C(NN)FD - Deep Learning Modelling of Multi-Stage Axial Compressors Aerodynamics

Abstract The field of scientific machine learning and its applications to numerical analyses such as Computational Fluid Dynamics (CFD) has recently experienced a surge in interest. While its viability has been demonstrated in different domains across a wide range of engineering applications, it has not yet reached a level of robustness and scalability to make it practical for industrial applications in the turbomachinery field. The highly complex, turbulent, and three-dimensional flows of multi-stage axial compressors for gas turbine applications represent a remarkably challenging case. This is due to the high-dimensionality of the regression of the flow-field from geometrical and operational variables, and the high computational cost associated with the large scale of the CFD domains. This paper demonstrates the development and application of a generalized deep learning framework for predictions of the flow field and aerodynamic performance of multi-stage axial compressors, which is also potentially applicable to any type of turbomachinery. A physics-based dimensionality reduction unlocks the potential for flow-field predictions for large-scale domains, re-formulating the regression problem from an unstructured to a structured one and reducing the number of degrees of freedom. The relevant physical equations are used to define a multi-dimensional physical loss function, which considers the flow field, gradient of the flow field, radial distributions and integral quantities. Compared to “black-box” approaches, the proposed framework has the advantage of physically explainable predictions of overall performance, as the corresponding aerodynamic drivers can be identified on a 0D/1D/2D/3D level. An iterative architecture is employed, improving the accuracy of the predictions, as well as estimating the associated uncertainty at each level of the predictions. A quantifiable measure of the confidence in resolving the relevant flow features is provided, without a significant increase in computational cost. The model is trained on a series of dataset including manufacturing and build variations, as well as different geometries, compressor designs and operating conditions, representative of typical aerodynamic assessments. This demonstrates the capability to predict the flow-field and the overall performance in a generalizable manner, for different geometries and compressors across their maps, with accuracy comparable to the CFD benchmark.