Abstract
The global drive toward renewable energy is imposing challenging operating requirements on power turbines. Flexible load-leveling applications must accept more frequent and demanding start-stop cycles. Full transient analyses are too computationally expensive for real-time simulation across all operating regimes so monitoring relies on sparse physical measurements. Alone, these sparse data lack the fidelity for real-time prediction of a complex thermal field. A new hybrid methodology is proposed, coupling data across a range of fidelities to bridge the limitations in the individual analyses. Combining several fidelity methods in parallel: low-order models, corrected by real-time physical measurements, are calibrated with high-fidelity simulations. The multifaceted hybrid approach enables the real-time speed of low-order analysis at high resolution. This paper series develops the critical enabling features of the hybrid method. Fast three-dimensional thermal simulation is fundamental to the methodology. Low-order network models enable the real-time thermal calculation of regions inaccessible to monitoring and facilitate clearance and stress simulation necessary for flexible turbine operation. A novel automated construction method is presented, allowing complex full turbine thermal networks to be built with ease. Developed in Tensorflow, the thermal networks directly support graphics processing unit acceleration and neural network integration for seamless data transfer within the hybrid system. Capturing flow and material physics from high fidelity data, the hybrid network method demonstrates comparable accuracy at greatly reduced computational cost. The method is validated using real-machine data capturing a period of flexible transient operation.