This paper demonstrates the application of artificial intelligence-driven turbomachinery design, its numerical performance predictions and their numerical validation. A common problem in the industrial application of turbomachinery is that readily available turbomachines are not necessarily matching the desired performance targets (performance characteristics) required for a specific application. Many machines operate under off-design conditions and hence are not operating at maximum efficiency. Traditional numerical analysis and response-driven optimization methods are ineffective and still too time-consuming and are particularly sensitive to changing performance targets. Most commercially available optimization algorithms are based on maximizing or minimizing a response function, for instance the standard error from a desired target performance characteristic of a turbomachine, by changing design variables.

This work uses a newly developed artificial intelligence-based approach that is not dependent on the specific design target using the turbomachinery design software AxSTREAM from SoftInWay. Here a neural network was trained within a constraint design space by many samples of design variables and their respective numerical performance predictions. For the numerical verification of the designs the solver Simcenter STAR-CCM+ from Siemens was used. Subsequently the trained neural network was applied to generate a set of design parameters that satisfied the physically feasible desired target performance characteristics very fast. This trained neural network enabled an effective reversal of the traditional iterative design process where now the desired target performance characteristics became the input and the geometry became the output, turning it into a generative inverse design process. This method was applied to generate a centrifugal compressor design within a given geometrically and physically constraint design space. A specific desired target performance characteristic was chosen. The generated designs and results are presented in detail.


Numerical Performance Predictions of Artificial Intelligence-Driven Centrifugal Compressor Designs

Division / Sponsors: 
Heat Transfer Division (HTD), Fluids Engineering Division (FED)

Heat Transfer Summer Conference
Paper Proceedings: https://asmedigitalcollection.asme.org/HT/HT2020/volume/83709

18th International Conference on Nanochannels, Microchannels, and Minichannels
Paper Proceedings: https://asmedigitalcollection.asme.org/ICNMM/ICNMM2020/volume/83693

Fluids Engineering Division Summer Meeting
Paper Proceedings: https://asmedigitalcollection.asme.org/FEDSM/FEDSM2020/volume/83716

HT-ICNMM2020 - Virtual Conference, SHTC - Virtual Conference, FEDSM2020 - Virtual Conference

Supporting the need among researchers worldwide for additional online and multimedia content, ASME now offers the ASME Conference Video Collection. This platform will collect and provide long-term access to conference presentation videos, encompassing all technology areas within ASME's technical conference and events program. About 30,000-40,000 engineers across all mechanical engineering and related disciplines participate in ASME conferences each year. This includes authors of approximately 6,000 to 8,000 technical papers published by ASME, the majority of whom also submit video presentations for our virtual conferences. The video collection will be available as an add-on to existing product packages or as a separate subscription.