Announcing our presentation at the 2017 ASME Turbo Expo Conference

Anthony Jarrett will be presenting a paper at the 2017 ASME Turbo Expo Conference (Charlotte, North Carolina) on June 29th. The paper describes one of the modules of LPTi's XactLIFE system, and is titled Validation of a Gas Turbine Thermodynamic Model Without Accurate Component Maps, authored by Anthony Jarrett and Ying Chen.                                                              

The abstract is reproduced below. Find us at the conference to hear all the details!

The authors have developed an engine performance model for use within a physics-based analysis tool to predict gas turbine engine life. The model employs a multivariate optimization method to solve the gas turbine thermodynamic equations, and incorporates a calibration phase to capture the behavior of individual engines without requiring accurate component maps.

To validate this approach, a database of test cell data for a turboprop engine has been used. The data consists of approximately 80 engine tests; each one with five operating points. Using a cross-validation method, each engine was uniquely calibrated using four of the operating points, and then validated using parameters from the fifth operating point. To benchmark the calibration process, these analyses were repeated without the calibration stage.

The un-calibrated outputs showed a lack of both precision and accuracy, due to imprecision in the component maps, and variation from engine to engine. In contrast, the calibrated outputs of compressor discharge temperature (CDT), compressor discharge pressure (CDP), and turbine inlet temperature (TIT) were predicted within 1% error for more than 95% of all cases.

Although most of the key thermodynamic parameters were predicted accurately, we have found that the shaft power calculation demonstrates some significant deviations from the test cell data. This has been attributed to the formulation of the turboprop thermodynamic model, and ongoing work is attempting to mitigate this issue. This understanding of the characteristic engine algorithms will provide valuable guidance in selecting suitable engine parameters as inputs and references.