Abstract

Thermoacoustic oscillations occur in combustion chambers when heat release oscillations lock into pressure oscillations. They were first observed in lamps in the 18th century, in rockets in the 1930s, and are now one of the most serious problems facing gas turbine manufacturers.

Most analysis of thermoacoustic oscillations has been linear and has considered the point at which the oscillations become unstable. Nonlinear studies have usually assumed that, once linearly unstable, a thermoacoustic system grows to limit cycle oscillations. Recent experiments, however, show that limit cycle oscillations in thermoacoustics are the exception rather than the rule. These experiments motivate the current study.

In this study, a pre-mixed flame in a tube is modelled using a level-set approach. The flame dynamics is coupled to the acoustics. This system, although relatively simple, exhibits much of the elaborate nonlinear behaviour found in experiments, such as bi-periodic, quasi-periodic, multi-periodic and chaotic oscillations. This raises questions, and some answers, about how thermoacoustic oscillaions should be modelled in the nonlinear regime.