See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/329761440

Automatic Preform Shape Optimization for the Stretch Blow Molding

Process

Presentation · May 2005

CITATIONS READS

0 71

4 authors, including:

Francis Thibault Alain Malo

National Research Council Canada National Research Council Canada
151 PUBLICATIONS   1,075 CITATIONS    10 PUBLICATIONS   109 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Interactive Real-Time Surgery Simulator View project

BlowView Die Swell Prediction View project

All content following this page was uploaded by Francis Thibault on 19 December 2018.

The user has requested enhancement of the downloaded file.

 Automatic Preform
Shape Optimization for
the Stretch Blow Molding
Process
F. Thibault, A. Malo, B. Lanctot and
R. DiRaddo

Industrial Materials Institute

National Research Council Canada
May 3rd, 2005
 Outline

• Project Objectives and Summary

• PET Material Characterization
• Basic Preform Shapes and Parameterization
• Process Constraints
• Optimization Strategy
• Case Study on Industrial Part
2
 Project Objectives

Development of a Design Advisor for Two-Stage

Reheat Stretch Blow Moulding

Intelligent Preform Design Design Optimization

Container Performance

Processing Parameters

Preform Shape

3
 Project Overview

Performance Specifications
• Weight
Performance Container Characteristics
• Topload, burst, vacuum Optimization Thickness profile
(if needed)
• Thickness bounds

Mathematical rules
&
past designs Yes Preform Design ? No

Preform design advisor (PDA) Existing preforms

Initial guess

Preform Shape Optimization Process Optimization

Operating conditions
Preform tooling • Preform temperature profile
Fine tuning
• Thickness profile (if needed) • Stretch profile
• Shape (geometry) • Pressure profile (P1,P2)
• Cooling conditions
4
PET Material Characterization

5
 Mechanical Behavior of
PET
Strain rate effect

Strain hardening
To effect

Current models: (a) Ogden model, (b) Mooney-Rivlin model

• Represent well the strain hardening, but not the strain rate effect
• To be used for the materials where the strain rate effect is not important
6
 Visco-Hyperelastic Model
• New model developed at IMI:
S   pC1  2{M 1  M 3 ( II C  3)}δ  2{M 2  M 3 ( I C  3)}(C1 ) 2
t 
t
  ( M 4  M 5 II C ) exp(  )εd
M6
where: 0

- S : 2nd Piola Kirchhoff stress tensor, C: Right-Cauchy deformation tensor

- ε: Strain rate tensor, Mi: PET model parameters
PET 9921
200%/s

100 %/s

200 %/s
100 %/s

200 %/s
100 %/s

7
 Basic Preform Shapes

80-85% of SBM Preforms

8
 Preform Parameterization
Z

IDF
Finish (I) Lf
Lc Tc
Transition/Taper LT
(II)
TTB IDTB
Lp

Body (III) LB

IDBB
TBB
End Cap (VI) OREC

TEC

9
 Process Constraints
Z

IDF
Lf
Stretching rod
Lc Tc • IDBB, IDTB  Drod + RCl
constraint
LT
• IDBB < IDTB

Lp
TTB IDTB • ODBB < ODTB

• ODTB < Z
} Injection moulding
demouldability
constraints
LB

To avoid preform
IDBB • IDTB < ODBB inter-penetration - nesting
TBB (not essential)
OREC

TEC

10
 Constrained Optimization
6 Design variables (X1 to X6)
Design objective:
Z
nel e,k
IDF Min  container(X)  
 k T
t  t ( y ) 2

Lf k 1 

Lc Tc by manipulating design variables

LT = (X1-Lf) * Lbtrans X1,min’ < X1’ < X2,max’ [-]
Lbottle X2,min < X2 < X2,max mm
X2 X3 Drod + RCl < X3 < IDF mm
X1 X4,min < X4 < X4,max mm
Drod + RCl < X5 < IDF mm
LB X6,min < X6 < X6,max [-]

subject to process constraints

X5 < X3
X4 X5 (2X4+X5) < (2X2+X3)
OREC=(2X4+X5)/2 (2X2+X3) < Z
X3 < (2X4+X5)
TEC = X6 * X4 11
 Preform Shape Optimization

Initial Preform Stretch Blow Update Preform Shape

Design Moulding Process & Thickness

Position & Shape

Dxi, Dyi, Thicki, i=1,…,5

DOT CONMIN
($ commercial $) (public domain)
Visco-hyperelastic material model

Objective: target a bottle thickness distribution 12

SBM Simulation Tool

13
 Case study: Husky container
Temperature profile before blowing stage -> known
140
120

Temperature (C)
100
80
60
40
20
0
0 10 20 30 40 50 60 70
Axial Position (mm)

Blowing stage conditions

250 4.5
Eastman
Stretching rod displacement

P2 4
PET CB12 200 3.5

Pressure (MPa)
3
150
(mm)

2.5
2
100
1.5

50 1
P1 Stretching Rod 0.5
8oz Spice Jar Pressure
0 0
0 0.5 1 1.5 2 2.5 3
Target container thickness: 0.47 mm Time (sec)
14
 Optimization Results
71 simulations (5 min per sim.) (obj. function+gradients)
0.18 19
Objective Function
Preform Weight
18.5

0.15
18

17.5
Objective Function - StdDevRef (mm)

0.12

17

Preform Weight (g)

0.09 16.5

16

0.06
15.5

15
0.03

14.5

0 14
0 1 2 3 4 5
Optimization Iteration
15
 Preform Shape & Thickness
Optimization Results
2.52 mm

1.33

1.93 mm

0.22

Initial Design Iter. #1 Iter. #2 Iter. #3 Iter. #4 Iter. #5 16

 Container Thickness
Results

0.8
Initial Design
Uniform Target = 0.466 mm
Iteration #1
0.7 Iteration #3
Iteration #5

0.6
Bottle Thickness (mm)

0.5

0.4

0.3

0.2
-140 -120 -100 -80 -60 -40 -20
Distance along Bottle axis (mm)
17
 Container Thickness
Results

0.8

Uniform Target = 0.466 mm

Optimal Design
0.7
Husky Design

0.6
Bottle Thickness (mm)

0.5

0.4

0.3

0.2
-140 -120 -100 -80 -60 -40 -20
Distance along Bottle axis (mm) 18
 Preform Shape &
Thickness Comparison
3.3
Husky preform shape
3.1

2.9

2.7
Preform Thickness (mm)

2.5

2.3

2.1

1.9

1.7 Initial Design

Unconstrained Optimization
Husky Design
1.5
Constrained Optimization

1.3
-65 -55 -45 -35 -25 -15 -5
Distance Along Preform Axis (mm)

Optimized preform shape

19
Software Tour

20
 Concluding Remarks

• New preform shape optimization algorithm

integrating:
– Process constraints (stretching rod, preform and
core rod demouldability)

• Validated against industrial part (good agreement)

• Integrated into a new design advisor software of

AMCOR Pet Packaging (Manchester, MI)
21
Injection Moulding
Process Illustration

22
 Relation between Axial Ratio
and Preform Length

LF

X1  lp  LF  Lb / λ ax  LF  X  LF
'
1

X1’

X  f ( X1 )
'
1
View publication stats
23