Lean Blowout Detection Techniques for Partially Premixed Flames in a Dump Combustor

Abstract

Lean premixed or partially premixed combustion in gas turbine combustors is one of the popular strategies for achieving very low NO x levels. However, lean combustion is extremely susceptible to lean blowout (LBO). LBO is a significant safety hazard for aeroengines and necessitates an expensive and time-consuming shutdown and restarting procedure for stationary engines. To avoid blowout, usually current combustors are therefore typically operated with a fuel–air mixture that is significantly richer than the LBO limit (ΦLBO). These techniques found to be good in terms of safety margin but produces higher NO x due to high temperature of flame. Online LBO prediction techniques are required for real-time control of combustors to maintain a continuous flame with smaller margins from the LBO limit. Experiments carried out on a gas-fueled swirl-stabilized dump combustor at different degrees of premixing and flow rates. LPG and methane are used as fuel for the combustor. Combustion behavior at different degrees of premixing has been studied. Existing LBO detection techniques are evaluated, which necessitates the development of new LBO techniques. Two optical signal-based detection techniques have been discussed here—flame color-based technique and symbolic time series analysis (STSA)-based technique. Both the methods performed well for a wide range of parametric variation.