Pith. sign in

REVIEW 3 major objections 6 minor 49 references

Physics-Constrained Machine Learning for Chemical Engineering

T0 review · 3 major / 6 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Embedding physical laws in machine learning advances chemical-engineering models, but four bottlenecks block adoption.

desk verdict A clear, conventional perspective restating the PCML taxonomy and challenges; no new results, but an honest and useful overview for chemical engineering. read the letter →

arxiv 2508.20649 v1 pith:U73XMZ4L submitted 2025-08-28 cs.LG cs.SYeess.SY

Angan Mukherjee , Victor M. Zavala (Department of Chemical , Biological Engineering , University of Wisconsin-Madison , Madison , USA) This is my paper · ORCID
classification cs.LGcs.SYeess.SY
keywords physics-constrainedmachinelearningphysics-informedneuralnetworksprocesssystemsengineeringchemicalhardconstraintsuncertaintyquantificationhybridmodelingbenchmarking
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This perspective argues that embedding physical laws into machine learning models—not just fitting data—is the route to models that are reliable, generalizable, and interpretable for chemical engineering. It organizes the field into three strategies: soft constraints that penalize physics violations during training, hard constraints that force predictions to satisfy mass and energy balances exactly, and hybrid models that couple physics and learned components bidirectionally. The paper claims a concrete trade-off: soft constraints are flexible but never guarantee conservation; hard constraints guarantee physics but do not scale; hybrids are expressive but hard to train. The stake is practical: reactor design, process control, and experimental design all depend on models that must not silently violate conservation laws, especially with sparse or noisy data. Its reactor comparison shows a physics-constrained model tracking noisy transient data with tighter uncertainty bands than a pure neural model—arguing this is the payoff the field can generalize once the four named bottlenecks are solved.

What carries the argument

The load-bearing construct is the taxonomy of constraint enforcement: soft-constrained training adds a physics penalty to the data loss; hard-constrained training imposes physics as equality constraints on ML outputs, enforced either by sequential projection onto the constraint manifold or by solving an all-at-once constrained optimization (simultaneous projection) with interior-point solvers; hybrid models couple ML and physics bidirectionally. The sequential-versus-simultaneous distinction carries the argument: it governs whether models scale, handle complex hybrid structures, and satisfy constraints exactly.

What would settle it

Run one standardized benchmark on real, noisy operating-plant data—matched data volume and compute budget—comparing soft-constrained, hard-constrained, and well-regularized pure ML models. If PCML shows no advantage in accuracy, conservation-law violation rate, or data efficiency, the paper's central promise is falsified; if a hard-constrained simultaneous model trains a deep network on millions of points without difficulty, its scalability-ceiling claim is falsified.

Watch

Extended reading notes

Core claim

The paper's central claim is that physics-constrained machine learning (PCML) can yield predictive models that are accurate, physically consistent, and interpretable, especially with limited or noisy data. Structurally, it categorizes PCML into soft-constrained, hard-constrained (sequential or simultaneous projection), and general hybrid models. It argues that hard constraints via simultaneous nonlinear programming give exact conservation but currently cannot scale; sequential projection is flexible but slow and unstable; soft constraints are convenient but not guaranteed. It claims PCML improves surrogate modeling and control, and that physical constraints narrow prediction uncertainty, as

Load-bearing premise

The paper's case rests on an editorial judgment that its three-way taxonomy (soft, hard, hybrid) faithfully represents the PCML field and that the four named challenges are the true bottlenecks; if the surveyed literature is unrepresentative, the proposed research agenda loses its foundation.

Editorial extensions

If this is right

  • Hard-constrained PCML enforcing mass and energy balances exactly would give process models conservation guarantees by construction, rather than relying on data to imply them.
  • Physically consistent uncertainty quantification, via variational inference on physics-constrained neural ODEs, would enable PCML in closed-loop experimental design and stochastic predictive control.
  • A domain-agnostic benchmark suite with curated noisy, multiscale datasets and diagnostics for physical violation would replace ad-hoc case studies and make PCML claims testable.
  • Scalable training for simultaneous projection—parallel, GPU, and distributed optimization—is the precondition for multiscale PCML linking molecular structure to process economics.
  • Sequential projection methods remain practical for large datasets and deep networks but are limited to unidirectional ML-to-physics structures; hybrid bidirectional models require simultaneous training.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial inference: the identifiability discussion implies a meta-use for PCML—treating different physics priors as competing model architectures and letting noisy data discriminate among them would turn constraint selection into hypothesis testing.
  • Editorial inference: if the scalability critique is right, reduced-space or Schur-complement linear algebra tailored to neural-network Jacobians inside interior-point solvers is a natural next step the paper signals but does not develop.
  • Editorial inference: the reactor example suggests physics narrows uncertainty bands, but raises an untested question—whether physically constrained uncertainty estimates become overconfident when the embedded physics model is wrong.
  • Editorial inference: the absence of benchmarks implies current 'reliability gains' claims may be inflated by favorable case selection; a neutral benchmark would likely shrink, but not erase, the reported PCML advantage over well-regularized pure ML.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This perspective article argues that physics-constrained machine learning (PCML), which fuses mechanistic models with data-driven components, can improve the reliability, generalizability, and interpretability of chemical-engineering models. The authors propose a general modeling notation (Eq. 2.1-2.2), a taxonomy of soft-constrained, hard-constrained (sequential and simultaneous), and general hybrid approaches, and a survey of applications in surrogate modeling, real-time control, and uncertainty quantification. They illustrate a claimed advantage of PCML over standalone ML with a reactor example (Figure 4). The paper concludes by listing four open challenges: model identifiability, physically consistent uncertainty quantification, missing benchmarking standards, and scalable computation with multiscale/heterogeneous data.

Significance. The paper fills a useful niche as a perspective for the chemical-engineering community: it organizes a broad literature into a compact notation and taxonomy, and it explicitly names the lack of standardized benchmarks and physically consistent UQ as bottlenecks. This is an honest and potentially catalytic framing. The manuscript does not present new derivations, simulations, or data; its value lies in synthesis and agenda-setting. Given that no code or reproducible artifacts accompany the paper, the specific quantitative claims (e.g., the reactor comparison in Figure 4) must be evaluated for transparency. The enumerated challenges are well aligned with current debate, and the recommended future directions are concrete enough to guide research. For a perspective, the central claims are plausible, but the paper would be stronger if it distinguished more sharply between established evidence and open conjecture.

major comments (3)
  1. [§2.1, Eq. (2.4)] The physics loss as written is degenerate. Eq. (2.2)-(2.3) define ŷ = φP(u, θ), so the expression Lp(θ) = (ŷ(u,θ) - φP(u,θ))ᵀ(ŷ(u,θ) - φP(u,θ)) is identically zero for every θ. This makes the soft-constrained objective in Eq. (2.5) vacuous and undermines a central distinction of the proposed taxonomy. The physics residual should be defined with respect to the physical equations themselves (e.g., residuals of mass/energy balances or PDEs evaluated at the NN outputs), not as the difference between a symbol and its own definition.
  2. [§2.3, Figure 4] The reactor example is used to support the claim that PCML improves generalization and uncertainty relative to standalone ML, but the presentation lacks the details needed to assess or reproduce it. No reactor model equations are given; no data set size, noise distribution, or noise variance is reported; the NN architecture (depth, width, activation) is not specified; and the training procedure (optimizer, epochs, loss weights, variational-inference settings) is absent. No quantitative error or uncertainty-width metrics are provided. Without these, the figure is anecdotal. At minimum, cite or provide a documented implementation, or explicitly label the figure as a schematic illustration rather than a systematic comparison.
  3. [§3, Benchmarking Tools] There is an unresolved tension between the paper's opening claim and its own benchmarking assessment. Section 3 begins 'Although PCML has shown to improve the accuracy, generalizability, and interpretability of ML models,' but the Benchmarking Tools paragraph states that 'test problems remain fragmented' and 'rarely capture the complexities of real-world process systems.' The manuscript thus asserts field-wide benefits while conceding that tools to measure them fairly do not exist. For a perspective this tension is acceptable if the claim is framed as 'suggestive, case-study-level evidence,' but the current wording overstates what the surveyed evidence demonstrates. I recommend softening the claim and explicitly stating that the enumerated benefits are established only on specific curated examples.
minor comments (6)
  1. [§1, Introduction] Typo: 'Although there approaches for both soft-constrained...' should read 'Although there are approaches...'
  2. [§2.2, Hard-Constrained PCML] Typo: 'requires solving a more complex problem that that the ML component explicitly as constraints' should be '...than that of the ML component explicitly as constraints' or rephrased.
  3. [§2.2, Soft-Constrained PCML] Extra closing parenthesis in 'weakly incorporated as additional penalty terms [26]))'. Also, the symbol φp for the physics loss is used without defining the residual space; see the major comment on Eq. (2.4).
  4. [§2.1, Eq. (2.1)] The data set notation D = "u, y" would be clearer as D = {u_i, y_i}_{i=1}^N, especially since u and y are later treated as variables.
  5. [References] Reference [44] has a malformed author list ('Waegeman Willem Schmid Nina, Fernandes del Pozo David and Hasenauer Jan'). Please correct to standard journal format.
  6. [§1, Taxonomy] 'A couple of merging paradigms' should likely be 'emerging paradigms'.

Circularity Check

1 steps flagged · score 3.0 of 10

One formal definition is self-defeating: the physics loss is identically zero because the model output is defined to equal the physics model output; otherwise the perspective is not circular.

  1. self definitional [Section 2.1, Eq. (2.4)]
    "The physics loss function, expressed as Lp(θ), penalizes the violation of known physical constraints (represented by φp) as: Lp(θ)=(ŷ(u,θ)−φ_P(u,θ))^T(ŷ(u,θ)−φ_P(u,θ)) (2.4)"

    By Eq. (2.2c), the PCML output is defined as ŷ = φ_P(u,θ). Substituting into Eq. (2.4) gives Lp(θ) ≡ 0 for every θ. Thus the 'physics loss' does not penalize anything; the soft-constrained objective (2.5) becomes λd·Ld(θ) only. The claim that soft-constrained PCML weakly enforces physics via a penalty term reduces, by construction, to unconstrained data fitting. This is a self-definitional collapse in the formal framework, though it does not undermine the paper's broader survey or its hard-constrained examples.

full rationale

This is a perspective paper, not a derivation, so most circularity patterns do not apply: no parameter is fitted and then renamed a prediction, no uniqueness theorem is imported, and the survey claims are supported by a broad literature (e.g., Raissi et al., Karniadakis et al.) rather than by the authors' own results alone. The self-citations [7,10,15,28,45,49] are used as examples of existing hard-constrained methods and UQ demonstrations; the central message would still stand if those references were replaced by independent studies, and at least one (ref [45], PLOS Comput Biol) is externally published. The acknowledged absence of standardized benchmarks (Section 3) is a limitation and an internal tension with the confident opening assertion, but it is not circular reasoning. The one concrete circularity is the formal definition of Lp in Eq. (2.4): because ŷ is defined as φ_P in Eq. (2.2c), Lp is identically zero, so the paper's algebraic statement of soft-constrained PCML contains no physics penalty. This is a definitional collapse in the paper's own equations, and while it does not reduce the entire perspective to its inputs, it is a genuine self-definitional issue. Overall circularity is therefore low-to-moderate.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The perspective does not introduce new entities or fitted parameters. Its claims rest on assumptions about process model structure, neural network expressiveness, and the fidelity of embedded physics, all standard for the PCML literature.

assumptions (3)
  • domain assumption Chemical processes can be adequately represented by differential and algebraic equations (DAEs) that encode conservation laws and thermodynamics.
    Section 2.1 assumes the real system is governed by DAE-type physical models, which is the standard modeling paradigm in chemical engineering but is an idealization for many real processes.
  • standard math Machine learning components, such as neural networks, have sufficient approximation power to represent unknown model parts.
    The paper's premise that neural networks can be used as flexible model components relies on universal approximation properties, though it does not state the theorem explicitly.
  • domain assumption The physics constraints used in PCML are known and correctly specified by the modeler.
    All cited methods assume that the physical laws embedded are accurate. The paper itself notes that models are often approximate, but the core of PCML relies on this assumption.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Physics-Constrained Machine Learning for Chemical Engineering." pith.science (2026). https://pith.science/paper/U73XMZ4L

@misc{pith2026250820649,
  author       = {Pith},
  title        = {Pith review of: Physics-Constrained Machine Learning for Chemical Engineering},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/U73XMZ4L}},
  note         = {Machine review of arXiv:2508.20649}
}
read the original abstract

Physics-constrained machine learning (PCML) combines physical models with data-driven approaches to improve reliability, generalizability, and interpretability. Although PCML has shown significant benefits in diverse scientific and engineering domains, technical and intellectual challenges hinder its applicability in complex chemical engineering applications. Key difficulties include determining the amount and type of physical knowledge to embed, designing effective fusion strategies with ML, scaling models to large datasets and simulators, and quantifying predictive uncertainty. This perspective summarizes recent developments and highlights challenges/opportunities in applying PCML to chemical engineering, emphasizing on closed-loop experimental design, real-time dynamics and control, and handling of multi-scale phenomena.

Figures

Figures reproduced from arXiv: 2508.20649 by the authors.

Figure 1
Figure 1. Summary of the state-of-the-art PCML models and applications in process systems engineering. Soft [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Topology of different types of PCML structures. (a) [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Approaches for hard-constrained PCML using (a) sequential and (b) simultaneous projection layers. [PITH_FULL_IMAGE:figures/full_fig_p006_3.png] view at source ↗
Figures from the paper (1 more)
Figure 4
Figure 4. Figure 4: ML and PCML modeling of a transient reactor system. (a) Time-series profiles showing true outlet [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

49 extracted references · 41 canonical work pages

  1. [1]

    Physics- constrained deep learning for high-dimensional surrogate modeling and uncertainty quantifi- cation without labeled data

    Yinhao Zhu, Nicholas Zabaras, Phaedon-Stelios Koutsourelakis, and Paris Perdikaris. Physics- constrained deep learning for high-dimensional surrogate modeling and uncertainty quantifi- cation without labeled data. Journal of Computational Physics, 394:56–81, 2019

    work page 2019

  2. [2]

    Raissi, P

    M. Raissi, P . Perdikaris, and G.E. Karniadakis. Physics-informed neural networks: A deep learn- ing framework for solving forward and inverse problems involving nonlinear partial differential equations. Journal of Computational Physics, 378:686–707, 2019

    work page 2019

  3. [3]

    Scientific machine learning through physics–informed neural networks: Where we are and what’s next

    Salvatore Cuomo, Vincenzo Schiano Di Cola, Fabio Giampaolo, Gianluigi Rozza, Maziar Raissi, and Francesco Piccialli. Scientific machine learning through physics–informed neural networks: Where we are and what’s next. Journal of Scientific Computing, 92(3):88, 2022

    2022

  4. [4]

    When physics meets machine learning: a survey of physics-informed machine learning

    Chuizheng Meng, Sam Griesemer, Defu Cao, Sungyong Seo, and Yan Liu. When physics meets machine learning: a survey of physics-informed machine learning. Machine Learning for Compu- tational Science and Engineering, 1(1):20, 2025

    work page 2025

  5. [5]

    A novel temperature prediction method without using energy equation based on physics-informed neural network (pinn): A case study on plate- circular/square pin-fin heat sinks

    Kitti Nilpueng, Preecha Kaseethong, Mehrdad Mesgarpour, Mostafa Safdari Shadloo, and Som- chai Wongwises. A novel temperature prediction method without using energy equation based on physics-informed neural network (pinn): A case study on plate- circular/square pin-fin heat sinks. Engineering Analysis with Boundary Elements, 145:404–417, 2022

    work page 2022

  6. [6]

    P . O. Sturm and A. S. Wexler. Conservation laws in a neural network architecture: enforcing the atom balance of a julia-based photochemical model (v0.2.0). Geoscientific Model Development, 15(8):3417–3431, 2022. 13 https://zavalab.engr.wisc.edu/

    work page 2022

  7. [7]

    Development of steady-state and dynamic mass and energy constrained neural networks for distributed chemical systems using noisy transient data

    Angan Mukherjee and Debangsu Bhattacharyya. Development of steady-state and dynamic mass and energy constrained neural networks for distributed chemical systems using noisy transient data. Industrial & Engineering Chemistry Research, 63(32):14211–14239, 2024

    work page 2024

  8. [8]

    Lueg, Victor Alves, Daniel Schicksnus, John R

    Laurens R. Lueg, Victor Alves, Daniel Schicksnus, John R. Kitchin, Carl D. Laird, and Lorenz T. Biegler. A simultaneous approach for training neural differential-algebraic systems of equations. arXiv preprint arXiv:2504.04665, 2025

    arXiv 2025

Show all 49 references
  1. [9]

    Jakobsen

    Andres Carranza-Abaid and Jana P . Jakobsen. Neural network programming: Integrating first principles into machine learning models. Computers & Chemical Engineering, 163:107858, 2022

    2022

  2. [10]

    Development of mass, energy, and thermo- dynamics constrained steady-state and dynamic neural networks for interconnected chemical systems

    Angan Mukherjee and Debangsu Bhattacharyya. Development of mass, energy, and thermo- dynamics constrained steady-state and dynamic neural networks for interconnected chemical systems. Chemical Engineering Science, 309:121506, 2025

    2025

  3. [11]

    Constante Flores, and Can Li

    Hao Chen, Gonzalo E. Constante Flores, and Can Li. Physics-informed neural networks with hard linear equality constraints. Computers & Chemical Engineering, 189:108764, 2024

    2024

  4. [12]

    Schweidtmann

    Giacomo Lastrucci and Artur M. Schweidtmann. Enforce: Nonlinear constrained learning with adaptive-depth neural projection. arXiv preprint arXiv:2502.06774, 2025

    2025 arXiv

  5. [13]

    Schweidtmann

    Giacomo Lastrucci, Tanuj Karia, Zo ¨e Gromotka, and Artur M. Schweidtmann. Picard-kkt-hpinn: Enforcing nonlinear enthalpy balances for physically consistent neural networks. arXiv preprint arXiv:2501.17782, 2025

    2025 arXiv

  6. [14]

    Hardnet: Hard-constrained neural networks with universal approximation guarantees

    Youngjae Min and Navid Azizan. Hardnet: Hard-constrained neural networks with universal approximation guarantees. arXiv preprint arXiv:2410.10807, 2025

    2025

  7. [15]

    On the development of steady-state and dy- namic mass-constrained neural networks using noisy transient data

    Angan Mukherjee and Debangsu Bhattacharyya. On the development of steady-state and dy- namic mass-constrained neural networks using noisy transient data. Computers & Chemical En- gineering, 187:108722, 2024

    2024

  8. [16]

    Enforcing analytic constraints in neural networks emulating physical systems

    Tom Beucler, Michael Pritchard, Stephan Rasp, Jordan Ott, Pierre Baldi, and Pierre Gentine. Enforcing analytic constraints in neural networks emulating physical systems. Phys. Rev. Lett., 126:098302, Mar 2021

    2021

  9. [17]

    Ashfaq Iftakher, Rahul Golder, and M. M. Faruque Hasan. Physics-informed neural networks with hard nonlinear equality and inequality constraints. arXiv preprint arXiv:2507.08124, 2025

    2025 arXiv

  10. [18]

    Andreas W ¨achter and Lorenz T. Biegler. On the implementation of an interior-point filter line- search algorithm for large-scale nonlinear programming. Mathematical Programming, 106(1):25– 57, 2006

    2006

  11. [19]

    An overview of simultaneous strategies for dynamic optimization

    Lorenz T Biegler. An overview of simultaneous strategies for dynamic optimization. Chemical Engineering and Processing: Process Intensification, 46(11):1043–1053, 2007. 14 https://zavalab.engr.wisc.edu/

    2007

  12. [20]

    Neural networks meet physics: A survey of physics-informed approaches to modeling and simulation

    Karthika Nasir, Rahul Menon, and Sneha Iyer. Neural networks meet physics: A survey of physics-informed approaches to modeling and simulation. TechRxiv preprint TechRxiv:174612233.30190684/v1, 2025

    2025

  13. [21]

    B-pinns: Bayesian physics-informed neural networks for forward and inverse pde problems with noisy data.Journal of Computational Physics, 425:109913, 2021

    Liu Yang, Xuhui Meng, and George Em Karniadakis. B-pinns: Bayesian physics-informed neural networks for forward and inverse pde problems with noisy data.Journal of Computational Physics, 425:109913, 2021

    2021

  14. [22]

    fpinns: Fractional physics-informed neural networks

    Guofei Pang, Lu Lu, and George Em Karniadakis. fpinns: Fractional physics-informed neural networks. SIAM Journal on Scientific Computing, 41(4):A2603–A2626, 2019

    2019

  15. [23]

    Jagtap, Ehsan Kharazmi, and George Em Karniadakis

    Ameya D. Jagtap, Ehsan Kharazmi, and George Em Karniadakis. Conservative physics- informed neural networks on discrete domains for conservation laws: Applications to forward and inverse problems. Computer Methods in Applied Mechanics and Engineering, 365:113028, 2020

    2020

  16. [24]

    Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang

    George Em Karniadakis, Ioannis G. Kevrekidis, Lu Lu, Paris Perdikaris, Sifan Wang, and Liu Yang. Physics-informed machine learning. Nature Reviews Physics, 3(6):422–440, 2021

    2021

  17. [25]

    Perspectives on the integration between first-principles and data-driven modeling

    William Bradley, Jinhyeun Kim, Zachary Kilwein, Logan Blakely, Michael Eydenberg, Jordan Jalvin, Carl Laird, and Fani Boukouvala. Perspectives on the integration between first-principles and data-driven modeling. Computers & Chemical Engineering, 166:107898, 2022

    2022

  18. [26]

    Dae-pinn: a physics-informed neural network model for sim- ulating differential algebraic equations with application to power networks

    Christian Moya and Guang Lin. Dae-pinn: a physics-informed neural network model for sim- ulating differential algebraic equations with application to power networks. Neural Computing and Applications, 35(5):3789–3804, 2023

    2023

  19. [27]

    When and why pinns fail to train: A neural tangent kernel perspective

    Sifan Wang, Xinling Yu, and Paris Perdikaris. When and why pinns fail to train: A neural tangent kernel perspective. Journal of Computational Physics, 449:110768, 2022

    2022

  20. [28]

    Physics- constrained neural ordinary differential equation models to discover and predict microbial com- munity dynamics

    Jaron Thompson, Bryce M Connors, Victor M Zavala, and Ophelia S Venturelli. Physics- constrained neural ordinary differential equation models to discover and predict microbial com- munity dynamics. bioRxiv, 2025

    2025

  21. [29]

    Zavala, Carl D

    Victor M. Zavala, Carl D. Laird, and Lorenz T. Biegler. Interior-point decomposition approaches for parallel solution of large-scale nonlinear parameter estimation problems. Chemical Engineer- ing Science, 63(19):4834–4845, 2008. Model-Based Experimental Analysis

    2008

  22. [30]

    Niket Sharma and Y. A. Liu. A hybrid science-guided machine learning approach for modeling chemical processes: A review. AIChE Journal, 68(5):e17609, 2022

    2022

  23. [31]

    Physics-informed neural networks for hybrid modeling of lab-scale batch fermentation for β-carotene production using saccharomyces cerevisiae

    Mohammed Saad Faizan Bangi, Katy Kao, and Joseph Sang-Il Kwon. Physics-informed neural networks for hybrid modeling of lab-scale batch fermentation for β-carotene production using saccharomyces cerevisiae. Chemical Engineering Research and Design, 179:415–423, 2022

    2022

  24. [32]

    Physics-informed neural networks for heat transfer prediction in two-phase flows

    Darioush Jalili, Seohee Jang, Mohammad Jadidi, Giovanni Giustini, Amir Keshmiri, and Yasser Mahmoudi. Physics-informed neural networks for heat transfer prediction in two-phase flows. International Journal of Heat and Mass Transfer, 221:125089, 2024. 15 https://zavalab.engr.wisc.edu/

    2024

  25. [33]

    A physics-informed assembly of feed- forward neural network engines to predict inelasticity in cross-linked polymers

    Aref Ghaderi, Vahid Morovati, and Roozbeh Dargazany. A physics-informed assembly of feed- forward neural network engines to predict inelasticity in cross-linked polymers. Polymers, 12(11), 2020

    2020

  26. [34]

    cv-pinn: Efficient learning of variational physics-informed neural network with domain decomposition

    Chuang Liu and HengAn Wu. cv-pinn: Efficient learning of variational physics-informed neural network with domain decomposition. Extreme Mechanics Letters, 63:102051, 2023

    2023

  27. [35]

    Physics-informed neural networks with domain decomposition for the incompressible navier–stokes equations

    Linyan Gu, Shanlin Qin, Lei Xu, and Rongliang Chen. Physics-informed neural networks with domain decomposition for the incompressible navier–stokes equations. Physics of Fluids , 36(2):021914, 2024

    2024

  28. [36]

    Physics-based neural networks for simulation and synthesis of cyclic adsorption processes

    Sai Gokul Subraveti, Zukui Li, Vinay Prasad, and Arvind Rajendran. Physics-based neural networks for simulation and synthesis of cyclic adsorption processes. Industrial & Engineering Chemistry Research, 61(11):4095–4113, 03 2022

    2022

  29. [37]

    Learning the solution operator of parametric partial differential equations with physics-informed deeponets.Science Advances, 7(40):eabi8605, 2021

    Sifan Wang, Hanwen Wang, and Paris Perdikaris. Learning the solution operator of parametric partial differential equations with physics-informed deeponets.Science Advances, 7(40):eabi8605, 2021

    2021

  30. [38]

    Physics-informed neural operator for learning partial differential equations

    Zongyi Li, Hongkai Zheng, Nikola Kovachki, David Jin, Haoxuan Chen, Burigede Liu, Kam- yar Azizzadenesheli, and Anima Anandkumar. Physics-informed neural operator for learning partial differential equations. ACM / IMS J. Data Sci., 1(3), May 2024

    2024

  31. [39]

    Grace C. Y. Peng, Mark Alber, Adrian Buganza Tepole, William R. Cannon, Suvranu De, Savador Dura-Bernal, Krishna Garikipati, George Karniadakis, William W. Lytton, Paris Perdikaris, Linda Petzold, and Ellen Kuhl. Multiscale modeling meets machine learning: What can we learn? A...

    2021

  32. [40]

    A review of physics-informed ma- chine learning methods with applications to condition monitoring and anomaly detection.arXiv preprint arXiv:2401.11860, 2024

    Yuandi Wu, Brett Sicard, and Stephen Andrew Gadsden. A review of physics-informed ma- chine learning methods with applications to condition monitoring and anomaly detection.arXiv preprint arXiv:2401.11860, 2024

    2024 arXiv

  33. [41]

    Physics-informed recurrent neu- ral network modeling for predictive control of nonlinear processes

    Yingzhe Zheng, Cheng Hu, Xiaonan Wang, and Zhe Wu. Physics-informed recurrent neu- ral network modeling for predictive control of nonlinear processes. Journal of Process Control , 128:103005, 2023

    2023

  34. [42]

    Abdolvahhab Rostamijavanani, Shanwu Li, and Yongchao Yang. A study on data-driven iden- tification and representation of nonlinear dynamical systems with a physics-integrated deep learning approach: Koopman operators and nonlinear normal modes. Communications in Non- linear Sc...

    2023

  35. [43]

    Physics-informed deep koopman op- erator for lagrangian dynamic systems

    Xuefeng Wang, Yang Cao, Shaofeng Chen, and Yu Kang. Physics-informed deep koopman op- erator for lagrangian dynamic systems. Science China Information Sciences, 67(9):192201, 2024

    2024

  36. [44]

    Assessment of uncertainty quantification in universal differential equations

    Waegeman Willem Schmid Nina, Fernandes del Pozo David and Hasenauer Jan. Assessment of uncertainty quantification in universal differential equations. Philosophical Transactions of the Royal Society A, 383(2293):20240444, 2025. 16 https://zavalab.engr.wisc.edu/

    2025

  37. [45]

    Thompson, Victor M

    Jaron C. Thompson, Victor M. Zavala, and Ophelia S. Venturelli. Integrating a tailored recurrent neural network with bayesian experimental design to optimize microbial community functions. PLOS Computational Biology, 19(9):1–25, 09 2023

    2023

  38. [46]

    Deepxde: A deep learning library for solving differential equations

    Lu Lu, Xuhui Meng, Zhiping Mao, and George Em Karniadakis. Deepxde: A deep learning library for solving differential equations. SIAM Review, 63(1):208–228, 2021

    2021

  39. [47]

    Learning constrained parametric differentiable predictive control policies with guarantees

    J ´an Drgo ˇna, Aaron Tuor, and Draguna Vrabie. Learning constrained parametric differentiable predictive control policies with guarantees. IEEE Transactions on Systems, Man, and Cybernetics: Systems, 54(6):3596–3607, 2024

    2024

  40. [48]

    Continuous-molecular targeting for integrated solvent and process design

    Andr ´e Bardow, Klaas Steur, and Joachim Gross. Continuous-molecular targeting for integrated solvent and process design. Industrial & Engineering Chemistry Research, 49(6):2834–2840, 2010

    2010

  41. [49]

    Ikegwu, Panzheng Zhou, Reid C

    Ugochukwu M. Ikegwu, Panzheng Zhou, Reid C. Van Lehn, Victor M. Zavala, and Aurora del Carmen Mungu´ıa-L´opez. A fast computational framework for the design of solvent-based plas- tic recycling processes. Computers & Chemical Engineering, 199:109148, 2025. 17

    2025

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.