Abstract

In this paper, we use various regression models and Artificial Neural Network (ANN) to predict the centrifugal compressor performance map. Particularly, we study the accuracy and efficiency of Gaussian Process Regression (GPR) and Artificial Neural Networks in modelling the pressure ratio, given the mass flow rate and rotational speed of a centrifugal compressor. Preliminary results show that both GPR and ANN can predict the compressor performance map well, for both interpolation and extrapolation. We also study the data augmentation and data minimzation effects using the GPR. Due to the inherent pressure ratio data distribution in mass-flow-rate and rotational-speed space, data augmentation in the rotational speed is more effective to improve the ANN performance than the mass flow rate data augmentation.

Volume Subject Area:
Heat Transfer and Thermal Engineering
Keywords:
centrifugal compressor, Gaussian process regression, neural network, artificial intelligence, machine learning
Topics:
Artificial neural networks, Compressors
This content is only available via PDF.
Copyright © 2019 by ASME. This work was authored in part by a U.S. Government employee in the scope of his/her employment. ASME disclaims all interest in the U.S. Government’s contribution.
You do not currently have access to this content.