Montana State University
Academics | Administration | Admissions | A-Z Index | Directories

Montana State Universityspacer Mountains and Minds
MSU AcademicsspacerMSU AdministrationspacerMSU AdmissionsspacerMSU A-Z IndexspacerMSU Directoriesspacer

Contact Me
Steven Shaw
Associate Professor
Tel: (406) 994-5982

ECE Department
Montana State University
610 Cobleigh Hall
Bozeman, MT 59717

Steven R. Shaw: Research Interests

I am generally interested in extracting information from physical systems using novel measurements and analysis procedures. This theme is closely supported by several classical areas in electrical engineering, including system identification, control, modeling, optimization, instrumentation and circuit design.

Recent Research Presentations

Research Projects

Current sensing and non-intrusive fuel cell monitoring
currentsens.jpg Demonstration of far-field current sensing using a regularly spaced and oriented array of Hall-effect sensors. Relatively simple analog electronics can be used to extract current measurements from a wire bundle using an array of Hall effect sensors, even if the sensors do not encircle the conductors. The best coefficients are derived from the singular value decomposition of the current to sensor gain matrix.
fcmonitor.jpg A non-intrusive load monitor (NILM) adapted to measure the interactions between a molten carbonate fuel cell system and attached load.
Micro Mirror Control
micromirror.jpg Two MEMS mirrors mounted and wire-bonded to a TO-5 package. These mirrors can provide dynamic focus and aberration control in miniature optical instruments, potentially including confocal microscopy of tissues by catheterized insertion.
MEMS mirrors MEMS mirrors in a optical profilometer stand, showing a portion of the sensing and control circuitry under development in the laboratory.
Tubular Solid Oxide Test Facility
siemens.jpg Construction is wrapping up on a new off-site test facility to house a Siemens tubular solid oxide stack. This photograph shows the unit in the process of installation.
High Temperature Energy Conversion Fuel cell Controls and Instrumentation
multisourceboard.jpg A multi-source fuel cell control board, using phase-interleaved off-the- shelf power electronic modules with a modified current sharing control scheme. Such controls may provide an extra degree of freedom that can be exploited to increase efficiency, prolong stack life, and introduce stack diagnostics under load.
transientrecogcontrol.jpg A FPGA board used to implement transient recognition control for fuel cell systems. Transient recognition control applies cluster weighted modeling to estimate long-term behavior of electrical transients in real-time. The real-time modification, or sequential cluster weighted modeling, involves replacing the "to be received" portion of the input pattern with the conditional expectation. The resulting computations are relatively computationally intensive for use in closed-loop, but feasible.
FCsimcontrol.jpg Analog electronics used to create a two-loop control system that allows fuel cell materials engineers to electronically scale a cell, repeat unit, or even the middle cells in a short stack to simulate the electrical terminals of stack with an arbitrary number of cells and area. This instrumentation allows system designers to study interactions between prototype cells and the full-scale loads they are proposed to power. This avoids the present necessity of doing cell development in series with system design. We believe this approach offers greater insight into electrical terminal interactions than either finite element cell / stack models or analytical models.
inconel.jpg Machining a mirror-surface for a fuel cell test platen on an Inconel 600 bar with the Compass 250/1000 lathe and carbide tooling. A recess will be bored in a subsequent operation to accommodate a tube, which is held in place with an autogenous TIG weld.
SOFC1.jpg Inconel 600 platens at temperature, with SOFC cell under test.
PEM Degradation and Control Projects
PEM.jpg PEM degradation apparatus, showing custom 200+ channel 14 bit data acquisition system. This apparatus was used for continuous around-the clock monitoring of 160 PEM membranes, under transient loading conditions, at a rate of 2kHz from installation until failure.
hyperbaric.jpg A hyperbaric chamber used to explore pressure as a control variable for a pressurized PEM stack. The bags of PET pellets are being used to move the dominant pole by reducing the system volume.
Degradation oriented control and diagnostics
SOFC4.jpg Anode supported SOFC
Fresh cell ready for testing.
SOFC5.jpg Cell on platen
An SOFC about to be sandwiched between inconel 600 platens, with nickel foam.
SOFC1.jpg Inconel 600 platens at temperature, with SOFC cell under test.
SOFC3.jpg Failed SOFC on test platen
After observing degradation at the electrical terminals under test, this cell shows an annular pattern of failure in an early SOFC experiment. The profound degradation of this cell is accidental, as a result of experimental setup.
scan1.jpg Connection on SOFC
Low magnification SEM cross-sectional image showing electrical connection between mesh current collector and solid oxide fuel cell
scan13-sm.jpg Electrically induced SOFC failure
Mechanical failure at the electrode in an anode-support SOFC, after a few minutes under electrical degradation conditions.
SOFC7-13.jpg Cathode failure
More electrically induced delimination of an anode-supported SOFC, with cathode failure. Failure at cathode is probably secondary.
Nonlinear Least Squares
Converging Parameters Converging Parameters
When a user selects an iterative estimation algorithm, criterion, and model to fit measured data, he or she is really specifying a nonlinear dynamic system with states that consist of the parameters of the problem. The hope is that the user-selected initial guess (initial value) lies within the basin of attraction of the global, optimal solution. The figure shows one such basin of attraction for a benchmark optimization problem. Initial guesses (states) that will never converge are shown in black. My group has developed methods that significantly enlarge the basin of attraction for certain problems, making associated system identification less sensitive to the initial guess.
NSF MRI Spectrometer Development
spectrometer1.jpg This two-board electronic package compromises high-performance analog electronics (top) and a DSP that can be dropped into a loop that uses the interference of a reference laser for control, e.g. the mirror position control loop of a Fourier-transform infrared spectrometer (FTIR). The DSP estimates the time-variation of the non-ideal components, such as sensor drift, and can use these estimate to improve the accuracy of the position control. In addition, even in step-scan mode, the DSP can create synthetic feedback signals that enhance the analytical frequency range of the spectrometer for time-resolved spectroscopy experiments.
High Performance Solid Oxide Fuelcell Powered Flight
HPDflighttest.jpg Preparing to collect baseline data on electric powered flight under specific flight profiles using a conventional remote control plane equipped with an autonomous data acquisition system.
USAF2.jpg "Wind tunnel" and test stand used for propeller, drive, and controller testing for solid-oxide fuelcell powered flight.
USAF1.jpg Off-the-shelf brushless DC motor controller, modified for use with SOFC stack, shock mounted to the test stand cage. Prop is in a pusher configuration.


View Text-only Version Text-only Updated: 11/28/2007
© Montana State University 2007