<p>Acoustic detection and characterisation of laser induced breakdown is an attractive proposition in laser ignition systems in which condition monitoring is necessary but where optical access for monitoring purposes is impractical. Presented is a signal processing method based on wavelet decomposition for the non-invasive detection of acoustic emissions resulting from laser induced breakdown in an atmospheric pressure combustion test rig, representative of a single combustion chamber in a sub 15 MW industrial gas turbine. The probability and consistency of laser induced breakdown is determined from the acoustic signal and used to characterize the operating conditions and identify abrupt and incipient or slowly developing faults.</p>