Event Detail

Event Type: 
Applied Mathematics and Computation Seminar
Friday, March 6, 2015 - 12:00 to 13:00
KEAR 112

Speaker Info

Thermal Sciences Division, National Energy Technology Laboratory (NETL), U.S. Department of Energy

When developing and applying measurement systems, it is important to consider the
analysis and characterization approaches in the specific context of the measurement
hardware, system to be measured, and the desired outcome. The measurement might
be for determining a physical quantity of interest in the system, or it may be toward
assessing a qualitative state of a system. In advanced systems, multiple measurements
including calibration and characterization routines are usually needed to get results. This
necessitates a propagation of uncertainty analysis. However, all the system variables
which can impact the measurement are rarely all known or even well defined in the real
world. This talk discusses practical considerations in the application of advanced
measurement systems using 2 case studies at NETL with a focus from system analysis
to measurement result.
In the first case, the measurand was the heat flux into a boiler water wall and accuracy
was important. A network of temperature point measurements was implemented and
heat flux was determined based on Fourier’s law of heat conduction. Uncertainty
quantification methodology is discussed and results determined with a typical as-
received device measurement uncertainty of about +/- 30%. A calibration system was
designed and built to characterize and correct for suspected systematic errors with an
expected improvement in heat flux uncertainty to better then +/- 5%. Boiler testing
results showed an improvement in precision for adjacent heat flux pairs close to the
expectation, with a similar improvement in accuracy expected. The measurements also
show interesting drift characteristics during some testing which is thought to have been
caused by uneven ash/slag deposits on the boiler walls.
In the second case, a measurement system was developed and deployed to diagnose
the arc distribution in vacuum arc remelting furnaces. Identifying constricted arc column
conditions at a time scale pertinent to alloy solidification was important. A method based
on the Maxwell-Ampere law and magnetic field measurements external to the system
was used. To characterize system parameters, a 3D Magnetostatic finite element model
(FEM) of the furnace under various arc position states was constructed and
implemented. Then the FEM output data set was fit to a simplified equation based on the
Biot-Savart law in order to decrease computational times in a real time measurement
system scenario. Testing on an industrial furnace showed that significant arc
constrictions persisted during melting and several characteristic arc distributions were
observed. Results also suggested that there were more arcs in the furnace at a given
time then could be uniquely located using the number of sensors deployed.