Learning thermoacoustic interactions in combustors using a physics-informed neural network

Abstract

We introduce a physics-informed neural network (PINN) method to study thermoacoustic interactions leading to combustion instability in combustors. Specifically, we employ a PINN to investigate thermoacoustic interactions in a bluff body anchored flame combustor, representative of ramjet and industrial combustors. Vortex shedding and acoustic oscillations appear in such combustors, and their interactions lead to the phenomenon of vortex-acoustic lock-in. Acoustic pressure fluctuations at three locations and the total flame heat release rate serve as the measured data. The coupled parameterized model is based on the acoustic equations and the van der Pol oscillator for vortex shedding. The PINN was applied in the combustor, where the measurements suitable for a future machine learning application were not anticipated at the time of the experiments, as is the case in the vast majority of available data in the literature. We demonstrate a good performance of PINN in generating the acoustic field (pressure and velocity fluctuations) in the entire spatiotemporal domain, along with estimating all the parameters of the model. Therefore, this PINN-based model can potentially serve as an effective tool in improving existing combustors or designing new thermoacoustically stable and structurally efficient combustors.

References in corpus (20)

• Adam: A Method for Stochastic Optimization
• B-PINNs: Bayesian Physics-Informed Neural Networks for Forward and Inverse PDE Problems with Noisy Data
• hp-VPINNs: Variational Physics-Informed Neural Networks With Domain Decomposition
• Physics-informed neural networks for inverse problems in nano-optics and metamaterials
• PPINN: Parareal Physics-Informed Neural Network for time-dependent PDEs
• Deep Learning of Vortex Induced Vibrations
• A Physics Informed Neural Network for Time-Dependent Nonlinear and Higher Order Partial Differential Equations
• On the convergence of physics informed neural networks for linear second-order elliptic and parabolic type PDEs
• Uncovering near-wall blood flow from sparse data with physics-informed neural networks
• Transfer learning based physics-informed neural networks for solving inverse problems in engineering structures under different loading scenarios
• Tackling the Curse of Dimensionality with Physics-Informed Neural Networks
• Geometric MCMC for Infinite-Dimensional Inverse Problems
• Self-Adaptive Physics-Informed Neural Networks using a Soft Attention Mechanism
• Approximation of the acoustic radiation impedance of a cylindrical pipe
• $\Delta$-PINNs: physics-informed neural networks on complex geometries
• Physics-Informed Echo State Networks
• Physics-informed neural networks for predicting gas flow dynamics and unknown parameters in diesel engines
• Stability analysis of thermo-acoustic nonlinear eigenproblems in annular combustors. Part I. Sensitivity
• Learning characteristic parameters and dynamics of centrifugal pumps under multiphase flow using physics-informed neural networks
• Separable PINN: Mitigating the Curse of Dimensionality in Physics-Informed Neural Networks

Cited by in corpus (1)

• A Digital Twin for Diesel Engines: Operator-infused Physics-Informed Neural Networks with Transfer Learning for Engine Health Monitoring