linear (Series1)
600

400

200

Instron tensile stiffness (N/mm)

0
0 200 400 600 800 1000
73
 TSI checkout with IPST Robot
TSI vs IPST V^2
14
IPST in-plane specific stiffness (km/s)2

Sample ID
12

y = 1.0821x - 0.1875
42 lb kraft liner 2
R = 0.9782
33 lb medium
10
26 lb medium
56 lb liner WC

16 lb medium 8 Series1
Linear (Series1)
18 lb medium
OCC 42lb liner MD values
20 lb medium 6
CD values
3m mylar
yellow copy
inkjet mylar 4

WAM 33#
blotter
2
2 3 4 5 6 7 8 9 10 11 12
TSI index
Correlation looks ok for a variety of various liners medium, 74
copy paper and plastic film
This is the spreadsheet analysis using Excel “Solver” to fit the model
RCT = C x (SCT)b x (Et)(1-b) i.e., a “McKee” style equation for RCT
using a combination of compression strength (SCT) and buckling load
( ~ Et which in turn, is TSI x BW)
model: q38

RCT = (STFI)^b(Et)^(1-b)

Et = TSIxBW

C = 0.35, b = 0.36 C= 0.349467355

b= 0.362403921
STFI (lb/in)) Ring Crush (lb) ET Model error^2
18.6 61.48 688.8574233 65.02243389 12.54883787
22.43 76.06 837.696652 78.83173899 7.682537056
22.069
30.813
101.15
118.4
1066.474227
1172.550736
91.41300701
109.5963138
94.80903252
77.50488992 RCT  0.35(SCT)0.36 (TSICD )0.64
31.231 122.4 1350.911968 120.5385411 3.465029404
33.467 129.6 1378.359952 125.192706 19.42424044
46.08 158.41 1640.9707 157.1106897 1.688207376
35.829 126.49 1421.312619 130.8608967 19.10473791 Model to calculate an
53.631 161.49 1652.173551 166.7145053 27.29545532
26.209 98.46 1136.489074 101.3153034 8.152757588 equivalent RCT
40.786 160.2 1556.65121 145.3418956 220.7632653
21.702 74.34 875.2949972 80.10608446 33.24773006
34.443 125.36 1285.666476 121.0113364 18.91087478
16.051
28.64
61.98
98.77
680.5922305
1063.086323
61.16781024
100.264588
0.6596522
2.233793158
Data is for a
22.178
14.288
87.82
39
961.1309916
661.461174
85.70075308
57.58578752
4.491207499
345.4314977
series of
17.155
7.539
65.7
15.15
882.722197
281.03
73.96061475
26.46449812
68.23775609
128.0178676
linerboards and
8.599
9.981
21.9
25.47
331.8952
394.284
30.86274479
36.35707584
80.3307942
118.5284204
medium of a wide
14.975
22.193
44.99
74.09
534.2069
747.9719
51.11413764
73.05644945
37.5050618
1.068226737
range of basis
RSQ= 0.980453481
1331.101873 sum error
weights

75
Got data ? Send it to me to analyze !  Roman@gatech.edu
This shows that the r2 improves from 0.94 to 0.98 if we use a non-linear bending
model for RCT: the value is that we get a better prediction for RCT using SCT,
BW and TSI_CD
Ring Crush vs SCT

250
y = 3.591x - 1.3082 RCT  0.35(SCT)0.36 (TSICD )0.64
R2 = 0.9371
200
Ring Crush (lb)
Ring Crush (lb/6 in)

150 Bending model

Linear (Ring Crush

100 (lb))
y = 0.8825x + 12.246 Linear (Bending
2
R = 0.9805 model )
50

0 MSE linear = 2.26 lbs

0 50 100 150 200
MSE bending model =
SCT (lb/in) or 0.35 SCT (TSI)
.36 .64
1.59 lbs
76
 Will this always work ?
• If just using Et = D x (TSI_CD) x BW the assumption of stiffness being
proportional to failure may not hold if:
1. stress-strain curve is changed by increased or decreased
ductility of paper
2. addition of filler, starch, impregnation with polymer, radical
change in fiber species

Comparable stiffnesses but different failure loads can occur…

TSI_CD RCT – refined linerboard

Compression RCT 2 – waxed linerboard
load

RCT 1 – linerboard

Compression Displacement 77
Relevance to Industry, you, me, the
world and everything else..
Current requirement to address the manufacture of recyclable boxes for food and
produce transport meeting FDA specs must be produced locally –demand
cannot decline with a growing population.
Chinese corrugated production volume surpassed US already in 2008
Overseas product and shipping containers are not a threat to this
market: food transport and packaging

Doublewall BC box for

Single wall B flute box for
furniture shipping
PC monitor
Mullens here are bogus 78
Corrugated is not going away
Awwright !!
Let’s make boxes !!
Yeeaahh!!

Hasta la
Vista, RCT
!!

Thank you:
Abitibi-Bowater

Questions, problems, testing, research… 

79
Roman@gatech.edu
 Thank you

“serving the paper industry since 1929…to survive is to do

research, to thrive is to implement…”
80