Dhiren Modi

1
Active Control of Flow Progression in the Vacuum
Infusion Process Using Real-Time Flow Simulations

• Dhiren Modi

SAMPE-UK Ireland Chapter Student Competition
Date 22/11/05
2
Outline

• Introduction
• Objectives
• Analytical study of pressure profile in the VI
  process
• Proposed Control Strategy
• Experimental Results
• Conclusions

3
Polymer Composites- Applications
4
Vacuum Infusion (VI) Process

• Without Flow Enhancement Media

• With Flow Enhancement Media

5
Issues in VI Process

• Process Optimisation
• Fill-time Minimisation
• Void Reduction
• Resin Wastage Minimisation
• Complex Nature of Flow
• Preform Compaction
• Flexible Mould
• Flow Induced Changes in Compaction
• Multi-scale (Fiber vs. Tow) Flow
• Reinforcement Heterogeneity
• Control of Flow Important but Difficult !

6
Objectives

• To Investigate the Pressure Profile in the VI
• process
• To Develop and Validate a FULLY AUTOMATED
• FLOW CONTROL SYSTEM for VI process

7
Analytical Solution

• Rectilinear Flow
• From Continuity Equation,
• Using Darcys Law

• Iterative Solution Procedure

Vfo and B - Fitting Coefficients
8
Rectilinear Pressure Profile
Fitting Coefficients from Saturated Expansion
9
Analytical Solution

• Radial Flow
• From Continuity Equation,
• Using Darcys Law

• Iterative Solution Procedure

Vfo and B - Fitting Coefficients
10
Radial Pressure Profile
Fitting Coefficients from Saturated Expansion
11
Objectives

• To Investigate the Pressure Profile in the VI
• process
• To Develop and Validate a FULLY AUTOMATED
• FLOW CONTROL SYSTEM for VI process

?
12
Proposed Control Strategy
Start
Perform Numerical Simulations
13
Prerequisite

• Flow sensing system
• Cost-effective
• Less intrusive
• Linkable with controls hardware

• Simple and Cheap Solution
• Webcams !
• FireWireTM Camera (100 Each)!
• Daisy-chaining Capability
• Better Performance

14
Challenges

• Image Analysis
• Filter Design
• Data Processing
• Mesh Correlation
• Gate Scheme Selection
• Hardware Implementation

15
Mesh Correlation
16
Demonstration Model
17
Gate Scheme Selection
An Injection Scheme with Dmin is Selected!
18
Gate Scheme Selection
19
Virtual Experiments
Mean- 1E-08 m2, S/D- 2.29E-09 m2
20
Uncontrolled Experiments(Random Mat)
Vent
Unfilled Region, When Resin Reaches at the Vent
(Centre of the mould)
21
Controlled Experiments (Random Mat)
Vent
Unfilled Region, When Resin Reaches at the Vent
(Centre of the mould)
22
Control Efficiency (Random Mat)
23
Modified Preform lay-up
24
Uncontrolled Experiments(Modified Lay-up)
Vent
Unfilled Region, When Resin Reaches at the Vent
(Centre of the mould)
25
Controlled Experiments(Modified Lay-up)
26
Controlled Experiments(Modified Lay-up)
Vent
Unfilled Region, When Resin Reaches at the Vent
(Centre of the mould)
27
Control Efficiency (Modified Lay-up)
28
Conclusions

• The pressure gradient in the saturated region of
  the mould can be similar for the VI and the RTM
  processes for many of the low-compliant
  reinforcements.
• A new control scheme was developed and
  demonstrated.
• The control system is able to identify the
  macroscopic flow deviations and take appropriate
  control action for efficient infusion.

29
Acknowledgements

• Supervisors
• Dr. Michael Johnson
• Prof. Andrew Long
• Prof. Christopher Rudd

30
Active Control of Flow Progression in the Vacuum
Infusion Process Using Real-Time Flow Simulations

• Dhiren Modi

SAMPE-UK Ireland Chapter Student Competition
Date 22/11/05