Quantification of spatial resolution of in-cylinder chemical species tomography

Alex Tsekenis, Hugh McCann

Research output: Contribution to conferencePoster

Abstract

In response to stringent emission regulations and increased demand for vehicles with higher fuel economy, a large amount of research is being conducted on high efficiency and low-emission compression-ignition (CI) technologies. Novel combinations of internal combustion engines and fuel blends are being developed and applied.

For such engine technologies, the spatial distributions of reacting chemical species such as hydrocarbons are a primary determinant of their performance. In-situ visualisation of hydrocarbon features and their spatial distributions is of direct interest to current engine research.

Manchester’s Chemical Species Tomography (CST) using Near-IR absorption spectroscopy is a fast and minimally-intrusive method to visualise the in-cylinder distribution of target species. The method can be applied on commercial fuels. A number of applications have been reported using hydrocarbons as the target species, visualising fuel distribution during engine operation.

The minimally-intrusive nature of the technique places constraints on the number of projections that can be collected using commercially available optical components. Consequently, tomographic image reconstruction becomes an ill-posed problem. With the addition of factors impacting the signal to noise ratio of the measurements, the reconstructed images are of finite and non-uniform spatial resolution. As the features of interest in a practical engine application span a broad range of sizes and concentration levels, it is crucial to rigorously characterise the system’s spatial resolution.

We are presenting a novel experimental method for the rigorous characterisation of the spatial resolution of the CST system described above. We believe this method can be adapted for use in other hard-field tomographic modalities. The method has been specifically designed to address the non-uniform spatial resolution across the imaging space of a system that utilises a sparse and irregular imaging array. The characterisation is objective and repeatable.

The method employs a rectangular propane plume with length equal to the bore of the engine and hence the imaging space. The feature is angularly translated inside the imaging space at predetermined positions, with one of the longitudinal edges crossing the centre of the imaging space at all times. Tomographic image reconstruction of the plume is performed at each angular position.

Analysis of the collected data for each position follows with plotting the edge spread function (ESF) of the diametrical edge, followed by the line spread function (LSF) and finally the modulation transfer function (MTF). An amplitude threshold is set on the MTF, and the highest occurring frequency at that threshold is translated back into the spatial domain to yield a spatial resolution figure across the edge for that angular position. The spatial resolution figure from each position is used to plot a profile across the imaging space using a colour grading scheme.

We will discuss the experimental challenges arising when attempting to create a well defined, planar gaseous flow to be used in these experiments, and how regimes of non-laminar flow impact the acquired data from the modality employed. The gaseous feature boundaries are visualised with refraction index photography and compared with initial tomographic image reconstructions of the same feature.
Original languageEnglish
Publication statusPublished - 2011
EventGordon Conference on LASER Diagnostics in Combustion - Waterville Valley Resort, NH, United States
Duration: 14 Aug 2011 → …

Conference

ConferenceGordon Conference on LASER Diagnostics in Combustion
CountryUnited States
Period14/08/11 → …

Fingerprint Dive into the research topics of 'Quantification of spatial resolution of in-cylinder chemical species tomography'. Together they form a unique fingerprint.

Cite this