An Extensible Chemical Sensor Simulation Environment

1
An Extensible Chemical Sensor Simulation
Environment

• Jerome Soller, Ph.D., Mike Jones, Michael
  Johnson, Ph.D., Wendell Duncan, Ph.D.
• CogniTech Corporation
• Joel Dubow, Ph.D., Sandeep Pandita, Chung Yul Cha
• University of Utah

2
Requirements

• Compose simulations from components
• Source code / resource management
• Represent / propagate uncertainties
• Reusable chem/bio specific simulation components
• Integration with
• HLA-compliant simulations
• Vapor dispersion models
• Other DoD assets

3
Performance Specifications
4
Comparison with Other Options
5
Source Code View
6
Integration with Transport Models

• DTRA Hazard Prediction and Assessment Capability
  (HPAC)
• A source of chemical agent concentration inputs
  to sensor models
• Deployments of HPAC
• Standalone workstation capability
• Client/Server capability with CORBA interface
• Transport Component of CogniTech Sensor
  Simulation Environment
• Interfaces Java client of CORBA HPAC server

7
Example Model LIDAR

• LIDAR model uses adjustable optical, electronic,
  atmospheric, other parameters to predict
• direct detection noise power
• carrier / noise
• signal / noise
• instantaneous signal / noise
• probability of detection

8
Visual Composition Interface
9
Inside a Component
10
Calculated Detection Probabilities / Numerical
Uncertainty
11
Example Joint Chemical Agent Detector (JCAD)

• Point Detector
• Input is
• natural atmosphere (Concentration modulated by
  fluctuations, turbulence, and interferents)
• laboratory
• Output is typically a change in frequency
• Need atmosphere model/data and target gases
• The JCAD system is an array of coated Surface
  Acoustic Wave Devices (SAW)
• Need to predict how it will work in real world
  conditions
• Need to effectively design new arrays in response
  to new threats

12
Sketch of Saw Detector
13
Components of SAW Model
SAW Detector Components
substrate
interpret
Coating
modulation
GLM
resonance
polymer
polymer
PZT
frequency
Chemometric
quartz
FPOL
phase
Bayes
PVDF
Oxide
amplitude
ANN
LiNbO3
SnO2
thermal
other
enzyme
Output
Input
constant
Dynamic
Atmosphere
14
Observed Linear and Non Linear SAW
Behavior- schematic

• Non-Linear Behavior (two types)

• Linear Behavior

Change in Frequency
SXFA
Change in Frequency
PIB
FPOL
Gas Concentration
Gas Concentration
Gas DMMP, Simulates GB Nerve AgentMaterials
the coatings of the SAW sensor Derived from
Grate, J. W. Patrash, S. J. Kaganove, S. N.
Wise, B. M. Anal. Chem. 1999, 71, 1033
15
What Has been Accomplished by the University of
Utah

• Model dfreq vs dconc for published SAW sensors
  and their polymer coatings for various gases
• Models for temperature, pressure and substrate
  property variation of detector performance
• Estimate of intrinsic errors in device
• Models and protocols for experimental validation
  of observed non linearities based on
• adsorbtion
• diffusion
• self assembly
• clustering in strong acid / base SAWs

16
Upcoming Development

• Completion September 2004
• BOM vol. 1-compliant simulation integration tools
• Additional search capabilities
• Enhanced transport models interfaces
• Sensor/detection model components for
• SAW
• IMS
• LIDAR/DIAL
• Chromatographic detectors
• Testing
• Navy TEAMS Facility deployment

17
Conclusion

• Sensor Simulation Environment enables
• Life cycle management for chem/bio simulations
• Composability
• Representation and propagation of uncertainty
• Integration with a wide variety of systems
• These enable
• Detector system design
• End to end analysis of capabilities and
  limitations

18
Acknowledgments

• This project was supported by the Office of Naval
  Research contracts N00014-02-C-0169 and
  N00014-01-M-0131.
• Mr. Thomas Holland, the Director of the Naval
  Surface Warfare Center Dahlgren Division Testing
  Experimentation, Assessment Modeling and
  Simulation (TEAMS) facility
• Dr. Wendy Martinez, Program Officer -
  Computational Decision-Making, Office of Naval
  Research
• Assistance was provided by the DTRA HPAC team

19
Contact information

• Jerome Soller, Ph.D.
• CogniTech Corporation
• 1060 East 100 South
• Suite 306
• Salt Lake City, Utah 84102
• Phone [phone]
• Fax [phone]
• E-mail soller_at_cognitech-ut.com
• Web site www.cognitech-ut.com