GTI2016: Filling in the missing link on combustion monitoring between laboratory and real gas turbine

Meggitt Sensing Systems and Combustion Bay One work together on monitoring of combustion dynamics in gas turbines. The latest progress will be presented at the conference GTI2016 in Berlin (presentation no. 026). This presentation focuses on the latest efforts made to automate the recording and documentation of even small changes of the burner operation conditions, and correlate these to the dynamic measurements.

Key elements of the combined static-dynamic synchronous acquisition (click to download the presentation)

Synchronous capture of GT combustor operating conditions and main combustion dynamics parameters combining quasi-static and fast-response measurements in a well-documented and compact manner
L. Pfefferkorn*, C. Wieland*, F. Giuliani*, H. Reiss†
*CBOne, Austria †Meggitt Sensing Systems, Switzerland
 

Date and location: GTI2016, Berlin, Thursday 29th September 2016 at 11:30 

The focus is put on the precise combustion control using CBOne's burner test rig and methods, and detailed description of the latter using MSS' measurement chain including fast-pressure sensors and accelerometers, as well as associated data acquisition systems.

The experimental set-up consists of an atmospheric burner, confined in a casing on which the dynamic sensors are placed. The flame can be excited acoustically using a siren, where both amplitude and frequency can be varied independently from each other. The equipment and the methodology used was presented at ASME TE 2016 [GT2016-56166].

First, the controls of the siren were improved from manual operation to automated operation based on pre-programmed nominal test sequence profiles, so that specific tests can be repeated in a well reproducible way. Furthermore, the siren control produces a log file on the actual real-time siren's operating conditions. Second, the set-up is instrumented with flowmeters, pressure gauges and thermosensors to describe precisely the operating conditions of the burner. In order to make a good use of the available acquisition chain for the dynamic values, with 16 entries acquired at sampling rates up to 98kHz. Parallel to this multiple quasi-static measurements are recorded and then recombined with the dynamic data, resulting in a structured measurement of well-time stamped static and dynamic data. Different data management strategies are tested to keep the data flow low and facilitate the post-processing, in view of a practical use on field applications.

The post-processing for the reconstruction method of the static parameters is described and illustrated by display versus dynamic parameters. It is shown how slight changes in operating conditions and/or in acoustic excitation can provoke dramatic changes on the flame's behaviour. The practical features highlighted with this method are the compactness of the data, and the ease of identifying the causes of detectable changes in flame dynamics, in combination with reduced testing time.