Session: 04-12 Flashback and Blowoff III

Paper Number: 128722

128722 - Image-Based Flashback Detection in a Hydrogen-Fired Gas Turbine Using a Convolutional Autoencoder 

Flame flashback is a major concern in hydrogen-fired gas turbines. In order to determine the flashback propensity of a hydrogen burner, several burner design tests at different operating points and fuel blends are performed at the test facility of Siemens Energy. A camera monitors the flame in the combustion chamber and the occurrence of flame flashback events in the image recordings becomes clearly visible.  We develop a data-driven approach to detect flame flashback events based on the camera images at 100% hydrogen operation. This fuel regime is simply identified by the control system of the test cell, since it is predefined in the test procedure.  At 100% hydrogen concentration, all images feature identical characteristics since the pure hydrogen flame is not visible for the camera. Simultaneously, the highest susceptibility to flashback is attained in this regime. We use both facts as well as the good suitability of image data for performing Machine Learning tasks to train our model to detect anomalies. Here, anomalies correspond to flashback events. Flashback is captured by a Convolutional Autoencoder (CAE) using the reconstruction error associated with a dynamic threshold as a measure of anomaly. This newly developed dynamic threshold overcomes the difficulties in the generalization capability of the CAE that could not be solved by advanced image processing techniques. Along with the CAE, the compressed representation (the latent space of the CAE) detects the position of flame flashback events. Our methodology proves that it is possible to detect flame flashback using only flame images.

Presenting Author: Paul Porath Technische Universität Berlin

Presenting Author Biography: Paul Porath has completed his master's degree in mechanical engineering at the TU Berlin in 2022. During this time, he worked as a working student at Siemens AG in the combustion department, where he was mainly responsible for CFD simulations. After graduation, he decided to pursue an academic path at the TU Berlin in Prof. Ghani's DMF group, where he has been researching turbulent premixed hydrogen flames for the past year.

Authors:

Paul Porath Technische Universität Berlin
Vikas Yadav Technische Universität Berlin
Lukasz Panek Siemens Energy AG
Abdulla Ghani Technische Universität Berlin