A Co-Simulation Approach for Transient Thermal Clearance Analysis of a Gas Turbine Whole Engine Model

Abstract Gas turbine analysis is an interdisciplinary field involving many different phenomena including but not limited to fluid flow, heat transfer, and structural mechanics. In order to utilize the most optimal workflows in each of these fields, engineers often use different in-house or commercial off-the-shelf (COTS) software tools depending on the physics being analyzed. When the different physics interact, manual operations to transfer data between software tools require customized scripts and can take many iterations to yield a converged result. The industry typically relies on automation scripts to facilitate this process. Such scripts require constant maintenance, as well as extensive verification and validation efforts. In this paper, we highlight a co-simulation workflow that eliminates manual file transfer, reduces design turnaround time, and allows for more design iterations. This approach involves co-simulation between a finite element method (FEM)-based thermo-mechanical gas turbine whole engine model (WEM) and a 1D flow model representing the secondary flow paths. The models are connected using the Functional Mock-up Interface (FMI) industry standard. First, we will describe the WEM and the 1D flow models and define the co-simulation interfaces that will enable the exchange of data. We will focus in particular on the 3D thermo-mechanical transient responses that have an impact on the 1D flows which then also influence the temperatures of the solid. Through the results obtained, we will describe the benefits of this approach over conventional approaches in the gas turbine industry today. This paper demonstrates the potential for increased accuracy in results through co-simulation by exchanging comprehensive transient information such as seal clearance and mass flow. Additionally, we showcase an enhanced data exchange process that eliminates the need for manual file transfers, resulting in faster design analysis turnaround times by reducing the number of iterative solves required.