Ben's Website

Serious musings

Hyperspectral Spectroscopy

researchundergraduw

I worked with Professor Scott Sanders and his group in the ERC throughout my undergrad. Our research looked to create sensors that quickly measure combustion systems. These tools allow engine designers to verify engine performance and critique their design metrics. These methods can also be applied to other dynamic systems; eg. rocket engines. Here's an overview:

Creation of a Fiber Optic Thermometer for Widespread Commercial Use in Internal Combustion Engines

Our research engines have windows in the cylinder walls to allow direct observation of combustion but many other engines do not have this convenience. LaVision, GmbH. has developed an optical spark plug to provide optical access to an engine while retaining normal spark plug function. With LaVision, we've developed a hyperspectral sensor to measure engine combustion in any spark-ignited engine (truck, car, jeep, lawnmower, etc.).

High Speed Grating Spectrometer

Combustion can be a dirty process and as the engine runs soot can accumulate in the optical path. There are a few ways to overcome this, this design simply threw power at it (up to 0.5 W broadband). The sensor generated broadband light (1333-1373nm) centered on the 'R' H2O absorption branch and sent it through a sample. The concentration, temperature, and pressure of water in the sample caused some wavelengths to be absorbed, altering the broadband signal. This alteration is recorded by a 14kHz infrared linescan camera attached to a grating spectrometer. Comparing the altered signal to simulated water absorption measurments allows the concentration, temperature, and pressure of the water in the sample to be determined.

Visualizing Fiber Mode Movement in Multi-mode Fiber

Light traveling through multi-mode fiber interferes with itself, causing variations in intensity along and at the exit from the fiber. The specific pattern is a function of the wavelength and position of the fiber.

A common problem in our applications is 'catching' all of the light sent through a sample. Beamsteering (the bending of light due to inhomogeneous engine conditions) limits the amount of light that we can capture in the output fiber. Enlarging the diameter of the fiber is the simple solution but mode noise is encountered in diameters beyond 10μm. Mode noise results from the light interfering with itself and dominates other noise sources in the system, as shown in the animation (right) and picture. A single-mode fiber would lack the black and white variations in intensity, appearing to be uniformly illuminated. This investigation with former M.S. student Renatta Bartula tried to qualify the mode movement but was unable to remove the mode noise through post-processing.