FlamePINN-1D: Physics-informed neural networks to solve forward and inverse problems of 1D laminar flames
0
🧠
Sign in to get full access
Overview
- The paper proposes a framework called FlamePINN-1D to solve both forward and inverse problems in combustion studies using physics-informed neural networks (PINNs).
- The framework is tested on three cases of increasing complexity: freely-propagating premixed (FPP) flames with simplified models, FPP flames with detailed models, and counterflow premixed (CFP) flames with detailed models.
- For forward problems, FlamePINN-1D aims to solve flame fields and infer unknown eigenvalues, while for inverse problems, it aims to reconstruct fields and infer unknown parameters from sparse, noisy observations.
- The results validate the capabilities of FlamePINN-1D across various flames and working conditions, and offer advantages over traditional methods.
Plain English Explanation
Combustion studies and applications often involve two types of problems: forward problems and inverse problems. Forward problems focus on predicting the behavior of a combustion system, while inverse problems aim to infer the underlying properties of the system from observed data.
The FlamePINN-1D framework proposed in this paper offers a way to solve both forward and inverse problems in a unified manner. It uses a machine learning technique called physics-informed neural networks (PINNs) to combine the governing equations of combustion systems with data-driven approaches.
The framework is tested on three increasingly complex cases of 1D laminar flames: simple freely-propagating premixed (FPP) flames, more detailed FPP flames, and counterflow premixed (CFP) flames. For forward problems, the goal is to solve the flame fields and infer unknown properties, like laminar flame speeds. For inverse problems, the goal is to reconstruct the continuous flame fields and infer unknown parameters, such as transport and chemical kinetics, from sparse, noisy observations.
The results show that FlamePINN-1D can effectively handle these various combustion problems, offering advantages over traditional methods. For example, it is differentiable and mesh-free, meaning it doesn't have the same discretization errors as traditional numerical simulations. It's also easier to implement for inverse problems, which are often more challenging to solve.
Technical Explanation
The FlamePINN-1D framework proposed in this paper aims to solve both forward and inverse problems in combustion studies using physics-informed neural networks (PINNs). PINNs are a type of machine learning model that can incorporate the governing equations of a physical system into the neural network architecture, allowing for superior generality and few-shot learning ability compared to purely data-driven approaches.
The framework is tested on three cases of increasing complexity: Case 1 are freely-propagating premixed (FPP) flames with simplified physical models, while Case 2 and Case 3 are FPP and counterflow premixed (CFP) flames with detailed models, respectively. For the forward problems, FlamePINN-1D solves the flame fields and infers the unknown eigenvalues, such as laminar flame speeds, under the constraints of the governing equations and boundary conditions. For the inverse problems, FlamePINN-1D reconstructs the continuous flame fields and infers the unknown parameters, such as transport and chemical kinetics, from noisy sparse observations of the flame.
The results across the various flames and working conditions strongly validate the capabilities of FlamePINN-1D. Compared to traditional methods, FlamePINN-1D has several advantages: it is differentiable and mesh-free, exhibiting no discretization errors, and it is easier to implement for inverse problems. The inverse problem results also indicate the possibility of optimizing chemical mechanisms from measurements of laboratory 1D flames.
The paper also discusses some proposed strategies, such as hard constraints and thin-layer normalization, that are proven to be essential for the robust learning of FlamePINN-1D.
Critical Analysis
The paper presents a promising approach to solving both forward and inverse problems in combustion studies using physics-informed neural networks (PINNs). The results demonstrate the effectiveness of the FlamePINN-1D framework across various flame configurations and working conditions.
One potential limitation of the research is the focus on 1D laminar flames, which may not fully capture the complexity of real-world combustion systems. While the authors note that the framework can be extended to higher-dimensional flames, further validation on more realistic scenarios would be beneficial.
Additionally, the paper does not discuss the computational efficiency of the FlamePINN-1D framework compared to traditional numerical simulations. As the inverse problems can be computationally intensive, the scalability and runtime of the approach could be an important consideration for practical applications.
Overall, the research presents a valuable contribution to the field of combustion modeling and highlights the potential of integrating machine learning with governing equations to tackle both forward and inverse problems in a unified way. Readers are encouraged to critically evaluate the research and consider its implications for their own work or the broader field of combustion science and engineering.
Conclusion
The FlamePINN-1D framework proposed in this paper demonstrates a promising approach to solving both forward and inverse problems in combustion studies using physics-informed neural networks (PINNs). By combining the governing equations of combustion systems with data-driven methods, the framework exhibits superior generality and few-shot learning ability compared to purely data-driven approaches.
The results across various 1D laminar flame configurations validate the capabilities of FlamePINN-1D in solving forward problems to infer unknown eigenvalues and inverse problems to reconstruct continuous fields and infer unknown parameters. The framework's differentiability, mesh-free nature, and ease of implementation for inverse problems offer advantages over traditional numerical simulation methods.
While the research focuses on 1D laminar flames, the potential for extending the framework to higher-dimensional and more realistic combustion scenarios holds promise for advancing the field of combustion modeling and optimization. Further exploration of the computational efficiency and scalability of the approach could also be valuable for practical applications.
Overall, the FlamePINN-1D framework represents a significant step forward in the integration of machine learning and governing equations to tackle the diverse challenges in combustion studies and applications.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
🧠
0
FlamePINN-1D: Physics-informed neural networks to solve forward and inverse problems of 1D laminar flames
Jiahao Wu, Su Zhang, Yuxin Wu, Guihua Zhang, Xin Li, Hai Zhang
Given the existence of various forward and inverse problems in combustion studies and applications that necessitate distinct methods for resolution, a framework to solve them in a unified way is critically needed. A promising approach is the integration of machine learning methods with governing equations of combustion systems, which exhibits superior generality and few-shot learning ability compared to purely data-driven methods. In this work, the FlamePINN-1D framework is proposed to solve the forward and inverse problems of 1D laminar flames based on physics-informed neural networks. Three cases with increasing complexity have been tested: Case 1 are freely-propagating premixed (FPP) flames with simplified physical models, while Case 2 and Case 3 are FPP and counterflow premixed (CFP) flames with detailed models, respectively. For forward problems, FlamePINN-1D aims to solve the flame fields and infer the unknown eigenvalues (such as laminar flame speeds) under the constraints of governing equations and boundary conditions. For inverse problems, FlamePINN-1D aims to reconstruct the continuous fields and infer the unknown parameters (such as transport and chemical kinetics parameters) from noisy sparse observations of the flame. Our results strongly validate these capabilities of FlamePINN-1D across various flames and working conditions. Compared to traditional methods, FlamePINN-1D is differentiable and mesh-free, exhibits no discretization errors, and is easier to implement for inverse problems. The inverse problem results also indicate the possibility of optimizing chemical mechanisms from measurements of laboratory 1D flames. Furthermore, some proposed strategies, such as hard constraints and thin-layer normalization, are proven to be essential for the robust learning of FlamePINN-1D. The code for this paper is partially available at https://github.com/CAME-THU/FlamePINN-1D.
Read more6/14/2024
0
Equation identification for fluid flows via physics-informed neural networks
Alexander New, Marisel Villafa~ne-Delgado, Charles Shugert
Scientific machine learning (SciML) methods such as physics-informed neural networks (PINNs) are used to estimate parameters of interest from governing equations and small quantities of data. However, there has been little work in assessing how well PINNs perform for inverse problems across wide ranges of governing equations across the mathematical sciences. We present a new and challenging benchmark problem for inverse PINNs based on a parametric sweep of the 2D Burgers' equation with rotational flow. We show that a novel strategy that alternates between first- and second-order optimization proves superior to typical first-order strategies for estimating parameters. In addition, we propose a novel data-driven method to characterize PINN effectiveness in the inverse setting. PINNs' physics-informed regularization enables them to leverage small quantities of data more efficiently than the data-driven baseline. However, both PINNs and the baseline can fail to recover parameters for highly inviscid flows, motivating the need for further development of PINN methods.
Read more9/2/2024
0
Data-Guided Physics-Informed Neural Networks for Solving Inverse Problems in Partial Differential Equations
Wei Zhou, Y. F. Xu
Physics-informed neural networks (PINNs) represent a significant advancement in scientific machine learning by integrating fundamental physical laws into their architecture through loss functions. PINNs have been successfully applied to solve various forward and inverse problems in partial differential equations (PDEs). However, a notable challenge can emerge during the early training stages when solving inverse problems. Specifically, data losses remain high while PDE residual losses are minimized rapidly, thereby exacerbating the imbalance between loss terms and impeding the overall efficiency of PINNs. To address this challenge, this study proposes a novel framework termed data-guided physics-informed neural networks (DG-PINNs). The DG-PINNs framework is structured into two distinct phases: a pre-training phase and a fine-tuning phase. In the pre-training phase, a loss function with only the data loss is minimized in a neural network. In the fine-tuning phase, a composite loss function, which consists of the data loss, PDE residual loss, and, if available, initial and boundary condition losses, is minimized in the same neural network. Notably, the pre-training phase ensures that the data loss is already at a low value before the fine-tuning phase commences. This approach enables the fine-tuning phase to converge to a minimal composite loss function with fewer iterations compared to existing PINNs. To validate the effectiveness, noise-robustness, and efficiency of DG-PINNs, extensive numerical investigations are conducted on inverse problems related to several classical PDEs, including the heat equation, wave equation, Euler--Bernoulli beam equation, and Navier--Stokes equation. The numerical results demonstrate that DG-PINNs can accurately solve these inverse problems and exhibit robustness against noise in training data.
Read more7/16/2024
0
Physics-informed neural networks for parameter learning of wildfire spreading
Konstantinos Vogiatzoglou, Costas Papadimitriou, Vasilis Bontozoglou, Konstantinos Ampountolas
Wildland fires pose terrifying natural hazards, underscoring the urgent need to develop data-driven and physics-informed digital twins for wildfire prevention, monitoring, intervention, and response. In this direction of research, this work introduces a physics-informed neural network (PiNN) to learn the unknown parameters of an interpretable wildfire spreading model. The considered wildfire spreading model integrates fundamental physical laws articulated by key model parameters, essential for capturing the complex behavior of wildfires. The proposed machine learning approach leverages the theory of artificial neural networks with the physical constraints governing wildfire dynamics, such as the first principles of mass and energy conservation. Training of the PiNN for physics-informed parameter identification is realized using data of the temporal evolution of one- and two-dimensional (plane surface) fire fronts that have been obtained from a high-fidelity simulator of the wildfire spreading model under consideration. The parameter learning results demonstrate the remarkable predictive ability of the proposed PiNN in uncovering the unknown coefficients in both the one- and two-dimensional fire spreading scenarios. Additionally, this methodology exhibits robustness by identifying the same parameters in the presence of noisy data. The proposed framework is envisioned to be incorporated in a physics-informed digital twin for intelligent wildfire management and risk assessment.
Read more6/24/2024