Generation of Gas Turbine Component Maps on Basis of Operational Data

Abstract

This paper describes how operational data from heavy duty gas turbines can be used for component map generation. The main aspects described are the data evaluation and validation process, the applied degradation correction methodology and the component map generation using calibrated streamline curvature methods for compressor and turbine. The operational data storage system of heavy duty gas turbines can be used, in case the customer agreed to provide the data, for fleet statistics, degradation behavior investigation or component map generation. The update of existing component maps using operational data is mainly necessary for older gas turbines types since the available numerical gas turbine models do not always represent the current state of knowledge. The process of generating component maps based on operational data requires several steps which are explained in detail in this paper. The first step is the data evaluation and validation part. This step is based on a full thermodynamic evaluation including an evaluation of systematic and random uncertainties for all required performance parameters. This generated dataset is then validated using a combination of a Kalman Filter based single fault isolator and a fuzzy logic based multiple fault isolator. A short performance evaluation of this data validation system is given as well. After the validation part the operational dataset is corrected for aging effects regarding compressor and turbine performance in order to get the new and clean component characteristic. Subsequently, a validated and aging corrected high quality database for the component map generation is available. The applied steps as well as a direct comparison for the compressor efficiency prior and post aging correction are displayed. In the following steps, already existing streamline curvature methods (SCM) for compressor and turbine are adapted to the generated dataset using a probabilistic based calibration process. The applied optimization technique is identical for compressor and turbine, but two different approaches for the calibration of the loss modeling have been implemented. The compressor SCM is calibrated with a minimum set of modified loss parameters which are modeled as a function of load. For the turbine, the modifications of the loss coefficients are constant over load. This requires an increased set of loss parameters for calibration compared to the compressor. The calibration results for both components are presented and discussed in detail. The calibrated SCMs can now be used for the component map generation in order to yield high quality component maps in accordance with current fleet experience even for older gas turbine frames.