Deep learning modelling of manufacturing and build variations on multi-stage axial compressors aerodynamics

Giuseppe Bruni,Sepehr Maleki,Senthil K. Krishnababu

2024-09-06

Abstract:Applications of deep learning to physical simulations such as Computational Fluid Dynamics have recently experienced a surge in interest, and their viability has been demonstrated in different domains. However, due to the highly complex, turbulent and three-dimensional flows, they have not yet been proven usable for turbomachinery applications. Multi-stage axial compressors for gas turbine applications represent a remarkably challenging case, due to the high-dimensionality of the regression of the flow-field from geometrical and operational variables. This paper demonstrates the development and application of a deep learning framework for predictions of the flow field and aerodynamic performance of multi-stage axial compressors. A physics-based dimensionality reduction unlocks the potential for flow-field predictions, as it re-formulates the regression problem from an un-structured to a structured one, as well as reducing the number of degrees of freedom. Compared to traditional "black-box" surrogate models, it provides explainability to the predictions of overall performance by identifying the corresponding aerodynamic drivers. This is applied to model the effect of manufacturing and build variations, as the associated performance scatter is known to have a significant impact on $CO_2$ emissions, therefore posing a challenge of great industrial and environmental relevance. The proposed architecture is proven to achieve an accuracy comparable to that of the CFD benchmark, in real-time, for an industrially relevant application. The deployed model, is readily integrated within the manufacturing and build process of gas turbines, thus providing the opportunity to analytically assess the impact on performance with actionable and explainable data.

Machine Learning,Computational Engineering, Finance, and Science,Fluid Dynamics

What problem does this paper attempt to address?

The paper attempts to address the issue of aerodynamic performance variations in multi-stage axial compressors due to changes in geometric and operational variables during manufacturing and assembly processes. Specifically: 1. **Challenges in Simulating Complex Flow Fields**: Traditionally, Computational Fluid Dynamics (CFD) faces significant challenges in simulating the complex 3D turbulent flow fields of multi-stage axial compressors in the turbomachinery industry. 2. **Real-time Prediction in Industrial Applications**: A deep learning framework is proposed that can predict the flow field and aerodynamic performance of multi-stage axial compressors in real-time, with accuracy comparable to CFD benchmarks. 3. **Interpretability and Scalability**: By using a physics-based dimensionality reduction method, unstructured problems are transformed into structured ones, thereby reducing degrees of freedom and enhancing the interpretability of predictions. 4. **Impact of Manufacturing and Assembly Variations**: The study investigates the impact of variations in the manufacturing and assembly processes (such as gaps and surface roughness) on overall performance, particularly the significant impact of these variations on CO₂ emissions. The research aims to develop a method capable of quickly assessing the performance of each engine build, avoiding the high computational cost and long time cycles required by traditional CFD analysis, thereby providing practically feasible data support.

Deep learning modelling of manufacturing and build variations on multi-stage axial compressors aerodynamics

C(NN)FD -- a deep learning framework for turbomachinery CFD analysis

Application of Machine Learning to Forced Response Predictions of an Industrial Axial Compressor Rotor Blade

Steady-State Transonic Flowfield Prediction via Deep-Learning Framework

Machine learning algorithm functional on environmental sustainability assessment in turbomachinery sector: Application on centrifugal compressors

Towards Predictive Maintenance of a Heavy-Duty Gas Turbine A New Hybrid Intelligent Methodology for Performance Simulation

Design Optimization of Multi-Stage Axial-Flow Compressor Using Machine Learning Models

Combining Thermodynamics-based Model of the Centrifugal Compressors and Active Machine Learning for Enhanced Industrial Design Optimization

Fluid Dynamics of Axial Turbomachinery: Blade- and Stage-Level Simulations and Models

Efficient Performance Prediction of a One-and-half Stage Axial Compressor Based on a Streamline-Curvature Throughflow Method and Stage Characteristics

The Current State of High-Fidelity Simulations for Main Gas Path Turbomachinery Components and Their Industrial Impact

AI Assisted High Fidelity Multi-Physics Digital Twin of Industrial Gas Turbines

Field Reconstruction and Off-Design Performance Prediction of Turbomachinery in Energy Systems Based on Deep Learning Techniques

Performance Prediction and Design Optimization of a Transonic Rotor Based on Deep Transfer Learning

Integrated Graph Deep Learning Framework for Flow Field Reconstruction and Performance Prediction of Turbomachinery

Off-Design Performance Prediction of a S-CO2 Turbine Based on Field Reconstruction Using Deep-Learning Approach

Centrifugal Compressor Performance and Flow Path Generation With Artificial Intelligence

Aerodynamic Prediction on the Off-Design Performance of a S-CO2 Turbine Based on Deep Learning

Deep learning to develop zero-equation based turbulence model for CFD simulations of the built environment

Development of a Throughflow-Based Simulation Tool for Preliminary Compressor Design Considering Blade Geometry in Gas Turbine Engine

Transition Modeling for Low Pressure Turbines Using Computational Fluid Dynamics Driven Machine Learning