CONSTRUCTAL LAW, BIOMIMICRY, AND TOPOLOGY OPTIMIZATION THROUGH THE LENS OF GENERATIVE AI
DOI:
https://doi.org/10.59277/CLC.2024.09Keywords:
Morphological Freedom, Thermal Management, Nature-Inspired Design, Generative AIAbstract
Efficient heat management in the context of semiconductor miniaturization demands unconventional solutions to navigate constrained design spaces. The emergence of unsupervised learning as a versatile design companion capable of processing multi-disciplinary datasets presents promising avenues for streamlined bio-inspired solutions for fundamental conduction problems. As a case study, we tackle the pivotal area-to-point problem that led to the formulation of the Constructal Law.
This study proposes a novel methodology harnessing Denoising Diffusion Probabilistic Models (DDPM) to seamlessly integrate features from topology optimization, constructal theory, and biomimetic structures. We use DDPM as a digital non-equilibrium complex system to generate geometric patterns for conductive heat sink design. These patterns are sampled to address the original formulation of the area-to-point problem.
This study emphasizes the significance of morphological freedom in addressing multi-objective problems, reinforcing arguments first formulated within the constructal paradigm. Our framework facilitates low-cost sampling of intricate shapes capable of serving multiple temperature objectives by synthesizing principles from different fields. We show the emergence of dendritic structures to solve distribution problems in an unsupervised learning scenario, drawing a parallel between information and energy flows.
This research underscores the transformative potential of Generative AI in blending design features across disparate disciplines, a potent tool for developing conductive heat sink solutions beyond deterministic optimization approaches.
References
(1) Bejan A., Evolution in thermodynamics, Appl. Phys. Rev., 4, 1 (2017).
(2) Bejan A., Constructal-theory network of conducting paths for cooling a heat generating volume, Int. J. Heat Mass Transf., 40, 4, pp. 799–816 (1995).
(3) Marck G., Nemer M., Harion J.L., Russeil S., Bougeard D., Topology optimization using the SIMP method for multiobjective conductive problems, Numerical Heat Transfer, Part B: Fundamentals, 61, 6, pp. 439–470 (2012).
(4) Li B., Hong J., Yan S., Liu H., Ge L., Generating optimal heat conduction paths based on bionic growth simulation, International Communications in Heat and Mass Transfer, 83, pp. 55–63 (2017).
(5) Cai H., Guo K., Liu H., Xiang W., Liu C., Multiobjective optimization of area-to-point heat conduction structure using binary quantum-behaved PSO and Tchebycheff decomposition method, Canadian Journal of Chemical Engineering, 99, 5, pp. 1211–1227 (2021).
(6) Sohl-Dickstein J., Weiss E.A., Maheswaranathan N., Ganguli S., Edu S., Deep Unsupervised Learning using Nonequilibrium Thermodynamics, ICML 2025, Forty-Second International Conference on Machine Learning, Vancouver Convention Center, 13th – July 19th.
(7) Ho J., Jain A., Abbeel P., Denoising Diffusion Probabilistic Models, Adv. Neural Inf. Process. Syst., 2020, Jun.–Dec.
(8) Jarzynski C., Equilibrium free energy differences from nonequilibrium measurements: a master equation approach, Phys. Rev. E. Stat. Phys. Plasmas Fluids Relat. Interdiscip. Topics, 56, 5, pp. 5018–5035 (1997).
(9) Song J., Meng C., Ermon S., Denoising Diffusion Implicit Models.
(10) Shen S.C., Buehler M.J., Nature-inspired architected materials using unsupervised deep learning, Communications Engineering, 1, 1, pp. 1–15 (2022).
(11) Bejan A., Vascular flow design and predicting evolution, International Communications in Heat and Mass Transfer, 155, p. 107517 (2024).
(12) Bejan A., Constructal design evolution versus topology optimization, International Communications in Heat and Mass Transfer, 141 (2023).
(13) Bejan A., “Evolutionary design: Heat and fluid flow together,” International Communications in Heat and Mass Transfer, 132, p. 105924 (2022).
(14) Ignuta-Ciuncanu M.C., Stärk H., Martinez-Botas R., Bio-Inspired Design of Conductive Heat Sinks Using a Generative Autoencoder Framework, https://ssrn.com/abstract=4576761.
