Joint tracking in friction stir welding

1

• Joint tracking in friction stir welding
• Paul Fleming
• Vanderbilt University Welding Automation
  Laboratory

2
Introduction

• This research presents methods for the monitoring
  of tool alignment relative to the joint-seam in
  Friction Stir Welding, as well as techniques for
  implementing automatic seam-tracking for Friction
  Stir Welding

3
Friction stir welding

• Material joined by a rotating tool which
  traverses along joint line
• Joint types include lap, T and butt

R. S. Mishra and Z. Y. Ma, Materials Science
Engineering R-Reports, 2005, 50(1-2), III 78.
4
Goal of this research

• Develop system capable of detecting the lateral
  position of the FSW with respect to a desired
  position such as centered about the weld seam
• Developing system which utilizes above estimator
  in a feedback control system which maintains a
  desired lateral position
• This is Through the Tool Tracking (TTT)?
• Patent pending serial number Serial No. is
  12/130,622

5
Alignment of FSW tool

• Alignment implies the lateral offset of the FSW
  tool relative to a desired position or path, such
  as the joint seam.
• Effects of misalignment vary between joint types

6
Purpose of research

• The estimator itself could be used as an
  in-system quality check misalignment can cause a
  number of quality flaws and in some joint-types
  (such as blind T-joints) it may not be possible
  to determine lateral position by visual
  inspection
• TTT would provide automated seam-tracking, which
  could allow for tracking of linear and non-linear
  weld seams.

7
Force as a feedback signal

• Forces collected during the weld are used as the
  feedback signal to determine lateral position
• Force signals have already been used in FSW
• Discover metallurgical defects
• Detect gaps in sample fit-up
• Implement load-control
• Estimate tensile strength

8
Experimental Case Blind T-Joints

• Experiment to determine ability to predict
  lateral offset by collected force signals
• 30 welds are run with a varying lateral alignment
• Forces (X,Y,Z and Mz) are recorded throughout
  each weld

9
Results Recorded forces (axial)?
10
Results Recorded forces (axial)?
Voids
11
Results Recorded forces (traverse)?
12
Results Recorded forces (traverse)?
Voids
13
Results Collected Forces

• Examining recorded forces indicate develop of
  lateral position estimator very likely possible
• Attempt to implement position estimator using
  machine learning techniques, treat forces as
  input data and known lateral position as target

14
Position estimation

• Desire to construct an estimator which can
  predict offset position given gathered forces.
• Many possible choices linear or non-linear
  regression, regression tree, SVM
• General regression neural network chosen

15
Neural Networks

• Neural networks are non-linear statistical data
  modeling tools.
• They can be used for classification and
  regression problems

http//en.wikipedia.org/wiki/ImageArtificial_neur
al_network.svgfile
16
General Regression Neural Network

• GRNN is an artificial neural network which
  estimates continuous variables using probability
  density functions
• Converges to conditional mean regression surface

D. F. Specht, IEEE transactions on neural
networks, 1991, 2(6), 568 - 576
17
GRNN performance
Predicted Offset Position
Actual Offset Position
18
Continuous monitoring of weld

• After learning the GRNN using training data, then
  applied network to weld runs where the lateral
  offset was changed several times throughout the
  weld

19
Monitoring lateral position over time
Void Free Region
20
Monitoring lateral position over time
21
Research into Monitoring Capabilities

• Presented research demonstrates effectiveness of
  technique for determining offset position in
  open-air T-joints
• Current research seeks to apply the same
  technique to lap-joints
• Additional research topics
• Enhancing the signal
• Applicaitions...

22
(No Transcript)
23
On-line seam-tracking

• The system is envisioned in two-varieties
• In one, it is assumed that it is capable for an
  estimator block to be developed which can
  determine the magnitude and direction of lateral
  offset. A controller maintains the desired
  offset throughout the weld
• In another case, a signal is maximized at a
  certain position (such as the axial force in this
  experiment around the centered position). In
  this case, the system weaves back and forth to
  gain the center position.

24
Incorporating load control

• Load control is a component of some FSW systems.
• The seam tracking system, which uses forces as
  its input signal, could be made to include load
  control by operating in two alternating stages
• Use seam-tracking to move tool to desired offset
• Use load control to obtain desired axial force as
  known location

25
Future research

• Future research for both monitoring and control
• Monitoring
• Improving the offset monitoring system and
  applying to more joint types
• Tracking
• Development and testing of systems which
  automatically track the weld seam

26
Thank you

• Questions?