Investigation of Ignition Probability in a Gas Turbine Combustor Using Laser Ignition

Abstract

For an aircraft gas turbine engine, the ignition performance is usually expressed in terms of the range of flight conditions over which stable combustion can be established. At present, the size of an aircraft’s gas turbine combustor is governed predominately by its altitude relight performance and is principally derived from existing empirical design rules. These indicate the volume required to give adequate primary zone airodynamic loading to achieve the desired relight performance. A possible means of improving altitude relight performance is to relocate the point of ignition away from the combustor wall to a more favourable location within the combustion chamber. One way of achieving this is through the use of laser ignition. Reported in the paper is an initial programme of work in which the possibility of using laser ignition has been investigated in a gas turbine research combustor operating at several inlet conditions. Comparative results show that, when sited at the same location, laser ignition gave no noticeable improvement in ignition performance when compared to a standard surface discharge igniter (SDI). However, by using laser ignition to locate the ignition site away from the combustor wall, the range of combustor mass flow and AFR at which ≥75% ignition probability could be achieved was increased by approximately 33%. In conjunction with the experimental study, an ignition probability model, based on the local magnitude of a Karlovitz stretch factor, has been developed to identify suitable regions within the combustor in which to apply laser ignition. The Karlovitz parameter gives an indication as to whether or not a flame kernel will propagate successfully and has been used in conjunction with flow field and scalar distributions generated by Computational Fluid Dynamics (CFD) to yield a 2D map of ignition probability. However, the sensitivity shown by the model to the accuracy of the CFD predictions meant that reliable estimates of optimum ignition sites could only be obtained using experimental fuel and turbulence intensity distributions.