Development of Abdomen Compression Measurement Sensors

1
Development of Abdomen Compression Measurement
Sensors

• T. Shams, N. Rangarajan, J. Rowe, H. Conner
• GESAC, Inc

2
Outline

• Usefulness of compression as injury measure
• some limitations of current methods
• Exploring alternative measurement methods
• Hall sensors
• packaging, calibration, response
• Shape sensors
• Flex sensors
• packaging, calibration, response
• Discussion
• Current work

3
Measuring Abdomen Compression

• Compression measure important in abdomen injury
  assessment
• Maximum compression, V.C, Vmax.Cmax
• (Cavanaugh, Viano, Rouhana, etc)
• Current measurement methods
• Pressure (Mooney)
• Stringpots (e.g Thor)
• Fluid resistance (Rouhana)
• Limitations of current methods
• relies on measuring deflections at a points
• may miss location of maximum deflection
• reliability under oblique loading may not be
  optimum
• no reliable method for measuring in children

4
Exploring Alternative Methods

• Looked at several alternative methods
• Hall sensors
• They can measure relative rotations of a small
  section up to /- 40 deg
• Number of sensors can be used to measure
  deformation of linear strip
• Shape sensor
• Measure displacement at end of flexible beam due
  to delay in transmission of light beam
• Resistive flex sensors
• Depends on change of resistivity when a flex
  sensor is bent
• Can be used to measure average curvature of small
  sections

5
Hall Sensor-Description

• Sensor is small - lt 0.5 cm
• Voltage output proportional to relative distance
  between magnet and sensor
• high level signal
• function of distance or angle
• Easily available
• Can be programmed
• Sensitivity
• Range
• Temperature coefficients

6
Hall Sensor-Mounting Calibration

• Evaluated sensor response for various geometries
• Relative location
• Relative angle
• Decided on hinge mechanism for mounting sensor
  magnet
• Developed calibration fixture for obtaining
  calibration data

7
Hall Sensor-Calibration Fit

• Shows good linear fit between 25 deg and 25
  deg
• Correlation gt 0.99
• Shows excellent cubic fit between 40 deg and
  40deg
• Correlation gt 0.9999
• Normally, will program best range sensitivity
  for individual sensors
• Excellent repeatability
• variation lt 0.1

8
Hall Sensor-Packaging for Abdomen

• Built bands with 3-7 sensors
• Used flexible strips with low stretchability
• Fit into groove cut into abdomen foam
• Tested with disk and rod impactors

9
Hall Sensor-Quasi-Static Response

• In quasi-static loading, voltage output from
  sensors at different locations reflected local
  curvature
• Output lagged behind LVDT but reached peaks at
  same time
• Calculated deflection using calibration values
  similar to LVDT

10
Hall Sensor-Dynamic Response

• In dynamic loading, similar situation
• Initial and final lag
• Computed peak below external measurement
• Peak also appears more smoothed out

11
Hall Sensor-Limitations

• Problems
• Proper sizing and mounting of hinges
• Found adhesive that would work with PVC material
  and Urethane strip
• Mounting of strip
• Strip had lag in following foam deformation
• Tends to move away from foam after impact
• Flexibility of strip requires additional
  tension-interferes with foam stiffness

12
Shape Sensor-Description

• Available from Measurand, Inc (Canada)
• Has processing box attached
• Tested with angular calibration fixture

13
Shape Sensor-Calibration Limitations

• Shows reasonable linear fit between 90 deg and
  90 deg
• Limitations
• Requires multiple sensor array to cover perimeter
  of abdomen
• Much more expensive
• Requires separate processing box, especially for
  high speed applications
• Previous user experience indicated special
  procedures for using with soft foam substrates

14
Flex Sensor-Description

• Resistive flexible sensor
• Resistive layer painted, usually on Mylar backing
• Conductive sections painted on one side
• Resistance proportional to amount of bending
• Obtained from electronic stores
• Used in data gloves
• Inexpensive
• Longer strips can be made

15
Flex Sensor-Calibration Procedure

• Calibration
• Using various radii wooden templates
• Get voltage output as function of curvature (or
  radius)
• End point at location of solder tabs can cause
  problems

16
Flex Sensor-Calibration Fit-1

• Calibration graph
• Each segment appears fairly linear after initial
  low slope
• ( 0.1 (1/in) curvature)
• Linearity depends on uniformity of conductive
  sections
• Better fit over longer segments

17
Flex Sensor-Calibration Fit-2

• Multi-segment strips show some variation between
  segments
• Quadratic (with flat as zero) shows best fit
• R2 0.99

18
Flex Sensor-Preliminary Testing

• Tested using small foam components
• Horizontal vertical orientations of sensors
• Quasi-static
• Impact speeds 1 3 m/s
• Impactor mass 3 5 kg
• External displacement measured by LVDT

19
Flex Sensor-Preliminary Results

• Preliminary results show
• Peak deflection and peak time predicted within
  /- 5
• Unloading occurs more rapidly
• With two strips, the peak deflections show
  similar time histories

20
Flex Sensor-Testing with Infant Dummy

• Testing with Aprica 3.4 kg infant dummy
• Disk and cylindrical impactors
• Tested in horizontal and vertical configurations
• Tested with two or three strips

21
Flex Sensor-Results with Infant Dummy

• Comparison with LVDT
• Small initial lag
• General agreement in time
• Peak underestimated
• Faster unloading
• Two parallel strips show good agreement

22
Flex Sensor-Offset Testing

• Offset impacts with rod
• Expected variation with distance
• No internal stringpot to measure deflection
  independently

23
Discussion-1

• Both Hall sensors and Flex sensors show promise
  as possible instruments for measuring dynamic
  compression
• end conditions need to be addressed
• Hall sensors
• with proper mounting, show good calibration fit
  (cubic fit) and repeatability (R2 gt 0.9999)
• problem with maintaining contact with abdomen
  surface
• still need proper procedure for stringing array
  of sensors into linear strip

24
Discussion-2

• Flex sensors
• can be obtained as strip- eliminating
  difficulties in construction
• calibration fit not as precise as Hall (quadratic
  fit) - R2 0.99
• good repeatability
• problems in securely attaching additional wire
  contacts along length
• ends tend to rebound faster, making unloading
  appear faster
• smaller strips ( 4.5 in 9 in) are appropriate
  size for child abdomens
• can be mounted in horizontal and vertical
  arrangements

25
Current Work

• Selecting optimum length and number of segments
  for use in different abdomen sizes including
  infant dummy
• Verifying measurements under oblique and offset
  impacts
• Improving computation procedure with variable end
  conditions

26
Acknowledgment

• We would like to thank
• Toyota Motor Corporation, Japan
• for funding this work