Surrogate Model-Based Optimization of a Dual-Shield Total Temperature Probe for Aero-Engine Applications

The design of high-precision total temperature probes for aero-engines is constrained by the massive computational cost of high-fidelity simulations. This paper overcomes this barrier by introducing a surrogate model-based optimization framework for a dual-shield probe. A computationally efficient data-driven framework is established, merging conjugate-heat-transfer Computational Fluid Dynamics (CFDs), a Support Vector Regression (SVR) model, and a Genetic Algorithm (GA), which collectively replace the traditional costly design loop. The surrogate model’s exceptional predictive fidelity is confirmed, and this approach obtains improvement in measurement accuracy, successfully reducing the temperature deviation and meeting the stringent requirement. Finally, the demonstrated framework is geometry-agnostic, establishing a generalizable and cost-effective strategy for the rapid design of high-performance thermometric components in gas turbine systems.